The human brain features a complex network of visual pathways with massive feedforward and feedback connections across different visual areas. However, little is known about the impact of these feedback connections on neural activity in the human brain. We approach this using a multimodal framework combining iEEG and fMRI while the subjects view 1000 natural scene stimuli. For each iEEG electrode in the early visual area (V1/V2), we compute a Pearson correlation map with all the fMRI vertices, across the 1000 stimuli, giving us a time by vertices correlation matrix. This provided us with a brain-wide temporally evolving map revealing how visual information evolves in space and time. Not only do we observe significant correlation of iEEG broadband (70-170 Hz) activity with fMRI vertices in early visual areas, corresponding to the location of the electrodes, but interestingly, we also observe that these correlations spread across ventral and dorsal visual areas ~150-500 ms.
These results were also consistent with the predictions of a task-optimized deep neural network, further revealing that with time, activity in early visual areas aligns better with the features represented in high-level layers of this network. Overall, our multimodal framework enables us to evaluate the impact of feedback in early visual areas at high spatiotemporal resolution, reflecting the rich dynamics of interaction across different brain regions.

※2026/1/7までに登録された場合、所外からの参加可。
参加登録フォーム：
https:https://docs.google.com/forms/d/e/1FAIpQLSeJiZg9rviS2DRfWNigbd3GFv9RflYEYxTf56DMKACrYgJk0Q/viewform