Abstract
Introduction: Asymmetric prior probabilities and reward schedules induce systematic biases in sensorimotor choices; the proportion of more likely or valuable choices increases while response times for those choices decrease. However, it remains unknown to what extent perception may also be affected. Methods: On each trial, two spots (diameter: 0.6 deg, peak luminance: 4.7 or 6.3 cd/m2) appeared on either side of fixation (±6 deg) against a noisy (SD: 8.2 cd/m2) background (mean: 37.6 cd/m2). Five observers (3 naïve) were asked to make a 2AFC saccadic choice to the brighter spot. For five observers, we biased this choice using asymmetric spatial probability (75% rightward or leftward); for two observers, we also used an asymmetric reward schedule (for a total of 7 cases). Upon saccade initiation, the spots and central cross were extinguished and the latter replaced by a test spot visible for 250 ms. Observers were then asked to report whether the earlier saccadic target or the later test pedestal was brighter in a 2IFC perceptual choice. We constructed oculometric curves and derived the saccadic point-of-subjective-equality (PSE) for favored and unfavored locations to quantify overall motor bias. We also constructed psychometric functions to quantify perceptual gain (the relationship between perceived brightness and signal strength) and uncertainty (the slope of the psychometric functions) at favored and unfavored locations. Results: As expected, the experimental manipulations induced motor biases in all cases (mean PSE shift: 2.0 d' units). They also increased perceptual gain (mean: 21%) and uncertainty (mean: 11%) in a correlated manner (r2=0.53; p2 = 0.36, pConclusion: Our findings are consistent with a neural mechanism in which motor and perceptual decisions share a common early weighting factor that scales both visual signals and their limiting noise.
Supported by: NASA Human Research Program (Space Human Factors Engineering).