Abstract
The phenomenon of “motion capture” posits that coherent object motions are estimated after discarding high spatial frequencies. Here we show that “captured” dot motion signals are not discarded, but are instead combined with the low-frequency carrier signal. Although low-frequency carrier gratings influence the perceived speed of superimposed high-frequency dot elements (consistent with capture), the speed of the dot elements also affects the perceived speed of the carrier grating. Stimuli were vertical gratings (0.5 cpd; 40% peak contrast) with superimposed random dots (20 dots/deg^2; 40% peak contrast) that moved coherently in a Gaussian window. Dot and grating speed were varied independently. Seven frames of motion were shown at 10 Hz. Naïve subjects compared test stimuli to a remembered standard (2.67 deg/s). Subjects were instructed to judge either dot or grating speed exclusively. For grating trials, superimposed dot motions of 2.00, 2.67 or 3.33 deg/s yielded average PSEs of 3.21, 2.91, and 2.71 deg/s of grating motion, respectively. This is consistent with dot speed having a weight of about 0.25 in the estimation of grating motion. For dot trials, carrier gratings with speeds of 2.00, 2.67, or 3.33 deg/s yielded average PSEs of 3.43, 2.89 and 2.53 deg/s of dot motion, consistent with a 0.40 weight for grating speed in the estimation of dot speed. We conclude that motion mechanisms estimate coherent motion by integrating across a wide spectral range (with limited task-dependent tuning). Motion capture phenomena probably do not depend on selective capture (or inhibition) of high-frequency signals by low-frequency carriers.