Entorhinal/dentate excitation of CA3: A critical variable in hippocampal models

We investigate how the strength of entorhinal cortical inputs during traini ng affects learned performance using computer simulations of a minimal comp utational model of hippocampal region CA3. After the model learns two parti ally overlapping sequences, it is tested on two contradictory prediction pr oblems - disambiguation and goal-finding. Relative to total activity, the a ctivity level of entorhinal inputs during learning profoundly affects perfo rmance on each task. The optimal input levels differ for the two sequence p rediction problems, but a small region of overlap exists where both tasks c an usually be performed successfully. This sensitivity to relative input ac tivity suggests critical tests of the model. (C) 2000 Elsevier Science B.V. All rights reserved.