Doiron Lab

Our theoretical neuroscience group is interested in how the biological “nuts and bolts” (i.e. neurons and synapses) of the brain coordinate their activity to generate rich brain dynamics and perform complex computations. We tackle this broad topic from several directions, but all of them leverage an indisputable fact about the brain. The responses of channels, synapses, neurons, and networks of neurons all show significant variability. By this we mean that these responses are unreliable, with a stochastic character from moment to moment, and trial to trial in experiments. What would seem to be a nuisance to many, is an opportunity for us. We use the structure of this variability – its magnitude, its timescale, and its dimension – to build and constrain circuit models of cortex. This gives us better models to probe the complex processing and learning mechanics that make brains so fascinating.

From a mathematical perspective doing theoretical neuroscience requires a suite of techniques. In our group we liberally use nonequilbrium statistical mechanics, theories underlying stochastic differential equations, nonlinear dynamics, information theory, as well as modern techniques from machine learning and data science. The intersection of these tools allows us to uncover the salient aspects of neuronal interaction, so as to provide an interpretable and mechanistic theory of brain.

Finally, our group has a philosophy that I have nurtured for over 20 years: the best theory is done hand-in-hand with experiments. Our lab has many collaborations with world class experimental groups, both here at the University of Chicago, as well as all over the world. These groups use cutting-edge techniques to probe the brain at a vast array of scales: from synapses to whole brains, from millisecond responses to the long term storage of memories. Just like in the word ‘Team’ there is no ‘I’ in ‘Theory’ – collaboration is the scaffold of our science.