No evidence for dichoptic motion sensing: a reply to Carney and Shadlen

Georgeson and Shackleton (1989, Vision Research, 29, 1511-1523) confirmed the existence of dichoptic apparent motion (AM), but argued that its basis was the spatio-temporal correspondence of visible features ("feature tracking"), not early motion sensors. Direction selectivity in motion sensors was probably purely monocular. The key evidence came from AM of missing-fundamental (MF) gratings. Monocular MF motion was always reversed, implying motion sensors responding to the third harmonic (3f) component. Dichoptic MF motion was in the correct direction at the higher contrasts and lower drift rates, and overall was highly correlated with judgements of pattern structure, suggesting feature-tracking. Carney and Shadlen (1992, Vision Research, 32, 187-191) criticized some of our methodology and the theoretical interpretation. Their central argument was that dichoptic AM for sine-waves did reflect dichoptic motion sensing, but was less reliable at higher contrasts. Hence forward motion of the dichoptic MF pattern should occur only where sine-wave (3f) motion-sensing declines. We discuss their critique, and find little support for it. We also present additional data on dichoptic AM for 3f and 5f gratings, showing that Carney and Shadlen's predictions were not upheld. Feature tracking remains the most plausible account of dichoptic AM.