Abstract
In order to perceive and interact with the visual world local image characteristics (e.g., orientation, spatial scale) arising from V1 must be appropriately distributed to both dorsal and ventral pathways. While significant evidence exists suggesting the importance of V4 for establishing the link between simple form analysis and the perception of complex object based scenes (Pasupathy & Connor 2002) little attention has been given to how similar processes are accomplished in the motor (dorsal) pathway. In this investigation we examined the relative precision and congruence of perceptual and motor estimates of luminance defined object boundaries. Twelve participants were asked to (a) point at, or (b) estimate the size of circles presented along their AP axis. Circles ranged in size from 1–4 cm and provided either a clear edge boundary (white/black transition) or were defined by a 2-d Gaussian such that the circle center was white and degraded to black at 2sd. Overall, participants demonstrated decreased RT associated with both increasing target size and edge salience. Further, and in accord with the seminal work of Fitts' (1954), participant movement times decreased with increasing target size. However, movements made to the Gaussian-blurred images were performed more rapidly. In order to more directly examine size estimates across perceptual and motor systems, an effective target width was calculated for pointing movements to all targets and the 95% confidence interval was compared to the mean perceived target size. While both motor and perceptual system scaled to target size the perceived target estimates were larger for blurred objects; the endpoint variability of the motor system showed not sensitivity to boundary blur. Thus it appears that the ventral and dorsal systems use edge features arising from early visual areas differentially when ‘constructing’ visual objects.
Natural Sciences and Engineering Research Council of Canada (GB, MH).