Joseph Emerson¹ (), Karen Navarro², Cheryl Olman^1,2,3; ¹University of Minnesota, Graduate Program in Neuroscience, ²University of Minnesota, Department of Psychology, ³University of Minnesota, Center for Magnetic Resonance Research

Human visual perception relies on contextual information to make inferences about spatially localized features in visual scenes. Similarly, neural representations of local features in primary visual cortex (V1) are shaped by broad spatial context through long-range lateral connections and feedback from higher-order visual areas. However, it is unclear exactly how and to what extent lateral and feedback connectivity individually contribute to contextual modulation of neural responses in V1. Ultra-high-field fMRI have enabled non-invasive imaging of cortical layers in humans, which can be exploited to examine the cortical origins of neural signals underlying blood-oxygenation-level-dependent (BOLD) contrast. We analyzed data from five participants using 7T fMRI at 0.6 mm isotropic resolution to measure the influence of visual context on BOLD response profiles across cortical depth in V1. Participants viewed sine-wave grating disks embedded in large surround gratings with matched spatial frequency and contrast. Segmentation cues were provided by either an offset in relative orientation or an offset in relative phase between target and surround gratings for a total of three contextual conditions plus a surround-only condition to measure the effects of cortical feedback in the absence of feedforward input. Our analysis isolated the effects of orientation-tuned surround suppression (OTSS) from figure-ground modulation (FGM). Consistent with contextually-driven responses measured in mice and monkeys, we found significant modulation of BOLD signal in target-selective voxels in the absence of feedforward input. While we found strong signatures of FGM in superficial and deep layers, we did not find significant modulation of the BOLD signal due to OTSS. Our results suggest that the mechanisms responsible for OTSS have a weaker impact on the BOLD signal. We conclude that a large proportion of the BOLD signal measured in V1 depends on feedback from higher-order visual cortex, which is reflected in contextually-dependent changes in laminar profiles.

Acknowledgements: NIH Grant 5T32MH115886-04, NIH Grant 5R01MH111447-05, NIBIB P41 EB027061, NIH Grant P30 NS076408, NIH Grant 1S10OD017974-01, NIH Grant NS123482