For more information.
Abstract:
The classic view of the striatal circuit in learning and decision making is that corticostriatal inputs encode specific actions or stimuli, and a homogeneous reward prediction error provided by dopamine neurons serves to modify the strength of those corticostriatal synapses, altering the behaviors which are most likely to subsequently occur. However, due to technical limitations, it has been difficult to test these ideas rigorously. To address this gap, my lab has been recording and manipulating activity in genetically and anatomically defined inputs to the striatum during a range of learning and decision making tasks. In the first half of the talk, I will describe recent imaging of populations of dopamine neurons during a decision making task in virtual reality. We found that in addition to conveying reward prediction error, individual dopamine neurons also convey specialized and spatially organized information regarding specific behavioral variables. This works raises important questions regarding the potential functions of non-reward signals in the dopamine system. In the second half of the talk, I will describe neural representations of drug-seeking in corticostriatal inputs to the striatum, and examine how changes in these representations may reflect and generate drug motivation.