Detection of visual stimuli in correlated noise

The hypothesis of detection by temporal probability summation (TPS) as characterised in Mortensen [(2007). An analysis of visual detection by temporal probability summation, submitted for publication] is critically evaluated, considering data from an experiment of Roufs and Blommaert [(1981). Temporal impulses and step responses of the human eye obtained psychophysically by means of a drift-correcting perturbation technique. Vision Research, 21, 1203-1221], who derived the impulse and the step response for a sustained type of channel. The assumed approximate linearity of the channel is discussed with respect to recent findings from neurophysiological investigations. The data are shown not to be compatible with the TPS-postulate. Further, a model of a cell assembly, consisting of a homogeneous set of neurons, is presented that allows for a natural interpretation of random fluctuations in case of temporal peak detection (TPD). The model may be discussed with respect to TPS as well as to TPD; in any case, the model allows to integrate some results concerning the effects of attentional focussing on the detection process.