Abstract
To generate adaptive movements we must generalize what we have previously learned to novel situations. The generalization of adapted movements has typically been framed as a consequence of neural tuning functions that overlap for similar movement kinematics - what might be considered bottom-up generalization. However, as is true in many domains of human behavior, generalization can also be framed as the result of deliberate decisions about how to act (top-down generalization). Here we attempt to broaden the scope of theories about motor generalization, hypothesizing that part of the typical motor generalization function can be characterized as a consequence of top-down decisions concerning the subjective similarity of different movement contexts. We tested this proposal by having participants make explicit similarity ratings over both traditional kinematic contextual dimensions (movement direction) and more abstract contextual dimensions (target shape), and perform a visuomotor adaptation generalization task where trials varied over those dimensions. Across five experiments, we measured the relationship between subjective similarity ratings and motor generalization. In some cases this link was rather strong, though it was determined by both task-relevance and explicit instruction. These results support a broadening of the descriptive framework used to understand the generalization of motor behaviors and support a more careful deployment of instructions in generalization studies.
Significance Statement Generalization describes the transfer of knowledge from one context to another, and is typically thought to result from a higher-order inference process. However, in the motor adaptation domain, generalization is thought to arise from neural representations tuned to low-level kinematics. To bridge these differing views, we measured peoples’ subjective similarity judgements of different task contexts during sensorimotor adaptation. We found that motor generalization was closely linked to participant’s subjective judgements, and that explicit instructions about the consequential dimension of different contexts further shaped generalization. These findings emphasize that in addition to low level kinematic considerations, top-down inferences about which action to take in a given context should be considered as another key component of motor generalization.
Competing Interest Statement
The authors have declared no competing interest.
Footnotes
Competing Interest Statement: The authors declare no competing interest.
New experiment 5 added. Figure 6 added