Differences in prefrontal and motor structures learning dynamics depend on task complexity: A neural network model

  • Authors:
  • S. E. Lew;H. G. Rey;D. A. Gutnisky;B. S. Zanutto

  • Affiliations:
  • Instituto de Ingeniería Biomédica, Facultad de Ingeniería, Universidad de Buenos Aires, Paseo Colon 850, C1063ACV Capital Federal, Buenos Aires, Argentina and IBYME-CONICET, Buenos ...;Instituto de Ingeniería Biomédica, Facultad de Ingeniería, Universidad de Buenos Aires, Paseo Colon 850, C1063ACV Capital Federal, Buenos Aires, Argentina and IBYME-CONICET, Buenos ...;Department of Neurobiology and Anatomy, University of Texas Medical School at Houston, Houston, TX 77225, USA;Instituto de Ingeniería Biomédica, Facultad de Ingeniería, Universidad de Buenos Aires, Paseo Colon 850, C1063ACV Capital Federal, Buenos Aires, Argentina and IBYME-CONICET, Buenos ...

  • Venue:
  • Neurocomputing
  • Year:
  • 2008

Quantified Score

Hi-index 0.01

Visualization

Abstract

Neurons in the basal ganglia (BG) of monkeys learning a simple visual discrimination (VD) task show faster changes in activity than those in the prefrontal cortex (PFC). This motivated the hypothesis that changes in the BG activity can ''lead'' those in the PFC. Given that the PFC is a key player in the learning of complex tasks, we tested the former hypothesis by using a neural network model that learns simple and complex contingencies as VD and delayed matching to sample (DMTS) tasks. Even though the model accounted for the results in the VD task no such ''lead'' was observed in the DMTS task. We propose that when the task requires learning high-order contingencies, such as in the DMTS case, motor structures quickly select the subset of responses allowing the subject to obtain reward, but learning in the cortico-BG loop progresses in a concurrent way in order to maximize reward.