AUTHOR=N'Guyen Steve , Thurat Charles , Girard Benoît TITLE=Saccade learning with concurrent cortical and subcortical basal ganglia loops JOURNAL=Frontiers in Computational Neuroscience VOLUME=8 YEAR=2014 URL=https://www.frontiersin.org/journals/computational-neuroscience/articles/10.3389/fncom.2014.00048 DOI=10.3389/fncom.2014.00048 ISSN=1662-5188 ABSTRACT=

The Basal Ganglia (BG) is a central structure involved in multiple cortical and subcortical loops. Some of these loops are believed to be responsible for saccade target selection. We study here how the very specific structural relationships of these saccadic loops can affect the ability of learning spatial and feature-based tasks. We propose a model of saccade generation with reinforcement learning capabilities based on our previous BG and superior colliculus models. It is structured around the interactions of two parallel cortico-basal loops and one tecto-basal loop. The two cortical loops separately deal with spatial and non-spatial information to select targets in a concurrent way. The subcortical loop is used to make the final target selection leading to the production of the saccade. These different loops may work in concert or disturb each other regarding reward maximization. Interactions between these loops and their learning capabilities are tested on different saccade tasks. The results show the ability of this model to correctly learn basic target selection based on different criteria (spatial or not). Moreover the model reproduces and explains training dependent express saccades toward targets based on a spatial criterion. Finally, the model predicts that in absence of prefrontal control, the spatial loop should dominate.