Age-related variations of visuo-motor adaptation beyond explicit knowledge

Abstract

Visuo-motor adaptation suffers at older working age. The age-related decline of behavioral adjustments is accompanied by reduced explicit knowledge of the visuo-motor transformation. It disappears when explicit knowledge is kept constant across the age range, except for particularly high levels of explicit knowledge. According to these findings, at older adult age both the acquisition of explicit knowledge and its application for strategic corrections become poorer. Recently it has been posited that visuo-motor adaptation can involve model-free reinforcement mechanisms of learning in addition to model-based mechanisms. We tested whether age-related declines of reinforcement learning can also contribute to the age-related changes of visuo-motor adaptation. Therefore we enhanced the contribution of reinforcement learning to visuo-motor adaptation by way of introducing salient markers of success and failure during practice. With such modified practice conditions, there were residual age-related variations of behavioral adjustments at all levels of explicit knowledge, even when explicit knowledge was absent. The residual age-related variations were observed for practiced target directions only, but not for new target directions. These findings are consistent with an age-related decline of model-free reinforcement learning as a third factor in the age-related decline of visuo-motor adaptation. Under practice conditions, which spur model-free reward-based learning, this factor adds to the decrements of the acquisition of explicit knowledge and its use for strategic corrections.

Keywords: after-effect; explicit knowledge; model-based learning; reinforcement learning; visuo-motor rotation.

FIGURE 1

Mean proportion of hits (A), and mean initial direction errors (B) during practice of the virtual putting task. Error bars mark standard errors of the means.

FIGURE 2

Mean adaptive shifts (A), after-effects (B), and explicit shifts (C) in deg, shown separately for short and long target amplitudes, practiced and new target directions, and both age groups. Error bars mark standard errors of the means.

FIGURE 3

The regressions of individual adaptive shifts for practiced (A) and new target directions (B) and of individual after-effects for practiced (C) and new target directions (D) on individual explicit shifts. Filled circles mark the individual data points of young participants, open circles those of older participants. The zero points on the abscissae and ordinates are marked by vertical and horizontal lines, respectively.