It is well established that older adults experience deficits in motor control, but how aging impacts motor adaptation remains a subject of considerable debate. Some studies suggest prominent learning impairments; however, others have shown no age-related differences. These mixed results may be explained by the fact conventional adaptation tasks engage both implicit and explicit learning processes, with aging having a differential impact on each process. Moreover, previous studies have typically involved small sample sizes reducing sensitivity to detect possible effects of aging.

To address these two concerns, we re-examined this question using methods that isolate either implicit or explicit learning processes in visuomotor rotation tasks (Experimental protocols were approved by UC Berkeley's IRB). In Experiment 1, we used clamped feedback to isolate implicit adaptation and collected a large sample using a web-based crowdsourcing approach (n = 100; 50/group). Participants reached to one of four targets and during the perturbation block, the cursor always following an invariant trajectory, rotated by 45º from the target position (non-contingent). Despite being asked to ignore the cursor, participants in both groups exhibited implicit adaptation in the opposite direction of the cursor. Strikingly, the results show that implicit adaptation was enhanced with age: Whereas the younger participants adapted 17.8° (sd=1.6°), the older participants adapted 24.4° (sd=1.5°) (t(97) = 3.0, p = 0.004). In Experiment 2, we used delayed feedback to isolate explicit re-aiming (n = 200; 100/group). Specifically, participants reached to one of three targets, receiving endpoint feedback, with the cursor appearing after an 800 ms delay. During the perturbation block, the cursor position was rotated by 60º from the actual hand position (contingent). Strategic re-aiming was attenuated in the older participants. At the end of the perturbation block the younger participants had adapted 33.3° (sd=3.6°) whereas the older participants had adapted 20.2° (sd=3.7°) (t(98) = 2.6, p = 0.01).

We complemented these experiments with a meta-analysis, focusing on 60 studies that have reported data relevant to the question of the effect of aging on motor adaptation. We focused on two key dependent variables: Late adaptation, a measure that likely reflects the contribution of both implicit and explicit processes, and the aftereffect, a measure of implicit adaptation. Across studies, the aftereffect was  enhanced in older adults (0.4 (95% CI = [0.2, 0.6]; positive = greater in older adults). Late adaptation showed the opposite effect, with this measure attenuated in older adults (mean effect size [CI]: -0.4 [-0.6,-0.2]). Assuming this attenuation is likely driven by an explicit re-aiming deficits in older adults, the results of the meta-analysis converge with our experimental results, pointing to a differential effect of age on implicit and explicit motor learning processes.