Implications of plan-based generalization in sensorimotor adaptation

Abstract

Generalization is a fundamental aspect of behavior, allowing for the transfer of knowledge from one context to another. The details of this transfer are thought to reveal how the brain represents what it learns. Generalization has been a central focus in studies of sensorimotor adaptation, and its pattern has been well characterized: Learning of new dynamic and kinematic transformations in one region of space tapers off in a Gaussian-like fashion to neighboring untrained regions, echoing tuned population codes in the brain. In contrast to common allusions to generalization in cognitive science, generalization in visually guided reaching is usually framed as a passive consequence of neural tuning functions rather than a cognitive feature of learning. While previous research has presumed that maximum generalization occurs at the instructed task goal or the actual movement direction, recent work suggests that maximum generalization may occur at the location of an explicitly accessible movement plan. Here we provide further support for plan-based generalization, formalize this theory in an updated model of adaptation, and test several unexpected implications of the model. First, we employ a generalization paradigm to parameterize the generalization function and ascertain its maximum point. We then apply the derived generalization function to our model and successfully simulate and fit the time course of implicit adaptation across three behavioral experiments. We find that dynamics predicted by plan-based generalization are borne out in the data, are contrary to what traditional models predict, and lead to surprising implications for the behavioral, computational, and neural characteristics of sensorimotor adaptation. The pattern of generalization is thought to reveal how the motor system represents learned actions. Recent work has made the intriguing suggestion that maximum generalization in sensorimotor adaptation tasks occurs at the location of the learned movement plan. Here we support this interpretation, develop a novel model of motor adaptation that incorporates plan-based generalization, and use the model to successfully predict surprising dynamics in the time course of adaptation across several conditions.

Implicit visuomotor adaptation remains limited after several days of training

“…This was made possible by asking participants to report their intended aiming location (i.e. the direction that they intend to move in order to compensate for the perturbation), by verbally specifying a location relative to the target (Taylor et al, 2014, Bond andMcDougle et al 2015;Brudner et al 2016;McDougle et al 2017;Butcher et al 2017), or by tapping the surface of a touch screen Hutter and Taylor, 2018). In these experiments, participants reported their intended aiming location and then make a reaching movement.…”

Explicit strategies in force field adaptation

“…Since we instruct our participants to abandon the strategy at the beginning of the clamp phase, the two-state model would expect the explicit state to become disengaged and for behavior to abruptly switch to reflect only the implicit state, which should be in field A-direction and decay from there. Thus, whereas No Push channels during initial clamp not being at the same level as No Push channels in late field B can be explained by generalization of implicit learning being locally tied to the explicit movement plan (Hirashima and Nozaki 2012;Day et al 2016;Sheahan et al 2016;McDougle et al 2017;Schween et al 2018), the gradual drift that we observe at the beginning of the clamp phase indicates additional properties of underlying learning, such as a combination of two implicit and one explicit timescale (Miyamoto et al 2014), or a gradual attenuation of strategic aiming rather than an immediate one.…”

“…Results obtained by these methods suggest that explicit learning may underlie a range of behaviors that were previously explained by extensions of implicit internal models, such as savings (Morehead et al 2015), structural learning (Bond and Taylor 2017), and context-dependent learning by abstract cues (Hegele and Heuer 2010;Schween et al 2019), thus highlighting its importance for motor learning (see Krakauer et al 2019 for a recent review).…”

“…The position-dependent component has subsequently been attributed to a different, model-free learning mechanism(Haith et al 2011). Considering that recent studies have found proposed model-free reinforcement learning to depend on explicit strategies(Shmuelof et al 2012;Codol et al 2018;Holland et al 2018), this can be taken to suggest that the strategic component of learning should depend on position and, in this respect, resemble off-target aiming typical of cursor rotation experiments(Benson et al 2011;Taylor et al 2014;Bond and Taylor 2017). Alternatively, strategies could also depend on velocity.…”

Of hands, tools, and exploding dots: How different action states and effects separate visuomotor memories

“…This capacity has been investigated in dual adaptation experiments, where different cues are linked with different -often conflicting -sensorimotor transformations to determine the extent with which the cues enable the formation of separate visuomotor memories (Ghahramani and Wolpert 1997;Seidler et al 2001;Imamizu et al 2003;Osu et al 2004;Bock et al 2005;Woolley et al 2007Woolley et al , 2011Hinder et al 2008;Hegele and Heuer 2010;Howard et al 2012Howard et al , 2013Howard et al , 2015Ayala et al 2015;van Dam and Ernst 2015;Sheahan et al 2016Sheahan et al , 2018Heald et al 2018). Whereas earlier studies have predominantly found that the posture or initial state of the body act as sufficient cues (Gandolfo et al 1996;Ghahramani and Wolpert 1997;Seidler et al 2001;Howard et al 2013), a number of more recent studies observed that distinct movement plans effectively separate memories for sensorimotor transformations (Hirashima and Nozaki 2012;Howard et al 2015;Day et al 2016;Sheahan et al 2016;McDougle et al 2017;Schween et al 2018), even when these plans are not ultimately executed (Sheahan et al 2016).…”