Active perception relies on the concerted activity of populations of neurons in the cerebral cortex. These neuronal populations communicate via dense local and long-range connections, forming complex cortical circuits. While we are beginning to understand the elementary building blocks of such circuits during sensory processing, the circuit-level mechanisms by which top-down signals modulate sensory activity are largely unknown.The core of my proposal is to dissect the cortical circuits underlying top-down guided visual perception during active behaviour. The primary innovation of this proposal is to seek large-scale recordings from single neurons of identified type across all cortical layers in an actively behaving animal performing a well- controlled perceptual task.In Aim 1 we will develop novel paradigms to perform in the mouse visual perception experiments that achieve the ‘gold-standard’ in experimental control set by the behaving monkey. We will train mice in psychophysical tasks that offer exquisite control over the visual stimulus and the behavioural response, yield hundreds of trials per session, can be easily combined with recordings of neuronal activity and allow the manipulation of top-down factors.In Aim 2 we will conduct, for the first time, a detailed, circuit-level analysis of top-down modulations of sensory processing in visual cortex. Specifically, we will identify inhibitory interneurons in the neuronal population with optogenetic techniques and ask how behavioural relevance modulates activity for different cell types.This work will bring us closer to a functional explanation of active perception and will bridge the gap between two hitherto disconnected fields: it will open up questions of Cognitive Neuroscience for the field of Cellular Neuroscience.