Abstract
Background. Action measures are commonly used to assess the accuracy of perception. However, actions can be calibrated to any stable perceptual experience because stability affords perceptuo-motor adaptation and prediction. Accuracy at action tasks therefore does not require accurate perception. To illustrate this, we measured dynamic pointing after locomotor adaptation. Method. We adapted participants to altered visual feedback during walking in an immersive panoramic virtual environment. Specifically, we doubled the visual gain of walking for one set of participants and adapted the second set to a normal gain. Following adaptation participants were tested in a dynamic pointing task while walking without visual feedback. That is, they started each trial pointing to an obliquely positioned target in the visual world and then tried to track the stationary (remembered) target with their hand while walking with eyes closed. Results. Participants adapted to a normal gain tracked the egocentric location of the targets relatively accurately. However, those adapted to doubled visual gain rotated their hand too rapidly even though they had never practiced pointing in the adapting environment. In contrast, explicit estimates of egocentric distance to similar targets did not differ as a function of adaptation condition, but underestimated egocentric distance in near space by a factor of about 0.75. Evidently adaptation altered spatial updating during locomotion without altering perceived target distance. Conclusions. A great deal of evidence suggests that egocentric distance is misperceived. Locomotor action measures, such as pointing, can be accomplished successfully if spatial updating of egocentric position is calibrated by experience. Thus any misperception of egocentric distance could be masked by such measures. Dynamic pointing tasks are informative about spatial updating, but they are not directly informative about the absolute scaling of perceived space. Actions can be calibrated to successfully reflect perceptuo-motor expectancies even in the presence of inaccurately scaled perceptions.