Marcin Leszczynski

People tend to prefer human-made over AI-made artworks. Interestingly, this preference remains even if the images are identical and only the stated authorship differs. The mechanism of this bias remains unclear. Here, we explored whether viewers explore the same images differently depending on whether they believe the images were created by a human or AI. We made three observations. First, replicating previous studies, images labeled as AI-made were evaluated significantly less positively than the same images labeled as human-made (the effect sometimes referred to as negative AI-bias). Second, global eye-movement metrics - fixation count, total fixation duration, and gaze dispersion – did not differ between the two contexts, indicating that overall viewing behavior was not modulated by authorship beliefs. Third, trial-level analyses revealed that aesthetic appreciation was positively associated with longer fixation duration and more dispersed gaze patterns. Importantly, this relationship depended on context: gaze dispersion was a strong positive predictor of aesthetic appreciation in the AI condition, whereas this association was attenuated when images were believed to be human-made. Together, these findings show that although authorship beliefs do not alter global eye-movement statistics, they nevertheless shape how visual exploration contributes to aesthetic evaluation. Even when images are identical, believing that they are AI-generated elicits both reduced aesthetic appreciation and a distinct, more exploratory viewing pattern that becomes more tightly linked to such aesthetic appreciation. This reveals that authorship beliefs modulate the perceptual mechanisms underlying aesthetic experience.

We sample visual scenes with short gaze fixations separated by saccades. While low-level integration is known, semantic integration of foveal vision across multiple fixations remains unclear. We hypothesized that the brain responds to changes in semantic information from one fixation to the next, and therefore postulated a neural signal associated with semantic novelty for each saccade. Novelty was measured using a deep network on foveal vision. Novelty modulated frontal and occipital fixation-related potentials in human EEG during natural viewing of full-length movies (3.4×10 ⁶ saccades). Intracranial recordings in humans (9.0×10 ⁴ saccades) and non-human primates (3.3×10 ⁴ saccades) revealed broadband high-frequency activity modulations in ventromedial visual and frontal brain areas. This modulation was stronger for movies than static images, and frontal modulation preceded occipital modulation, suggesting top-down effects. This modulation of fixation-related activity with novelty suggests that foveal representations are integrated across saccades to construct scene representations during natural viewing in primates. Significance statement: Primate vision relies on eye movements to sample the world, yet how the brain integrates high-level meaning across these “snapshots” remains a mystery. By combining deep-learning models with massive electrophysiological datasets from humans and macaques, we show that the brain responds to changes in semantic content between fixations. We identified a “semantic novelty” signal that modulates neural activity across the primate cortex, appearing in frontal areas before a saccade is completed. These findings suggest visual processing is not a series of independent glimpses, but a continuous process integrating foveal information at the semantic level. This work bridges biological vision and artificial intelligence, providing a new framework for understanding how the brain maintains a coherent understanding of a dynamic world.

Natural vision is an active, predictive process guided by expectations about when and where information will appear. Yet how gaze is shaped in dynamic, multisensory environments remains poorly understood. Using immersive virtual reality with eye-tracking, we examined oculomotor behavior during naturalistic navigation. Participants cycled through a virtual city while avatar cyclists, first heard as overtaking them from behind via spatialized auditory cues, later became visible as they passed. Auditory cues triggered anticipatory gaze shifts to expected locations, indicating that eye movements were guided by auditory predictions rather than reactive visual responses. Violations of auditory–spatial expectations elicited longer fixations. Critically, removing auditory cues impaired predictive gaze orienting, delayed gaze orienting and increased collisions with obstacles. These findings demonstrate that auditory input fundamentally shapes predictive models guiding visual exploration and adaptive behavior in dynamic environments, underscoring the multisensory basis of active perception in real-world interactions.

The discrete events of our narrative experience are organized by the neural substrate that underlies episodic memory. This narrative process is segmented into distinct units by event boundaries, which facilitate a replay process that acts to consolidate each event into a narrative memory. High-frequency oscillations (HFOs) may synchronize neural activity during these processes. We use intracranial recordings from participants viewing and freely recalling a continuous, audiovisual stimulus. We find that hippocampal HFOs increase following event boundaries and hippocampal-cortical coincident HFOs (co-HFOs) occur in cortical regions that underlie event segmentation (inferior parietal, precuneus, lateral occipital, and inferior frontal cortices). Event-specific co-HFO patterns that occur during event viewing reoccur following event boundaries for the subsequent three events and during recall. This is consistent with models that support replay as a mechanism for memory consolidation. Therefore, HFOs may coordinate activity across brain regions that facilitate event segmentation, encode memory of discrete events, and bind representations to assemble memory of a coherent, continuous experience.

Sensory stimulation of the brain reverberates in its recurrent neural networks. However, current computational models of brain activity do not separate immediate sensory responses from this intrinsic dynamic. We apply a vector-autoregressive model with external input (VARX), combining the concepts of ‘functional connectivity’ and ‘encoding models’, to intracranial recordings in humans. This model captures the extrinsic effect of the stimulus and separates that from the intrinsic effect of the recurrent brain dynamic. We find that the intrinsic dynamic enhances and prolongs the neural responses to scene cuts, eye movements, and sounds. Failing to account for these extrinsic inputs leads to spurious recurrent connections that govern the intrinsic dynamic. We also find that the recurrent connectivity during rest is reduced during movie watching. The model shows that an external stimulus can reduce intrinsic noise. It also shows that sensory areas have mostly outward, whereas higher-order brain areas have mostly incoming connections. We conclude that the response to an external audiovisual stimulus can largely be attributed to the intrinsic dynamic of the brain, already observed during rest.

Sensory stimulation of the brain reverberates in its recurrent neural networks. However, current computational models of brain activity do not separate immediate sensory responses from this intrinsic dynamic. We apply a vector-autoregressive model with external input (VARX), combining the concepts of “functional connectivity” and “encoding models”, to intracranial recordings in humans. This model captures the extrinsic effect of the stimulus and separates that from the intrinsic effect of the recurrent brain dynamic. We find that the intrinsic dynamic enhances and prolongs the neural responses to scene cuts, eye movements, and sounds. Failing to account for these extrinsic inputs, leads to spurious recurrent connections that govern the intrinsic dynamic. We also find that the recurrent connectivity during rest is reduced during movie watching. The model shows that an external stimulus can reduce intrinsic noise. It also shows that sensory areas have mostly outward, whereas higher-order brain areas mostly incoming connections. We conclude that the response to an external audiovisual stimulus can largely be attributed to the intrinsic dynamic of the brain, already observed during rest.

Natural vision is an active sensing process that entails frequent eye movements to sample the environment. Nonetheless vision is often studied using passive viewing with eye position held constant. Using closed-loop eye-tracking, with saccade-contingent stimulation and simultaneous intracranial recordings in surgical epilepsy patients, we tested the critical role of eye movement signals during natural visual processing in the hippocampus and hippocampal-amygdala circuit. Prior work shows that saccades elicit phase reset of ongoing neural excitability fluctuations across a broad array of cortical and subcortical areas. Here we show that saccade-related reset systematically modulates neuronal ensemble responses to visual input, enables phase-coding of information across the saccade-fixation cycle and modulates network connectivity between hippocampus and amygdala. The saccade-fixation cycle thus emerges as a fundamental sampling unit, organizing a range of neural operations including input representation, network connectivity and information coding. One-Sentence Summary Saccade-fixation cycle: a fundamental sampling unit, organizing input representation, information coding and network activation

Summary: Sensory stimulation of the brain reverberates in its recurrent neural networks. However, current computational models of brain activity do not separate immediate sensory responses from this intrinsic dynamic. We apply a vector-autoregressive model with external input (VARX), combining the concepts of “functional connectivity” and “encoding models”, to intracranial recordings in humans. This model captures the extrinsic effect of the stimulus and separates that from the intrinsic effect of the recurrent brain dynamic. We find that the intrinsic dynamic enhances and prolongs the neural responses to scene cuts, eye movements, and sounds. Failing to account for these extrinsic inputs, leads to spurious recurrent connections that govern the intrinsic dynamic. We also find that the recurrent connectivity during rest is reduced during movie watching. The model shows that an external stimulus can reduce intrinsic noise. It also shows that sensory areas have mostly outward, whereas higher-order brain areas mostly incoming connections. We conclude that the response to an external audiovisual stimulus can largely be attributed to the intrinsic dynamic of the brain, already observed during rest.

Our continuous visual experience in daily life is dominated by change. Previous research has focused on visual change due to stimulus motion, eye movements or unfolding events, but not their combined impact across the brain, or their interactions with semantic novelty. We investigate the neural responses to these sources of novelty during film viewing. We analyzed intracranial recordings in humans across 6328 electrodes from 23 individuals. Responses associated with saccades and film cuts were dominant across the entire brain. Film cuts at semantic event boundaries were particularly effective in the temporal and medial temporal lobe. Saccades to visual targets with high visual novelty were also associated with strong neural responses. Specific locations in higher-order association areas showed selectivity to either high or low-novelty saccades. We conclude that neural activity associated with film cuts and eye movements is widespread across the brain and is modulated by semantic novelty.

Our continuous visual experience in daily life is dominated by change. Previous research has focused on visual change due to stimulus motion, eye movements or unfolding events, but not their combined impact across the brain, or their interactions with semantic novelty. We investigate the neural responses to these sources of novelty during film viewing. We analyzed intracranial recordings in humans across 6328 electrodes from 23 individuals. Responses associated with saccades and film cuts were dominant across the entire brain. Film cuts at semantic event boundaries were particularly effective in the temporal and medial temporal lobe. Saccades to visual targets with high visual novelty were also associated with strong neural responses. Specific locations in higher-order association areas showed selectivity to either high or low-novelty saccades. We conclude that neural activity associated with film cuts and eye movements is widespread across the brain and is modulated by semantic novelty.