Shining light on hierarchical cortical interactions in the mouse visual cortex.
The neocortex is organized in a hierarchy of functionally specialized areas. This architecture influences our understanding of sensory, motor and cognitive processes and inspires machine learning algorithms. Several theories suggest how the multi-level network of the neocortex might allow us to build complex representations of the world, use prior experience to interpret current events, predict the consequence of our actions or learn from experience.
However, the computations that the hierarchical organization of the neocortex network endows us with remain largely unknown. This in large part due to the lack of understanding of organizing rules in cortico-cortical connections that could inspire and constrain theories of cortical computation. I will present recent experiments from my group where we use optical recordings and circuit mapping methods to unveil the organizing rules of inter-area networks in the mouse visual cortex. We found the long-range cortical connections are organized with great cell-type and tuning-dependent specificity, forming large distributed networks with precise connectivity.
Our observations shine light on the computations implemented by long-range cortical networks, constraining existing theories of cortical computations.