Computational nonlinear dynamics model of cognitive bistability for simulating interrupted stimulus and long range correlations

Abstract

Simulation results of bistable perception due to ambiguous visual stimuli are presented which are obtained with a nonlinear dynamics model using delayed perception–attention–memory (PAM) coupling. Like the related Synergetic model of Ditzinger and Haken (1989) the present approach is based on attention fatigue within a perception–attention loop. In addition, however it includes a feedback delay time of 40 ms in agreement with Lamme(2003) in order to obtain estimates of absolute time scales. Percept reversals are induced by attention fatigue with an adaptive attention bias which balances the relative percept duration. The recursive PAM equations may be mapped to corresponding reentrant Thalamo-Cortical–Hippocampal loops representing the attentional top-down modulation according to Itty and Koch(2001) and memory effects. Periodic stimulus simulations as a function of stimulus off-time yields the reversal rate variation in surprisingly good quantitative agreement with classical experimental results of Orbach et.al. (1966) when selecting a fatigue time constant of 1 – 2 s. Coupling of the attention bias to the perception state introduces memory effects if the bias time constant is less than 20 s. They are quantified through the Hurst parameter H, exhibiting significant long range correlations (H larger than 0.5), i.e.the fractal character of the reversal time series (Fürstenau 2007, 2009) in agreement with recent experimental results of Gao et.al.(2006). Transition times of 150 – 200 ms between conscious perception states, mean percept dwell times of 3 – 5 s and the dwell time statistics (standard deviation/mean ca.= 0.5) exhibit good agreement with experimental values reported in the literature (e.g. Lamme(2003), Borsellino et.al.(1972), Levelt (1967), Ito et.al.(2003)).