Neurophysiology of vision and decision processes
Perceptual decisions are influenced by cognitive and motivational context. It is increasingly recognized that some of this influence involves processes at the level of sensory neurons. An attractive computational account of these influences is that of probabilistic inference. We are interested in how neurons in the macaque visual cortex reflect both feed-forward sensory information and top-down information about the animal’s behavioral context and motivational state, and how these combined signals are used by the brain to guide behavior.
To address these questions we use a variety of approaches to quantify the animals’ cognitive states and record from neuronal populations across layers in the early and mid-level visual cortex. While we identify signatures characteristic of top-down processing in the visual cortex, some behavioral contexts influence the decisions largely independently of the visual activity, contrasting with predictions of the computational account. Together, our findings constrain computational accounts of perceptual decision-making that include top-down signals.