On the role for binocular cues in the fast extraction of egocentric distance

Abstract

Previous research has shown that humans can estimate the distance to previewed targets (floor-level, 3-5 m distant) with near perfect sensitivity (slopes near 1), even when targets are viewed in single brief glimpses of 100 ms or less. However, performance at brief viewing durations is marked by a bias toward underestimation that is relieved with longer viewing durations. Does the high sensitivity to distance at brief viewing durations depend on binocular viewing, and should the underestimation at shorter viewing durations be attributed to a relatively long time course for the availability of binocular cues? To address these questions, participants verbally estimated target distance under monocular and binocular viewing conditions with viewing durations of 74 and 1200 ms. Viewing durations were administered in blocks with block order manipulated between groups. Bias and sensitivity each depended on block order, suggesting that verbal reports may be particularly susceptible to carryover effects. Analyses thus focused on first block performance. Sensitivity to distance was generally high but depended on viewing duration: participants were oversensitive with 1200-ms trials (slope=1.39) but not with 74-ms trials (slope=.72). Importantly, in both cases sensitivity was undiminished by monocular viewing. Underestimation bias was high with 74-ms trials (-45%) but nearly eliminated with 1200-ms trials (-9%). Experiment 2 controlled for carryover effects by administering monocular trials first in single block followed by a block of binocular trials (74 ms only). Performance was similar to that of Experiment 1 and, critically, sensitivity did not differ between viewing conditions. In sum, the results of these experiments indicate that neither sensitivity nor bias correction depends on binocular viewing. Overall, the data suggest that binocular cues do not play a role in performance within the manipulated distance range.