Spatiotemporal patterns of neural activity and their role in perception

PROJECT SUMMARY Individual neurons within the visual cortex exhibit variable responses when repeatedly presented with the same stimulus, and these fluctuations predict whether or not a faint target will be detected. In the prior round of funding, this team discovered that “intrinsic traveling waves” (iTWs) occur spontaneously in visual area MT of the awake marmoset. Previous studies had observed iTWs, but since the animals were anesthetized, it remained unknown (1) if iTWs occur in awake animals and (2) what role, if any, iTWs play in perception. The team found that iTWs create periods of both elevated and suppressed spiking activity. They also found that iTWs conjointly modulate both stimulus-evoked spiking responses and perceptual sensitivity in a visual detection task. A large-scale spiking network model developed by the team predicts that iTWs are a result of the time delays from axonal conduction along the long-range horizontal fibers common across visual areas. These results raise the possibility that iTWs are a new, dynamic mechanism of gain control. In this competitive renewal, this experiment- and theory-driven team is well positioned to conduct experiments to test this role for iTWs by uncovering their connection to specific patterns of connectivity in the visual system. This work will progress in three Specific Aims. Aim 1: Test the hypothesis that iTWs fall into distinct motifs that regulate neural response gain and perception in a feature-selective manner. The team’s network model predicts that clustered, feature-specific connectivity among cortical columns in area MT will create iTWs that traverse subnetworks of like-tuned neurons. They will test this theory-driven prediction in recordings from Area MT. Aim 2: Test the prediction that iTWs vary in their impact and timing across layers of the cortical column. Using combined laminar and ECoG array recordings, they will examine the impact of iTWs in the supragranular, granular, and infragranular layers of area MT. Aim 3: Test competing published models of iTWs that make opposite predictions about their role in behavior. One class of models predicts that iTWs drive pairwise correlations in ongoing undriven activity (so-called “noise correlations”), which impair sensory discrimination. iTWs are thus, according to these models, a source of information-limiting correlations that need to be diminished to improve sensory processing. The model of iTWs developed by this team predicts, instead, that iTWs make no measurable contribution to pairwise correlations in undriven activity and that they improve sensory processing by regulating the gain of sensory evoked responses, an effect analogous to the benefits that stem from attention-dependent regulation of gain. Taken together, the proposed experiments will elucidate the neural mechanisms underlying iTWs and probe their roles in both detection and discrimination of stimuli. In addition, these experiments will test a new hypothesis about the role of long-range horizontal fibers that link similarly tuned neural populations. This team will test if these pathways, which are themselves sculpted by the natural statistics of the visual environment, guide iTWs and in so doing provide a novel means of selectively sampling visual features that are typically found in the environment.