Information processing in neural systems

The ability of the nervous system to respond to external stimuli and to initiate actions represents the outcome of computations performed at many spatial and temporal scales. Through collaborations among members of the Centre we are developing experimental protocols and computational models that span these various levels of investigation. A common theme in this work is the use of mathematical techniques such as system identification, machine learning, and signal processing that relate neuronal responses to sensory stimuli or motor outputs. For example, the capacity of a single neuron to encode sensory information is affected by intracellular signaling cascades, which are in turn triggered by the release of neuromodulators. Recent studies of these processes in the electric fish by one member of the Centre (Chacron) has led to a collaboration with another member (Cullen) to study the link between the spiking responses of primary vestibular afferents and head movements in the alert monkey. This research therefore forms a bridge from the molecular to the behavioral level by extracting computational principles that are common to multiple species and multiple sensory systems. A similar collaborative effort between two labs (Cook and Pack) is currently underway to understand the neuronal correlates of conscious perception of time in the primate brain, and this work is being used to formulate mathematical models of information processing in mental disease.