Visual processing of socio-emotional cues is shaped by both emotional and physical stimulus features. Building on prior evidence that physical salience can modulate emotional salience and can lead to a negative bias in emotion recognition, we investigated how variations in image contrast affect emotional face detection using a signal detection framework.
Ninety-four healthy participants in a preregistered (osf.io/qx89e) ongoing study completed an Emotional Face Detection Task in two blocks. In one block, they should answer with yes and no whether the face displays a negative facial expression, and in the other whether the face has a positive facial expression. In each block, emotional (happy or fearful) and neutral faces were shown in three conditions: low, unchanged, and high contrast. In addition, electrodermal activity and heart rate were recorded as measures of autonomic activity.
Participants presented an overall conservative response bias, indicating cautious responding, and greater sensitivity in discriminating positive from neutral compared to negative from neutral faces. Importantly, different patterns were found depending on emotional valence: For negative faces, high contrast induced a less conservative bias, and less sensitivity compared to unchanged and low contrast, indicating a tendency towards a negative bias. For positive faces, high contrast induced an inverse pattern: more conservative bias and greater sensitivity. Autonomic responses are currently analysed and will be presented at the conference.
Our findings indicate that physical contrast does not produce a uniform perceptual salience effect on emotional face detection but instead contrasting patterns of emotional processing depending on valence.

There is evidence that visual perception is influenced by anxiety. This study investigates how anxiety, measured as a state, trait, or through active induction, modulates conscious visual perception in a breaking Continuous Flash Suppression (b-CFS) task, a visual masking method that delays the access of stimuli to conscious awareness. 44 participants are compared for state /trait anxiety and undergo a threat-of-shock manipulation to actively induce anxiety. Neutral faces are presented under b-CFS in both upright (ecologically meaningful) and inverted (not meaningful) orientation to check whether anxiety specifically affects ‘holistic’ processing. Crucially, we separate detection (“did you see anything?”) from identification (“what did you see?”) sensitivity to differentiate responses often conflated in the b-CFS literature, analysing both using objective Signal Detection means. Our results reveal that trait but not state anxiety significantly increases detection sensitivity under b-CFS. We found no selective effect of anxiety on stimuli with holistic properties but reproduced the Face Inversion effect commonly reported in the b-CFS literature. Our results show that trait anxiety alters Detection sensitivity, suggesting an increase in neural signal transmission in primary visual areas before advanced categorisation takes place. Furthermore, detection and identification responses seem clearly discernible under b-CFS and interact with the face inversion effect. We conclude Detection and Identification measure different stages of visual awareness and need to be individually assessed in research moving forward.

Numerous studies have demonstrated rapid (< 100 ms) visuo-cortical differentiation of threat-associated faces. This may be due to low-spatial frequency (LSF) visual information originating from magnocellular pathways. Yet it remains unclear whether potentially magnocellular fear signals extend beyond evolutionarily prepared emotional faces and whether they are subject to short-term neuroplasticity. If so, spatial frequency characteristics should modulate processing of faces with newly acquired threat-relevance. Furthermore, it is unknown whether sub-bands of the visual spectrum are associated with autonomic arousal. Using a differential fear-conditioning paradigm, this study tested whether early visual attentional capture, indicated by the P1 event-related potential component, prioritizes LSF information of threat-associated faces with neutral expressions. Additionally, it was tested whether such effects would be paralleled by threat differentiation in the skin conductance response (SCR). For contingency aware participants, stimulus ratings confirmed successful fear conditioning and participants showed a selective left-hemispheric enhancement of the P1 in response to LSF threat-faces. By contrast, CS differentiation in the SCR was not modulated by spatial frequencies but by stimulus duration, with longer CS presentations resulting in larger SCR to threat compared to neutral faces. For contingency unaware participants, trial-by-trial amplitudes of P1 and SCR were positively correlated. Data support the notion that magnocellular-cortical pathways adapt quickly to novel threat-associations and facilitate rapid threat retrieval even for perceptually neutral faces. However, at least in the short term, these signals do not necessarily associate with anticipatory arousal in SCR.

As inherently social beings, humans rely on social attention to prioritize relevant social stimuli in order to interpret our surroundings and anticipate others’ behavior. Evidence from eye-tracking studies suggests that social cues automatically and rapidly capture visual attention, indicating a reflexive attentional mechanism that operates even in complex scenes, under brief presentation conditions, and in the presence of highly salient non-social distractors. However, complementary neural evidence supporting reflexive social attention remains elusive. Because of its early timing (50 – 90 ms) and V1 origin, the C1 component observed in EEG recordings is considered as an index of the earliest stage of cortical visual processing. Thus, the goal of the present study was to investigate the influence of social attention on the initial stages of the visual processing hierarchy. To this end, a total of 38 healthy participants viewed brief presentations of naturalistic images containing social information in one of the four visual quadrants while eye movements and EEG were recorded. Analysis of the C1 component revealed more positive amplitudes for social information presented in the lower relative to the upper visual hemifield, consistent with the retinotopic organization of the primary visual cortex and its characteristic polarity inversion across visual hemifields. These findings provide compelling evidence for early prioritized sensory processing of social information and support the theory of reflexive social attention, which operates largely independently of top-down influences.