Sara Ershadmanesh

I am interested in hybrid, model-based and model-free, learning and decision-making. To understand human behavior related to these systems of decision-making, it is important to understand the way in which they are themselves meta-controlled. Thus, during my Phd, I studied the relationship between meta-cognitive ability and learning style and also manipulation of model-based/ model-free behavior via different methods of self-monitoring. Furthermore, I worked on a hybrid strategy of learning others’ confidences in different social contexts.  In my postdoc, I examined Noradrenergic  and Dopaminergic manipulation of meta-control over model-based and model-free decision-making. I collected behavioral (human) data for this purpose. Additionally, I created a new meta-cognitive efficiency measure using the RL framework for learned value-based choices. This measure outperformed old methods on both real and simulated data.