Abstract
Generalization represents the ability to transfer knowledge, e.g., motor skills, from one context to another. In daily life, motor learning often occurs in an environment where various stimuli compete for limited attentional resources. Here, we examined whether divided attention during learning alters the generalization of visuomotor adaptation. We formed No-Load and High-Load groups to perform a dual-task paradigm: a visuomotor adaptation task (45° CCW rotation) and a concurrent attention demanding rapid serial visual presentation task (RSVP) in which a stream of five inverted or upright ‘T’ in different colors appear sequentially. The High-Load group was required to detect a conjunction "T" target (color and orientation) and report the number of detected T’s (1, 2 or 3) after each trial by a button response. The No-Load group was told to ignore the RSVP stream and instead made a button response after each trial. Both groups first trained (45° CCW) in one direction (Adaptation) and then had to reach to other untrained directions (Generalization). Results showed that the RSVP accuracy in the High-Load group was 61.3% ± 4.9 (chance level was 33%), suggesting that the High-Load group allocated some attention to the RSVP task, limiting available attentional resources for the adaptation task. We found similar levels of Adaptation (i.e., reaching accuracy) for both groups, but the No-Load group had better Generalization (i.e. accuracy to untrained directions) than the High-Load group. Specifically, we demonstrated that divided attention reduced gain and sharpened tuning of the generalization function. Thus, dividing attention during learning restricts the range and gain of generalization of visuomotor adaptation, highlighting a critical role for attention in motor learning. This result might suggest that attention can modulate subpopulation(s) of neurons in motor areas, which have directional tuning in or near the training direction.
Meeting abstract presented at VSS 2013