Cortical feedback dynamics in the alpha frequency range support coherent natural vision
Tue-P12-Poster II-102
Presented by: Lixiang Chen
Our eyes constantly receive complex visual information about the environment across the whole visual field. To form coherent visual experiences, the brain must integrate these inputs by their spatiotemporal regularities. Here, we tested whether and how the integration processes are mediated by oscillatory neural activity using spectrally resolved EEG and spatially resolved fMRI measurements. We hypothesized that spatial integration is governed by low-frequency (alpha/beta) rhythms linked to top-down modulations of sensory processing. In separate EEG and fMRI experiments, participants viewed short video clips (3s) depicting everyday situations through circular apertures in the right and left visual fields. We manipulated the spatiotemporal congruency of the videos shown through the two apertures, so that they could or could not be integrated into a coherent percept. Using frequency-resolved decoding on the EEG data, we found that cortical representations shifted from feedforward-related gamma activity when videos were spatiotemporally inconsistent to feedback-related alpha activity when they were spatiotemporally consistent, suggesting a critical role of feedback processes in integrating spatially distributed inputs into a coherent percept. Further, combining the EEG data with spatially resolved fMRI recordings in a multimodal representational similarity analysis framework, we demonstrate that alpha-frequency feedback is directly associated with representations in early visual cortex, suggesting that rhythmic feedback reformats early visual representations of coherent natural inputs for efficient processing in downstream regions. Together, our results elucidate a mechanism for multiplexing feedforward and feedback information in cortex that enables coherent real-world vision.
Keywords: cortical alpha rhythms; top-down modulation; multivariate pattern analysis; EEG-fMRI fusion