Abstract
The accuracy of open-loop manual pointing to a perceptually mislocalized visual target increases linearly with hand-to-body distance, approaching complete accuracy at full extension with errors equal to the perceptual mislocalization for the hand close to the body (Psychonomics, 1999). Three experiments measured the time course of this distance dependence in pointing and height-matching to a small target 80 cm distant at true eye level. A 50 deg-long line pitched 30 deg topbackward or 20 deg topforward, 25 deg eccentric to the median plane in otherwise total darkness, induced systematic errors in the target's perceived elevation. During the first 5 minutes of each experiment the distance dependence was measured with both inducing line and target present, showing 12 deg manual errors in elevation with the hand in the midfrontal plane, 6 deg errors at 40 cm, and less than 2 deg errors at full arm extension for 30 deg inducing pitch. Following the first 5 minutes, (1) in Expt. 1 the inducing line was extinguished, and errors at all hand-to-body distances decayed to baseline from the induced level along the same near-exponential function (time constant=3 minutes); (2) in Expt. 2 the target was extinguished and the inducer remained, and the distance-dependent manual differences during the earlier 5-minute period persisted; and (3) in Expt. 3 when both the inducing line and target were removed, the manual errors all went to the large values measured during the earlier 5-minute period when both target and inducer were present. No important differences were obtained between pointing and height-matching. The same patterns of response were found with both inducing pitches. Thus, the distance-dependent accuracy of the sensorimotor response to a perceptually mislocalized target is not simply a disconnection between perception and action, but depends strongly on the presence of the inducing visual field.
Support: NIH grant EY10534.