Dataset and analysis code for the intracranial EEG study of visual discomfort induced by striped patterns

This repository contains the dataset and analysis code for the study "Intracranial mapping identifies high-frequency cortical dynamics underlying visual discomfort."

The dataset includes intracranial EEG (iEEG) recordings obtained from nine patients with non-photosensitive epilepsy undergoing clinical monitoring with depth electrodes. During recordings, participants viewed high-contrast striped gratings designed to elicit visual discomfort alongside control patterns (Pattern Glare Test). The dataset consists of log-transformed, baseline-normalized mean power across frequencies ranging from 55 to 1000 Hz, including high gamma (55–80 Hz), ripples (80–250 Hz), fast ripples (250–600 Hz), and very fast ripples (600–1000 Hz). Analyses were designed to address four hypotheses:

1. Aversive (3 cpd) patterns will elicit a greater increase in occipital high-gamma–band power than control (0.5 cpd) patterns.

2. Individuals reporting higher subjective visual discomfort will exhibit stronger occipital LFP responses, assessed via continuous brain–behaviour correlations with visual discomfort scores.

3. Aversive stimuli will evoke high-gamma activity that propagates beyond the occipital cortex to temporal, parietal, insular, and limbic regions.

4. Neural activity induced by aversive stimuli will also be evident at frequencies above the high-gamma range, including ripples, fast ripples, and very fast ripples.

The results indicate that individual differences in visual discomfort susceptibility are reflected in extrastriate cortical activity: patients with higher discomfort scores exhibited stronger and more sustained post-stimulus increases in high-frequency power in BA 18 across multiple frequency bands. At the group level, aversive patterns elicited significantly greater high-gamma power in BA 18 and, at later latencies, BA 19. Beyond occipital cortex, aversive stimulation recruited lateral temporal regions with sustained responses consistent with ventral stream engagement. Insular and limbic regions showed no consistent condition-specific modulation following baseline normalization. High-frequency activity across ripple, fast ripple, and very fast ripple bands mirrored the high-gamma pattern in occipital and temporal cortex, though responses at higher frequencies were weaker and more spatially sparse.