Ayberk Ozkirli¹ (), Angela Ingrid Renton^1,2, Debajyoti Sengupta¹, Jesse Livezey³, David J. Klein³, Yaqing Su³, Adam Hanina³, Dimitri Van De Ville^1,2; ¹Neuro-X Institute, Ecole Polytechnique Fédérale de Lausanne (EPFL), Geneva, Switzerland, ²Department of Radiology and Medical Informatics, University of Geneva (UNIGE), Geneva, Switzerland, ³Dandelion Science Corp, USA, and Dandelion Science Sarl, Switzerland

Recent work demonstrated that spatial-frequency–tagged natural-scene videos elicit steady-state visual evoked potentials (SSVEPs) which act as robust neural markers of visual function in age-related macular degeneration (Renton et al., 2025). Building on this work, we sought to determine whether electroencephalography (EEG) or magnetoencephalography (MEG) provides a superior signal-to-noise ratio for detecting these frequency-tagged responses, and to capture how artificial scotomas modulate these visual cortical neural responses in healthy controls (N=20). To this end, neural activity was recorded using simultaneous MEG and EEG while participants performed a novel gaze-contingent two-alternative forced-choice (2AFC) letter-discrimination task (T vs. L). On each trial, the target appeared at one of 36 visual field locations relative to the participant’s real-time gaze position. Participants completed the task with free-gaze under three central-scotoma masking conditions (None, small, large). Throughout the experiment, full-screen videos composed of either natural-scene or artificially-generated texture were displayed in the background, with low- and high-spatial-frequency information distinctly tagged at 5 and 6 Hz (counterbalanced across trials). We found that SSVEP signal-to-noise ratios were substantially higher when detected with MEG compared with the simultaneously acquired EEG, with a medium effect size when averaged across experimental conditions. These results demonstrate a clear advantage for MEG in capturing frequency-tagged responses in early visual cortex. Moreover, increasing scotoma size was associated with stronger correlations between occipital and occipitoparietal sensors (medium-to-large effect sizes), consistent with prior evidence of cortical reorganization and compensatory recruitment under central vision loss. These findings highlight the efficacy of combining gaze-contingent stimulation, naturalistic video-embedded frequency tagging, and neuroimaging to sensitively probe the neural dynamics of visual function and plasticity. This approach offers a promising framework for individualized mapping of functional reorganization and for developing neurofeedback-based vision-rehabilitation interventions in clinical populations.

Acknowledgements: This work was funded by the the Swiss Innovation Agency (Innosuisse, Project Enlighten, No. 116.587 IP-LS).