Implicit visuomotor adaptation remains limited after several days of training

Abstract

Learning in sensorimotor adaptation tasks has been historically viewed as solely an implicit learning phenomenon. However, recent findings suggest that implicit adaptation is heavily constrained, calling into question its utility in motor learning, and the theoretical framework of sensorimotor adaptation paradigms. These inferences have been based mainly on results from single bouts of training, thus, it is possible that implicit adaptation processes supersede explicit compensation strategies, such as explicitly re-aiming their intended movement direction, over repeated practice sessions. We tested this by dissociating the contributions of explicit re-aiming strategies and implicit adaptation over five consecutive days of training. Implicit adaptation plateaued at a value far short of complete learning. We sought to determine if these constraints on implicit adaptation extend to another sensorimotor task, mirror reversal. Implicit adaptation was inappropriate for mirror reversal and became suppressed during training. These findings are consistent with a handful of recent studies suggesting that implicit adaptation processes, as studied in sensorimotor adaptation paradigms, cannot fully support long-term motor skill. New and NoteworthyIn this set of studies, we find that implicit adaptation cannot fully account for learning in adaptation tasks, such as the visuomotor rotation and mirror reversal tasks, even following several days of training. In fact, implicit adaptation can be counterproductive to learning. These findings question the utility of implicit adaptation processes to motor learning more broadly.

Contextual effects in sensorimotor adaptation adhere to associative learning rules

“…However, a number of lines of evidence have highlighted how implicit adaptation, when isolated from other learning processes, is a curiously stereotyped learning system that can even, given certain experimental constraints, produce outputs that are counterproductive to performance (Mazzoni and Krakauer, 2006; Morehead et al, 2017). For example, participant’s initial response to a mirror reversal visual perturbation is in the opposite direction to the optimal solution, as if the visual feedback is rotated (Wilterson and Taylor, 2020). These observations have motivated recent efforts to reframe sensorimotor adaptation as the result of more rigid “direct policy updating” rather than the updating of a flexible forward model (Hadjiosif et al, 2020).…”

“…When properly isolated from other processes that contribute to learning in typical adaptation tasks, such as explicit cognitive strategies (McDougle et al, 2016), implicit adaptation appears to be curiously limited: It does not exhibit savings (Avraham et al, 2020; Haith et al, 2015; Morehead et al, 2015), is generally insensitive to both the magnitude and consistency of errors (Bond and Taylor, 2015; Hutter and Taylor, 2018; Avraham et al, 2019), and appears to be impervious to information relevant to task performance (Mazzoni and Krakauer, 2006; Morehead et al, 2017; but see Leow et al, 2018 and Kim et al, 2019). Moreover, as shown in several recent studies, implicit adaptation fails, at least initially, in producing appropriate changes in behavior in response to other simple perturbations like mirror reversals (i.e., when visual feedback is mirrored across an unseen axis with respect to the veridical limb position), a task that should be natural for a canonical forward model to solve (Hadjiosif et al, 2020; Telgen et al, 2014; Wilterson and Taylor, 2020).…”

De novolearning versus adaptation of continuous control in a manual tracking task

“…It has been shown that people use re-aiming strategies—or more generally, cognitive strategies—to compensate (at least in part) for visuomotor rotations ( Mazzoni and Krakauer, 2006; de Rugy et al, 2012; Taylor et al, 2014; Morehead et al, 2015 ) and force fields ( Schween et al, 2019 ). In principle, re-aiming enables people to compensate for arbitrary re-mappings of their environment, including large (90°) visuomotor rotations ( Bond and Taylor, 2015 ) or mirror-reversed visual feedback ( Wilterson and Taylor, 2019 ). However, implementing re-aiming is cognitively demanding and time-consuming process that significantly increases reaction times ( Haith et al, 2015; Leow et al, 2017; McDougle and Taylor, 2019; Fernández-Ruiz et al, 2011 ).…”

Did we get sensorimotor adaptation wrong? Implicit adaptation as direct policy updating rather than forward-model-based learning

“…and explicit contributions to learning under mirror--reversal by collecting aiming reports (Wilterson & 363 Taylor, 2019) (though the veracity of such aiming reports for assessing implicit learning has recently been 364 called into question (Maresch & Donchin, 2019)). Several other studies have also shown that, although 365 error tends to increase during the first few trials of exposure to a mirror reversal, errors start to decrease 366 later in learning (Lillicrap et al, 2013;Wilterson & Taylor, 2019). This reversal may be attributable to 367 learning a revised sensitivity derivative linking observed error to a corrective update (Abdelghani & 368 Tweed, 2010; Kasuga et al, 2015) or to downregulation of implicit adaptation in favor of an explicit 369 solution.…”