**Faculty Candidate Search**

Networks of excitatory and inhibitory neurons form the basic computational units in the cortex. Neurons in such networks encode and process information by asynchronously emitting action potentials.  Theoretical and experimental works have shown that this asynchronous activity stems from collective network dynamics based on a balance between excitation and inhibition. How balanced networks implement complex, behaviorally relevant computations however remains an outstanding open question. In this presentation, I will argue that dynamics of excitatory-inhibitory networks can flexibly switch between different computational modes depending on behavioral demands. 

In the first part of the talk, I will show that model excitatory-inhibitory networks can display two different types of asynchronous activity depending on the strength of feedback coupling. These two types of activity correspond to two different computational modes: a transmission mode in which the network reliably propagates information, and a classification mode in which the network pre-processes signals for non-linear computations such as categorization and decision-making. In the second part of the talk, I will examine neural activity recorded in the ferret auditory cortex while the animal either passively listened or actively discriminated stimuli. Using population analyses, I will show that the type of information encoded in the neural activity strongly depends on the behavioral state of the animal.