During oculomotor delayed response tasks, the activity of monkey prefrontal neurons is known to encode the parameters of memorized locations. Computational models have provided explicit mechanistic hypothesis to test directly in neurophysiological data the mechanistic basis of this mnemonic representations. In this talk, I will present results that address the stability of the neuronal code during the memory delay, the architectural substrate that supports it, the basis for behavioral imprecisions in the monkey's location reports, and evidence for temporal stabilizing mechanisms and their behavioral consequences. By guiding the analysis of neurophysiological data with mechanistic cortical network models, and validating predictions with behavioral data, a consistent picture of prefrontal network dynamics underlying spatial working memory is emerging.
Biography
Albert Compte received his PhD in Statistics Physics from the Autonomous University of Barcelona in 1998 and his postdoctoral training as a neuroscientist in the laboratory of Xiao-Jing Wang at Brandeis University (Massachusetts, USA). Since 2002, he has directed his research group on brain mechanisms of cognitive function, first at the Institute of Neurosciences of Alicante, and later at IDIBAPS (Barcelona). Dr. Compte combines computational simulations of biological neural networks with data from neurophysiological recordings and from behavioral and neuroimaging experiments in humans to decipher the neural bases of elementary cognitive capacities such as working memory and their dysfunctions in neurological and mental illnesses.