Abstract
The frontal cortex is involved in switching between perceptual states and switching between motor states like fixation and making saccades. We wonder if these phenomena share the same decision circuitry. To test this idea we compared stochastic properties of timing of flips in binocular rivalry to those of saccadic timing reported by our lab last year. Concerning rate (reciprocal latency) distributions of sequences of saccades, Beintema et al. (VSS 2003) showed this distribution to be less skewed and to have a lower median for first saccades than for later saccades. He found a better fit to an adapted Beta distribution called Beta prime than to a Gamma distribution. Beta prime distributed rates can be explained by assuming hard competition between two neural signals (one in favour of fixation; the other in favour of making a saccade) rising toward their thresholds (Van den Berg, NeuroScience 2001). We investigated both rate and latency distributions of sequences of perceptual flips, and compared the quality of fit to both the Gamma and the Beta prime distribution. Results on three subjects indicate that the traditional Gamma-to-latency fit performs less well than the three remaining combinations. Also, rate distributions of first flips show a lower median but equal skew compared to those of subsequent flips. To make a more critical distinction between the four distributions, we also investigated quality of fit to the theoretical reciprocal Gamma and Beta Prime distribution (e.g. a good fit of latencies to a reciprocal Beta Prime distribution indicating Beta Prime distributed rates). The results of this procedure supported our finding that latencies do not fit a Gamma distribution well, and furthermore point toward the Beta Prime distribution as the best fit to our rate data, suggesting remarkable similarities between state switches in binocular rivalry and saccadic search.