Mistaken Identity: Temporal-Order Biases in the Perception of Speed

Abstract

Biases in the perceived speed of a moving stimulus have been observed for various spatial configurations of a motion stimulus (eg Norman et al, 1996 Perception 25 815 – 830). In general, psychophysical studies are designed to remove any possible source of bias due to the temporal order of presentation. Here, we examined biases in speed perception between the first and second temporal interval in a 2IFC procedure. The stimulus was a circular patch (∼ 2 deg) containing 50 translating dots. A standard speed was always presented in the first interval (∼ 6 deg s⁻¹), with one of 7 test speeds in the second interval. A range of ISIs (0 – 2.3 s) and stimulus durations (100 – 900 ms) was tested. We also dissociated spatial location (left/right of fixation point) and direction of dot travel (left/right and vertical/horizontal) of the two stimuli, whilst maintaining the temporal separation.

In all cases we found an increase in perceived speed from 1st to 2nd interval and further, the effect was independent of direction, ISI, duration, and the spatial position of the 1st and 2nd stimuli. We can suggest no simple account of the results based upon neural fatigue in motion detectors, temporal sampling windows, or representational momentum. Tentative models are currently being explored with three interval stimuli.