Face familiarity elicits reliable neural signatures during perception, yet how it is represented when faces are recalled from memory by using visual imagery and how these signals relate to perceptual representations, remains unclear. We combined perceptual and imagery conditions with electroencephalographical measures and multivariate pattern analysis. In a perceptual session we presented the faces of four famous persons while in the imagery session we used a cued-recall paradigm and presented abstract symbols, associated to them. By using time resolved multivariate decoding we tested if familiarity information is shared across stimulus categories, and if perceptual face familiarity representations overlap with processing during cued-recall, related to imagery. The face perception condition showed robust familiarity information over occipito-temporal electrodes. The symbolic cues also carried similar familiarity information, demonstrating cross-category familiarity processing between symbols and faces. During cued-recall, familiarity information emerged over posterior brain regions and evolved into a prolonged and late information component over right-anterior electrode sites. Importantly, cross-domain temporal generalization analysis revealed a strong overlap between perception and imagery: perceptual face familiarity information mapped onto later recall-related processing, with a pronounced and prolonged right-anterior dominance and a shorter overlap over left-posterior channels. Together, these results suggest that familiarity processing, similarly to category representations, is supported by a common representational format across perception and imagery, transitioning from dynamic posterior representations, consistent with perceptual analysis, to a stable and prolonged representation over anterior part of the right hemisphere, that is compatible with the maintenance of imagery.

Regulating the space between ourselves and others is a fundamentally embodied process, integrating bodily signals, motivational tendencies, and emotional responses. Loneliness, a significant risk factor for mental and physical health, has been linked to disruptions in these processes, yet most research has relied on static stimuli rather than dynamic, ecologically valid approaches.

This pilot study explored whether loneliness and interoceptive body trust shape cardiac orienting responses to interpersonal proximity. 19 participants completed a stop-distance task in immersive virtual reality where avatars approached at varying speeds while cardiac activity was recorded continuously, followed by standardised questionnaires.

Initial correlational analyses revealed that both trait loneliness and interoceptive body trust independently associated with peak cardiac deceleration amplitude. Given their negative intercorrelation, we tested a mediation model. The analysis revealed a speed-dependent suppression effect in fast trials: loneliness directly predicted shallower deceleration, suggesting allostatic blunting of attentional orienting towards social stimuli, while simultaneously predicting lower body trust, which in turn predicted deeper deceleration, reflecting unmodulated amplification of the orienting response through a separate interoceptive pathway. These opposing effects cancelled at the total level, masking both processes entirely. These findings suggest that loneliness is associated with cardiac orienting through two mechanistically distinct routes, and that interoceptive body trust plays a key role in how the social environment is physiologically registered.

In line with open science principles, the full pipeline was built using open-source tools (e.g. BBSIG Pipeline, EDIA Toolbox, and BIDS-formatting), and all analysis code is publicly available on GitHub.

Fear conditioning paradigms offer a framework for examining associative learning across neural, physiological, and subjective systems. Although neural indices of conditioned stimulus (CS) discrimination are well established, their relation to psychophysiological and subjective measures remains unclear. In a large sample (n > 250), we examined how neural CS discrimination, assessed via functional magnetic resonance imaging (fMRI), relates to skin conductance responses (SCR), fear-potentiated startle (FPS), and self-reported fear and US-expectancy.

Across outcome measures, greater CS discrimination was generally associated with increased activity in brain regions implicated in threat processing, including dorsolateral prefrontal cortex (dlPFC), dorsal anterior cingulate cortex (dACC), anterior insula, and thalamus. This pattern was most pronounced during early extinction following a 24-hour delay, and less evident during the explicitly instructed acquisition phase. In contrast, during (early) extinction, greater CS discrimination was associated with reduced activity in limbic and prefrontal regions commonly linked to fear learning and regulation. This inverse relationship was particularly robust during early extinction, suggesting a shift in neural dynamics as learned threat responses are updated. Finally, positive associations between CS discrimination and activity in striatal regions (including caudate, putamen, pallidum, and nucleus accumbens) emerged selectively for US-expectancy ratings during early extinction, indicating a potential role of striatal systems in expectancy-based updating processes.

Together, these findings highlight phase-dependent and measure-specific relationships between neural and peripheral indices of fear learning, underscoring the complexity of cross-system integration in human threat conditioning.

The reaction time-based Concealed Information Test (RT-CIT), a measure designed to detect concealed knowledge, has demonstrated strong validity in both laboratory and forensic settings. However, the potential use of countermeasures remains a concern, motivating the implementation of response deadlines and the inclusion of additional physiological measures. In the present study, we report data from two preregistered experiments examining the effects of response deadlines on behavioral and autonomic responses in an adapted version of the RT-CIT. Participants (N = 81 and N = 42) memorized details of an imagined criminal act illustrated with photographs. They subsequently completed an RT-CIT with short (1 s) and long (2 s) response deadlines, presented in a counterbalanced order, while pupil size was recorded. During the task, participants were instructed to deceptively deny recognition of previously memorized crime-related details (probes), while truthfully responding to designated target items and unfamiliar but equally plausible neutral details. Consistent with prior findings, probes elicited slower responses and higher error rates than neutral items. A comparable pattern was observed in pupil responses, with greater dilation for probes than neutral items, independent of physical stimulus characteristics. For both measures, individual effect sizes varied as a function of task order, with stronger effects observed in the condition presented first, particularly when participants began with the short response deadline. These findings indicate that pupillometric responses closely parallel behavioral effects, supporting their utility as a reliable indicator of concealed information, even under the rapid stimulus presentation typically used in RT-CIT applications.

States of systemic inflammation are accompanied by a wide range of sickness symptoms, including pain, nausea, and fatigue, with implications for health-related quality-of-life. While sickness symptoms originate from immune signaling, their subjective experience may depend on how individuals perceive and interpret internal bodily signals – an ability referred to as interoception. However, it remains unclear to what extend individual differences in interoceptive ability shape the perception of inflammation-induced symptoms. We herein implemented the translational model of experimental endotoxemia to induce a transient inflammatory response along with flu-like sickness symptoms in healthy volunteers.

We analyzed data from N=50 healthy volunteers who participated in a preregistered randomized controlled experimental endotoxemia study. All were intravenously injected with low-dose lipopolysaccharide (LPS; 0.4 or 0.8 ng/kg body weight). Objective physiological markers and self-reported bodily and affective symptoms were repeatedly assessed during the study day. Interoceptive ability was measured using validated questionnaires.

Our results indicated interoception as significant predictor of both self-reported bodily and affective symptoms. Gender significantly affected self-reported symptoms, with more pronounced symptoms in female participants. Exploratory moderation analysis revealed that interoceptive ability significantly moderated the relationship between objective physiological markers and self-reported symptoms.

Thus, interoceptive ability shapes the perception of inflammation-induced bodily and affective symptoms. Individuals with greater interoceptive ability exhibited a stronger coupling between objective physiological markers and self-reported symptoms, suggesting that interoception may determine how immune signals are translated into conscious symptom experience. Interoception emerged as a key mechanism contributing to individual differences in symptom perception, independent of gender and LPS-dose.

Individuals with body image disturbances often experience severe stress when confronted with images of their own bodies. Body exposure (BE) is a common therapeutic intervention aimed at improving body image and psychological well-being. It is unclear at this point, however, whether BE interventions may also modulate the physiological response to psychosocial stress as another potential benefit of BE.

In the ongoing study, N = 59 body-dissatisfied women (meanage=22.4; dissatisfaction measured continuously using the Eating-Disorder-Examination Questionnaire) were randomized to either three sessions of therapist-guided BE focusing on positively (n=21) or negatively (n=20) evaluated body parts. Three sessions of Progressive Muscle Relaxation (n=18) served as control group. Subsequently, regardless of the prior intervention group, all participants underwent the Trier Social Stress Test with (n=29) or without (n=30) a concurrent BE through a mirror. We assessed heart rate, heart rate variability, salivary cortisol throughout the sessions, complemented with questionnaire assessment.

We hypothesized that the valence (positive, negative) of BE as well as the concurrent BE during the TSST influences the subjective and physiological stress response. Preliminary results show a significant increase in subjective stress (F=59.58, p<.001) and cortisol levels (F=5.60, p=.045) in response to the TSST regardless of intervention and exposure types with no significant differences in cortisol responses across intervention and exposure types. Full results will be presented and discussed at the conference. Our findings might provide insights into the effectiveness of BE in reducing body-image related stress.

The environment is full of changes that our visual system continuously monitors, yet only some of these changes reach our awareness. The differences between conscious and unconscious mismatch responses remain insufficiently explored, particularly in neuroimaging. Furthermore, previous research often confounds brain activity related to awareness with that of task relevance and reporting, resulting in a potential overestimation of effects. Therefore, in the current study, we used a modified version of an inattentional blindness (IB) paradigm, combined with delayed reporting, to examine the effects of awareness on mismatch processing while controlling for the aforementioned confounds. In detail, participants viewed roving oddball sequences consisting of vertical and horizontal rectangles while completing a demanding foreground task. BOLD responses were obtained in two physically identical phases that differed only with respect to the participants' awareness: 1) unawareness of the oddball stimuli through IB, and 2) awareness of the oddball stimuli through informing the participants about them. We observed a significant main effect of mismatch in sensory areas, specifically the left lateral occipital cortex, but no interaction with awareness. No further areas indicated mismatch processing or a modulation by awareness. Therefore, awareness does not seem to impact visual mismatch processing when task relevance is controlled. We also investigated the underlying mechanisms of mismatch responses using neurocomputational modeling. Model comparison provided strong evidence in favor of binary precision-weighted prediction error (pwPE) over adaptation, and three control models.

Clinician-assigned diagnostic categories and patient-reported symptom patterns do not always align neatly in routine mental healthcare. The present study examines whether symptom-based patient profiles can be identified in a transdiagnostic internalizing sample, how these profiles relate to therapist-assigned ICD-10 diagnoses, and whether they contribute to the explanation of psychotherapy outcome.

Routine outpatient psychotherapy data will be analyzed, including therapist-assigned ICD-10 diagnoses and symptom questionnaire data collected in clinical practice. Patients with depression- and anxiety-related diagnoses will be included. Baseline symptom profiles will be derived using clustering approaches applied to symptom dimensions. The analyses will quantify the degree of correspondence between symptom-derived profiles and ICD diagnoses, including overlap, concentration, mismatch, and cross-diagnostic spread. In addition, the analyses will examine whether identified profiles primarily reflect overall symptom severity or whether they capture more distinct symptom configurations. Finally, the association between symptom profile membership and treatment outcome will be tested, with particular focus on whether symptom profiles account for outcome variance beyond diagnosis and baseline severity.

This study addresses a clinically relevant question in routine psychotherapy research: whether symptom-based profiling provides information beyond standard diagnosis and helps clarify heterogeneity in internalizing disorders. The findings are intended to inform ongoing discussions about diagnostic alignment, symptom-based case characterization, and the clinical usefulness of routine self-report data.

Single-trial EEG analyses are increasingly used to derive measures of neural activity that preserve trial-level variability and support analyses of individual differences. However, it is critical to determine how much of the variance reflects within-person differences across trials and whether the amount of within-person fluctuation across trials differs across data analytic choices. In this study, we assessed the amount of variability in single-trial N170 and Late Positive Potential (LPP) estimates across a defensible multiverse of quantification and parameterization decisions. Using data from an emotion classification task (N=98, happiness condition), we implemented 18 pipelines: two baseline durations (-200ms, -100ms), three reference schemes (common average (CAV), linked mastoids, and current source density (CSD)), and three parameterization strategies (conventional peak-picking, ICA-based peak-picking, and Residue Iteration Decomposition (RIDE)) to estimate single-trial N170 and LPP amplitude and latencies. Within-person variance was quantified as the intraclass correlation coefficients (ICCs) derived from linear mixed-effects models. Across pipelines, ICCs were generally low, particularly for latency and the LPP component, indicating that most variance occurs at the within-person level rather than between individuals. ICA-based peak-picking with CAV yielded the highest ICCs for N170 amplitude, while CSD-based pipelines reduced dependability. These findings demonstrate that conclusions about single-trial EEG variability may be influenced by quantification and parameterization decisions, and component characteristics. A comprehensive evaluation of the robustness of variance decomposition across component quantification and parameterization choices can provide guidance for future within-person reliability analyses, and aid the interpretation of future between- and within-person related research questions based on single-trial EEG analyses.

Learning about others' personality traits is crucial for building personal relationships and guiding future interactions. This social learning process can be conceptualized as a reduction in prediction errors (PEs), defined as the difference between expectations about a person and subsequent feedback. Pupil dilation has been established as a marker of various non-social learning processes, but its role in abstract social learning remains unclear.

In our study, participants performed a feedback-based learning task in which they learned about the personality traits of several target individuals ("profiles") based on trial-by-trial feedback, while eye movements and pupil size were recorded. On each trial, participants estimated how a given person would rate themself on a positive trait item and received that person's actual self-rating as feedback, allowing trial-wise PE computation. One profile was, unbeknownst to participants, constructed to match their own personality ratings provided at the beginning of the experiment, creating a self-similar profile.

Behaviorally, linear mixed-effects models (LMMs) revealed a significant decrease in absolute PEs across trials, indicating successful trait learning. Absolute PEs were significantly smaller for the self-similar profile than for other profiles, whereas the slope of PE reduction did not differ reliably between conditions. Pupillometrically, we assessed whether pupil dilation scaled with PE magnitude in the feedback phase. LMM analyses showed that larger absolute PEs were associated with stronger pupil dilation on a trial‑by‑trial basis.

These findings suggest that pupil dilation parametrically tracks prediction error magnitude during social learning and may serve as a physiological index of personality-trait learning about others.

Infants’ active manipulation of objects in the mouth has been historically framed as incidental to teething (Iverson, 2010). However, developmental research shows it has a functional role in heart rate regulation (Dur & Gözen, 2021) and stimulates the orofacial articulators, encouraging vocal production (Fagan & Iverson, 2007). Thus, mouthing lies at the intersection of multiple interacting brain-body systems and may be related to developing cognitive abilities. We sought to determine its potential mechanistic role in coordinating these systems.

We analyzed two unique datasets of caregiver-infant play (Dataset 1: N = 30; Dataset 2: N = 5; ages 9–15 months), that included behavioral coding of mouthing events, audiovisual data, and eye gaze data from head-mounted eye trackers. In Dataset 2, infant heart rate was also measured.

In Dataset 1, infants were more likely to vocalize during or shortly after mouthing than at baseline (14.0% vs. 8.3% probability) and to gaze at caregivers’ faces (18.6% vs. 12.2%). Ongoing analyses from Dataset 2 show similar patterns (24% probability of post-mouthing vocalization, n = 4; 44% probability of viewing caregiver face during mouthing, n = 1). Most notably, heart rate data from 3 infants in Dataset 2 revealed slower heart rate during mouthing (M = 128 bpm) compared to baseline (M = 131 bpm).

For infants, mouthing behavior supports social attention, vocal production, and sustained visual attention. We propose mouthing to be an organizing behavior for the developing body that modulates developing cognitive abilities across multiple domains by lowering heart rate.

Genetic variations in the oxytocin and dopamine systems are known to influence social processes, thus an influence on empathy is conceivable. Two frequently examined single nucleotide polymorphisms, the oxytocin receptor (OXTR) gene polymorphism (rs53576) and the COMT-gen polymorphism (rs4680), might influence dispositional empathy in humans, but there are many inconsistent results. A gen x gen interaction could explain heterogeneous results.

Between 2022 and 2025 N = 487 students, 386 female and 101 male, between the age of 18 and 48, participated in the study. Genetic analyses were performed using buccal cell swabs. Empathy was measured by the Saarbrücker Persönlichkeitsfragebogen, which consists of four subscales: perspective taking (PT), fantasy (FS), empathic concern (EC) and personal distress (PD). Gender will be considered as a covariate.

An ANOVA revealed no effect on the total empathy score (sum of PT, FS and EC), so a MANOVA was carried out to analyze potential effects on the subscales separately. A significant gen x gen interaction effect was found for the EC-subscale. Carrier of the AA-genotype (rs53576) showed increased EC-scores, depending on the number of copies of met-alleles (rs4680), whereas carriers of the G-allele were not affected by COMT genotypes at all. No such effect was found for any other subscale.

The main limitations of the study were the imbalanced and in some cases small group sizes due to rare genotype combinations. Potential explanations are discussed.

A novel approach to explore the neurocognitive underpinnings of lying is to investigate the mental processes during deception preparation, rather than its execution. In our previous EEG study (Voltz et al., 2026, NeuroImage), participants answered autobiographical questions of four semantic categories. They were instructed to lie to questions of one predefined category (25% of trials) and to otherwise tell the truth. A preceding cue indicated the category and therewith the need to lie or to be honest regarding the subsequent target question, which was not exactly known before. There, we found increased fronto-central P2, P3a, Contingent Negative Variation (CNV) and parietal P3b amplitudes along with reduced centro-parietal alpha power following lie compared to truth cues, indicating enhanced attentional and general preparatory processes prior to actual, infrequent deception. In the present study (N = 48 healthy adults), we applied a similar design but reduced the number of categories to three and presented the individual lie category twice as often as each of the two truth categories, leading to balanced frequencies of lie and truth trials (both 50%). Second, we used a jittered, more natural cue-target onset asynchrony. Now, P2 was larger after perceptually rare truth cues, while the P3a lie effect was not significant. Importantly, we conceptually replicated the lie effects on P3b, CNV and alpha power, which appear as especially robust markers of preparing deception. Thus, deception preparation might be a promising indicator for future lie detection, whereas markers of earlier attentional processing might only occur when lying is relatively rare.

Phantom limb pain (PLP) affects up to 80% of amputees and remains a major clinical challenge despite multimodal treatment approaches. Its neural mechanisms are not yet fully understood but are thought to involve alterations in cortical oscillatory activity. Evoked phantom limb pain, in which pain can be deliberately induced through phantom movements, provides a controlled paradigm to investigate the underlying neural dynamics of PLP.

Transcranial alternating current stimulation (tACS) represents a promising approach to modulate such oscillatory patterns, suggesting potential therapeutic benefits. However, substantial interindividual variability in intrinsic oscillatory frequencies may limit its efficacy, highlighting the need for individualized stimulation protocols.

In this pilot project, task-based EEG data from evoked PLP (n = 9) are analyzed to compare three methods for estimating individual peak frequencies: maximum power (peak picking), center of gravity, and Gaussian fitting. These approaches are applied to the alpha, beta, and gamma frequency bands to evaluate their suitability for deriving individualized stimulation parameters. Preliminary testing in a single participant, stimulated at 24 Hz based on the estimated individual peak frequency, showed a reduction in evoked PLP following real tACS compared to sham stimulation.

These findings aim to inform the development of robust methods for identifying individualized stimulation frequencies and guide the design of personalized tACS protocols for the modulation of disorder-related oscillatory activity in PLP.

Low-intensity transcranial focused ultrasound (TUS) as a non-invasive tool has gained popularity due to various studies in basic and applied sciences as well as related guidelines and standards. In a preregistered double-blind within-subjects study, our group used TUS on the right prefrontal cortex, which previously enhanced self-reported global positive mood, decreased negative self-reports of mental conflict (anxiety/worrying), and modulated associated midfrontal activity as reflected in functional magnetic resonance imaging. This study’s large-scale dataset allowed us to publish recently on further TUS effects on physiology and behavior (N = 152), measured during a well-established virtual T-maze task for the investigation of motivational conflicts regarding whether participants execute approach versus withdrawal while allowing to record related electroencephalographic data such as midfrontal theta (MFT). In brief, we showed that TUS significantly decreased conflict-related MFT, which significantly explained increases in approach and decreases in withdrawal. Now, we were interested whether these global effects might differ based on individual moderating variables. Our large-scale dataset offers various data for such exploratory analyses (e.g., on demographics and personality). We did not find significantly relevant moderators but detected patterns by trend that can possibly be of relevance for future work. Eventually, this might provide more personalized neuromodulation for research and applications such as regarding emotional and motivational mental health.

Alterations in fear learning and fear inhibition are thought to play a central role in the development and maintenance of anxiety- and stress-related disorders. Individual differences in these processes have been linked to anxiety-related traits (ART), e.g. trait anxiety, intolerance of uncertainty, and neuroticism. However, evidence on the relationship between ART and conditioned fear responses is heterogeneous. Cumulative progress in this field has further been limited by a fragmented literature with studies focusing on single traits, a narrow set of outcome measures or isolated experimental phases. Additionally, small sample sizes have often limited the reliable detection of individual differences. The current study investigated the effects of ART on associative differential fear conditioning (acquisition, extinction and renewal) in 277 participants. Conditioned responses were assessed using multimodal outcome measures, including fear and US expectancy ratings, skin conductance responses (SCR), fear-potentiated startle (FPS) and functional magnetic resonance imaging (fMRI). The results indicated no associations between ART and conditioned responding in SCR and FPS during any experimental phase. In contrast, associations emerged for subjective ratings during extinction and renewal. These effects were observed for responses to CS+ and CS− but not CS discrimination. Region-of-interest analyses revealed limited neural associations with reduced CS discrimination only in the dorsal anterior cingulate cortex during early renewal. Taken together, the results suggest that higher levels of ART are associated with a generally heightened subjective evaluation of threat and US expectancy and do not reflect an aberrant fear learning pattern.

There is a systematic link between aberrant cardiac responses to stress and impaired physical and mental health. According to the vagal tank theory, it is important to consider resting, reactivity, and recovery phases when examining cardiac parasympathetic responses to stress. Prior research mainly compared rest and reactivity. Thus, we examined patterns of parasympathetic activity during recovery after a psychosocial stressor.

Overall, 56 young adults (22.91±3.32 years, 64.3% female) completed a 5-minute resting baseline, a 10-minute psychosocial stressor (the Maastricht Acute Stress Test), and a 20-minute recovery period. Continuous ECG was recorded using MindWare, and heart rate variability (RMSSD) was calculated as an index of cardiac parasympathetic activity in one-minute intervals.

Most participants (n=48) showed peak RMSSD values during the recovery period, indicating parasympathetic rebound above baseline. Cluster analysis identified three recovery patterns differing in rebound magnitude. Comparisons of cluster characteristics revealed significant differences for maternal care, self-reported psychological, and physiological health (both rated on a Likert scale). Post-hoc comparisons suggested lower maternal care in the group with the greatest rebound, and poorer psychological and physiological health in the cluster with the smallest rebound.

Taken together, the current results suggest that the parasympathetic system rebounds after suppression through a stressor. Three rebound patterns were identified, with self-reported health emerging as a key predictor for rebound magnitude. These findings emphasize the importance of post-stress recovery, as it may have implications for long-term health. Further research is needed to clarify the factors contributing to individual differences in rebound capacity.

Age-related cognitive decline in individuals over 60 years and patients diagnosed with mild cognitive impairment (MCI) is consistently associated with deficits in lure discrimination, a process dependent on hippocampal pattern separation (PS. Here, the Mnemonic Similarity Task (MST) can be used to directly target the hippocampal Pattern Separation (PS). Several studies found a positive impact of targeted memory reactivation (TMR) on PS. However, limits in subject selection and testing duration prompted us to conduct a modified MST to research whether TMR can influence PS on a longer scale in cognitively healthy and impaired individuals.

Participants complete two consecutive encoding sessions in which 180 object stimuli are categorized semantically, followed by an initial recognition probe requiring lure discrimination responses ("old", "similar", or "new"). Over five consecutive nights, 50% of encoded items are aurally reactivated during verified N2 and N3 sleep stages, facilitated by real-time automated sleep scoring via a consumer-grade EEG headband. Recall assessments are administered following the first and final TMR sessions to capture longitudinal performance trajectories. Sleep macroarchitecture and microstructure are characterized via full 32-channel polysomnography during a baseline night, the first TMR night, and a post-intervention follow-up night.

The primary outcome measure is differential discrimination performance for reactivated versus non-reactivated items across the three recall phases. As a prospect, the setup could be used as a form of at-home applicable, non-invasive symptomatic treatment in the early stages of dementia.

Every time we learn a new skill or store an experience in memory, our brain undergoes functional and structural changes. Advances in neuroimaging, such as diffusion-weighted MRI (DW-MRI), allow to non-invasively track microstructural changes by analyzing the motion of water molecules within human brain tissue. Using these methods, recent studies have demonstrated that repeated encoding and retrieval of an object–location association task leads to the rapid formation of a physical memory trace in the parietal cortex. Notably, this trace fulfills all criteria for a memory engram within just 90 minutes post-learning: active during encoding, engendering a change in the underlying substrate, and active during retrieval. Whether memory engrams differ depending on learned information remains an unanswered question.

In this study, 79 participants completed an object-location learning task with repeated encoding and retrieval of image pairs depicting either animate objects (n=20), inanimate objects (n=20), indoor scenes (n=19), or outdoor scenes (n=20). DW-MRI was measured immediately after the learning task until up to 60 minutes later. We identified regions that fulfill all engram criteria and additionally allow for decoding of learned material from microstructural change patterns. We found material-specific engrams in the early and higher-order visual areas, posterior parietal cortex, and the cerebellum. Our findings suggest that memory engrams in humans are distributed across multiple interacting brain systems, depending on the specific learning content.

Electronic gambling machines (EGMs) are a popular but harmful form of gambling (Allami et al., 2021; Peters, 2025). While playing EGMs, participants show increased reaction times (RTs) after positive reinforcement (Delabbro & Winefield, 1999; Murch et al., 2024). This effect, called the post-reinforcement pause (PRP) has been observed for several decades (Ferster & Skinner, 1957). However, the underlying process are still poorly understood, especially in EGMs. Computational modelling provides a powerful tool for the identification and decomposition of such latent decision-making components.

In the current study, we reanalysed the behavioural data of 46 participants playing on a EGM for up to 200 trials. (Paliwal et al., 2014). Using hierarchical Bayesian modelling, we fit and compared multiple drift-diffusion models (DDMs) on the data. Each model contains covariates for the type of outcome of the previous trial (true loss, near win, loss-disguised-as-win, true win) for one of the different DDM parameters: non-decision time (τ), boundary separation (α), initial starting point (β), and the drift-rate (δ). Initial results indicated that the α-model provided the best fit. This suggests that the PRP might be related to increased response caution (Verbruggen et al., 2017). We discuss these findings in light of previous research efforts and their meaning with regard to theoretical considerations of the PRP.

Die präzise Unterscheidung zwischen Links und Rechts ist eine kritische kognitive Funktion. Fehler können im OP zu sogenannten „Never Events“ führen, vermeidbaren Behandlungsfehlern mit schweren Folgen für Patienten. Da gängige Tests zur Links-Rechts-Verwechslung (LRV) die komplexen Anforderungen chirurgischer Umgebungen kaum abbilden, führt diese Studie den Never Events in Surgery Test (NEST) ein: ein ökologisch valides Verhaltensparadigma zur Messung von LRV im angewendeten Kontext.

Teilnehmer (N = 108) absolvierten den NEST sowie etablierte Verfahren (BLRDT, LRCT, LRQ) zur Messung von LRV. Die Ergebnisse zeigen, dass die Leistung im NEST signifikant von der Lagerung der Übungspuppe abhing: Während die Fehlerraten in Bauchlage am niedrigsten waren (p < .05), bot die Rückenlage die kürzesten Reaktionszeiten (p < .001), ohne die Fehlerhäufigkeit signifikant zu erhöhen.

Zentral ist die nachgewiesene kognitive Dissoziation: Zwar korrelierten die Reaktionszeiten aufgabenübergreifend (als Ausdruck allgemeiner Verarbeitungsgeschwindigkeit), doch die Fehlerraten der klassischen Tests konnten die Fehler im NEST-Paradigma nicht vorhersagen. Dies legt nahe, dass die LRV-Anfälligkeit in angewandten Kontexten ein unabhängiger, kontextspezifischer Risikofaktor ist, der von traditionellen Assessments nicht erfasst wird. Die Befunde fordern einen Paradigmenwechsel hin zu maßgeschneiderten, kontextspezifischen Ansätzen in Training und Personalauswahl, um die Patientensicherheit nachhaltig zu erhöhen.

Goal directed action planning substantially develops around the preschool years, but the neural mechanisms underlying children’s action processing are largely unknown. Previous research in adults suggests that action processing recruits different subregions of left inferior frontal cortex (IFC) (Papitto et al., 2020). Here, we investigate children’s action planning employing an End-state comfort task (ESC) while measuring fNIRS. The ESC refers to the tendency of avoiding an uncomfortable hand position at the end of a goal-directed movement. For example, starting an action (grasping a cup) in an uncomfortable position and finishing a second action (rotating the cup) in a comfortable position (Rosenbaum et al., 1990). N=59 children placed a cup from point A to point B while either rotating it or not. The cup’s opening faced either upwards or downwards. Participants were presented with a total of 20 trials, four in each condition. We measured neural responses using fNIRS in a 15 second planning phase, where the cup was presented but out of reach, followed by an execution phase, in which the cup was moved. Based on adult data, we hypothesized that action planning abilities modulate the activation of regions in left IFC, with stronger activation for rotation compared to non-rotation. Preliminary results confirm an effect of condition modulated by behavioral performance: for children who show ESC compliant behavior, we find stronger activation for rotation than no-rotation conditions in the right IFC. The data will be discussed against the backdrop of developmental changes in action planning abilities and its neural correlates.

Hierarchical predictive processing (HPP) conceptualizes perception as inference: priors are tested against sensory evidence across hierarchical levels to minimize uncertainty. In computational psychiatry, aberrant precision-weighting of priors has been implicated in psychosis, yet studies isolating single prior attributes have yielded inconclusive neurobehavioral mechanisms. Recent work suggests that aperiodic, rather than periodic, activity may mediate active inference in hierarchical predictive coding (HPC). Here we probed neurobehavioral correlates of predictive hierarchies and their modulation by psychosis.

We recorded EEG from 43 neurotypical individuals performing a random-dot kinematogram task under three prior conditions. Low-level priors were induced via implicit learning of tone–motion associations; high-level priors via explicit, temporally stable false beliefs in motion-altering glasses; baseline trials contained no directional predictions. Psychosis proneness was assessed by questionnaires.

Drift diffusion modelling indicated hierarchical priors primarily modulated drift rates, rather than starting points, of Bayesian evidence accumulation. Contrary to our hypothesis, cluster-based permutation tests of occipital time–frequency representations revealed no reliable modulation of oscillatory power. Exploratory spectral parameterization showed that aperiodic, but not periodic, EEG components were modulated by priors. We observed descriptive evidence for an anti-proportional modulation of hierarchical priors by psychosis proneness.

Modelling analysis suggests that hierarchical priors impose dynamic, rather than static biases, consistent with precision-weighted integration of priors and sensory evidence. Selective modulation of aperiodic activity implicates excitatory-inhibitory balance in active inference, aligning with dendritic HPC models. Testing whether overly precise high-level priors compensate for underweighted low-level priors in psychosis pathogenesis will require future work.

Salivary sampling is an accessible and validated method for assessing how hormones respond to experimental manipulation. However, steroidal hormones like testosterone are released in a pulsatile fashion, which can confound measurement accuracy. In this study, we tested whether pulses of testosterone can be detected in 10 salivary samples collected across a 90-minute period, reflecting a typical experimental session procedure. Following detrending of testosterone values across participants (women = 42, men = 23), candidate pulses were identified as positive residuals that met three criteria: they were local maxima, exceeded expected assay measurement error (2 × CV%), and surpassed individual-level variability (2 × MAD). Approximately half of the sample (n = 31) exhibited one testosterone pulse during the session, including 13 of 23 men and 18 of 42 women. These findings highlight the importance of accounting for pulsatile hormone dynamics in salivary research, and implications for methods of smoothing and detecting such pulses are discussed.

Emotional news influence our judgements, even when their veracity is questionable. Previous studies have demonstrated that faces paired with affective information in the form of news headlines are evaluated similarly regardless of source credibility, showing the dominant role of emotion in person perception. It remains unclear whether this bias is already present at the level of encoding - whether emotional news headlines preferentially capture visual attention regardless of credibility.

The current pre-registered experiment investigated the distribution of visual attention between news contents that vary in valence and in the credibility of their source. Thirty participants viewed stimuli displaying various news stories at the same time, while their eye movements were recorded. Specifically, headlines differed in social affective information about depicted persons (positive, negative, neutral) and source credibility (high, low). Participants freely explored the news pages and subsequently completed a memory test and provided explicit social judgements about the persons.

Eye-tracking data confirmed that participants noticed the different sources. Results showed that negative emotional headlines were preferentially attended to regardless of credibility. Negative and positive information enhanced face memory and social judgments were strongly influenced by the headline content, independent of the credibility of the source. These findings indicate that emotional content biases information processing from the earliest stages of attentional selection, with cascading effects on memory and person evaluation. Critically, source credibility does not seem to modulate this effect at any of the stages explored, highlighting a potential mechanism underlying the persistence of social-emotional misinformation effects.

Social interactions and social support can have a buffering effect on anxiety, stress, and related physiological responses. However, social support is a multidimensional construct, with different types of support (informational, emotional, esteem, instrumental, and negative) potentially leading to distinct effects on outcome variables like anxiety and mood. To overcome these limitations, this study investigated the effects of social support types on state social anxiety during social interactions as well as general and depressive mood directly after social interactions in everyday life of healthy participants from the general population. Using smartphone-based Ecological Momentary Assessment (EMA), participants reported on social interaction characteristics, interaction partner and received support, as well as their anxiety and mood levels based on six survey prompts per day over a period of five consecutive days. Additionally, during a 12-hour continuous daily ECG recording, heart rate and heart rate variability were measured using mobile ECG sensors worn around the chest. The final sample consisted of N = 93 participants (Mage = 25.45, 61 female) with 1545 observations of social interactions. Preliminary behavioral results indicate that while emotional and esteem support had a positive effect on general mood, with greater informational and negative support participants experienced higher depressive mood after social interactions. Similarly, esteem support was associated with lower anxiety, while higher informational and negative support were linked to increased anxiety during social interactions. This study highlights the benefits of EMA research and characterizes the types of everyday-life social support that can mitigate or aggravate prevalent symptoms of anxiety and depression.

Decisions can be made within domains, where options share similar attributes, and across domains, where options differ fundamentally. Current theories of value-based decisions, particularly the common-currency theory, posit that the brain translates the various attributes of different options into a common value, enabling a standardized value-comparison mechanism. We challenge this assumption by proposing that the need to compute integrated values depends on the type of decisions. First, we assume that overt information gathering relies on distinct search strategies in within- compared to across-domain decisions (specifically, the frequency of option-wise comparisons is higher in the latter). Furthermore, we hypothesize that integrated values in value-coding areas (especially ventromedial prefrontal cortex, vmPFC) are represented differently in within- and across-domain choices.

To test this hypothesis, we developed a value-based, multi-attribute decision-making task that included both types of choices. N = 50 healthy participants were instructed to complete the task while their fMRI signals and gaze data were recorded. As predicted, participants showed more attribute-wise searches (p<<0.01) in within-domain choices and more option-wise searches in across-domain choices (p<<0.01). However, overall behavioral performance did not differ significantly between decision types (p=0.08). We demonstrated that vmPFC reflects the subjective value difference between options only in within-domain choices and to a significantly higher extent than in across-domain choices (WD: p<0.05, and AD: p=0.9). Together, these findings suggest that people employ distinct search and choice strategies depending on the decision type, and that an integrated subjective value representation of each option may not be necessary for all types of decisions.

It has been suggested that action representation relies on a basic structuring level according to which the action is organized (Bauer et al., 1998). Here, we ask whether violations in action structure have differential effects on action processing depending on an action’s internal structure.

In an EEG study, we presented N=35 adults and N=45 preschoolers with two types of daily actions: In sequentially structured actions, sub-units are relatively flexible in order, i.e., one action step is completed before the next action step starts but the order of action steps is variable (e.g., mixing an apple spritzer). Dependent structures refer to actions that are fixed in order as the execution of one action step directly depends on the completion of another step (e.g., filling a glass with water that needs to be turned first). For both types of actions, we systematically manipulated their structure by switching the order of adjacent action steps within the sequence.

Results suggest that order manipulations have more pronounced effects for dependent compared to sequential structures with a late (600-850ms) positivity effect over anterior regions in adults – partially mirroring previous findings on structural action manipulations (Maffongelli et al., 2015; 2018) - whereas effects in children are found in posterior regions. As the positivity effect has been interpreted as a marker of structural complexity more generally (Kaan, 2007), selective modulation in the dependent condition might indicate differences in the structural representation between constrained and unconstrained action sequences with different neural circuits underlying processing in children and adults.

Modulation of alpha- and beta-frequency oscillations recorded over sensory cortices has been linked to the top-down weighting of sensory information relative to competing information from other modalities. Here, we tested whether alpha-/beta-tACS over the occipital cortex would improve action performance in a virtual reality-based hand-target matching task involving visuo-proprioceptive conflict, where conflicting visual movement feedback needed to be ignored in favour of proprioception. Participants had to match an auditory target rhythm with either their unseen real hand (“RH”) grasping movements or the movements of a virtual hand (“VH”) that displayed their actual movements after a constant time delay. Consequently, visual movement feedback was either task-relevant (VH task) or a distractor (RH task). Using a cross-over, double-blind, within-subject design, we applied low-intensity (2 mA peak-to-peak) transcranial alternating current stimulation (tACS) at either 10 Hz (“alpha”) or 20 Hz (“beta”) to the occipital cortex, or a sham stimulation, during this task. Neither alpha- nor beta-tACS significantly affected hand-target matching performance; however, fixation variability in the RH condition decreased significantly during beta-tACS relative to sham stimulation. Therefore, despite not directly modulating manual performance, occipital beta-tACS may have facilitated the ignoring of mismatching visual movement feedback when necessary.

Collaboration increasingly takes place in digitally mediated settings such as screen-based remote work. This raises critical questions about how remote interaction, relative to co-presence, shapes shared task representations and neurophysiological alignment (i.e., inter-brain synchrony) during collaborative problem solving (CPS). We present an ongoing multimodal hyperscanning study with an intermediate sample of 27 dyads (54 participants) examining how interaction context influences interpersonal alignment across behavioral, physiological, and neural levels. For the brainstorming task analyses reported here, complete subjective data were available for 24 dyads (48 participants; M_age=28.5±10.0; 23 women, 25 men; mean relationship closeness = 5.98±1.52 on a scale from 0 to 7). Dyads completed two collaborative tasks under both co-present and remote conditions in a counterbalanced within-subjects design: a cooperative maze task requiring time-critical coordination without verbal communication, and a structured verbal brainstorming task based on the Object Characteristics Task and Alternative Uses Task (OCT/AUT), targeting collaborative creative thinking. Simultaneous electroencephalography, functional near-infrared spectroscopy, eye tracking, and peripheral physiology were recorded alongside behavioral and subjective measures. Preliminary analyses showed no significant subjective differences between interaction contexts for the maze task. In the brainstorming task, however, remote interaction was associated with lower affective valence during the AUT task phase (W = 297, p_FDR = .042, r_rb = .40) and lower perceived cooperation across the OCT/AUT tasks (W = 166, p_FDR = .010, r_rb = .53). These preliminary findings suggest that digitally mediated interaction may particularly affect the subjective experience of verbal creative collaboration, while leaving tightly constrained nonverbal coordination comparatively unchanged.

Chronic pain is linked to alterations in resting-state EEG activity, particularly increased power in low-frequency bands, proposed as neurophysiological markers of neuropathic pain. However, it remains unclear how these alterations manifest during stimulus processing under varying affective contexts. We investigated neural responses to thermal stimulation combined with olfactory stimuli in healthy controls (HC, n = 22) and idiopathic small fiber neuropathy (SFN, n = 26) patients. Individually calibrated warm and painful stimuli were combined with pleasant, unpleasant, or neutral odors in a 2 x 2 x 3 design. EEG (64-channel) was recorded simultaneously with fMRI (reported elsewhere), and time-frequency power (1-30 Hz) was analyzed using four separate linear mixed models (delta, theta, alpha, beta). Significant main effects of odor were observed across all frequency bands (p ≤ .005). A main effect of temperature emerged in the alpha band (F = 103.59, p < .001), while no effect of group was found. Group x temperature interactions were present across all bands (p < .001), alongside significant three-way interactions (p < .001). Post-hoc analyses showed that in SFN patients, painful versus warm stimulation increased alpha power (z = 9.21, p < .001). HC showed the opposite pattern, with decreased alpha power. Under painful heat, pleasant odors elicited stronger responses than unpleasant odors in both groups, more pronounced in SFN patients (alpha: z = 13.19 vs. 6.91, p < .001). Our findings suggest altered thalamocortical dynamics and enhanced affective pain modulation in SFN, indicating the role of olfaction in non-pharmacological pain modulation warrants further investigation.

Understanding the neural basis of individual differences in cognition and lifestyle is a central goal of brain–behavior research. However, M/EEG findings on associations between neural activity and cognitive abilities are mixed (e.g., Fruehlinger et al. 2025, Ociepka et al. 2023).

In the present preregistered study, we tested whether resting-state MEG features predict fluid and crystallized intelligence, as well as lifestyle variables including mental and physical health, social and intellectual engagement, and socioeconomic status, using openly available data from 350 healthy adults aged 18–88 years. We extracted conventional spectral MEG features and approximately 6000 time-series descriptors and used spatially informed partial least squares regression with cross-validation to predict behavioral outcomes.

Conventional MEG features showed moderate accuracies for cognitive and lifestyle variables, but predictive performance was poor, with predicted R² values close to or below zero. Aperiodic-adjusted alpha power was among the strongest predictors for fluid (r = .43, R² = -.09) and crystallized intelligence (r = .22, R² = -.38), and physical health (r = .27, R² = -.35). A few time-series metrics (e.g., scaling measures such as the Hurst exponent) reached generally higher accuracies, for example, with fluid intelligence (r = 0.53, R² = 0.15), yet explained only minimal variance. Moreover, partial confound analyses suggest that model predictions of fluid intelligence and physical health were influenced by age-related variance.

Together, these findings suggest that the resting-state MEG markers examined here provide limited predictive value for cognition and lifestyle in healthy adults. Future work should evaluate nonlinear approaches and more explicitly model potential confounds.

This preregistered study investigated the influence of visuotactile co-stimulation on adaptive action under visuoproprioceptive conflict. We used a virtual reality based, recurrent hand-target phase matching task in which participants had to align the grasping rhythm of either their real hand movements, occluded from view, or the movements of a virtual hand presented on screen (controlled via a data glove, with an added delay) with a target oscillation (auditory sine). Participants were instructed to align either the real or the virtual hand, with instructions alternating at regular intervals. We applied brief, non-painful electrical impulses to the real hand, together with a virtual rod touching the virtual hand on screen. The study used a nested 2x3 within-subject design: The first factor was which hand (real or virtual) had to be aligned with the auditory rhythm. The second factor was the type of stimulation, which consisted of visual and tactile stimuli in three combinations: synchronous stimulation, asynchronous stimulation and no (tactile) stimulation.

Based on previous work suggesting that “embodiment” of a fake hand via visuotactile co-stimulation helps bind discrepant visual and proprioceptive hand position information together (in passive scenarios), we hypothesized that synchronous co-stimulation of the virtual and real hands would improve performance in the virtual hand condition, whereas asynchronous stimulation would improve performance in the real hand condition. Furthermore, we expected visuotactile stimulation to affect how efficiently participants would switch between instructed tracking modalities.

Data collection has been completed, analysis is ongoing; the final results will be presented in the poster.

Apathy is a core symptom of schizophrenia, which often precedes onset and serves as a predictor of everyday functioning. Targeted treatment is limited due to an incomplete understanding of the underlying behavioral and neurocognitive mechanisms. Effort-based decision making (EBDM) paradigms may be used to investigate these mechanisms and alterations across the schizophrenia spectrum.

In this ongoing study, we therefore implemented an event-related fMRI version of an accept-reject EBDM task. In this task, participants were asked to perform fast-paced button presses (i.e. efforts) in return for different rewards. The design featured 16 combinations of effort (30%, 50%, 70%, 90% of individual maximum button presses) and monetary reward (€0.05, €0.10, €0.15, €0,20). We recruited healthy subjects (n = 10; target: N = 60) and assessed schizotypy (SPQ) and apathy (AES) traits. We used a general linear model to identify neural correlates of choice and effort anticipation. To estimate performance parameters, we applied a drift diffusion model (DDM).

Preliminary results show that individuals with higher SPQ scores exhibit lower acceptance rates for high-effort choices, regardless of reward amount. Apathy scores correlate negatively with acceptance rate and positively with SPQ scores. fMRI data reveal that SPQ and AES correlated negatively with BOLD response in nucleus accumbens. Our DDM results support altered effort-reward processing in individuals with higher schizotypal traits.

Taken together these preliminary findings suggest that elevated schizotypal traits and apathy are associated with a reduced willingness to exert effort for reward, potentially reflecting blunted motivational signaling in the nucleus accumbens and altered effort–reward integration mechanisms.

Background:

Adolescence is a period of rapid neurodevelopment during which psychiatric symptoms may emerge. However, symptom-specific markers show inconsistent associations with cognition and brain structure and rarely generalise longitudinally.

Methods:

Using data from the Adolescent Brain Cognitive Development study, we derived a transdiagnostic mental load measure integrating multiple symptom domains. For each time-point separately, we employed Principal Component Analysis followed by Gaussian Mixture Modelling to identify four data-driven mental health burden groups. We defined a stable subgroup with consistent group membership across Years 1–3. We examined differences in cognition and structural brain measures at Year 2, with baseline and Year 4 included as control timepoints. To assess robustness, we additionally performed cross-sectional analyses defining groups independently at each timepoint.

Results:

Stable burden groups showed clear and consistent separation across all mental health variables at all time-points. The High mental health burden group showed the greatest elevations in psychotic-like experiences (PLEs), distressing PLEs, depression, anxiety, mania, suicidality, and adverse experiences. Higher mental health burden was associated with lower performance in vocabulary, memory, and processing-speed, alongside widespread reductions in whole-brain, cortical, and white-matter volumes at baseline and Year 2. These effects were strongest in adolescents with persistently high mental health burden and were replicated in cross-sectional analyses. At Year 4, brain-behaviour associations were attenuated, likely reflecting developmental reorganisation and reduced sample size.

Conclusions:

Our study demonstrates that a transdiagnostic measure of global mental health burden provides a scalable, developmentally sensitive marker of early psychiatric vulnerability, overcoming limitations of symptom-specific approaches.

Background. Obsessive-compulsive disorder (OCD) is a neuropsychiatric disorder with symptoms of unwanted repetitive thoughts and behaviors that affects 1-3 % of children and adolescents. Theories suggest that OCD is a neurodevelopmental disorder with early risk factors such as altered brain development leading to OCD pathophysiology. In this regard, studies started to investigate structural brain alterations in prefrontal, orbitofrontal, and cingulate regions in brain measures such as volume, or later cortical thickness and surface area, highlighting the cortico-striato-thalamo-cortical circuitry as central to symptom manifestation. Specifically, altered gyrification might reflect early neurodevelopmental cortical deficits. To date, previous studies on gyrification reported differing results and found de- or increased frontal gyrification, including the orbitofrontal cortex (OFC) and only focused on adult OCD. However, investigating adolescent samples is important to gain insights into the neurodevelopmental origins of OCD. Therefore, our aim was to investigate the gyrification of adolescents with OCD vs. typically developing adolescents focusing on the OFC.

Methods: Using structural magnetic resonance imaging, we assessed n=67 adolescents with OCD (15.64 ± 2.10 years) and n=93 sex- and age-matched TD adolescents (15.39 ± 2.32 years). We performed quality control and analyzed the local gyrification index (lGI) using Freesurfer and OFC ROIs of the Desikan-Killiany-atlas.

Results: Preliminary analyses revealed smaller lGI related to OCD symptomatology in the OFC, implicated in the cortico-striato-thalamo-cortical circuitry.

Conclusion: Results highlight early neurodevelopmental risk factors for OCD symptomatology in cortical gyrification.

Social transmission of stress refers to physiological and subjective stress responses elicited by observing another individual undergoing socio-evaluative stress (Engert et al., 2018). Although growing evidence suggests that stress responses can be transmitted between individuals, the literature is characterized by substantial heterogeneity in study populations, paradigms, and measurement approaches, limiting comparability across studies (Mahreinecke et al., 2025). As no quantitative synthesis of effect sizes has been conducted, this systematic review and meta-analysis aimed to synthesize evidence in adult observers and examine determinants of social stress transmission.

A comprehensive literature search was conducted across PsychINFO, PubMed, Embase, and Web of Science databases, using search terms related to stress transmission and stress biomarkers. Studies were included if they examined vicarious stress and/or stress resonance in adult observers of the Trier Social Stress Test (TSST) or comparable paradigms. Fourteen studies met the inclusion criteria for the systematic review, and ten were eligible for meta-analysis.

Across studies included in the systematic review, vicarious and/or resonant stress responses were observed in adult observers, reflected in physiological (e.g., cortisol) and subjective stress measures. Observer-related characteristics, such as state and trait empathy, were associated with stress transmission, although findings were inconsistent across studies. Sex differences were not consistently observed, suggesting that stress may be transmitted similarly in male and female observers. The relationship between observer and demonstrator emerged as relevant, revealing stronger stress transmission in romantically close dyads and those sharing social identity.

These findings highlight the importance of observer- and demonstrator-related factors in modulating social stress transmission.

Individuals value non-instrumental information that cannot be used to alter outcomes, even if it comes at a cost. This study investigated perfectionism-related individual differences in the willingness to pay for the early resolution of uncertainty via non-instrumental information regarding the outcomes of small lotteries. Perfectionism was assessed along the dimensions of Evaluative Concerns Perfectionism (ECP) and Personal Standards Perfectionism (PSP). Two studies were conducted, one online (N = 82) and the second in the lab (N = 88). In each trial, participants observed a simple coin flip lottery with each side associated to either winning or losing small amounts of money. Participants placed bids to indicate their willingness-to-pay for learning the outcome immediately, measuring the subjective value of information. We analyzed effects of lottery conditions, ECP, PSP and their interactions on bidding behaviour. Results showed that bids were lowest in lotteries with negative valence (i.e. two negative outcomes) and highest in lotteries with positive valence (i.e. two positive outcomes). Bids increased with increased expected value and range of lottery outcomes. In the lab sample, but not the online sample, higher PSP predicted higher bids in lotteries with increasing expected value. Individuals high in ECP and low in PSP showed increased bids in lotteries with increasing range. These results show that perfectionism sub-types have differential modulating effects on information-seeking, which appear to be amplified when placed in a real laboratory environment.

Humans continuously process multisensory input to guide decisions in a complex environment. This raises questions about how different sensory modalities are integrated into coherent, generalizable knowledge structures. While previous research has focused on multisensory integration in perceptual contexts, it remains unclear how this integration unfolds during more abstract cognitive processes, such as category learning and rule inference. We address these questions by investigating how multimodal sensory information is integrated into hierarchical category representations. Participants engage in a category learning task that uses naturalistic stimuli depicting everyday coffee shop scenes, each defined by three feature dimensions: auditory (music), static visual (interior design), and dynamic visual (activity). Each feature dimension can take on four different values, which are ranked on a scale from 1 to 4. Participants repeatedly choose the higher-valued coffee shop scene in a two-alternative forced-choice task, receiving feedback to guide learning.

Previous research has shown that sleep benefits rule inference in decision tasks, including information integration category learning. We test whether and how sleep supports the abstraction of rules across modalities by contrasting performance in participants who stay awake vs. sleep after training. Sleep showed a significant benefit for out-of-sample rule inference compared to a time interval spent awake. Next, the role of aperiodic EEG activity and oscillatory markers on behavioral performance after sleep will be investigated to test whether sleep actively reorganizes knowledge representations to enhance integration across sensory modalities.

We spend substantial parts of our time awake in so-called offline states, where we are less responsive to the demands of our immediate surroundings. A possible benefit of this frequent disengagement could be increased memory consolidation during offline states. It has been shown that participants can remember more previously presented information when they engage in wakeful rest, as compared to a cognitively demanding task following encoding. Previous studies indicate this effect to be larger in memory impaired populations. By measuring electrophysiological activity (EEG) during the interval following information presentation, we aim to identify neurophysiological correlates of memory performance.

In this study we recruited 30 healthy older participants and a pilot sample of six stroke patients with memory impairments. We presented them with short prose passages, and for 12 minutes following the story presentation they were tasked with either (1) resting with closed eyes while remaining awake, (2) relaxing with their eyes closed while counting their breath (3) or engaging in cognitively demanding tasks. We employed a repeated measures design, with each participant experiencing all post-encoding conditions. Condition order was counterbalanced across participants. We recorded EEG activity throughout the whole experiment.

Preliminary findings show that memory performance of the healthy participants is independent of their post encoding activity. However, oscillatory neural activity during the post encoding interval is related to memory performance.

Baseline HRV has been shown to predict the responsiveness to placebo analgesia in healthy young males, with higher baseline HRV related to greater pain relief by the placebo treatment. We sought to investigate whether this relationship can be generalized to other forms of cognitive pain modulation and older adults. Besides placebo analgesia, cognitive distraction from pain is a common cognitive modulation of pain.

The present study employed a within-subject design with two sessions with healthy younger (18-30 years) and older participants (65+ years). In both sessions, resting-state HRV was measured with ECG during five minutes. Participants then completed a set of cognitive tasks and trained the distraction task, a n-back working memory task with low (0-back) and high (2-back) cognitive load conditions, before an individual pain calibration of a non-painful and a moderate pain intensity by transcutaneous electrical stimulation to the inner forearm. Afterwards, two blocks of the pain distraction task were conducted, embedded within three blocks of mere pain perception. During these tasks, participants received electrical non-painful and painful stimuli and rated them regarding their intensity and unpleasantness. Preliminary results indicate a negative association between baseline HRV and pain relief by cognitive distraction, contrary to prior placebo findings. This effect appeared to be primarily driven by males and could be stronger in older participants, whereas females showed no clear relationship. The Differences between low and high cognitive load conditions seemed to be relatively small. These findings could lead to a better understanding of cognitive pain treatments.

Interestingly, hunger-stimulating (orexigenic) and hunger-suppressing (anorexigenic) hormones/neuropeptides, and their receptors are present in taste cells. There is initial evidence that these metabolic molecules are locally secreted within the oral cavity already following the detection of specific basic taste compounds, with bitterness emerging as an interesting candidate. In this ongoing study, we examine whether modality‑specific intra‑oral stimulation via bitter and sweet taste (without ingestion) is sufficient to elicit intra-oral hormone release, alterations in blood glucose, and food-intake-related responses. Participants (18-35 years; 50% female, 50% male; BMI 18.6-24.9 kg/m²) attend the laboratory after an overnight 10−12 hour fasting. In a 2×2 mixed crossover design, each participant receives suprathreshold taste stimulation with either a bitter solution, quinine‑hydrochloride dihydrate, or a sweet solution, sucrose (between-subject factor) on one of two consecutive days (tastant day). On the other day (control day), they are exposed to a neutral solution, i.e., distilled water. The order of conditions (tastant, neutral) is counterbalanced across participants (within-subject factor). In each session, 0.2 mL of the respective solution is sprayed onto the anterior tongue four times, with 15‑minute intertrial intervals, after an initial neutral baseline stimulation. We repeatedly measure blood glucose, local salivary insulin, appetite-related variables, physical symptoms, and mood. Additionally, we assess actual food intake during an ad libitum test meal, food preferences, beverage choice, and selective attention. Results of intragroup comparisons (tastant versus control day) and intergroup comparisons (bitter versus sweet stimulation) will be reported. We will discuss the involvement of modality-specific tastes in metabolic regulation and eating behaviour.

Autism spectrum conditions (ASC) are associated with alterations in emotional face processing, particularly during the implicit detection of changes in facial expressions. These alterations are reflected in reduced visual mismatch negativity (vMMN) amplitudes in EEG and atypical occipital and frontal activation patterns in fMRI, suggesting an imbalance between top-down predictions and bottom-up sensory processing. Previous studies have indicated that personally relevant faces may normalize neural responses during explicit emotional face processing, but their impact on implicit emotional face processing remains unclear. In this simultaneous EEG-fMRI study, 21 autistic and 21 pairwise matched non-autistic adults completed a passive visual oddball task in which rare fearful expressions were presented among frequent neutral expressions, displayed by personally relevant and unfamiliar faces. Hemodynamic responses and event-related potentials were analyzed to assess general and implicit emotional face processing, and exploratory trial-by-trial ERP-fMRI covariation analyses were conducted. Personally relevant faces elicited enhanced hemodynamic responses in face-related regions, including the lateral occipital cortex, insula, and posterior cingulate, in both groups, with even stronger responses in autistic participants. ERP analyses further showed an Identity × Group interaction for P1 amplitudes, driven by more positive amplitudes for relevant versus unfamiliar faces in the non-autistic group only, suggesting differences in early face processing. However, personal relevance did not affect vMMN amplitudes or hemodynamic responses to deviant emotional stimuli. These findings suggest that personal relevance enhances general face processing in both groups, but may not affect neural responses to implicit emotional changes.

Psychosocial stress is known to alter prefrontal cortical activity, which may impair executive cognitive processes such as lexical retrieval or mental arithmetics. While cortical beta power has been proposed as a neural marker of cortical arousal and executive functions, its role during stress remains ambiguous. Here we investigated whether acute psychosocial stress affects executive functioning and whether cortical beta activity serves as a neural correlate of cognitive performance under stress.

Healthy adults (preliminary data: n = 10) completed a Trier Social Stress Test (TSST) session and a control session in a within-subjects crossover design. Resting-state EEG was recorded before and after each session. Salivary cortisol, ratings of stress perception, and state anxiety confirmed successful stress induction, with significantly higher values in the TSST compared to the control condition.

So far, stress significantly improved cognitive performance during the arithmetic task (p < .001), which was predicted by higher baseline beta power (p < .001). Additionally, increases in beta power in the TSST condition were associated with fewer arithmetic errors (p = .003). On the other hand, participants did not show any difference in semantic verbal fluency between sessions and no correlation with beta power was found at the current sample size. In sum, these findings suggest that beta-indexed cortical arousal may differentially influence arithmetic and verbal processing under acute stress.

Data collection is ongoing and complete results will be presented.

Chronic stress and aging are known to alter hypothalamic–pituitary–adrenal (HPA) axis functioning and pain processing, yet their combined influence on pain modulation remains insufficiently understood. This study examines how chronic stress (assessed via the Trier Inventory for Chronic Stress, TICS) and the cortisol awakening response (CAR) relate to experimentally induced pain in healthy younger and older adults. Participants were screened for chronic illness and medication use, completed the TICS, and provided saliva samples on three mornings at awakening and +30, +45, and +60 minutes to capture CAR dynamics as a marker of HPA-axis activity. Pain modulation was assessed using individually calibrated electrical stimulation to the outer hand, with thresholds for mild, moderate, and strong pain determined across different experimental conditions. Preliminary analyses indicate that higher chronic stress is associated with a reduced ability to modulate pain through cognitive distraction. We hypothesize that higher levels of chronic stress and a blunted CAR will be associated with impaired pain modulation, with stronger effects expected in older adults due to age-related changes in stress physiology. Detailed analyses of CAR measures will be presented at the poster session to further elucidate these relationships. These findings may contribute to a deeper understanding of how stress-related biological mechanisms shape pain experiences across age groups and may inform targeted prevention and intervention strategies.

Cross-cultural research suggested that individuals living in "carpentered" environments, characterized by straight lines, right angles, and rectangular structures, may be more susceptible to illusions, yet the robustness and underlying mechanisms of this association remain unclear. Urban environments exemplify such settings, where prolonged exposure could reinforce depth cue and perspective assumptions that increase susceptibility to illusions exploiting these visual cues. Using a twin sample to control for genetic confounds and shared early environment, we investigated whether current urban exposure shapes susceptibility to geometric illusions and whether this relationship is neuroanatomically manifested in hippocampal structures, given their role in visuospatial integration and scene conceptualization.

Twin pairs completed an online experiment assessing susceptibility to eight geometric illusions. Urbanicity was quantified using Geographic Information System data from the European CORINE Land Cover Atlas based on participants' current residential addresses. In a separate neuroimaging study, MRI data were examined to identify neural correlates of the association between urbanicity and illusion susceptibility.

Inter-illusion correlations were generally low. Significant positive associations were observed between urban fabric and susceptibility to both the Müller-Lyer and T illusions. Grey matter volume (GMV) in the hippocampus and parahippocampus was associated with susceptibility to these illusions, respectively.

Urban environments, characterized by rectangular geometries, may reinforce how the brain processes depth cues and perspective, increasing susceptibility to illusions that exploit linear structures. Hippocampal and parahippocampal GMV implicate scene-based spatial inference as a key mechanism. These findings suggest that susceptibility to geometric illusions reflects internalized depth assumptions calibrated by the geometry of our surroundings.

Naturalistic speech unfolds along the temporal and spectral dimension, affording

fine-grained predictions of its acoustics. Here, we ask whether the

brain predicts upcoming speech acoustics at a sub-phonemic level. Under a

hierarchical Bayesian prediction framework, acoustic signals could be

used to update the probability of upcoming lexical content.

Alternatively, the brain might predict acoustics and words concurrently,

relying on their joint probability distribution. To address these

questions, we analysed human neural activity recorded with intracranial

EEG, while participants listened to German and English conversations

(total n = 21 patients, 369 speech-responsive electrodes). We

capitalized on recently developed autoregressive text-to-speech modeling

(Fish-Speech1.5), to separately extract entropy (uncertainty across all

possible tokens) and surprisal (unexpectedness of actually perceived

token) for acoustic and lexical tokens separately. As expected,

electrodes over bilateral superior temporal cortices (STG) and left

inferior frontal gyrus (IFG) strongly responded to the speech signal.

Crucially, we found widespread encoding of acoustic surprisal and

entropy, above and beyond encoding of speech acoustics and word-level

suprisal and entropy (median ΔR²: 0.003, max ΔR²: 0.036). Moreover, in

40% of speech-responsive electrodes, neural responses reflected the

joint probability distribution of acoustic and lexical tokens, rather

than the posterior probability of word token given acoustics. Overall,

our results reveal continuous and concurrent predictive processing of

speech across representational levels, speaking in favor of a

joint-distribution model rather than a hierarchical Bayesian prediction

model.

A growing body of results suggests that the morphology of macroanatomical features, such as sulci, is linked to cognitive abilities. In the domain of reading, prior research has suggested a positive relationship between interruptions (gaps) in the occipito-temporal sulcus (OTS) and reading proficiency. Yet, studies examining this relationship have yielded mixed findings. While some studies report longitudinally accumulating effects of OTS gaps on reading performance (Bouhali et al., 2025), others find no effect in large homogeneous samples (Roell et al., 2022), or suggest the structural advantage is restricted to highly literate individuals (Cachia et al., 2018).

We aimed to clarify these discrepancies by examining the effect of OTS morphology on reading performance using a large, diverse developmental sample (n = 202, ages 6–21) from the publicly available Healthy Brain Network dataset, including individuals with developmental dyslexia.

We defined posterior and anterior OTS gaps in each hemisphere blind to participant diagnosis and linked OTS gap presence to performance on a standardized reading test.

Our findings show that the presence of an interrupted left anterior OTS was associated with significantly higher reading performance scores. However, this effect was only present in typical readers and when beginning readers were excluded. Conversely, in children with developmental dyslexia no significant performance difference between individuals with continuous and interrupted OTS morphology was found. These results may help reconcile inconsistencies in prior literature as they suggest that the functional advantage of anterior OTS interruptions may only manifest after reaching a specific level of reading proficiency.

Ketamine, a NMDA-receptor antagonist, receives attention as a pharmacological model of schizophrenia and as a rapid-acting antidepressant, although underlying mechanisms remain incompletely understood. In this within-subject placebo-controlled study, we investigated the effects of low-dose ketamine on resting-state EEG in a sample of 25 healthy participants. Recordings were obtained before and after ketamine (targeted plasma concentration 0.15 µg/ml), propofol (0.5 µg/ml), or placebo.

Here we focus on aperiodic changes of the EEG power spectrum as a proxy for excitation–inhibition (E:I) balance. Source-reconstructed EEG data were parcellated using the 100-region, 7-network Schaefer atlas, and analyzed along the sensory-to-association (SA) cortical hierarchy to determine the spatial distribution of E:I alterations.

Our results indicated that higher positions along the SA hierarchy are associated with greater shifts in E:I balance toward excitation; an effect further amplified under ketamine. Notably, the strongest excitation-related changes were observed in default-mode network regions.

We also aimed to understand the time-of-day dependency of ketamine’s effects by comparing morning versus afternoon sessions. Exploratory analyses suggest that ketamine’s impact on E:I balance may be modulated by circadian factors, with stronger afternoon effects of ketamine. Ongoing investigations now aim at systematically characterizing the interaction between circadian rhythms and ketamine’s neurophysiological and antidepressant effects.

This project also included propofol measurements, conceptually a pharmacological counterpoint to ketamine. We observed similar patterns of E:I balance shifts under ketamine and propofol, which questions the specificity of ketamine’s effects and raises caution when interpreting the aperiodic EEG spectrum as an E:I balance proxy.

Fibromyalgia syndrome (FMS) is characterized by augmented pain processing and impaired endogenous pain inhibition. The periaqueductal grey (PAG), a core hub of the descending pain modulatory system, interacts with brain regions involved in pain modulation, including the rostral anterior cingulate cortex (rACC), posterior insula, dorsolateral prefrontal cortex, hippocampus, and amygdala.

In this study, 18 patients with FMS and 17 healthy controls (HC’s) underwent fMRI during individually calibrated warm and painfully hot thermal stimulation. Thermal intensity was rated after each stimulus. To assess condition-dependent connectivity during painful hot stimulation, psychophysiological interaction analyses were performed using the PAG (MNI x, y, z = 0, −28, −6) as seed region. Small volume correction was applied within a priori regions of interest.

FMS patients rated warm stimuli as significantly more intense than HC´s, while painful hot ratings did not differ between groups. Compared with controls, FMS patients showed stronger PAG connectivity with the right posterior insula (z = 3.64; pFWE = .007), right rACC (z = 4.67; pFWE = .002), right prefrontal cortex (z = 5.69; pFWE < .001), and right hippocampus (z = 4.51; pFWE = .018). PAG–amygdala connectivity did not differ between groups.

Our findings suggest altered PAG-mediated pain modulation in FMS, reflected in increased coupling with sensory, cognitive, and limbic pain-processing regions during painful stimulation. Dysfunction of the descending pain modulatory network may contribute to FM pain maintenance.

A clearer understanding of the cognitive processes involved in deception is an essential step toward reliable lie detection. In two EEG experiments, we employed an instructed lying paradigm, wherein participants responded to autobiographical target questions from multiple semantic categories. They were instructed to lie about one category while responding truthfully to the others. Importantly, both deceptive and truthful responses could be given by either affirming or negating the question. In Experiment 1 (N = 32), we observed increases in RTs, P1, N1, P3, late positive complex (LPC) and midfrontal theta power in the infrequent lie trials (25%) compared to frequent truth trials (75%). For RT and theta, this lie effect was larger when participants gave affirmative rather than negative responses. In Experiment 2 (N = 48), lie and truth trials were presented equally often to better distinguish frequency effects from genuine deception-related processes. We replicated all previous findings regarding main and interaction effects, with the exception of the P1 lie effect. Additionally, the LPC effect was strongly reduced, implying that it – at least to a great extent – reflects attentional allocation to rare events in this paradigm. The combined findings indicate that lying about autobiographical information involves increased early attentional capture and later attentional allocation. Interestingly, the larger RT lie effect for affirmative responses was mirrored by midfrontal theta. Thus, to maximize the need for cognitive control and the observed differences between truthful and deceptive responses, future studies should investigate the generalizability of the affirmation-modulated lie effect to other lying tasks.

Background

Chronic pain modulates the activity of various brain regions, with emotional processing playing a crucial role in pain experience. Stress significantly influences chronic pain, as prolonged stress can dysregulate the HPA axis, leading to hyper- or hypocortisolism and thereby contributing to the pathogenesis of pain-related conditions. Accordingly, women with chronic pelvic pain (CPP) who report higher stress levels tend to experience more severe pain and impairment in daily life.

Purpose

Building on prior research, this study investigates how acute stress influences BOLD responses to experimentally induced acute pain in individuals with CPP compared to healthy women.

Methods

A total of N = 48 women (32 individuals with CPP and 16 healthy controls) are included in this study. Neural responses to acute stress and pain were assessed using the standardized “ScanSTRESS” fMRI paradigm in combination with two brief thermode-induced heat pain sequences.

Results

Preliminary findings indicate consistent amygdala activation in individuals with CPP across all pain conditions, potentially reflecting heightened anticipatory fear. Additionally, increased insula activation was observed in individuals with CPP compared to healthy controls when pain was induced after the stressor, suggesting enhanced sensitivity to the emotional and affective dimensions of pain.

Conclusion

These preliminary findings suggest that acute stress amplifies affective pain processing in individuals with CPP. In particular, increased amygdala and insula activation may reflect altered emotional appraisal of pain under stress. Understanding these mechanisms may help develop targeted interventions that address stress-related modulation of pain.

Approximately 80% of the world’s population consumes caffeine daily, often to facilitate effortful tasks. While caffeine is known to enhance alertness and psychomotor performance, its effects on motivational processes underlying effort-based decision-making remain unclear, particularly with regard to sex/gender differences and in vulnerable populations such as individuals with panic disorder. We investigated associations of acute caffeine intake (150mg vs. placebo) with motivational behaviour using an effort allocation task in a randomised, double-blind, crossover design with >36h caffeine-abstinence before sessions. In a sample of 53 healthy participants (33F/20M) and 29 individuals with panic disorder (25F/4M) with ≤300mg/week habitual caffeine intake, we analysed the motivational phases of invigoration and effort maintenance under varying reward magnitude (low vs. high number of points) and task difficulty (easy vs. hard trials). Preliminary results show that effort maintenance was higher for high and lower for low rewards, whereas higher difficulty was associated with reduced effort maintenance, despite the absence of real-world rewards. Caffeine intake correlated with increased invigoration, indicating enhanced motivational activation, while there was no significant association with effort maintenance. We found sex/gender differences, with men showing reduced effort maintenance and invigoration compared to women. Furthermore, effects of caffeine varied by sex/gender, suggesting sex/gender-specific modulation of reward sensitivity in response to caffeine intake. We did not observe overall group differences between healthy participants and individuals with panic disorder. These findings suggest that caffeine might primarily enhance initial motivational activation rather than sustained effort and highlights sex/gender-specific patterns in effort-based decision-making.

Implicit emotion regulation enables modulation of affective responses without conscious intent. Previous work suggests that such regulation engages large-scale cognitive control networks. However, anxiety alters perceptual sensitivity, emotional reactivity, and the efficiency of cognitive control, raising the question of how implicit regulation operates under heightened anxiety. One possibility is that anxiety disrupts regulation by compromising control-related processes at later stages. Alternatively, if control signals act on early processing stages, before a full emotional response is generated, regulatory effects may be preserved under heightened anxiety.

We examined these possibilities at behavioral and neural levels. Fifty-five participants completed a scrambled-sentence task (regulation vs. no-regulation), followed by passive viewing of negative and neutral images. Trait, state, and social anxiety were assessed and combined into a general anxiety proneness index. MEG data were analyzed using cluster-based permutation tests and mixed-effects models.

Implicit regulation shifted affective evaluations of negative stimuli, with effects increasing with higher general anxiety proneness. No comparable modulation was observed for arousal ratings. At the neural level, implicit regulation was associated with enhanced early occipital responses and attenuated later parietal activity. While general anxiety did not modulate these effects, occipital activity increased with higher state anxiety.

Together, these findings suggest that implicit emotion regulation remains effective across anxiety levels and may appear more effective with increasing anxiety proneness. Importantly, situational anxiety was associated with stronger regulatory influences at early perceptual stages, whereas dispositional anxiety predicted stronger downstream effects on affective evaluation, suggesting the potential utility of implicit emotion regulation under heightened anxiety.

Human memory is fundamentally constructive: rather than recording memories exclusively verbatim, the brain extracts and stores semantic representations that can generate vivid recollections of events that never occurred. The Deese-Roediger-McDermott (DRM) paradigm exploits this property, reliably inducing endogenous false memories for semantically related, non-presented words. Two prominent theoretical accounts, the Activation-Monitoring Theory and Fuzzy Trace Theory, converge on the proposal that false memory susceptibility is driven by the degree of overlap between the semantic memory representations of studied material and critical lures. What remains unresolved is whether the temporal pole, the brain's amodal semantic hub, encodes this overlap categorically or as a continuous, graded representational structure.

The present study addresses this question directly. Using 7T fMRI (1.5mm isotropic, multi-echo acquisition with TEDANA denoising) and whole-brain RSA searchlights, we will scan 44 participants performing an incidental semantic categorization task on words from 16 German DRM lists. Each list contains three lure types spanning the full associative hierarchy, representing a gradient across associative strength. We predict that temporal pole representational geometry reflects a continuous semantic gradient, and that individual differences in this neural overlap predict idiosyncratic false memory rates at the participant level. Beyond this primary test, we examine brain-model alignment between temporal pole representations and a battery of large language models, from static word embeddings to contextualized transformers and multimodal encoders, to characterize the computational principles underlying semantic memory organization and distortion.

Depression and anxiety disorders are highly prevalent, often characterized by recurrent courses and dysregulated physiological functioning. Respiration may serve as a transdiagnostic mechanism linking physiological and affective processes; however, controlled studies integrating both domains remain limited.

This study investigated the effects of a structured breathing-based intervention, Breathing Apperception Training (BAT), on clinical symptoms, respiratory regulation, and interoceptive processes in 70 patients with depression (50% with comorbid anxiety disorders and/or PTSD). In a randomized controlled longitudinal design, participants received a two-month BAT intervention or were assigned to a wait-list control group. Additionally, half of the sample completed a two-month follow-up assessment. Outcomes included clinician-rated and self-reported depression and anxiety, stress coping, and well-being, alongside physiological indices such as respiratory variability and breathing dynamics, pulmonary reserve volumes, and cardiac interoceptive accuracy.

Results showed significant improvements in clinician-rated depression, as well as within-session reductions in negative affect and anxiety. Follow-up analyses indicated small but sustained effects on well-being and body awareness. While group-level lung function and cardiac interoception remained stable, respiratory variability under stress suggested increased regulatory flexibility. Greater individual improvements in respiratory capacity were predictive for reductions in depressive symptoms and enhanced well-being. A minimum training dose of approximately 30 minutes per week was identified, with no additional group level benefit observed beyond this threshold.

Overall, BAT appears to enhance affect regulation in acutely depressed patients through individualized physiological adaptation, suggesting that clinical benefits are driven by improvements in respiratory functioning. These findings support BAT as a promising adjunctive intervention for depression.

Human cooperation often involves performing joint tasks, where success relies on how collective rewards are allocated among cooperating parties based on their individual performance and contribution to task outcomes. However, it remains unclear whether and how individual performance and contribution give rise to self-related biases in such allocation decisions. Here, we developed a novel performance-based social allocation task that manipulated how individual performance contributed to joint outcomes. Across two experiments, participants exhibited a robust self-allocation bias: they allocated more rewards to themselves and disproportionately disregarded their own performance, particularly when their performance did not causally contribute to the joint outcome. This self-allocation bias was amplified in individuals with stronger individualistic social preferences, as measured by social value orientation. At the neural level, self-relevant (versus self-irrelevant) allocation decisions were associated with increased activity in the medial prefrontal cortex extending into the anterior cingulate cortex, insula, and temporoparietal junction (TPJ). Moreover, the dorsomedial prefrontal cortex, lateral orbitofrontal cortex, and TPJ tracked trial-by-trial variations in relative performance as a function of contribution structure, independent of self-relevance. Together, these findings suggest that self-allocation bias in performance-based decisions is unlikely to arise from distorted neural encodings of performance. Instead, self-interest may shape how contribution-structured performance information is used in social-allocation choices, providing a more precise account of how self-serving behavior emerges in cooperative contexts.

DNA methylation markers of C-reactive protein (CRPm) have been proposed as an epigenetic indicator of chronic low-grade inflammation, a process implicated in major depression. Building on prior findings linking CRPm to early life adversity (ELA), recent life stress (RLS), and depressive symptoms, we examined whether CRPm is associated with structural brain alterations.

Using structural MRI data from a subsample of the Marburg-Münster Affective Disorder Cohort Study (MACS), including 381 patients with major depressive disorder (MDD) and 247 healthy controls (HC), we tested associations between CRPm and whole-brain gray matter volume (GMV). Significant clusters were extracted and further examined for associations with depressive symptom severity (Hamilton Depression Rating Scale, HAMD), ELA (Childhood Trauma Questionnaire, CTQ), and RLS (Life Events Questionnaire-Revised, LEQ-R).

Whole-brain analyses revealed that higher CRPm scores were associated with reduced GMV in clusters encompassing the right superior and medial temporal regions, orbitofrontal cortex, and left thalamus. Reduced thalamic volume showed weak negative associations with depressive symptom severity, ELA, and RLS, with lower volume being associated with higher scores in all three measures. However, these associations were attenuated and no longer significant after controlling for health-related variables, including BMI, smoking, and alcohol consumption.

These findings suggest that associations between inflammation-related epigenetic markers and brain structure may be partly explained by health-related behaviors rather than reflecting a direct pathway.

Background: Interoceptive processes are central to emotional experience and play a crucial role in maintaining mental health. Training programs based on mindfulness are a promising approach to modulate these processes. The present study aimed to investigate the effects of a two-week, app-delievered mindfulness-based breathing exercises on both behavioral and self-reported interoceptive indices, as well as on health-related outcomes.

Methods: In this pilot study, 25 healthy participants were randomly assigned to either the mindfulness intervention or a waitlist control group. Data were collected at baseline (t1) and after the two-week intervention (t2). Behavioral interoceptive performance was assessed using the Heartbeat Counting Task (HBC), and self-reported interoception was measured via the Interoceptive Attention Scale (IATS) and the Interoceptive Accuracy Scale (IAS). Additionally, health-related outcomes including perceived stress (PSS), emotion dysregulation (DERS), and anxiety (STAI) were measured.

Results: Analyses revealed a mixed pattern; HBC accuracy increased descriptively in the intervention group, whereas no clear differences were observed in the self-reported interoceptive or health-related measures between groups.

Discussion: These findings suggest that short-term mindfulness training may primarily influence interoceptive performance, while more complex health-related outcomes might require longer or more intensive interventions. The results highlight the potential of mindfulness-based training to modulate interoceptive processes and its relevance for mental health.

Editorial quality control in scientific publishing relies on checks for scope, ethics, policy compliance, reporting completeness, figure clarity, and data transparency, which are usually performed manually therefore inconsistently. We describe a cross-organizational collaboration between Springer Nature and EMBOPress integrating complementary AI-assisted quality control systems within the Data4Rev platform: Multi-Modal Quality Checks (mmQC), an open library of multimodal quality checks, a collection of figure and document-level checks auditing for example the inclusion of error bars, explicit statistical reporting, definition of replicates, axis units, scale bars, and presence of data availability statements while remaining conceptually aligned with higher-level editorial assessment approaches used at Springer Nature and with prospective efforts on AI-assisted editorial solutions (e.g. Arpi).

The project aims to perform existing editorial checks earlier and systematically in the submission process. A secure intermediary architecture is proposed to bridge the Springer Nature and EMBOPress technical environments while preserving network and governance boundaries. Early observations indicate that combining these systems may reduce downstream rework, improve consistency of reporting checks, and allow editors to focus on scientific merit instead.

To support responsible deployment and transparency of the underlying AI-assistance, an open benchmarking framework has been built for mmQC using curated human-reviewed ground truth. This way quality, consistency and coverage across reporting checks are assessed, accommodating a transparent account of the applied editorial policies. This work positions AI-assisted quality control as a scalable editorial infrastructure component. It is a secure, evidence-based integration of automated checks into peer-review workflows, providing automated quality feedback to authors during submission.

In everyday life, decisions often unfold over time and depend on others’ choices. To make such decisions, people rely on simple cues in their environment. It remains unclear how uncertainty shapes which cues are used. Here, we hypothesize that under uncertainty, people rely on the most informative and stable cue to guide their decisions.

We developed a sequential 2×2 social game in which two players maintained their energy across three trials (“days”). Energy could be gained by foraging (cooperation) with probabilistic success or by looting (defection) the other player. Foraging success depended on randomly varying weather conditions, with higher success in good than in bad weather. Task features (energy, weather, day) served as decision cues, defined the decision problem, and enabled formalization as a finite-horizon Markov decision process (MDP), providing a normative benchmark. Participants interacted with three computer agents (All-Cooperate, Random, All-Defect) across 30 games; agent type was not signaled and had to be inferred from behavior. Data were collected in a pilot (N = 29) and a preregistered replication sample (N = 33).

Our analyses showed that weather was the dominant choice cue under uncertainty. Consistently, model comparison using out-of-sample predictive accuracy (PSIS-LOO) showed that both beliefs about the agent type and weather outperformed the normative MDP benchmark (ΔELPD > 170). While simplifying the environment, the best-performing model still captured key characteristics of the MDP (r = 0.38).

These findings suggest that uncertainty simplifies decision-making by focusing behavior on stable, reliable features, enabling efficient choices under limited information.

Objective: Trauma-exposed individuals often exhibit touch aversion, yet it remains unclear whether trauma alters the anxiolytic effects of social touch and their dependence on prior interaction experiences.

Methods: We recruited 195 healthy adults (96 women, mean age = 25.2 years) and quantified early-life adversity using the Childhood Trauma Questionnaire. Participants were randomly assigned to either a Positive Interaction Paradigm (PIP) or a non-social control condition. Preferred interpersonal distance and wanting and liking of slow and fast affective touch, were assessed. Anxiolytic effects were indexed via subjective tension ratings and psychophysiological responses during a threat-of-shock task. Medial and dorsolateral prefrontal cortex (mPFC, dlPFC) activity was recorded using functional near-infrared spectroscopy.

Results: PIP improved mood and reduced the preferred interpersonal distance. Across participants, slow touch was preferred over fast touch, while fast touch elicited the strongest mPFC and dlPFC activation. In the threat-of-shock task, touch reduced subjective tension globally. Trauma exposure was not significantly related to prefrontal responses or the wanting and liking of touch. However, independent of PIP, higher trauma levels were associated with reduced anxiolytic effects of touch, as reflected in subjective ratings and skin conductance responses.

Conclusion: Social touch engaged prefrontal mechanisms and reduced subjective tension, but these effects were not modulated by prior positive interaction. Trauma exposure was specifically associated with diminished anxiolytic responses to touch, independent of general touch preference or prefrontal responses. This pattern suggests that trauma is linked to altered processing of touch as a safety signal rather than a general reduction in touch valuation.

Background: Hair analyses are increasingly used to assess long-term endocrine activity. However, several methodological questions remain insufficiently researched. This includes how hair should be sampled to optimize data quality. In this within-subject study, we compared two common hair sampling methods: obtaining the required amount of hair from a single larger sample vs. accumulating hair from three smaller samples.

Methods: A total of 53 participants provided six hair samples at two time points, three months apart (12 samples in total). At each time point, three samples were gathered using a single-sample approach (one larger hair strand per sample), and three samples were gathered using an accumulation approach (each sample consisting of three smaller hair strands from different positions). Samples were analyzed for cortisol, cortisone, progesterone, dehydroepiandrosterone, and endocannabinoids (AEA, 1/2-AG, OEA, SEA, PEA) via liquid chromatography-tandem mass spectrometry. Sampling methods were compared regarding within-method variability, differences in mean analyte concentrations, and test-retest correlations.

Results: Accumulated samples showed significantly lower within-method variability compared to single samples for hair cortisol and cortisone and several other analytes, indicating greater consistency. Mean analyte concentrations and test-retest stability were largely comparable across sampling methods.

Conclusions: Our findings highlight the importance of sampling strategy for endocrine hair analyses, particularly hair cortisol and cortisone. As accumulating hair samples from several smaller samples resulted in lower within-method variability, this method is recommended to obtain more robust estimates of long-term hormone levels.

Post-COVID syndrome (PCS) is a multisystem condition characterized by persistent fatigue, post-exertional malaise, and frequent comorbid anxiety and depression. Transcutaneous auricular vagus nerve stimulation (taVNS) has been proposed as a non-invasive intervention targeting neurophysiological dysregulation. This study examined whether taVNS reduces anxiety and depressive symptoms in individuals with PCS more than sham stimulation, and whether potential effects are mediated by changes in fatigue.

In a randomized crossover design, 21 participants received both active and sham stimulation for four weeks each. Outcomes were assessed at baseline (T0), mid-treatment (T1), post-treatment (T2), and follow-up (T3). Analyses focused on immediate (T2–T0) and sustained (T3–T2) change scores for anxiety (HADS-A), depression (HADS-D), and fatigue (FSS, CFS). Linear mixed models included condition, period, and sequence as fixed effects, with random intercepts for participants.

Across all outcomes, no significant main effects of condition were observed, indicating that symptom changes did not differ between active and sham stimulation. Overall, anxiety and fatigue decreased from baseline to post-treatment; however, these improvements were not condition-specific. Significant period effects further indicated that symptom changes differed between study phases, with reduced improvements in the second period.

Given the absence of a treatment effect, mediation analyses were not supported. Overall, these findings provide no evidence that taVNS is superior to sham stimulation in reducing depression, anxiety, or fatigue in PCS.

Depression is increasingly conceptualized not only as a mood disorder but also a disorder of maladaptive cognitive processing, with rumination, or repetitive and self-focused thinking about one’s distress, playing a central role in symptom maintenance and relapse. Contemporary models distinguish rumination into brooding and reflection that, alongside adaptive regulation strategies such as cognitive reappraisal, recruit distinct large-scale brain networks. Simultaneously, depression has been linked to dysfunction in reward-related systems critical for processing motivational values and guiding behavior. How these cognitive-affective states interact with reward circuitry, however, remains poorly understood.

In the proposed study, we will investigate the habenula as a key interface linking cognitive processing with affective valuation. The habenula encodes negative motivational value and modulates dopaminergic and serotonergic systems, and its hyperactivity has been implicated in depressive symptomatology. We propose that distinct cognitive-affective states differentially engage habenula-centered networks, thereby influencing reward processing and affective valuation.

Using 7T fMRI, participants engage in an autobiographical rumination-reappraisal paradigm, in which they process a personal negative experience across structured blocks of brooding, reflection, and reappraisal. A habenula functional localizer task serves to identify participant-specific habenula regions sensitive to loss-related reward processing. Neural activation, connectivity, and affective trajectories are then jointly analyzed to characterize these state-dependent interactions between cognitive-affective states and reward systems.

We hypothesize that brooding enhances habenula and salience network activity, reflection engages cognitive control systems, and reappraisal downregulates habenula output while restoring reward-related circuitry. This approach will provide mechanistic insight into how cognitive processing shapes affective and motivational dysfunction in depression.

Background: Delay discounting (DD) and probability discounting (PD) tasks are frequently used to study neural correlates of decision-making. However, real-world decisions involve outcomes that are both uncertain and delayed. This fMRI study was designed to examine whether an integrated delayed-probabilistic discounting task (DPD) recruits different brain areas than isolated DD and PD, and which brain areas contribute to the evaluation of its components (delays, probabilities, monetary values).

Methods: Sixty healthy participants first completed an adaptive discounting task outside the scanner in order to model individual discounting patterns and to generate trials with predictable choice patterns for the fMRI session. Inside a 3T MRI, they then completed a discounting task with three trial conditions (DD, DD, DPD), as well as an n-back task to test for the impact of working memory related processes.

Results: Decision-making during DPD trials recruited different brain areas than the combined DD and PD conditions alone, including regions outside a sample-based n-back mask. Brain activation during the presentation of delays, probabilities and magnitudes was significantly associated with model-derived discounting parameters (subjective value, discounting factors), but not objective parameters (objective value, delay, probability).

Discussion: Integrated DPD recruits brain regions above and beyond the isolated DD and PD tasks. These differences are not limited to working-memory-related regions, suggesting DPD is more than increased WM load. In addition, model-based behavior parameters outperformed objective parameters in predicting brain activation during stimulus presentation. These findings highlight the need to integrate complex decision-making processes and the strength of computational models in achieving this.

Alcohol Use Disorder (AUD) and Major Depressive Disorder (MDD) are highly prevalent and frequently co-occurring mental health conditions, yet their shared and distinct neurobiological mechanisms remain poorly understood. While resting-state functional magnetic resonance imaging has revealed functional connectivity (FC) alterations in both disorders, most prior work relies on group-level comparisons that obscure substantial inter-individual heterogeneity. This project addresses this limitation by applying a normative modelling framework to characterize individualized deviations in FC associated with AUD, MDD, and their comorbidity. Using data from the German National Cohort (NAKO; N ≈ 30,000), AUD and MDD are operationalized via validated self-report measures (AUDIT-C; PHQ-9). Rs-fMRI data are preprocessed and parcellated to derive whole-brain FC matrices, and then used as input for a deep autoencoder-based normative model trained on healthy controls. This model estimates expected FC patterns as a function of key covariates (e.g., age, sex, brain volume), enabling the computation of subject-specific deviation scores. These deviation profiles are then used to compare patterns of atypical FC across individuals with AUD, MDD, and comorbid presentations, as well as to examine dimensional associations with symptom severity. We adopt a data-driven approach to investigate (i) if rsFC-based normative deviation patterns distinguish individuals with predominantly alcohol-related symptoms from those with predominantly depressive symptoms, (ii) if specific connectivity dimensions deviate from the normative distribution as a function of symptom severity, and (iii) if individuals with comorbid profiles exhibit deviations that resemble AUD, MDD, or reveal a distinct pattern. First results will be presented at the conference.

Benzodiazepines act as positive allosteric modulators of the GABA(A) receptor and affect several motor and cognitive functions. By engaging perceptual-motor as well as inhibitory control processes, the antisaccade task was used in previous studies to investigate effects of benzodiazepines on behavioral performance. Using a randomized, double-blind, placebo-controlled within-subjects design, this study combined eye-tracking with BOLD fMRI in order to examine neural correlates of these effects for the first time. N = 39 healthy participants completed an antisaccade task after administration of either 1mg lorazepam or placebo. On behavioral level, lorazepam led to reduced (anti-)saccadic peak velocity as well as increased (anti-)saccadic latency. On neural level, drug-induced activation reduction was found in a fronto-parietal-occipital network, including key oculomotor regions. This result was further supported by our finding of increased GABA(A) receptor density in the affected network. On individual level, decline in peak velocity under lorazepam was associated with decreased neural activation in several cortical regions, including medial frontal eye fields. No interactions between drug and saccade condition (prosaccade, antisaccade) were found. Our results therefore suggest GABAergic modulation of a more general saccade-related network rather than of specific components for inhibitory control processes. Future studies may rely on BOLD signal as a sensitive marker for benzodiazepine activity during saccadic eye movements.

Although fear generalization has been extensively examined, extinction generalization remains comparatively understudied, despite its importance for treatment outcomes, as patients are often confronted with stimuli that differ from those encountered during exposure therapy. Furthermore, the context dependency of fear and extinction generalization has rarely been explored using cue-in-context paradigms, even though stimulus generalization may vary depending on whether a context is perceived as safe or threatening. To address this, 50 healthy students participated in a functional magnetic resonance imaging study investigating the neural mechanisms underlying context-dependent fear and extinction generalization. On the first day, participants underwent fear acquisition and generalization in context A, followed by extinction learning and generalization in context B. One week later, extinction recall in the safe context B and fear renewal in a novel context C were assessed. Outcome measures included fear and expectancy ratings, skin conductance responses (SCRs), and neural activity within regions of interest. After participants successfully acquired and extinguished fear on day one, testing one week later revealed context-specific patterns in neural responses and SCRs. During fear renewal in context C, neural activity in the insula and dorsal anterior cingulate cortex showed a quadratic generalization gradient, while SCRs indicated fear generalization to stimuli most similar to the CS+. In contrast, such patterns were not observed in the safe context B during extinction recall. Taken together, these findings identify contextual threat as a relevant factor shaping fear generalization, with potential relevance for understanding relapse and improving exposure-based treatments.

The transition from wakefulness to sleep is a dynamic process with no clear definition of sleep onset. It is marked by loss of our ability to perceive, store and respond to external information. This study investigates differences in the disintegration of information processing components during the wake-sleep transition.

19 participants (9 female) were instructed to fall asleep while listening to an audiobook and responding to a tone via button press. After a 90-minute sleep cycle, participants were awoken to perform a recognition task for audiobook passages. This procedure was repeated up to 5 times, yielding 78 wake-sleep transitions. Sleep scoring on the continuous EEG was performed by two independent raters. Tone responsiveness and remembered audiobook were calculated for fixed intervals and compared between sleep stages. ERPs and EEG spectral power were compared between events linked to hits/misses and memory/no memory in different sleep stages.

Participants’ ability to encode audiobook content ceased earlier than their ability to respond to a tone (F5,389=18.58, p<.001). Memory started to decline during wakefulness, while tone responsiveness showed the greatest decrease from stage S1 to S2 sleep. Neocortical stimulus processing proceeded into S2 showing a preserved P2 ERP component. The breakdown of memory was accompanied by global and local changes in alpha and theta power and steepening of the slope of the aperiodic activity independent of sleep stage. During the wake-sleep transition, more complex cognitive processes break down before basic processes such as stimulus responses, supporting the view that falling asleep is not a unified process.

Background:

Social anxiety is a robust risk factor for substance use, but the processes through which this risk develops remain unclear. Substance expectancy has been discussed as one contributing factor, yet it is usually treated as an individual-level construct. How these expectancies are shaped in social contexts, and how they relate to behavioral control processes, is less well understood.

Methods:

A total of 553 participants completed measures of social anxiety (SIAS), peer influence (PPI), cannabis policy attitudes (CPA), and substance expectancy. Analyses focused on cannabis expectancy, which was separated into positive and negative components. Regression models were conducted controlling for cannabis use (CUDIT), and policy attitudes were further examined at the component level. Network analysis was used to assess the position of expectancy within the overall system.

Results:

Positive and negative cannabis expectancy were moderately correlated but showed different patterns of association. Positive expectancy was related to peer influence, social anxiety, and policy attitudes, whereas negative expectancy was associated with social anxiety and peer influence only. Policy effects were mainly driven by perceived benefits. In the network analysis, expectancy variables showed the highest centrality.

Conclusions:

Cannabis expectancy appears to consist of related but distinct components. One possible interpretation is that positive expectancy is more closely linked to reward-driven, relatively automatic processes, whereas negative expectancy reflects more stable, goal-directed evaluations. This perspective may help to frame future work on how acute stress shifts the balance between control processes and contributes to craving.

Impulsivity and compulsivity are transdiagnostic traits associated with attenuated goal-directed control. Yet it remains unclear whether they affect the same underlying processes or show distinct associations with markers of goal-directed control. Using a contingency degradation paradigm, we tested preregistered hypotheses on the shared and distinct contributions of impulsivity and compulsivity to behavioral and subjective sensitivity to instrumental contingency.

A total of 414 participants (M ± SD = 26.98 ± 5.56 years, 162 female, 248 male, 4 non-binary/genderfluid/genderqueer) completed an online contingency degradation task and the UPPS-P Impulsive Behavior Scale and Obsessive-Compulsive Inventory-Revised. Goal-directed control was indexed by two markers: behavioral sensitivity, reflected in response rates across contingency changes, and subjective sensitivity, reflected in explicit causality ratings. Linear mixed-effects models tested effects of instrumental contingency, questionnaire scores, and their interactions.

Overall, participants showed robust sensitivity to instrumental contingency, with response rates and causality ratings both tracking contingency changes. Higher compulsivity scores were associated with reduced behavioral sensitivity, reflected in a shallower relationship between instrumental contingency and response rate. Impulsivity was not associated with behavioral sensitivity. However, a significant impulsivity × compulsivity interaction was observed for subjective sensitivity, with individuals high on both traits exhibiting lower accuracy in explicit causal judgements.

These findings suggest that goal-directed control is broadly intact at the group level, but trait-related differences emerge across behavioral and subjective markers. Compulsivity selectively predicted behavioral adaptation, while reduced explicit causality sensitivity required elevated levels of both traits - supporting a distinction between these markers in transdiagnostic research.

Being watched by others can affect cognitive control processes, including the sensitivity of error monitoring, as reflected by two dissociable event-related potentials, the Ne/ERN and the Pe. While social observation has been associated with an increased Ne/ERN, effects on the Pe were sparse and inconsistent. The present preregistered study investigated effects of social observation on error monitoring under high levels of uncertainty by utilizing the target-masking paradigm. In this modified flanker task, the target is invisible on some trials, thus preventing a representation of the correct response. We hypothesized that social observation could act as a metacognitive cue that amplifies error signals or error confidence under these conditions. Social observation was manipulated by the presence or absence of a silent observer. Although error rates and error confidence (measured by trial-wise confidence ratings) were unaffected, distinct effects of social observation on the two event-related potentials were obtained. The Ne/ERN was generally enhanced under social observation, replicating previous studies. In contrast, the Pe was reduced under social observation, but this effect was restricted to trials with invisible targets. This pattern suggests that social observation generally enhances early error signals but prevents these signals from being conveyed to later stages of error monitoring under high uncertainty. However, exploratory analyses suggest these findings may depend on the choice of baseline, warranting further investigation.

A considerable part of the variation in general cognitive ability can be attributed to general processing speed. Previous findings suggest that more intelligent people process and evaluate information more quickly. However, it is still an open question how a higher processing speed contributes to cognitive performance. The neurocognitive process model assumes that more intelligent people benefit from a higher level of white-matter-tract integrity, facilitating greater functional connectivity between neuronal networks, underlying working memory and cognitive control, resulting in faster processing speed.

To test the neurocognitive process model, a comprehensive collection of behavioural and neurophysiological measurements was obtained from 215 individuals (113 females, 102 males; 18-40 years), covering measures of general (Berlin Intelligence Structure test) and fluid intelligence (Advanced Progressive Matrices), working memory capacity (binding, complex span), cognitive control (cued task-switching), and information processing speed (Posner task, Sternberg task, two-choice reaction time task).

The study comprises electrophysiological recordings during the cognitive control task, processing speed tasks, and rest as well as structural (T1, T2, DWI) and functional MRI measurement during the same cognitive control task and rest.

Here, we present the extensive dataset, which will be accessible via a ZPID repository. Moreover, we provide a brief overview of the latest open science practices for sharing data, emphasising practical strategies and the typical challenges associated with large neuropsychological datasets. By reporting on repository tools and licensing options, the contribution offers concrete guidance to promote transparent, reproducible research and to facilitate the FAIR (Findable, Accessible, Interoperable, Reusable) principles of scientific data.

Background: Expectations contribute to symptom improvement in depression and may influence treatment outcomes alongside pharmacological effects. However, the neural mechanisms underlying expectancy–treatment interactions remain poorly understood. Here, we investigated whether brain morphology is associated with expectation effects and their interaction with esketamine treatment in major depressive disorder (MDD).

Methods: We conducted a double-blind, randomized study using a fully balanced placebo design with the factors “treatment” (intravenous application of 0.25 mg/kg esketamine or placebo) and verbally induced “expectation” (high vs. low). A total of 151 MDD patients underwent structural T1-weighted MRI and repeated psychometric assessments pre- and post-intervention. MRI data were preprocessed using the CAT12 toolbox. Voxel-based morphometry (VBM) analyses were performed in SPM.

Results: Self-reported changes in depressive symptoms 40-minutes post intervention were associated with a significant interaction between expectation and treatment in gray matter volume in the right cerebellar Crus I (p < 0.007). Post hoc analysis of VOI values indicated a symptom decrease related to brain volume specifically in the high expectation placebo group (placebo group mean difference = −0.030), whereas the esketamine group showed no significant association to cerebellar structure (esketamine group mean difference = −0.002).

Conclusion: These findings suggest that cerebellar structure is linked to expectancy-related symptom improvement, particularly in placebo contexts, while playing a lesser role in esketamine-related effects.

The ability to learn and retain associations between events and their contexts is crucial for survival, particularly in threatening environments. Animal research demonstrates that context modulates retrieval and expression of defensive responses, with contexts represented either as a set of independent features (elementally) or as an integrated configuration (conjunctively). How such contextual representations become integrated into fear learning and memory consolidation in humans is unclear. To address this question, we used a novel virtual reality–based, context-dependent fear conditioning paradigm under free behavior across two days. The virtual environment comprised three distinct offices, each containing two identical lamps (yellow, blue). On Day 1, participants (N = 72) explored two contexts, where one lamp was paired with an aversive electro-tactile stimulation (US) and the other was not. Crucially, contingencies were reversed across contexts. On Day 2, participants were re-exposed to both conditioned contexts and a third, neutral one, allowing us to assess context-dependent retrieval of learned associations. Subjective ratings, autonomic, attentional, and proxemic measures were recorded. Preliminary results show successful context-dependent threat learning reflected in verbal and autonomic responses. Memory performance was strongly modulated by context, consistent with context-dependent retrieval. In the neutral context, US expectancy was reduced for both lamps, consistent with a dominance of configural over elemental representation. Moreover, the strongest memory retrieval was linked to the first acquired contingencies, indicating persistence of early learning. Together, these findings highlight the role of context and its representations in shaping defensive responses and suggest their relevance for clinical interventions targeting maladaptive emotional memories.

Habits and motor skills are assumed to be both automatic and acquired through extensive repetition of behavior. However, motor skills are still goal-oriented while habits are autonomous and therefore invariant to changes in the value of a response-associated outcome. The current online study aimed to investigate the relation and potential distinction between motor skills and habits in a modified serial reaction time task with an outcome devaluation manipulation.

Participants (N = 277) repeatedly responded to four different-colored boxes by pressing random or fixed sequences of six button presses to receive points. In devaluation blocks at the beginning and at the end of the experiment, two boxes’ colors indicated that pressing the learned sequence would lead to minus points, thus the value of the outcome changed. Expression of habitual behavior would be reflected in continued performance of the previously learned sequence for the devalued stimuli.

Reaction times for the fixed sequences but not for the random sequences were significantly shorter in blocks at the end of the experiment compared to blocks in the beginning. This indicates the learning of the fixed sequences probably involving chunking of motor responses. Furthermore, there was a significant effect of devaluation: Participants did not perform the learned sequence for devalued stimuli indicating goal-oriented behavior. The presence of motor skilled but not yet habitual behavior indicates that these represent two different forms of automatic behavior.

Attentional biases toward threat are a core feature of anxiety, yet their reliable measurement remains a major challenge, limiting their utility for clinical translation. Building on meta-analyses highlighting the importance of temporal constraints (Bantin et al., 2016; Bar-Haim et al., 2007), the stimulus onset asynchrony (SOA) between distractors and targets was experimentally manipulated (100 ms vs. 500 ms) in this preregistered study. In addition, a Dual Probe variant was included to evaluate potential improvements in psychometric properties compared to the Dot Probe paradigm. Electroencephalography (EEG) was used to extract the N2pc component as a neural index of selective attention, alongside classical reaction time (RT) bias measures. We hypothesized that RTs would exhibit larger effects and greater reliability when SOA is short (100 ms) while the N2pc excels at longer SOA (500 ms).

Results indicate that the N2pc captures attentional allocation to threat but can only be evaluated at longer SOAs (500 ms) using classical preprocessing pipelines. Furthermore, its reliability was higher in the Dot Probe compared to the Dual Probe paradigm, although effect sizes remained unaffected. In contrast, RT-based attentional bias emerged exclusively at short SOAs (100 ms) and only in the Dot Probe task. Notably, this effect did not translate into improved reliability, underscoring limited suitability for individual-differences research.

These findings indicate a fundamental trade-off: Optimizing task parameters requires prioritizing either neural or behavioral indices. Taken together, electrophysiological measures under longer SOAs appear more suitable for assessing attentional bias, although challenges for increasing reliability remain.

Emotional regulation deficits are involved in a wide range of psychopathologies. The recently developed Emotional Dysregulation Questionnaire (EDQ; Gill et al., 2021) was designed to allow for a more comprehensive assessment of emotional dysregulation than the widely used Difficulties in Emotion Regulation Scale (DERS). The EDQ is based on an eight-factor model of emotion regulation deficits and expands on the DERS by also covering specific maladaptive regulation strategies, maladaptive beliefs about emotions, and maladaptive cognitive, behavioural, and affective consequences of emotional arousal. The current study evaluated the reliability and validity of a German version of the EDQ in a non-clinical sample of 816 adult participants (M = 23.65 years, SD = 5.9, 71% female). The EDQ comprises 40 items, organized in 8 scales, rated on a 1–5 Likert scale. Psychometric evaluation included analyses of item difficulty, item-total correlations, internal consistency, and convergent validity. Item difficulty (mean ratings across participants) varied between 1.43 and 3.34 (M = 2.20, SD = 0.49), part-whole-corrected item-total correlations between .45 and .89, and internal consistency (Cronbach’s 𝛼) between .70 and .92. Convergent validity with other questionnaires on emotion regulation and mental well-being was high, e.g., r = –.31 with reappraisal in the Emotion Regulation Questionnaire (ERQ) and r = .54 with the Perceived Stress Scale (PSS). Overall, item analysis, reliability estimates, and convergent validity suggest high psychometric quality for our German translation of the EDQ. The poster will additionally present a confirmatory factor analysis and discuss fields of application in future research.

Aims: Adolescence is marked by an increase in the prevalence of mental health problems, alongside sex differences (e.g., higher rates of internalizing symptoms in females, more externalizing symptoms in males). Sex hormones such as testosterone, estradiol, and progesterone influence brain development, behavior, and emotional regulation. However, findings on associations with mental health in youth remain inconclusive, partly due to heterogeneity in study populations (e.g., in terms of age and sex) and research methodologies (e.g., hormone type, assessment method). This review systematically synthesizes current knowledge in the field and identifies important directions for future research.

Methods: A systematic search was conducted in PubMed, PsycINFO, Web of Science, and Embase. Eligible studies examined associations between biologically assessed sex hormones (e.g., blood, saliva, urine, hair) and mental health outcomes in community or clinical samples aged 6–25 years. Cross-sectional and longitudinal observational studies were included. Data on sample characteristics, hormone measures, outcomes, and covariates will be extracted. Mini-meta-analyses will be performed where comparable effect sizes are available.

Results: The search identified 9,402 records, of which 5,600 remained after deduplication. At the time of submission, 60% of the studies have been screened. Studies vary in design, outcomes, and hormone assessment methods. We will summarize associations between hormones and mental health by age, sex, and method, also comparing long-term (hair) versus short-term (saliva, blood) perspectives.

Conclusions: Findings will support early identification of adolescents at risk of mental health problems and inform prevention and intervention strategies. Implications for research and clinical practice will be discussed.

The Jumping to Conclusions (JTC) bias—a tendency to make decisions based on insufficient evidence—is well established in abstract tasks, yet it remains unclear whether it extends to social contexts that require inferring others’ intentions. The present study examines whether JTC emerges in such interactions and whether it differs when inferring cooperative versus competitive intentions.

To test this, we conducted two online studies (each N = 30, recruited via Prolific). Participants completed grid games in which they interacted with cooperative and competitive computer agents. On each trial, they chose their own movement, predicted the agent’s movement, and rated the agent’s cooperativeness. After a fixed number of rounds, they could continue interacting with the same agent or switch to a new one, thereby allowing us to quantify information gathering as a behavioral index of JTC.

Results showed that participants engaged in significantly more trials with cooperative than competitive agents, indicating earlier termination of interaction and reduced information gathering in competitive contexts. These asymmetrical patterns suggest that motivational factors influence belief updating. Together, these findings support the paradigm as an ecologically valid framework for studying biased decision-making in social contexts and raise the question of how such biases operate under more complex, ambiguous conditions.

Ongoing work introduces environmental ambiguity to further probe JTC, with the expectation that ambiguity will reduce information gathering and amplify asymmetrical responses. Overall, our work provides insights into how people navigate uncertainty when forming social judgments.

Perception of the visual environment requires the integration of incoming contextual information with stored memories and prior knowledge. Predictive processing theories propose that this integration is supported by the hierarchical and laminar architecture of the visual cortex and its interactions with the medial temporal lobe (MTL), particularly the hippocampus and entorhinal cortex (EC). To examine the neural mechanisms underlying this process, we acquired ultra-high-field 7T fMRI data using an occluder paradigm that dissociates memory signals from signals driven by concurrent contextual cues. In this two-day experiment, participants (N=33) first learned cartoon scenes depicting real-world locations with specific target objects. Twenty-four hours later, during fMRI scanning, they were shown the learned scenes with the target objects occluded and asked to mentally retrieve the missing objects. Analyses include layer-specific decoding and representational similarity analysis to identify distinct memory signals across cortical layers in early visual cortex and MTL subregions. Preliminary results showed that context information could be decoded from the hippocampal subiculum, whereas feedback, memory-related information could be decoded from the deep layers of the early visual cortex and hippocampal CA2/3. These findings provide new insight into how perceptual predictions arise from interactions between sensory input and memory-based representations.

While it is well established that emotional states influence decision-making, the reverse relationship, i.e., how the availability of choice modulates emotional states and corresponding autonomic responses, remains less well understood. The current study tried to bridge this gap and examined how active choice and temporal uncertainty jointly shape affective responses to rewarding and threatening events.

Participants completed a choice/no-choice task in which stimuli were presented at varying levels of temporal uncertainty. In choice trials, participants selected whether the trial outcome appeared predictably at a fixed timepoint or unpredictably within a time window, while in no-choice trials, they were confronted with only one of the options. Stimuli were either aversive (slightly painful electro-tactile stimulation) or appetitive (monetary reward). Heart rate, skin conductance, pupil dilation and gaze behavior were recorded alongside valence ratings.

Analysis of choice patterns indicates that participants did indeed select the uncertain option more frequently in reward trials, consistent with our hypothesis. Skin conductance and pupil size increased during threat and reward anticipation while heart rate showed a deceleration. However, these responses were not modulated by choice availability and only showed weak differences between threatening and rewarding conditions. In sum, the current experimental paradigm allows for elucidating whether and how agency modulates affective reactivity under temporal uncertainty. Furthermore, it shows promise in uncovering and explaining trait-like preferences in threat and reward processing.

Electroencephalography (EEG) is a powerful tool for investigating neural processes underlying cognition and neuropsychiatric disorders. Yet, variability in EEG preprocessing strategies restricts reproducibility and data integration across study sites and laboratories, particularly in larger research consortia. This poster introduces the CLEAN-EEG preprocessing pipeline, designed to standardize data processing and documentation across multiple sites. The CLEAN pipeline is implemented in MATLAB using EEGLAB. It comprises three modular, script-based stages: main preprocessing (including down-sampling, filtering, line noise removal, and channel interpolation), independent component analysis preparation and decomposition with flexible options for artifact rejection or neural component extraction, and component exclusion with support for automated classification and dipole fitting. Emphasis is placed on transparency through comprehensive logging and quality-control plotting, as well as on minimizing rank reduction to preserve data suitability for advanced analyses such as source localization and connectivity modeling. By providing clear, adaptable recommendations while ensuring detailed documentation of every step, the CLEAN pipeline aims to harmonize EEG preprocessing in large-scale, multi-center studies. This open and reproducible approach facilitates high throughput analyses, supports the training of researchers, and enables the rigorous integration of neurophysiological data across study sites, study designs, and populations.

The cone of gaze describes the range of gaze angles around direct eye contact within which a person is perceived as looking at the observer. Existing research on the cone of gaze in autism spectrum disorder (ASD) is inconclusive, calling for a closer examination of the underlying perceptual processes. In this study, 37 autistic adults and 39 neurotypical controls matched for age and IQ completed a computerized gaze judgment task while eye movements were recorded. Participants judged whether photographed individuals — whose gaze was varied in 1° steps from −10° to +10° horizontally, across 176 trials (8 blocks; 2 stimulus genders × 2 mirror versions × 2 repetitions × 22 gaze angles)— were looking directly at them. Points of subjective equality (PSEs) were calculated for each participant via logistic regression on their judgment data, to quantify the individual boundary of the cone of gaze. Group comparison of PSEs revealed no significant difference, indicating that the perceptual threshold for direct gaze detection is preserved in ASD. Multilevel modelling (MLM) of reaction times and eye-tracking data showed that autistic participants had significantly longer reaction times and greater viewing times on the eye region of the stimuli. This pattern of results suggests that while the underlying perceptual mechanism is intact, autistic individuals may rely on more effortful or compensatory visual strategies when processing social gaze, with implications for our understanding of social cognition in ASD.

Background:

Anxiety disorders are highly prevalent, yet only ~40% of patients show substantial improvement following cognitive behavioral therapy (CBT). Resting-state functional connectivity (rsFC), measured via rsfMRI, has been proposed as a biomarker of CBT-related changes, particularly in salience, cognitive-control, and interoceptive networks. However, large-scale transdiagnostic evidence across anxiety subtypes remains limited.

Methods:

rsfMRI data were analyzed (8 min) from the multi-center PROTECT-AD trial in patients with anxiety disorders (n = 195) and healthy controls (n = 95) at two time points (pre/post CBT; controls untreated) using a mixed repeated-measures design. ROI-to-ROI analyses focused on regions implicated in anxiety-related networks.

Results:

Patients showed reduced overall connectivity in the right posterior insula post-CBT relative to controls (T(279) = -3.89, pFDR = 0.000127). This effect was driven by the PD/AG subgroup (T(277) = -3.83, pFDR = 0.000157) and was not significant in SAD or SP

Edge-wise analyses revealed decreased connectivity in three ROI pairs in the SAD subgroup (vs. controls): right dlPFC–left amygdala (T(277) = -3.96), right dlPFC–right posterior insula (T(277) = -3.87), and right dlPFC–left posterior insula (T(277) = -3.76; all pFDR < 0.05.

Conclusions:

CBT-related rsFC changes differ across anxiety subtypes, with PD/AG showing alterations in insula-centered connectivity and SAD showing reduced coupling between cognitive-control and affective/interoceptive regions. These findings support disorder-specific neural mechanisms of treatment response.

The ability to exert goal-directed behavior is crucial for everyday life. Research on action control has focused on this through the investigation of perception-action links. A recent account is the Binding and Retrieval in Action Control (BRAC; Frings et al., 2020) framework, which suggests that the features of the stimulus, the response and its effect are bound together in an event file. Importantly, the framework stresses the relevance of the immediate past, as previous perception-action links have a direct behavioral impact on current action control processes. Previous research of the time frames preceding stimulus presentation has shown that processes in intermittent brain states interact with processes during action control through attention allocation and a preallocation of cognitive control. Our previous study showed that cognitive flexibility depends on previous re-iterant brain activity in the theta (4-7 Hz) and alpha (8-12 Hz) frequency bands prior to the onset of the stimulus. In order to further delineate the temporal succession and interaction of recurring brain states, hidden Markov modeling (HMM) is applied to source-reconstructed EEG data of N = 41 subjects in a task switching paradigm. Preliminary data of these analyses is presented. Given the conceptual similarity of the immediate past and Markovian processes, where the current state is determined only by the previous, HMMs might be able to disentangle states of retrieval and (re-)binding.

Actively excluding unwanted memories from awareness, or retrieval suppression, supports mental well-being. Yet how mood states and personality traits influence this process remains unclear. This study uses the Think/No-Think (TNT) paradigm to investigate how suppression of emotional memories is modulated by trait anxiety, state anxiety, subclinical depression, and stress.

Twenty-one participants learned cue–target word pairs of varying emotional valence (neutral, negative, positive) and later either retrieved or suppressed specific targets, followed by the final recall tests. For the behavioral analysis, the final recall test showed significant suppression (reflected in lower recall of to-be-suppressed items), with reliably greater effects for emotional than for neutral items.

As predicted, EEG markers of cognitive control (frontocentral N450 peak) and conscious recollection (left parietal LPP wave, 500–900 ms) showed significant TNT condition effects. No-Think trials elicited a reliably more negative N450 amplitude and shorter latency, with emotional items showing reliably longer latencies compared to neutral items. The LPP mean amplitude was significantly higher in the Think condition.

General linear models and Pearson’s correlations revealed that suppression-induced forgetting depends on emotional valence, and is shaped by negative affect & trait anxiety, and their influence on neural activity.

Anhedonia, a reduced ability to experience pleasure, is a central feature of depression and is associated with dysfunctions of the neural reward system, including blunted responses to positive stimuli and reduced stress-buffering effects of reward. Motivational anhedonia, in particular, is characterized by a reduced subjective valuation of rewards that require cognitive or physical effort, which in turn leads to diminished engagement in goal-directed behavior. Despite advances in behavioral and animal research, it remains unclear whether these motivational and neural deficits can be reliably measured in humans using effort-reward decision-making tasks, especially with respect to effort-dependent modulation of electrophysiological markers. Bridging this gap is essential for advancing a mechanistic account of altered reward and motivational processing in anhedonia. To address this, we developed a novel paradigm to examine event-related potentials (ERPs) — particularly the Reward Positivity (RewP) — across systematically varied reward magnitudes and effort demands. This allows us to directly compare neural reward responses across distinct motivational contexts and to investigate how prior effort shapes the encoding of objectively equivalent rewards. Our ongoing research further explores how individual differences in anhedonia modulate these neural signatures and effort-based decision-making. At the conference, we will present the paradigm, discuss its utility for EEG research, and share pilot data illustrating how effort-dependent reward processing can be captured electrophysiologically as a window into motivational anhedonia.

Sensory input though inherently noisy can be more or less reliable. Less reliable input, like the Necker cube, induces perceptual uncertainty, while reliable input, like disambiguated variants thereof, induces perceptual certainty. In EEG, this is reflected in low (uncertainty) and high (certainty) ERP (event related potential) amplitudes (“ERP Uncertainty Effects”). These findings are based on experimental paradigms in which only one stimulus variant was shown repeatedly with fixed blank screen ISI (interstimulus intervals) per condition. This leads to high predictability of induced perceptual (un)certainty and timing within conditions. Here, we tested the yet unknown influence of predictability of both stimulus ambiguity and timing on the ERP Uncertainty Effects.

In Experiment 1, we removed predictability of stimulus ambiguity by presenting ambiguous and disambiguated Necker lattices in random order but with fixed ISI of 0.4s. In Experiment 2, we removed predictability of stimulus timing by showing lattices with three ISI in jittered and non-jittered conditions (0.3s – jitter: ±0.01s, 1s – jitter: ±0.1s and 3s – jitter: ±0.2s). Both stimulus variants were shown in separate conditions ensuring predictability of stimulus ambiguity.

The ERP Uncertainty Effects were absent with randomization of stimulus ambiguity. In contrast, modulating stimulus timing did not affect the ERP Uncertainty Effects. Neither ISI length nor the presence or absence of a temporal jitter had an influence.

Predictability of stimulus ambiguity is a necessary precondition for perceptual certainty estimation. When predictability of stimulus ambiguity is given, however, it is robust to modulations of ISI duration and temporal predictability.

Resolving ambiguous pronouns is a hallmark of human language comprehension that requires integrating syntactic structure with commonsense world knowledge. Winograd Schemas isolate this capacity: each sentence pair uses a single-word substitution to shift the correct referent of a syntactically ambiguous pronoun, ensuring that resolution cannot rely on surface-level heuristics but demands genuine reasoning. Originally developed as an AI benchmark, these controlled sentence pairs offer untapped potential for cognitive psychology by providing items with a definitive correct answer, a property most anaphora resolution paradigms lack.

We administered 240 German Winograd sentences to 241 adults in a timed online paradigm and modeled trial-level accuracy using generalized linear mixed-effects models. Working memory capacity and fluid intelligence each independently predicted correct resolution, consistent with capacity-constrained accounts in which maintaining candidate antecedents and reasoning over relational structure draws on limited cognitive resources. Verbal fluency, subclinical autistic traits, and schizotypal traits showed no effects, suggesting that resolution depends primarily on executive and reasoning processes rather than on variation in commonsense knowledge. Item analysis revealed substantial difficulty heterogeneity, providing empirical selection criteria for future use.

To situate human performance within a broader computational framework, we benchmark large language models spanning different families, sizes, and reasoning capabilities on the same items, enabling direct item-level comparison of human and machine resolution. The validated schemas and behavioral findings now inform an ongoing fMRI study that, to our knowledge, constitutes the first attempt to decode successful pronoun resolution from neural activity.

Resting heart rate (HR) and heart rate variability (HRV) are established correlates of emotion regulation difficulties and internalizing symptoms such as depression and anxiety, making cardiac activity a promising transdiagnostic marker. Yet research on HR and HRV in transdiagnostic clinical samples — especially during cognitive demand — remains sparse. The present study examined HR and HRV changes from rest to cognitive effort in patients with internalizing disorders and their associations with transdiagnostic symptom dimensions. Three hundred patients diagnosed with obsessive-compulsive disorder, an anxiety disorder, or a depressive disorder, and 60 healthy controls completed a four-minute resting baseline followed by a gambling task, with continuous ECG recording. Symptom questionnaires assessed anhedonia, anxiety, depression, and compulsivity. We hypothesized that patients would show lower resting HRV and higher HR than controls, and that greater symptom severity across all four dimensions would be associated with reduced HRV. Additionally, we predicted a cognitive load effect, with HRV being lower during the task than at rest, reflecting increased autonomic demands. Findings indicate significant effects of both condition (rest vs. cognitive effort) and group on HRV, while HR showed no comparable effects. The relationship between HRV and cognitive effort will be discussed through a transdiagnostic lens. These results suggest that HRV may capture cognitive alterations cutting across diagnostic boundaries, supporting its utility as a transdiagnostic biomarker with potential applications in clinical assessment and targeted interventions.

According to Aaron Beck’s cognitive model, depressed individuals show a negative bias. Our goal was to investigate whether early face processing is altered according to this bias in persons with high depressive symptoms, since faces provide important cues for social interaction. Our participants conducted two facial oddball tasks where they saw a stream of happy, neutral, and sad faces, with a total of 300 faces per task. Unknown to the participants, 80% of the faces were in one gender, while 20% were in the other gender. In a previous study where participants rated the valence of each face via a button press, we found that participants with high depressive symptoms showed larger N170 amplitudes after rare sad faces, while healthy participants showed larger N170 amplitudes after rare happy and neutral faces, compared to their frequent counterparts. In the current study, we tried to replicate this finding in a German sample and added a condition where participants rated the gender, instead of the valence, of each face. Participants filled in a prescreening questionnaire and were selected for the study when they showed either low or high depressive symptoms. We present results on the N170 component, which is especially sensitive to face processing and presumably reflects the first time point where stimuli are reliably detected as faces by the brain, and discuss these results in the light of existing literature.

The relevance of eating behavior to both physical and mental health underscores the need to understand its regulation. Gastric interoceptive sensitivity, i.e. the perceptual threshold of satiation, is an important physiological marker of eating termination, while stress represents another relevant contributing factor. However, the effect of stress on the coupling of gastric interoceptive sensitivity and food intake remains poorly understood.

This study assessed gastric interoceptive sensitivity in a fasted state and self-determined snack consumption following a standard meal, both behaviorally, under high- and low-stress conditions. 61 healthy adults participated in a repeated-measures crossover design. Gastric interoceptive sensitivity was measured using the Magic Table (MT), in which participants consumed yogurt from a self-refilling bowl until perceiving satiation and fullness.

Multilevel analyses revealed that satiation and fullness perception thresholds were significantly associated with the amount of free snack consumption. Under stress, however, this association was no longer significant for satiation perception, while fullness perception remained significantly related to snack intake. No main effect of stress on total snack intake was observed.

These findings support the assumption that gastric interoceptive signals guide snack termination. Under stress, however, the influence of early satiation signals on snacking regulation appears to be attenuated, potentially reflecting reduced integration of gastric signals into food intake decisions. Future research on interventions targeting dysregulated eating may focus on strengthening the integration of satiation signals into eating regulation, particularly under stressful conditions.

Violence, danger, and death are pervasive themes in art, media, and storytelling, yet we know surprisingly little about the emotional processes underlying the fascination with such content. Research on morbid curiosity (the motivation to seek out information about threatening or macabre phenomena) has relied primarily on self-report and behavioral choice measures. To date, no study has examined whether individual differences in morbid curiosity are associated with distinct physiological responses during the anticipation and viewing of negative stimuli. We address this gap using a within-subjects paradigm combining active choice behavior with peripheral psychophysiology. Participants complete trait measures of morbid curiosity, epistemic curiosity, and benign masochism, followed by a laboratory session in which they view short negative and positive video clips preceded by verbal descriptions. In one block, participants actively choose between a negative and a matched positive video; in another, they select from unlabeled options, encountering predominantly negative content. Facial electromyography (corrugator supercilii, zygomaticus major, levator labii) captures valence-related facial affect, and electrodermal activity indexes autonomic arousal, both extracted separately for anticipatory and stimulus-evoked phases. This study constitutes the exploratory stage of a two-study cross-validation project: physiological correlates identified here will be tested confirmatorily in an independent sample. Early results will be presented at the poster, addressing whether morbid curiosity (1) is reflected in choice behavior and subjective ratings, (2) whether it predicts anticipatory and stimulus-evoked physiological responses, and (3) whether actively choosing negative content modulates these responses differently than incidental exposure.

Frustration is a key driver of aggression, but studying this relationship in controlled settings remains methodologically challenging. To address this, we developed a virtual grid aggression task to manipulate frustrative non-reward in form of violating a pre-existing reward expectation. The present study aims to examine the hypothesis that frustrative non-reward increases anger and aggressive behavior.

In a preregistered online study, participants (n=48) moved through virtual grids to reach a target, with the option to "shoot" a computer agent. Grids had either a target field only for the participant or only for the computer agent (no-target grids). Frustration was induced by increasing the frequency of no-target grids across three blocks (50% in trials 1–8 vs. 75% in trials 9–20). Data was analyzed using multi-level models with random intercepts for participants.

As expected, anger (scale: 0-10) was generally low but significantly higher after no-target grids (p<.001, M=2.05 vs. M=1.64) and during phases with a 75% no-target rate (p=.010, M=2.00 vs. M=1.77), with no significant interaction between these predictors. Behaviorally, the shooting rate was significantly higher following no-target grids (p<.001, M=43.32% vs. M=19.54%), though no main effect of target rate or interaction was observed.

These findings underline the critical role of frustration in eliciting both emotional (anger) and behavioral (aggression) responses. The study demonstrates that virtual grid paradigms provide a standardized and effective tool for investigating these processes. The generalizability of the observed aggressive behavior and future strategies to overcome methodological limitations will be discussed at the conference.

Stress habituation is often assessed using changes in reactivity scores, but this approach may miss dynamic reorganizations of underlying stress systems. Here, we present a network-based approach that conceptualizes habituation as dynamic reorganization rather than mere attenuation of stress responses.

We analyzed data from N=120 participants who completed the Trier Social Stress Test (TSST) on two days. In line with prior research, the TSST increased subjective stress, salivary cortisol concentrations, alpha-amylase activity (sAA), heart rate, and blood pressure, which were each assessed six times per test day. On average, stress responses habituated on Day 2 except for sAA. To obtain a better understanding of network (re)-organization between these variables, separate multilevel vector autoregressive models were estimated for both days to disentangle temporal networks (lagged within-person effects over time), contemporaneous networks (instantaneous within-person associations), and between-person networks.

Global network strength did not differ significantly between TSST1 and TSST2 for the temporal network (Δ=-0.326, 95%-CI[-0.995, 0.413]) or the between-person network (Δ=-0.065, 95%-CI[-0.946, 0.786]). By contrast, the contemporaneous network was stronger in TSST1 than in TSST2 (Δ=0.515, 95%-CI[0.180, 0.851]). This pattern may indicate increased within-time coupling between stress markers during the initial stress exposure and comparatively reduced coupling upon re-exposure. Descriptively, differences were most apparent in associations involving subjective stress and cardiovascular markers.

Overall, this work introduces a novel framework for investigating stress habituation as dynamic network reorganization. The findings should be considered exploratory and are intended to stimulate discussion on how to conceptualize habituation.

Background:

Adaptive behavior in uncertain and changing environments relies on forming and updating internal beliefs about latent states, a process called dynamic belief updating (DynBU). Dysfunctions in DynBU have been implicated in psychiatric disorders, particularly anxiety disorders (ADs), which involve biased learning, heightened threat expectations, and impaired adaptation. Evidence on DynBU in clinically diagnosed adult AD patients is limited, motivating computational modeling combined with fMRI to investigate underlying mechanisms.

Methods:

Adult patients with ADs (n ≈ 30) and matched healthy controls (n ≈ 30) completed the Predator Task, a predictive inference paradigm in which participants predict predator attack angles in a threatening virtual environment. Behavioral data were modeled using a reduced Bayesian framework to derive trial-wise learning rates and relative uncertainty. Model-derived regressors will be applied in task-based fMRI analyses. The task was administered in a fixed order to avoid spillover effects. fMRI data are currently being analyzed.

Results:

Preliminary analyses indicate no significant differences between patients and controls in task performance or computational parameters. Mean estimation errors were comparable across groups, and model-derived learning rates did not differ significantly (two-sided t-tests: fixed learning rate p = 0.165; adaptive learning rate p = 0.125), suggesting similar dynamic belief updating at this stage.

Conclusions:

Initial findings suggest adult AD patients may exhibit intact behavioral dynamic belief updating, with ongoing fMRI analyses potentially revealing neural differences. These data demonstrate the feasibility of applying computational modeling and neuroimaging to study belief updating in adult anxiety disorders.

Zahlreiche Studien belegen, dass transgender Personen im Vergleich zu cisgender Personen häufiger von psychischen Belastungen wie Depressionen, Angststörungen und Stress betroffen sind. Ein zentraler Bestandteil der medizinischen Transition ist die geschlechtsangleichende Hormontherapie (gender-affirming hormone therapy, GAHT), die von vielen transgender Individuen in Anspruch genommen wird. Während die bestehende Literatur zeigt, dass die GAHT langfristig die psychische Gesundheit verbessert, sind die genauen zeitlichen Verläufe und spezifischen Effekte insbesondere in der frühen Phase der Behandlung bislang unzureichend geklärt. Ziel der vorliegenden Studie war es, transgender Personen während der ersten sechs Monate der GAHT über fünf Messzeitpunkte prospektiv zu begleiten, um die anfänglichen Effekte der GAHT auf die psychische Gesundheit systematisch zu untersuchen. An der Studie nahmen 74 transgender Personen (48 trans Frauen) sowie 61 cisgender Personen (26 cis Frauen) teil. Die Studienteilnehmenden füllten verschiedene Fragebögen zur Erhebung der psychischen Belastung (u.a. Depression, Angst, Stress), der Geschlechtskongruenz und der Lebenszufriedenheit wiederholt aus. Die Daten wurden mittels eines linear mixed-effects Models ausgewertet. Unsere vorläufigen Ergebnisse zeigen nach sechs Monaten GAHT eine signifikante Reduktion der depressiven Symptomatik und der Ängstlichkeit sowie eine signifikante Verbesserung der Lebenszufriedenheit, der selbst wahrgenommenen Geschlechtskongruenz und der Körperzufriedenheit. Interessanterweise weisen trans Frauen eine stärkere Verbesserung der psychischen Belastungen (Depression, Angst, Stress) auf als trans Männer, wohingegen Geschlechtskongruenz und Körperzufriedenheit gleichermaßen steigen. Diese Befunde zu bereits früh einsetzenden Effekten der GAHT auf die psychische Gesundheit unterstreichen die Bedeutung der medizinischen Transition, jedoch gilt es, die unterschiedlichen Verläufe zwischen trans Männern und trans Frauen in dieser frühen Phase zu berücksichtigen.

Visual search is influenced not only by stimulus salience and goal-driven factors but also by prior experience with environmental regularities. Locations that frequently contain distractors can acquire a reduced priority weight within the spatial priority map, reflecting learned adjustments in attentional priority. The present study examined (1) whether priority weights adapt to gradually changing distractor regularities, and (2) how previously reduced priority weights influence subsequent prioritization when the same location becomes target-relevant. Participants performed an additional singleton task in which one location more frequently contained a color distractor. During acquisition, distractor probability ratios either increased (2:1 → 4:1 → 8:1; ascending group) or decreased (8:1 → 4:1 → 2:1; descending group). We hypothesized that distractor location learning (DLL) would emerge in both groups and differentially impact subsequent target location learning (TLL) at the same location. Both groups exhibited DLL, indicating that high-probability distractor locations were assigned a reduced priority weight even under weaker ratios. Critically, when this location became target-relevant, a significant TLL effect emerged, indicating increased priority weight. The magnitude of this effect did not differ significantly between groups, although descriptively it was larger in the descending group. These findings suggest that priority weights in the spatial priority map are flexibly adjusted based on experience. While locations can acquire reduced priority weights through DLL, these weights can subsequently shift toward increased priority when task demands change, indicating a flexible but experience-sensitive system.

Cognitive impairments represent a core feature of Schizophrenia Spectrum Disorders (SSD), persist across the lifespan, and are primary predictors of functional outcome. However, routine cognitive monitoring is limited by the length of neuropsychological tests, the need for trained examiners, and vulnerability to practice effects with repeated administration. Here, we investigate speech as a natural and accessible surrogate for neuropsychological testing in SSD using a longitudinal design.

Participants were recruited from a longitudinal, multicenter cohort study conducted in the Netherlands (2017–2026), including 301 individuals with SSD who contributed a total of 380 measurements. All participants completed the Brief Assessment of Cognition in Schizophrenia (BACS), and speech was recorded during a semi‑standardized interview.

Outcomes from five cognitive subtests, along with acoustic, syntactic, and lexical–semantic features, were condensed into five interpretable components using parallel analysis/component analysis. Associations between cognition and speech were assessed using linear mixed-effects models, adjusted for demographic variables and evaluated using AIC.

Models predicted the subtests of verbal fluency, working memory, and processing speed with good cross‑validated performance (MAE 2.85 and 3.22), corresponding to deviations of approximately 7%, well below the ±20% threshold. Speech features predicted cognition for the subtests of verbal fluency, working memory, and processing speed (ΔAIC = 9.8, 5.4, and 5.3), with syntactic complexity and lexical content emerging as the strongest predictors.

Speech‑based modeling thus achieves clinically meaningful, symptom-level prediction, providing a strong foundation for scalable, low‑burden tools for real‑time cognitive monitoring in psychosis.

During a conversation interpersonal synchrony is observed in different modalities, which enhances social attunement. The degree to which individuals share similar personality traits may influence the dynamics of interaction and in turn interpersonal synchrony. Thus, in the current study, we explored the potential effect of similarity of personality traits on physiological alignment operationalised as heart rate synchrony within conversational dyads.

To this end, we collected electrocardiogram (ECG) and speech data during task-based dyadic conversations in an ecologically valid appointment-scheduling paradigm with cooperation and competition conditions. Personality traits were assessed using the self-report NEO-FFI (German) questionnaire. We conducted a cross-correlation analysis of heart rate across different time windows to examine heart rate synchrony. The data collection is ongoing (target: 20 dyads) and the present analysis is based on three dyads.

Our preliminary findings indicate that the dyads with greater similarity with respect to personality traits such as neuroticism and extraversion showed higher heart rate synchrony. The observations indicate that the similarity in personality traits between the speakers may play an important role in the physiological synchrony, shaping interpersonal interaction and conversational dynamics. Planned analyses will also examine linguistic and acoustic features of speech, as well as turn taking dynamics during these interactions.

Background: Adolescence is characterized by increasing demands on independent emotion regulation (ER), which could be supported by the perception and processing of bodily signals. This study investigated whether interoceptive accuracy (IAcc) is associated with habitual reappraisal (aim I), and whether both variables jointly predict adolescents’ affective responses to social exclusion (aim II).

Method: For aim I, n = 77 adolescents (mean age = 14.47 years, range = 12 – 17) performed the heartbeat counting task for measuring IAcc and reported their habitual usage of Cognitive Reappraisal in the Emotion Regulation Questionnaire. For aim II, a subsample of n = 32 (mean age = 14.19 years, range = 12 – 16) completed three rounds of the Cyberball paradigm (initial inclusion, exclusion, re-inclusion) and provided mood ratings after each round.

Results: Higher IAcc was significantly associated with greater habitual reappraisal. In the subsample social exclusion evoked negative affect, while re-inclusion led to a positive affective recovery. Neither IAcc, reappraisal, nor their interaction predicted initial affective reactivity. In contrast, the interaction predicted affective recovery: higher IAcc was associated with smaller recovery when habitual reappraisal scores were low (-1 SD), but with greater recovery when reappraisal scores were high (+1 SD).

Discussion: On a trait level, we found a linear relationship between IAcc and ER. On a task level, bodily perception appears to be not uniformly favorable. Greater access to bodily signals might facilitate recovery when adolescents are able to adaptively reinterpret social experiences, but it appears to hinder recovery when such regulatory resources are limited.

CD38 is known to affect oxytocin secretion with possible consequences for socioemotional processes such as reactivity to both stressful and positive experiences. Whereas the C- and A-allele (CD38, rs3796863) have each been reported as risk-alleles for different psychological disorders on a genetic level, the A-Allele is typically associated with a higher social sensitivity. While it has been linked to higher responses to interpersonal stress it has also been connected to higher positive outcomes like gratitude and satisfaction in relationships.

To further investigate the reported gene-environment interaction, its direction and possible interindividual differences we measured 55 life events (LE) typically relevant for student samples, the CD38 rs3796863 genotype, and their relationship with life satisfaction in N=1080 (N=870 females) undergraduates. To account for sex-differences in life satisfaction and possible influences on oxytocin-linked socioemotional processes, participant sex was added as a variable.

Not surprisingly, participants who reported predominantly positive LE experience greater life satisfaction than those who reported predominantly negative LE.

However, in contrast to the expected gene-environment-interaction, we observed a three-way-interaction with participant sex, which suggests different sensitivities to negative LE for men (CC-genotype) and women (A-allele), especially for LE that occurred after childhood (over age of 13y).

Additionally, men carrying the CC-genotype reported the lowest life satisfaction, regardless of LE.

The results indicate that the interaction of the CD38 rs3796863 and LE is more complex than anticipated and might be different depending on interindividual and situational characteristics.

Possible explanations, mechanisms and implications are discussed.

Background: Increasing numbers of young adults use social media platforms such as Instagram and TikTok, and their potential association with mental health is a central topic in public and scientific debate. This study aims to investigate the link between social media and an increased risk of depressive symptoms, body dissatisfaction, low self-esteem, and addictive use patterns.

Methods: We are currently conducting an anonymous online survey in Switzerland targeting 400 young participants aged 14-25 years. The survey includes validated instruments assessing addictive social media use patterns, depressive symptoms, body appreciation, and self-esteem. In addition, detailed measures of duration, content, and context of social media use are collected (including the tendency to use social media to cope with negative emotions), as well as relevant control variables such as age, gender, socioeconomic status.

Results: Preliminary findings indicate positive associations between longer daily social media use and depressive symptoms, with small to moderate effect sizes. Further results on usage patterns, as well as the roles of self-esteem, body appreciation, and emotion-driven usage, will be presented.

Discussion: The findings contribute to a better understanding of how social media usage patterns and individual characteristics may be associated with increased vulnerability to depressive symptoms, with implications for prevention and clinical practice.

Open-label placebo (OLP) treatments show beneficial effects on various distress symptoms, however, studies on physiological measures remain limited. Furthermore, the role of personality traits in OLP responses, and their interaction with treatment expectations, has not yet been systematically examined.

This randomized controlled trial investigated the effect of an OLP intervention on psychological distress, hair cortisol concentrations (HCC), and salivary cortisol (sC) in university students preparing for oral exams. Here, the moderating role of dispositional optimism on OLP effects and its interaction with treatment expectations has been explored.

202 participants were randomly assigned to an OLP or control group. The OLP group took two placebo pills daily for four weeks leading up to the exam. Psychological distress measures (negative affect, test anxiety, subjective stress) were repeatedly assessed until the exam. HCC was measured pre- and post-intervention, while sC was measured on the exam day.

Results indicate that OLPs significantly reduced psychological distress and HCC compared to controls. Optimism moderated the OLP effect on HCC, with less optimistic individuals demonstrating the strongest reduction, independent of expectation. In contrast, no group differences were observed in sC levels, nor were these influenced by moderators. Optimism did not moderate OLP effects on psychological distress within the intervention month. However, a significant three-way interaction between group, optimism, and expectation emerged. Interestingly, a similar pattern was observed on the exam day. These findings suggest that OLPs alleviate the psychophysiological impact of real-life stressor, with optimism and expectation differentially shaping psychological and physiological OLP responses.

Premenstrual Dysphoric Disorder (PMDD) is an affective disorder thought to arise from an altered sensitivity to normal hormonal fluctuations during menstrual cycle. While serotonergic dysregulation is implicated, around 40% of patients do not respond to SSRIs, suggesting additional neurobiological mechanisms. PMDD shows symptom overlap with ADHD, a condition associated with dopaminergic dysregulation. Notably, 45–48% of women with ADHD meet criteria for PMDD. This raises the question of dopaminergic involvement and whether co-occurrence reflects premenstrual exacerbation (PME) or a distinct comorbid condition. Preliminary findings from an ongoing longitudinal study are reported. Using a four-time-point design (mid-follicular, peri-ovulatory, mid-luteal, premenstrual), N = 46 women were examined: controls (n = 17), PMDD (n = 9), ADHD (n = 6), and ADHD+PMDD (n = 14). Daily hormone measurements of Estrone-3-glucuronide (E3G), Pregnanediol glucuronide (PdG) and Luteinizing hormone (LH), state anxiety (STAI-S), cognitive failures (CFQ), and reward sensitivity (PVSS-21) were assessed. We used linear mixed models with hormone levels and intraindividual variability as covariates. PMDD showed elevated state anxiety across cycle phases with a significant peri-ovulatory reduction. Cognitive failures peaked mid-follicular and decreased by ovulation. ADHD+PMDD showed reduced reward sensitivity with a peri-ovulatory increase. ADHD showed no cycle-dependent variation. Diagnostic group, rather than hormones, predicted cycle effects. Findings suggest distinct mechanisms across groups. PMDD shows cycle-dependent affective and cognitive changes, whereas ADHD+PMDD shows a dissociable reward-related pattern. This divergence argues against a purely PME-based explanation and suggests distinct vulnerability profiles. The reward-specific pattern in ADHD+PMDD, absent in ADHD, points to an interaction between dopaminergic vulnerability and cycle-dependent hormonal sensitivity.

Emotion regulation (ER) is fundamental to mental health but has been shown to be affected by stress, with noradrenaline (NA) as a potential modulator. We used transcutaneous auricular vagus nerve stimulation (taVNS) to investigate the effect of NA on cognitive ER.

Sixty healthy participants completed an ER paradigm, while receiving taVNS or sham stimulation in a randomized cross-over within-subject design. ER performance was assessed via arousal, valence, and success ratings, and phasic, task-evoked pupil dilation. Tonic pupil diameter (PD) and salivary alpha-amylase served to index NA activity. Cortisol, heart rate, and heart rate variability were measured as additional stress markers.

Ratings and pupillary responses (PR) confirmed successful emotion induction and ER. taVNS relative to sham increased tonic PD during rest and baseline PD during the ER paradigm, indicating enhanced NA activity, while alpha-amylase analyses are still ongoing. We found no significant effects of stimulation on arousal, valence or ER success ratings. However, in female participants, taVNS did enhance early PRs across all ER conditions and late PRs during reappraisal and distraction, suggesting higher cognitive regulatory effort and impaired ER performance, respectively.

Together, these findings indicate subtle sex-specific effects of taVNS on cognitive ER, revealing a potentially NA-mediated ER impairment in women. However, taVNS effects were only detectable in implicit pupillometric rather than in explicit self-report measures. Future studies are needed to explore whether higher stimulation intensities might result in effects on ER performance that go beyond the subconscious level and to further elucidate the role of noradrenergic mechanisms in ER.

Cognitive performance varies across older individuals, yet the neural basis of this variability is poorly understood. Different neuroimaging modalities provide complementary perspectives: structural MRI metrics capture brain integrity, while task-based fMRI (tfMRI) reflects dynamic neural modulations. The two modalities are often analyzed independently and univariately, rather than in a multimodal, multivariate whole-brain manner. Hence, their joint and unique contributions to cognition remain unclear.

To address this gap and empirically extend a prior conceptual framework (Mooraj et al., 2025), we decomposed shared and unique contributions of brain structure and function to behavior in 91 older adults measured twice across five years. Whole-brain, repeated-measure, multivariate partial least squares (PLS) analyses were carried out on grey matter volume (GMV), tfMRI, and n-back working-memory (WM) performance. We used two approaches, unimodal (each modality individually related to behavior), and multimodal (both modalities jointly related to behavior).

Using unimodal PLS analyses, both GMV-behavior and tfMRI-behavior relations were significant. Comparisons showed minimal spatial overlap between these brain maps, though these modalities shared some variance explained in WM accuracy. However, tfMRI explained substantial unique variance beyond GMV, and was the only modality to predict future WM performance. We further examined joint statistical and spatial associations using a multimodal multivariate approach to simultaneously relate GMV and tfMRI to WM. Supporting the unimodal findings, a latent multimodal brain-behavior component linked tfMRI to WM, with no concurrent GMV contribution in those regions. These findings highlight the value of a functionally-anchored approach for understanding individual differences in cognitive performance in older age.

Do we learn differently about the moral character of other people when we are stressed? The process of inferring other’s morality has recently been elucidated through the lens of Bayesian learning, wherein beliefs about others are dynamically updated in proportion to their uncertainty. We hypothesized that acute stress increases belief volatility – reflecting heightened uncertainty and flexible updating. In this preregistered study, participants were randomly assigned to either a psychosocial stress (N = 64) or control condition (N = 64), before predicting and observing moral decisions of two agents with differing moral characters. Bayesian computational modeling revealed that acute stress increased belief volatility, but specifically for good agents. Critically, this effect emerged during peak cortisol reactivity, and individual differences in cortisol responses predicted the magnitude of the volatility increase – converging evidence that hypothalamic–pituitary–adrenal (HPA) axis activation plays a key role in moral inference under stress. Stress-related increases in belief volatility could not be explained by shifts in participants’ own moral preferences, as acute stress had no impact on moral decision-making in an independent task. Our findings reveal a cognitive-computational mechanism through which stress shifts moral inference – an effect that may aid coping with uncertainty but also contribute to social dysfunction in stress-related psychopathologies.

This study investigated whether experimentally induced affective and cognitive states can be detected from gastric myoelectrical activity, a physiological signal that has received comparatively limited attention in state-detection research despite its important role within the gut–brain axis. In a laboratory-controlled within-subject design (N = 77), participants underwent fear, disgust, relaxation, and cognitive load (2-back) conditions while electrogastrogram (EGG) signals were recorded, and frequency-based gastric features were extracted and compared using repeated-measures ANOVAs. Most gastric features did not vary significantly across conditions; however, fear was associated with increased total dynamic gastric power and dynamic dominant frequency without shifts in dominant frequency, reflecting selective modulation of gastric amplitude during emotional arousal.

The Comet Toolbox (https://github.com/mibur1/comet) is an open-source Python toolbox for multiverse analysis in functional connectomics. It was developed to address a central challenge in network neuroscience: analytical results often depend on numerous interdependent methodological decisions, spanning preprocessing, connectivity estimation, network construction, and statistical modelling. In the absence of clear ground truth, many of these choices are equally defensible but can lead to substantially different results. Multiverse analysis addresses this problem by systematically evaluating results across all defensible analytical specifications. Comet integrates functional connectivity methods with graph-theoretical analyses in a modular and extensible framework. Analytical forking paths can be specified at any stage of the pipeline and are automatically expanded into executable scripts representing distinct analytical universes. The toolbox supports both graphical and script-based workflows and includes optimised implementations of computationally demanding network measures. Comet promotes reproducibility by explicitly documenting decision spaces and summarising results across universes. Beyond its current functionality, the framework is designed to support future extensions, including the integration of psychometric modelling into multiverse workflows. Comet thus provides a practical infrastructure for transparent robustness assessment and reproducible network neuroscience.

Post-exertional malaise is a central symptom in myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS). Beside physical triggers, patients report that cognitive demands and emotional involvement can as well lead to long-lasting symptom exacerbation. The current study aims to understand brain metabolic and functional connectivity changes induced by performing a 2-back working memory task during continuous functional magnetic resonance imaging (fMRI), embedded in between pre- and post-scans of resting state fMRI, single voxel magnetic resonance spectroscopy (1H-MRS) in the midcingulate cortex and whole-brain 31P-MRS imaging. In the pilot data, task performance during 2-back vs. 0-back condition is accompanied by increased hemodynamic responses in midcingulo-insular salience and lateral frontoparietal control networks. Metabolite concentrations in the midcingulate cortex show good spectral quality and 31P-MRSI enables measurement of adenosine-triphosphate and phosphocreatine concentrations. The results encourage spectroscopic and connectivity analyses localized to the midcingulate cortex and its brain network embedding, to understand neurobiological signatures following working memory demands. The analytical plan for a cross-sectional comparison between ME/CFS patients and control participants will be discussed, to investigate and understand critical physiological alterations induced by cognitive exertion.

Over the past decade, numerous studies have applied machine learning techniques to classify schizophrenia from Magnetic Resonance Imaging (MRI) data. However, a key question remains: do such classifiers capture neural signatures of vulnerability to schizophrenia, or do they primarily reflect consequences of the disorder (e.g., medication effects or chronic disease processes)? This distinction, along with improving the trustworthiness of the classifiers, is critical for evaluating their potential as prognostic tools. We trained a 3D Convolutional Neural Network using an Evidential Deep Learning (EDL) framework on structural brain MRI data from the SchizConnect dataset (289 schizophrenia cases, 351 controls), achieving 83% accuracy, an AUROC of 0.865 and meaningful uncertainty quantification on an independent generalization dataset. We applied the model to 49,297 individuals in the UK Biobank, generating continuous “brain-predicted schizophrenia” scores. Genome-wide association analyses based on these scores revealed a substantial SNP-based heritability of 22.6% (SE = 1.6%) and identified 22 genome-wide significant loci. Genetic correlation analyses showed a significant but modest overlap with schizophrenia (rg = 0.075, p = 0.009), alongside comparable correlations with other psychiatric (major depression and bipolar disorder), cognitive (intelligence), socioeconomic (educational attainment), and cardiometabolic traits (blood pressure). Together, these findings suggest that current MRI-based classifiers may capture a broad, nonspecific dimension of brain health rather than a disorder-specific vulnerability to schizophrenia. Importantly, our results demonstrate how combining EDL with imaging genetics can help disentangle the biological signals underlying model predictions, providing a framework to systematically evaluate and refine neuroimaging-based classifiers for clinical applications.

Phantom limb pain (PLP) is a debilitating consequence of limb amputation, affecting around 80 percent of amputees, with half of them suffering from chronic pain, without any known cure. Even though the mechanism and selectivity of PLP are not well understood, a clear bidirectional relationship between pain and mental health has been previously shown. Moreover, some studies have shown that cognitive flexibility may be associated with emergence and impact of chronic pain.

We have recruited 22 amputees with phantom limb pain, who have undergone a series of pain and mental health questionnaires that measured different aspects of PLP, as well as levels of PTSD, depression, anxiety and stress. We have also conducted a neuropsychological set shifting task that measured cognitive flexibility performance.

When looking at the relationship between cognitive flexibility performance and clinical measures, we have found that cognitive flexibility was significantly correlated with both pain measures, such as average PLP level last day and last week, as well as with mental health aspects like PTSD and depression. Moreover, no significant task performance difference between our patient cohort and healthy controls was found, indicating that PLP did not lower cognitive flexibility.

Although no directionality or causality can be determined, this is the first study to show positive association between pain and mental health levels in PLP with cognitive flexibility measures, suggesting a possible role of the ability to adapt and reassess in PLP and mental health comorbidities.

Cognitive models suggest that concrete and abstract concepts differ in how they are mentally represented. Specifically, the representations of concrete words and concepts are grounded more in sensorimotor experiences, while abstract concepts relate to a more diverse set of experiences or are learned entirely through language. Functional brain imaging studies support this, by showing different patterns of brain activation elicited during the processing of abstract and concrete words. We hypothesized that the different cognitive representations may be reflected in differences in the information theoretic complexity and in the dimensionality of brain signals elicited while processing concrete vs. abstract words. Using ERPs elicited by 230 concrete and 230 abstract words from a publicly shared dataset (N = 75; Dufau et al., 2015), we find increased entropy in multiple measures for abstract as compared to concrete words. In addition, the dimensionality of brain signals elicited by abstract words is also higher. We then conducted a time-resolved analysis and observed that brain signal complexity was increased at two circumscribed time points, specifically at the onset and at the offset of the well-established N400 ERP effect for abstract words. We interpret this finding as indicating a reconfiguration of cortical networks upon processing abstract words, and support this proposal by demonstrating that multiple graph theoretical measures of functional connectivity show characteristic changes restricted to the exact same time windows.

Introduction:

Chronic psychosocial stress is widespread in modern societies and can dysregulate neurobiological stress systems, increasing health problems. Importantly, stress can also be transmitted between individuals through empathic processes. While visual and auditory pathways of stress transmission have been widely studied, growing evidence suggests that chemosensory signals may also contribute. Human body odor changes under stress, and others can detect and respond to these changes. We investigate whether stress-related body odors facilitate empathic stress transmission between romantic partners.

Methods:

The study is ongoing and aims to recruit 80 romantic couples. Participants attend two laboratory sessions. In the first session, observers’ olfactory abilities are assessed and exercise sweat samples are collected from the target during a 15-minute jogging task. In the second session, one partner (target) undergoes the Trier Social Stress Test (TSST), while the other partner (observer) watches via live video. During observation, participants are exposed to either their partner’s stress sweat or exercise sweat. Physiological, behavioral, and subjective measures are collected, including cortisol (saliva), ECG, norepinephrine and epinephrine (blood), facial electromyography, emotion recognition, ratings, and questionnaires.

Results:

Data collection and analysis are currently ongoing. Preliminary findings indicate that the TSST reliably induces both physiological and subjective stress responses in participants. Initial analyses further suggest that observers resonate with the target’s stress in a congruent direction, showing corresponding physiological and subjective stress-related changes.

Sleep is thought to consolidate newly acquired memories through the reactivation of learning-related neural representations during non-rapid eye movement (NREM) sleep. This reactivation is thought to support neuroplasticity, enabling structural and functional brain changes that stabilise memories over time. However, how sleep reactivation contributes to overnight structural brain changes remains elusive. Here, we aim to investigate the relationship between memory reprocessing during sleep and learning-induced structural plasticity.

Fifty-six participants (age 18–34, M = 24.3) completed an image localisation learning task with scenes and object stimuli prior to a full night of sleep recorded using high-density 128-channel EEG. Structural brain imaging (T1W- and DW-MRI) was acquired before and after sleep in both a learning and a no-learning condition to capture rapid plasticity. From the EEG recordings, sleep architecture and oscillatory characteristics were extracted, and multivariate pattern analyses quantified the reinstatement of category-specific representations across wake and sleep.

Participants showed reliable improvements in memory retention across learning repetitions (all ps < .001). Multivariate pattern analyses of EEG data recorded during wakeful learning indicated successful decoding of stimulus category (p < .001), with decoding strength positively associated with task performance (p = .040). Preliminary analyses suggest that category-specific representations are reinstated during NREM sleep, with ongoing analyses examining their association with overnight memory retention and structural brain changes.

These findings offer early evidence that sleep reactivation of learned content may be linked to structural brain reorganisation, advancing our understanding of how sleep supports memory consolidation.

Motor imagery is a crucial cognitive process for action planning and motor learning, yet its neurophysiological substrates are not fully elucidated. While self-report questionnaires such as the Vividness of Movement Imagery Questionnaire (VMIQ) are widely used to assess imagery ability, it is unclear whether questionnaires reflect objective neural markers. In the current study, we investigated the relationship between resting-state EEG oscillations and VMIQ scores in healthy adults.

Fifty-two participants (age = 23.62 ± 4.05, 28 female) completed the VMIQ-2, followed by resting-state EEG recordings comprising 3-minute eyes-open and 3-minute eyes-closed conditions. Spectral power within the mu (8–12 Hz) and beta (13–30 Hz) frequency bands were extracted over the sensorimotor cortex, and then correlations and regressions were made to VMIQ scores.

We found that eyes-open relative beta power showed a consistent positive association with imagery ability. Beta power during eyes-open rest significantly correlated with VMIQ scores (r = 0.34, p = 0.014) and explained 11.4% of the variance in single-predictor regression. A multiple regression model including eyes-open mu and beta power was significant, with beta emerging as the only reliable predictor. Hierarchical regression further demonstrated that beta oscillations accounted for an additional 10.2% of variance beyond mu rhythms.

Our findings suggest that resting-state sensorimotor beta activity reflects a trait-like neural marker of motor imagery ability. Individuals with higher beta synchronization during eyes-open may possess more stable or efficient sensorimotor representations supporting imagery processes. The results highlight the potential of resting EEG as a biomarker for individual differences in motor imagery capacity.

Borderline personality disorder (BPD) is characterized by severe emotional and interpersonal instability, self-harm, suicidal ideation, and limited response rates to psychotherapy. Previous single-dose studies reported beneficial effects of intranasal oxytocin (OT) on social cognition and behaviour in BPD. In addition, its anxiolytic properties indicate a potential role in facilitating therapeutic alliance and thereby augmenting psychotherapy efficacy. However, evidence on prolonged OT administration as an adjunct to standardized psychotherapy in BPD remains scarce.

In a randomized, double-blind, placebo (PL)-controlled parallel-group design, we examined effects of an 8-week intranasal OT treatment (20 IU, three times daily) in 213 inpatients with BPD receiving dialectic behavioural therapy (DBT). The present analysis focusses on a subset of secondary outcomes, examining changes in anxiety and depressive symptoms. Anxiety was assessed using state and trait measures (VAS, BSI, AAS), while depressive symptom severity was measured with BDI and HDRS. Data were assessed at baseline, at several time points during treatment, and at four follow-ups up to one year after discharge and were analysed using hierarchical linear mixed models.

The intent-to-treat sample comprised 182 patients (DBT-OT: n=92; DBT-PL: n=90; mean age 27.83±7.79 years). DBT treatment reduced anxiety and depressive symptom severity overall. Importantly, OT-administration significantly enhanced treatment effects on all anxiety measures, while no additional effects were observed for depressive symptoms.

These findings identify intranasal OT as a promising adjunct to standardized psychotherapy by underscoring its anxiolytic effect pattern and highlight the potential of combined psychobiological treatment approaches in BPD and other psychiatric conditions associated with interpersonal difficulties.

The present study aims to examine the potential combined impact of sleep and retrieval practice on learning outcomes in school-aged children. Sleep is known to play a critical role in memory consolidation by strengthening neural representations formed during learning. Retrieval practice has been shown to enhance long-term retention more effectively than passive study. It is therefore assumed that the combination of adequate sleep and retrieval practice may optimize learning outcomes.

To investigate this assumption, the study firstly examines whether vocabulary achievement differs as a function of two distinct learning strategies, namely retrieval practice and repeated study. Therefore, learning conditions are randomly assigned within the sample. Secondly, learning performance of fourth-grade vs. sixth-grade students is compared. Learning performance is assessed by cued recall approximately 45 minutes after encoding (pretest), on the following day (posttest), and again after a delay of four to five weeks (follow-up).

Thirdly, sleep is assessed during the night between pretest and posttest using a multimethod approach, including a sleep log, a Fitbit Alta HR fitness tracker, and a Somnomedics actigraph.

In combination, the study aims to investigate whether sleep duration is positively associated with vocabulary learning performance and whether this relationship might be moderated by age and assigned learning strategy. The findings are expected to provide insights into developmental differences in sleep-related learning processes and to inform evidence-based educational practices. Preliminary results will be presented and discussed at the upcoming conference.

Understanding how leaders process discriminatory workplace interactions requires moving beyond self-report to capture implicit cognitive and emotional responses. We present a multimodal neurophysiological study examining how discriminatory workplace scenarios modulate affective responses. Here, we present self-reported affective responses and initial cardiovascular results.

Multimodal data (EEG, ECG, EOG, and eye‑tracking) were collected from 60 participants with leadership experience viewing short video‑based workplace scenarios depicting discriminatory or neutral interactions. The stimulus set was validated in independent online surveys prior to the laboratory study to ensure reliable differentiation between conditions. Following each video, participants provided trial‑by‑trial ratings of valence and arousal using the EmojiGrid, as well as evaluative judgements of perceived fairness and acceptability.

Preliminary subjective analyses revealed clear differentiation between conditions. Discriminatory scenarios were rated as significantly more unpleasant (M = −0.144, SD = 0.061) than neutral scenarios (M = 0.058, SD = 0.047) and elicited higher arousal (discriminatory: M = 0.072, SD = 0.089; neutral: M = −0.005, SD = 0.070). Paired‑samples tests confirmed robust condition effects for both valence, t(52) = −22.03, p < .001, and arousal, t(52) = 6.97, p < .001, replicating prior online validation results. Exploratory ECG analyses indicated a trend toward higher heart rate during discriminatory vs. neutral scenarios, t(52) = −1.93, p = .060, accompanied by lower MeanNN, t(52) = 1.89, p = .065.

These findings demonstrate that the paradigm reliably elicits differentiated emotional responses to workplace discrimination. Future neuropsychological analyses will examine neural responses during stimulus exposure to characterize underlying cognitive and affective processes.

Some degree of invariance to geometric transformations is vital for reading, enabling us to recognise letters and words at different sizes, with some tolerance to transformations like rotation and shearing. However, complete invariance to affine transformations would be suboptimal, as characters like <d>, <b>, <p>, and <q>, which differ chiefly in terms of mirroring and rotation, would be difficult or impossible to distinguish. Indeed, findings for typical readers show that, rather than early retinotopic representations of letters and words being discarded, such information is maintained even in later, more anterior ventral visual processing, alongside invariant representations. Moreover, readers with dyslexia commonly show difficulties in distinguishing between letter pairs like <d> and <b>, consistent with suboptimal invariance. Here, we apply a novel approach, drawing from optimal transport theory to calculate a representational continuum transitioning from retinotopy to invariant letter shapes. In simulations, we confirmed that the best letter recognition performance should be observed in a “goldilocks” zone, with a high degree of invariance, but preserving some transformation-sensitive information. We then model data from two preregistered EEG studies, estimating the representational spaces for letters in typical readers and participants with dyslexia. Rather than a gradual transition, both groups show a rapid switch from more retinotopic representations at ~100 ms to invariant shape representations at ~150 ms. Surprisingly, rather than a “goldilocks” zone of invariance, both typical and dyslexic readers show maximal alignment with either complete retinotopy or complete invariance, across the full timecourse. We discuss possible reasons for this finding.

Reading is a complex process involving the phonological and orthographic processing of visual word forms. Previous findings suggest that these processes are impaired in individuals with dyslexia. However, less is known about the earliest stages of orthographic processing which involve letter shape processing. In this preregistered study, we use an alphabetic decision task to investigate whether early visual representations of letters differ in 20 participants with developmental dyslexia and 20 matched controls. We explore group differences in brain-model alignment using a Bayesian, hierarchical representational similarity analysis of electroencephalography (EEG) data. We apply a computational framework based on optimal transport theory, in which letter shape similarity is calculated as an edit (Wasserstein) distance between the distributions of pixels in images of letters. Importantly, a variant of this edit distance (Gromov-Wasserstein distance) provides a measure of letter shape similarity that is more invariant to geometric information like retinal location – a core feature of visual letter recognition. We investigate whether prior results in typical readers can be replicated and whether brain-model alignment differs between dyslexic and typical readers. Preliminary results suggest that, compared to typical readers, neural representations of letters align less well with the computational models in the earliest stages of visual processing (pre-120 ms) and show less sustained sensitivity to invariant representations in later stages (post-300 ms). Data collection is still ongoing, but these preliminary results indicate that early neural representations of visual letters might differ in individuals with dyslexia.

Background: Previous research has linked auditory-verbal hallucinations in psychosis to stronger expectations of perceiving speech in noise. Studies suggest this may facilitate speech perception under ambiguous listening conditions in clinical and subclinical cohorts. As previous work mostly relied on subjective detection without verifying content recognition, this study investigates whether psychosis-proneness was associated not only with enhanced speech detection, but also with more accurate content recognition.

Methods: Participants completed two speech-in-noise tasks online using noise-vocoded words. In a linear version, each word was presented from lowest intelligibility and incrementally increased until correctly identified or maximum clarity was reached. After each presentation, participants indicated whether they perceived speech and a concrete word; if so, they reported the word. This allowed for assessing trial-level estimates of subjective speech detection and objective speech recognition thresholds. In a randomized version, words were presented at randomly assigned vocoding levels (1-15 channels), counterbalanced across trials. Time-reversed stimuli served as noise trials in both tasks.

Results: Signal detection analyses and Bayesian mixed-effects modelling showed that psychosis-proneness was associated with reduced speech detection thresholds, but not with improved word recognition. Detection thresholds were similarly reduced for noise trials, reflected in increased false alarm rates.

Discussion: The dissociation between subjective detection and objective recognition suggests that psychosis-proneness is linked to altered decisional processes rather than enhanced perceptual encoding. Our results highlight the importance of distinguishing perceptual sensitivity from response bias when studying speech perception in the psychosis spectrum, as this bias may increase false perceptions when sensory evidence is weak.

Prediction is a core feature of language, which is widely studied across research domains. The Munich Sentence (MuSe) database enhances reproducibility by providing sentence completion norms for 619 German sentences, including cloze probabilities and entropy estimates from up to 232 participants. Sentence completions were collected in two online studies in which participants completed sentence beginnings with a single-word response after either hearing (auditory sample, N = 133) or reading (visual sample, N = 98) the sentence beginning. All responses were manually preprocessed to correct typos and spelling mistakes and to label grammatical errors, proper nouns, and singular and plural variants of the same response. In addition to the sentence norms, we provide trial-level data with participant-level demographic information and subclinical autistic and schizotypal trait measures. Together with open access R-Scripts or our webtool, this allows for tailoring the cleaning and norming steps to integrate individual difference measures. For a subset of 479 sentence beginnings, the database also includes professional audio recordings of sentence beginnings which can be flexibly combined with 531 recordings of unique sentence-final words and implemented in auditory language paradigms. All material is freely accessible via the Open Science Framework (https://osf.io/ktnze/overview) and the MuSe webtool (https://munichsentencedatabase.franziskaknolle.com/).

Spatial distraction during driving from ubiquitous in-vehicle navigation systems, phone usage, or head-down displays shifts attention away from the road and utilizes neural resources needed for safe driving. To investigate the mechanisms of spatially distracted driving, we designed a virtual environment driving task with or without reckless drivers going to/from the side streets, and an off-road spatial attention secondary task to divert drivers’ attention spatially. Accidents were preventable during attentive driving. Eight healthy young participants drove via an MR-compatible driving simulator and structural and functional magnetic resonance images were acquired. The driving environment contained 60 laps of a closed-loop city map with two intersections. Data was analyzed via an event-related general linear model (GLM). The spatial attention task activated the thalamo-cortical structures, basal ganglia, default mode network, dorsal attention network, inferior part of cerebellum, as well as visual, somatomotor, prefrontal, and parietal cortices. These are associated with visuo-spatial attention and processing, relaying sensory and motor signals, as well as employing working memory to implement the task. The accident prevention task led to the recruitment of insula, inhibitory control network, brainstem, and anterior cingulate cortex. This pattern suggests resource allocation to the salience network, inhibitory control, and increasing alertness. The combination of the aforementioned tasks employed those brain areas that are associated with visuo-motor coordination, relaying sensory and motor signals, and alertness such as the cerebellum, somatomotor and visual cortices, thalamus, and brainstem; while many frontal, limbic, cerebellar, and default mode network areas were suppressed, which may reflect high cognitive load.

This exploratory study investigates the relationship between the gut microbiome, vagal activity, and emotion regulation in patients with depression within the AdiDep study. Growing evidence suggests that the gut–brain axis plays a crucial role in mental health, with bidirectional communication pathways linking intestinal microbiota and central nervous system functioning. In particular, the vagus nerve is considered a key mediator in this interaction, influencing both physiological regulation and emotional processes.

The present study aims to examine whether alterations in gut microbiome composition are associated with changes in vagal activity, as indexed by heart rate variability (HRV), and how these factors relate to deficits in emotion regulation commonly observed in depressive disorders. A sample of clinically diagnosed depressive patients is assessed using microbiome sequencing, psychophysiological measures of vagal tone, and standardized self-report measures of emotion regulation.

Preliminary findings revealed significant associations between reduced microbial diversity, lower vagal activity, and impaired flexibility in emotion regulation. These results may contribute to a better understanding of the biological mechanisms underlying depression and highlight the potential of integrative treatment approaches targeting the gut–brain axis.

Overall, this study provides novel insights into the complex interplay between biological and psychological processes in depression and may inform future interventions aiming to improve emotional functioning through modulation of the microbiome and vagal pathways.

Alterations of the gut microbiome have been associated with the pathophysiology of depression, yet empirical findings remain highly inconsistent. While this heterogeneity is often attributed to methodological limitations such as small sample sizes and reliance on a single stool sample per participant, it may also reflect the fact that microbiome alterations are not uniformly present across all individuals with depression. To address this, we first present data from the American Gut Project indicating that microbiome alterations are particularly pronounced in individuals with comorbid obesity. Analyses focus on global indices (e.g., alpha diversity, Firmicutes/Bacteroidetes ratio) as well as taxonomic composition at the genus level. Building on these findings, we investigate similar questions in an ongoing prospective, naturalistic study in hospitalized patients with major depressive disorder. Specifically, we examine whether microbiome alterations vary according to metabolic status (normal weight vs. obese) and depressive symptom profiles, with a particular focus on atypical depression (e.g., mood reactivity, hypersomnia, increased appetite). To improve the robustness of microbiome assessment and to capture both stable and fluctuating components, six stool samples per participant are collected at intervals of two to three days each. Based on prior findings, we hypothesize that microbiome alterations are most pronounced in subgroups characterized by metabolic vulnerability, such as patients with comorbid adiposity, and may further vary with symptom profiles. This perspective may help reconcile inconsistent findings and supports a more individualized approach linking gut microbiome characteristics to clinical phenotypes and metabolic comorbidity.

Aims: Depression and self-harm are highly prevalent during adolescence in clinical and population-based samples. They are closely interconnected across biopsychosocial domains, and confer substantial risks for mental and physical health across the life course. Peer victimization is a common social stressor in childhood and adolescence, while inflammatory markers such as interleukin-6 (IL-6) and C-reactive protein (CRP) may reflect biological stress responses linking social adversity to psychopathological burden. We examined the temporal ordering and reciprocal interplay of these processes between late childhood and early adulthood.

Methods: Data came from a population-based birth cohort (n = 8,931) with repeated assessments spanning ages 8 to 25. Victimization was self-reported five times (ages 8 – 23). Inflammatory markers (CRP: four waves between 9 and 24; IL-6: assessed at ages 9 and 24) were derived from blood samples. Self-harm was assessed at four (ages 11 – 24), and depression at nine timepoints (ages 10 – 25). Path models within a structural equation modeling framework are used to examine direct and indirect pathways linking social stressors (victimization), inflammation (CRP and IL-6), and psychopathology (depression, self-harm) across development.

Results and Conclusion: Preliminary findings from correlation analyses suggest interconnected developmental dynamics between victimization and inflammation from late childhood into young adulthood. Associations with psychopathology vary across age and symptomatology, underscoring the importance of timing and developmental context. By integrating repeated social, biological, and psychological measures over 17 years, this study advances a life course perspective on the biopsychosocial processes shaping youth mental health and informs developmentally sensitive prevention strategies.

Although the severe reduction of visual mental imagery (VMI), labeled aphantasia, has constantly been associated with behavioral differences on perceptual tasks and face recognition deficits, the electroencephalographic markers of these differences remain underexplored. One explanation integrates VMI into a predictive coding network: Mental images act as top-down predictions for incoming visual input, shaping perception. In aphantasics this mechanism is proposed to be diminished as weak predictions result in altered precision-weighting of prediction error signals. Individual differences in VMI were therefore expected to be reflected in posterior alpha activity, indexing inhibitory gating and precision-weighting. Further, the amplitude of the N170 and N300 were hypothesized to be influenced by predictive processes in early and late face perception. To test these hypotheses, 23 aphantasics and 30 controls completed a verbal cueing task in which semantic cues prompted VMI of emotional facial expressions prior to the presentation of congruent or incongruent faces while a 61-channel EEG was recorded. Robust priming was observed with no group differences, indicating preserved performance in aphantasia. ERP analyses showed a VMI group difference in the N170 and N300 alongside a condition effect, but no VMI group × condition interactions. Critically, both effects attenuated when controlling for preceding activity (P1 amplitudes), supporting a temporal cascade rather than distinct template-matching differences. Time-frequency analyses revealed increased posterior alpha power (8–12 Hz) during the imagery interval in aphantasia, most prominently over left occipitotemporal regions, consistent with prior findings on its neural generators. Interpretation is limited by age differences between groups.

Memory consolidation transforms initially encoded experiences into retrievable memory traces through reactivation of neural activity patterns representing the experience. Such reactivation of encoding-specific representations is thought to often occur during offline periods such as wake rest. However, to which extent representations during offline periods indeed resemble encoding or already reflect a transformation remains poorly understood. One reason is that many studies implicitly assume representational stability across brain states. In these studies, classifiers are typically trained on a localiser task mainly based on visual stimulus presentation before applied to different task phases.

To test this assumption, we designed a localiser task mirroring the structure of an associative memory paradigm, comprising both an encoding and a retrieval phase. Classifiers trained on localiser data were then tested on a separate main memory task, likewise comprising encoding and retrieval. Using multi-variate pattern analysis in fMRI data (n=32), we compared classifiers trained on encoding-only (cross-decoding), retrieval-only, or jointly on encoding and retrieval.

Preliminary analyses indicate that, when transferred to retrieval data from the main task, jointly trained classifiers yield higher decoding accuracies than classifiers trained on either phase alone. This pattern was particularly apparent in visual and posterior parietal cortex, where joint training appeared to outperform cross-decoding. These findings suggest that including both phases in classifier training may be beneficial for studying representational dynamics during early memory consolidation. In ongoing work, we want to apply these classifiers to post-learning wake-rest data to test their similarity to encoding and retrieval representations and further detect memory reactivation.

Theoretical Background: Parental bonding is essential for a child's survival, wellbeing, and long-term social-emotional development. However, traumatic and negative birth experiences can impair parental mental health and adversely affect bonding. Physiologically, negative birth experiences have been associated with elevated psychobiological stress markers. This study aims to broaden our understanding of the psychobiological mechanisms linking negative birth experience to maternal and infant bonding by exploring the association (1) between prenatal birth expectations and third-trimester cortisol levels, (2) of prenatal expectations and cortisol with actual birth experience, and (3) of birth experience and postpartum cortisol with mother-infant bonding quality.

Methods: Using data from a large longitudinal study (N= 653), self-reports as well as biological markers from N= 99 mothers are analyzed. Pregnancy related anxiety and prenatal stress, as well as salivary cortisol levels were assessed during the third trimester of pregnancy. Participants were asked to report subjective birth experience and provided additional salivary cortisol samples, 3-4 months after birth. Subjective quality of postpartum bonding was assessed 9 and 12 months after giving birth.

Results: Statistical analysis of the data is currently in progress. The results will be presented at the conference.

Discussion: These findings could advance our understanding of how birth experience and biological stress markers shape parent-infant bonding, with implications for prevention, intervention, and policy.

Learning relational structures from sparse and scarce sequential observations is a hallmark of human cognition, yet its precise mechanistic description remains a challenge. Like other forms of memory, cognitive map learning depends on the hippocampus and its interactions with entorhinal cortex and neocortex, mediated by associative binding and long-term potentiation. Learned maps are subsequently consolidated through experience replay during sleep. Despite substantial progress at molecular, synaptic, cellular, and systems levels, the computational principles underlying the remarkable efficiency of the hippocampus with limited data remain poorly understood. Existing computational models capture important aspects of hippocampal function, yet tend to abstract away biological detail, underrepresent the role of multiple timescales present during both acquisition and consolidation, and often require substantially more data than humans need for equivalent performance. Here we present a critical comparison of existing models along these dimensions, and propose a new architecture that aims at addressing these gaps. Our approach draws on the capacity of recurrent neural networks to learn and reconstruct the dynamics of sequential experience, offering a potential bridge between biological plausibility and computational efficiency.

Age-related differences in cognitive pain modulation remain insufficiently understood, particularly with regard to cognitive abilities. Distraction from pain through cognitive engagement, a well-established modulation strategy, depends on prefrontal cortex (PFC) functioning, which is susceptible to age-related decline and may reduce the effectiveness of distraction-based pain modulation in older adults.

This study examined how age and cognitive abilities relate to the modulation of experimentally induced pain under varying cognitive load. Healthy younger and older adults (N = 76) completed repeated conditions with acute pain stimulation during passive processing (no distraction), low load (0-back), and high load (2-back). Individually calibrated transdermal electrical pulses were applied at non-painful and moderately painful intensities to the inner forearm. Subjective pain ratings (intensity, unpleasantness) and pain-evoked event-related potentials (N1, P2) were assessed across conditions. Cognitive abilities were assessed beforehand with four tasks of executive functioning and attention.

Data were analyzed using structural equation modeling (LCSM). Latent variables for pain and ERP responses were specified, and latent change score models captured within-person changes across cognitive load. Age and cognitive performance were included as predictors.

Preliminary results indicate no significant effect of cognitive load on subjective pain responses. Further analyses on neural correlates of pain modulation, age and cognitive functioning will be presented at the conference.

These findings contribute to ongoing discussions on cognitive pain modulation across the lifespan and suggest that distraction-based modulation mechanisms may remain preserved in healthy aging. Future work will further clarify the conditions under which cognitive resources support pain regulation in older adults.

Interoception, the perception and integration of internal bodily signals, has been proposed as a key mechanism underlying self-functioning and emotion regulation, with particular relevance for borderline personality disorder (BPD) and posttraumatic stress disorder (PTSD). However, systematic comparisons of different interoceptive domains across clinical conditions are lacking. In this study, we investigated subjective, behavioral, and neural indices of interoception in a sample of 116 women comprising patients with BPD (n = 36), PTSD (n = 44), and healthy controls (n = 36). Interoception was assessed using self-report, a heartbeat counting task assessing interoceptive accuracy and confidence, and heartbeat-evoked potentials (HEPs) derived from resting-state and task-based EEG as neural markers of cortical cardiac signal processing. Both clinical groups reported reduced interoceptive sensibility and lower confidence in heartbeat perception compared to healthy controls, despite no group differences in interoceptive accuracy. Electrocortical results revealed condition- and scalp-specific group differences, with consistently reduced posterior HEPs in PTSD, whereas BPD did not differ reliably from controls. Multimodal exploratory factor analyses demonstrated that interoceptive measures clustered primarily by assessment method, providing no evidence for a single cross-modal interoceptive trait. Discrepancies between subjective interoceptive beliefs and task-based interoceptive performance were associated with greater emotion dysregulation but showed only weak and inconsistent relations with neural indices. Together, these findings indicate that altered interoception in BPD and PTSD is driven mainly by subjective and metacognitive components rather than impaired cardiac perception per se, underscoring the importance of targeting interoceptive beliefs and confidence in clinical interventions.