Dynamic signaling of spatial context in macaque primary visual cortex

Abstract

Objects in the world are rarely seen in isolation. For primates to have a useful representation of the external world neural mechanisms are required that appropriately integrate and differentiate local features of the image with the surrounding visual context. This thesis investigates such mechanisms by studying the dynamic signaling of spatial context via the extra-classical receptive fields (eCRFs) of neurons in macaque primary visual cortex (V1).

Neurons in V1 are commonly classified as simple or complex based on their response modulation to optimized and spatially restricted drifting sinusoidal gratings. In the first set of experiments, I demonstrate that this classification is robust to both changes in stimulus contrast as well as context produced by spatially extended stimuli of various configurations.

Stimulation of neurons eCRFs often evokes suppression of the responses to stimuli within the classical receptive field (CRF). In the second experiment, I find that this suppression develops over time; the strength and onset latency of suppression is contrast dependent. Additionally, there are two components to the suppression: early in time the suppression is untuned for orientation and later in time a tuned-component of the suppression arises. The untuned component is found across all cortical layers while the tuned component is most prevalent in layers 2/3 and 4b.

In the third experiment, I use a subspace reverse-correlation stimulus paradigm to characterize neurons’ responses to gratings of various orientations and spatial positions in the eCRF with fine temporal resolution. I find three underlying component mechanisms to eCRF modulation: early orientation-tuned facilitation followed by orientation-untuned and then orientation-tuned suppression. Facilitation is found to arise from regions within and local to the CRF, with suppression at larger spatial extents. Further, the tuned-facilitation and tuned-suppression have similar peak orientations and ranges of orientation bandwidth.

The fourth experiment uses a paradigm similar to experiment three, this time to characterize the tuning of spatial meta-contrast masking in human perception. I find a dominant tuned (for orientation and spatial frequency) suppressive mechanism that is in general agreement with the results on V1 eCRF properties, suggesting a candidate neural basis for such perceptual masking.