Perceived Speed of Cone-Isolating Stimuli
Robert F. Dougherty, William A. Press, Heidi A. Baseler, Brian A. Wandell
Purpose. We measured the color responsivity of the neural pathways governing perceived speed.
Methods. We used a 2AFC, multiple staircase, speed discrimination design. Two drifting radial sinusoids were presented on opposite sides of fixation. The stimuli were centered at 6 deg eccentricity (0.5 cpd, 8 deg diameter, 0.2 sec FWHM Gaussian temporal envelope). One stimulus (the standard) was L-cone isolating and had a fixed cone contrast of 0.05. The test stimulus, whose color and contrast were varied across conditions, moved at 8 deg/sec. On each trial, the test and standard were presented randomly on either side of fixation, and observers indicated which stimulus was drifting faster. Across trials, the speed of the L-cone standard was adjusted according to the staircase until the perceived speed of the standard matched that of the test. Contrast detection thresholds for the same stimuli were measured using a 2AFC staircase.
Results. Extending reports from other groups, we found that increasing contrast increased perceived speed for all colored targets. M-cone isolating and L-M targets required about twice as much contrast as a speed-matched L-cone stimulus. S-cone stimuli required over 10 times as much contrast to match the L-cone stimulus speed. At 8 deg/sec (4Hz) superposition of cone isolating targets (e.g., L+M or L+S stimuli) increased the perceived speed compared to each cone signal alone. The relative color contrast levels required to equate speed differed substantially from the levels required to equate visibility.
Conclusions. Perceived speed depends systematically on cone contrast and type: L-cone stimuli are about twice as efficient as M-cone stimuli and over 10 times as efficient as S-cone stimuli at driving the motion system. At speeds of 8 deg/sec, motion signals from the 3 cone types summate with respect to perceived speed, showing no opponent-colors tuning. The color tuning governing speed perception differs from the tuning of detection threshold, suggesting that these behaviors are determined by separate neuronal populations. Supported by NEI EY03164