Abstract
The ocular following response (OFR) is an involuntary, short-latency eye movement evoked by the onset of rapid motion of a large visual image, and reflects neuronal activity in visual areas MT and MST. Neurons in these areas carry signals representing the true motion of 2D visual patterns (e.g. plaids – two superimposed gratings moving in different directions). When the component gratings in a plaid have different contrasts, the true motion of the pattern is unchanged but the tuning of MT neurons is shifted toward the direction of the higher contrast grating. Under these conditions, human observers' perception of motion is also biased, though less strongly. We asked whether the OFR similarly has a bias that depends on the relative contrast of plaid components.
We presented gratings and plaids moving in different directions to human observers while tracking their eye position. We fixed one component contrast at 0.4, while the other was varied in octave steps from 0.4 to 0.025. In a control condition, only one component grating with a contrast of 0.4 was presented. We presented each stimulus for 500 ms, after which subjects indicated the direction perceived. For equal contrast plaids and for gratings, the perceived motion and eye movement responses were aligned to the true motion of the stimulus. As the ratio of component contrasts increased, both eye movements and perceptual reports became more biased towards the direction of the higher contrast component. At high contrast ratios, both responses were indistinguishable from those evoked by gratings alone, even though the weaker component remained visible. At low contrast ratios, the bias in the OFR matched that of MT neurons and was greater than in perceptual reports. These differences in bias suggest that the signals contributing to perceptions and eye movements evoked by plaids are not identical.
This work was supported by NIH grants EY02017 and EY04440, and by the Robert Leet and Clara Guthrie Patterson Trust Fellowship.