Simulating attentional blink with a neocortical attractor model

Abstract

This study simulates attentional or blink of themind, a phenomenon where a human subject missesperception of a later expected visual pattern as twoexpected visual patterns are presented less than 500 milli-seconds apart. With an abstract computational model ofthe neocortex, a cortical patch is stimulated with asequence of patterns at varying time points. Patterns thatbecome active attractors are considered recognized.A neocortical patch is represented as a square matrix ofhypercolumns, each containing a set of minicolumns withsynaptic connections both within and across both mini-columns and hypercolumns [1]. Each minicolumn con-sists of locally connected layer 2/3 pyramidal cells withinteracting basket and bipolar/double bouquet interneu-rons. All neurons are implemented using the Hodgkin-Huxley multi-compartmental cell formalism and includecalcium dynamics, synaptic depression and saturatingAMPA, NMDA and GABA synapses. Active patterns areencoded with global connectivity of minicolumns acrosshypercolumns and they inhibit each other.Stored patterns were stimulated at sliding differential timepoints to create neural attractor interference measurablewith synthetic electrode traces. The effect correspondedwith human visual attentional studies. Additionally,studies on the inhibitory effect of the drug class benzodi-azopine on attentional in human subjects werecompared with neocortical simulations where an inhibi-tory GABA receptor conductance was increased. Simula-tions showed increases in the attentional duration,agreeing with observations in human studies. Figures 1and 2.