Abstract
The existence of multiple channels selective for spatial frequency in human motion perception has been shown by masking experiments (Anderson & Burr, 1989). We examined this in humans and praying mantises. The “signal” stimuli were vertically-oriented sinusoidal luminance gratings of either Low or High spatial frequency, moving either left or right with a temporal frequency of 8Hz. Signals were masked by temporally broadband noise, with no net energy in either direction. The noise was spatially band-pass, centered on either Low or High spatial frequency. Human subjects pressed a button to report motion direction; for mantises, a human observer blind to the stimulus judged the direction of the mantis optomotor response. The Low and High spatial frequencies were chosen for each species such that contrast sensitivity was roughly equal for both. For mantises, we used 0.04 and 0.2 cpd (2.3 octaves apart); for humans, we used 0.4 and 2 cpd (2.3oct), and also 0.2 and 4 cpd (4.3oct). For humans, at 2.3oct separation, contrast thresholds at both signal frequencies were increased by noise at the other frequency, with the Low threshold increased more by High noise. At 4.3oct, the Low threshold was still affected by High noise, but not vice versa. This implies the existence of at least two mechanisms: a broadband motion sensor which detects both 0.2 and 4 cpd, and a narrower-band, higher-frequency sensor which detects 4cpd but not 0.2cpd. For mantises, at 2.3oct separation, thresholds at both Low and High frequencies were increased by noise at the “other” frequency, but the High threshold was increased more by Low noise: the opposite to humans. Thus if mantises have >1 motion sensor, the lower-frequency sensor is narrower-band. Many aspects of motion perception are surprisingly similar in insects and humans, but this finding represents a novel difference.
Meeting abstract presented at VSS 2015