Abstract
Most studies of heading perception have relied on overt, conscious judgments of heading direction. Do the results obtained from these tasks apply to all visually-mediated behaviors or only to overt judgments? To assess this question, two heading-based tasks were developed in which the mapping of perception-to-action was identical. Participants assigned to the “drive” task steered through a simulated environment with instructions to run over distinctive target objects. If the driver believed that the current heading direction was to the right of a target, a leftward turn was effected by pressing and holding a mouse key. Conversely, a rightward turn was effected by pressing and holding another mouse key. If the driver believed that no turn was needed, then both mouse keys were released. The visual stimuli produced by each driver were recorded and presented to a participant assigned to the “judge” task under identical viewing conditions. Judges were given the task of determining, on a moment-by-moment basis, whether motion was to the left of, right of, or directly toward the target. A “left” judgment was indicated by pressing one mouse key, “right” by pressing another, and “directly toward” by releasing both mouse keys. The relation between registered heading and response was thus identical for both drivers and judges. The only difference was that drivers' actions affected the visual stimuli, whereas judges' responses did not. Under optimal conditions, the accuracy of drivers and judges was approximately equal. However, under conditions of visual blur, driving accuracy was reduced significantly more than judgment accuracy. The results suggest that the process of visual heading registration that accompanies active driving behaviors relies on high spatial frequency information more than that of passive judgment behaviors. The result suggests the need to study active tasks if we wish to better understand active visually-mediated behavior.