Our ability to comprehend the world and to interact with it relies on our brains’ ability to process sensory signals and transform them into a flexible representation that supports perception, learning, memory, and action. This processing seems effortless. Yet at its core lies an intricate network of nerve cells that extract important information from a continuous stream of redundant, distracting and ambiguous signals (e.g. finding a needle in a haystack).
The purpose of the lab is to study the network computations and biological mechanisms that underlie this sensory processing in the mammalian brain. How are sensory computations implemented at the level of neural circuits? How are these computations adapted to the animal’s environment and to the animal’s behavioral goals? Solving these questions not only provides insights into how our brains process information but also paves the way toward creating autonomous devices that interact effectively with their environment, and neural prostheses that can restore brain function.
We address these questions by studying the visual system of mice. Mice are well suited for studying the neural circuits of vision. They are small and smart; their visual system resembles in many ways that of humans; and they can be manipulated genetically to identify and study specific groups of nerve cells. Using fast and sensitive laser-scanning microscopes, micropipette or advanced microprobes recordings, we measure the activity of large groups of nerve cells while the animal explores a controlled visual environment. By relating the measured neural responses to what the mice see, we infer the visual computations performed by the network and build mathematical models of them. We further test these models by electrically or optically stimulating targeted groups of neurons and observing the activity of neighboring or distant cells. This approach lets us assess how different components of the circuit can underlie specific visual computations, and how these computations are tailored to the animal’s visual environment and behavioral goals. Ultimately, this research will provide insights into how we see, and into how biological neural networks are constructed to perform useful computations.