Rat Prefrontal Cortex Inactivations during Decision Making Are Explained by Bistable Attractor Dynamics
Alex T. Piet, Jeffrey C. Erlich, Charles D. Kopec and Carlos D. Brody, Neural Computation, 2017.
Decision making is a complex behavior involving many separate computations: signal detection, evidence accumulation, decision formation, decision memory, and motor execution. An important question is how and where each of these computations are performed.
Recent experimental work in the lab investigated the role of one rodent cortical area known as the Frontal Orienting Fields (FOF). Two studies found that reversible inactivations to the FOF resulted in side specific biases in decision making (Erlich, 2011; Erlich, 2015). Interestingly, the biases exhibited by the rats were unusual in that the psychometric curve had scaled vertically, not horizontally. That is unusual because it indicates that the bias was independent of trial difficulty.
We sought to investigate what computational role for the FOF could produce a difficulty-independent bias. To answer this question we utilized simple two-node attractor networks. These network can perform many of the decision making computations that we wanted to investigate, depending on how the networks are set up. By simultaneously fitting the networks to both control and inactivation data, we asked how well each of these computations could explain the data from Erlich 2015.
What we found is that when the networks were asked to accumulate evidence or categorize already accumulated evidence into decisions they could not fit the experimental data well. However, when the network was asked to remember an already formed decision, the networks could match the data very well. This network in a “post-categorization” role naturally accounts for a variety of other data from the lab about FOF tuning curves, and responses to fast timescale inactivations (Hanks, 2015).