A practical evaluation of synchronicity in action potentials with noise

Abstract

The focus of this paper is to examine the temporal pattern recognition capacities of a simplified action potential model. The model is similar to that proposed by Hopfield, but with no logarithmic scaling and the addition of a speed variable to set the pattern. There is also a more thorough specification of synchronicity between units, by explicitly identifying a window where approximate alignment causes output activation. Action potential models have been shown to easily solve a vector scaling issue referred to as the analogue match problem. Two further noisy extensions to this problem are proposed. The first incorporates noise into the signal intensity and the second introduces noise into the recognition vector. The simplified action potential model is shown to function in the short-term, using a fixed window, as long as the noise is in proportion to the signal intensity.