Conference Agenda

Introduction

Disorders of Consciousness (DoC) result from severe brain damage leading to impaired or absent consciousness. Diagnosing the depth of consciousness alteration and distinguishing Unresponsive Wakefulness Syndrome (UWS) from Minimally Conscious State (MCS) require reliable neurophysiological markers. DoC’s brain dynamics are often characterized by cortical slowing, and delta power has been linked to poor consciousness states and outcomes. We hypothesize that these changes reflect the expression of sleep-like slow waves (SW). To test this, we investigated individual slow waves (SW), to assess their variability across DoC states and etiologies.

Methods

We explored SW in a large high-density EEG dataset from putatively awake DoC patients (105 UWS, 79 MCS) undergoing the ‘Local-Global’ ERP paradigm which assesses brain's ability to process auditory patterns at an unconscious (‘local’) and conscious (‘global’) level. We examined the relationship between SW metrics (e.g. amplitude) and patients’ diagnostic, prognosis and functional evaluation.

Results

Our results reveal that DoC patients have more high-amplitude SW than healthy, awake individuals. MCS patients display steeper SW slopes than UWS and SW are wider and more frequent in patients showing a ‘global’ effect associated with conscious processing. SW properties also vary with etiology, with traumatic brain injury patients exhibiting more frequent, higher-amplitude, and steeper SW than those with anoxic injuries.

Conclusion

These findings suggest that SW metrics vary with clinically and neurophysiologically defined consciousness states, offering potential as fine-grained markers to enhance EEG-based DoC diagnosis. SW may also play a protective role in brain injuries and provide insights into DoC etiologies’ pathophysiology.

False awakenings, where participants dream they are in the sleep lab and perceive it as real, offer a unique lens to study consciousness by merging waking and dreaming states. Our study examines lab dreaming, false awakenings, and lucid dreaming, using virtual reality (VR) and sensory stimulation to enhance metacognitive awareness during sleep.

Participants practiced reality checks in a VR version of the sleep lab, designed with dream-like elements, which served as one of three overlapping environments: the virtual lab, the dreamed lab during sleep, and the real, physical lab. This multi-layered approach aimed to improve participants’ ability to identify their state across different realities. In a first condition, after a single 10-minute VR training, participants napped, and lucidity levels were assessed. In a second condition, the same pre-sleep VR training was paired with a vibration stimulus, which was later delivered during REM sleep via a targeted lucidity reactivation (TLR) protocol. VR training resulted in 56% lab dreams and 48% lucid dreams across both conditions, with 66.7% of lucid dreams originating in the dreamed lab. Notably, all false awakenings in the dreamed lab led to lucid dreams. There were no differences between stimulation and no-stimulation conditions.

This study demonstrates the potential of VR and sensory cues to influence conscious awareness in dreams, providing valuable insights into the mechanisms of dream manipulation and the boundaries of human consciousness across different reality levels.

Objective: Lucid dreaming is characterized by awareness of dreaming and enhanced volitional control, suggesting reactivation of executive functions typically inactive during REM sleep. This study investigates directed information flow between frontal and posterior brain regions during sleep, particularly lucid dreaming, using multivariate Granger causality (GC) analysis of EEG signals.

Methods: We analyzed previously collected EEG data from 27 participants with narcolepsy, including 21 frequent lucid dreamers. Participants underwent five 20-minute nap sessions with intermittent auditory lexical decision tasks, responding via facial muscle contractions. Lucid dreams were identified through subjective reports and objective signaling. EEG data were preprocessed, and directed GC measures between frontal and posterior electrodes were computed across frequency bands.

Results: Causal influence from frontal to posterior regions in the delta-theta bands decreased progressively from wakefulness to deeper sleep stages, mirroring changes in consciousness levels. During lucid REM sleep, there was a significant increase in frontal-to-posterior GC in the delta-theta frequencies, reaching levels comparable to N1 sleep. This increase was independent of behavioral responsiveness to the lexical discrimination task.

Conclusions: The findings provide direct evidence of increased directed functional connectivity from frontal to posterior brain regions during lucid dreaming, supporting the reinvolvement of the frontoparietal executive network. The dissociation between this signature of lucidity and task responsiveness indicates distinct neural mechanisms for lucidity and behavioral responses during sleep.

Parallels between dreams and psychedelic experiences, particularly the N,N-Dimethyltryptamine (DMT) state, are well-documented, with shared features such as immersive effects of vivid imagery and emotional intensity, disconnection from the environment and neurophysiological markers like reduced alpha power. Rapid Eye Movement (REM) sleep, the stage most associated with dreaming, is a prime candidate for examining these similarities. Despite this, no quantitative studies have systematically compared the neurophysiological signatures of DMT and the different sleep stages. Recent advances in automatic sleep-staging algorithms now enable continuous quantification of sleep dynamics through "hypnodensities," which estimate the probability of each sleep stage within a given time window based on EEG features. Leveraging these tools, we hypothesize that the neural signatures of DMT –from the perspective of a staging algorithm– align most closely with those of REM sleep compared to other sleep stages.

We collected high-density EEG data from 18 participants receiving continuous infusions of DMT (20mg bolus plus a 2.27mg/min infusion) or placebo for 30 minutes. Hypnodensities were inferred from the C3 electrode using YASA algorithm and compared between conditions.

DMT significantly reduced the probability of wake-like activity while enhancing REM-like activity. This REM-like activity diminished rapidly following the cessation of DMT infusion.

We suggest that DMT induces physiological processes akin to those occurring during REM sleep. This suggests an overlap between the mechanisms underlying dreaming and psychedelic states. Comparing to sleep EEG with dream reports will be essential to clarify their shared and distinct features.

We applied Dynamical Independence (DI) to source-reconstructed high-density EEG data to identify macroscopic processes decoupled from mesoscopic neurophysiology across sleep stages. Emergence was quantified by optimising dynamical dependence of macroscopic variables over spatial scales, examining their temporal and spectral organisation.

In the temporal domain, deep sleep (N2 and N3) showed the highest degree of emergence, indicated by strong decoupling of higher- from lower-scale processes, followed by early sleep onset (N1), REM, and then wakefulness. Deep sleep also displayed a diffuse optimisation landscape, reflecting less consistent dynamical structure. In contrast, wake and REM had fewer but more prominent local minima, indicating more constrained higher-order organisation. N1 exhibited moderate emergence, with large basins of attraction and minimal local minima. Structural similarity analyses (Gromov-Wasserstein) revealed N1 to be most distinct from wake and REM, while wake and REM were more similar at higher-order scales.

In the spectral domain, dynamical dependence peaks (lower emergence) in the alpha range for N1 and wake, and in the delta range for deep sleep and REM. Emergent organisation increased with deeper sleep in the delta band but declined in the beta band. Sigma-band activity, associated with sleep spindles, marked the onset of deep sleep and similarly showed reduced multiscale functional organisation.

Overall, deep sleep exhibited greater emergence than wakefulness and other stages associated with vivid experiences, albeit with less organised structure. These findings underscore the band-specific nature of emergent macroscopic organisation and clarify how fluctuations in consciousness may reflect distinct multiscale patterns in time and frequency domains.