ROLE OF WORKING MEMORY IN INTERLIMB GENERALIZATION OF NEWLY LEARNED SKILLS

Abstract

Newly acquired skill memory can generalize/transfer to the untrained arm. Such interlimb generalization of a learned skill has been shown to be symmetric in nature and is thought to be mediated by cognitive processes that emerge during skill learning. However, it is unknown whether engaging in other cognitively demanding tasks following skill acquisition can influence skill generalization. Our research goal was to uncover how a secondary task, involving working memory, interacts with a newly formed skill memory and influences subsequent interlimb generalization. To test this idea, we conducted a set of three experiments by recruiting right-handed young healthy individuals (N=92) who learned a novel motor skill (long or short training on a skilled reaching task) followed by performing a working memory or control task with the right arm. Finally, all individuals were tested for immediate or delayed (after 24 hours) interlimb skill generalization to the untrained left arm. We found significant immediate as well as delayed generalization in individuals who received long training on the motor skill task, irrespective of whether they performed working memory or control task. On the other hand, performing the working memory but not control task following short skill training impaired generalization when the untrained arm was tested 24 hours later. These findings indicate that short training reflecting early stages of skill learning and the subsequent skill memory stabilization are dependent on working memory such that the underlying neural interactions mediating these processes can have implications for skill generalization. NEW AND NOTEWORTHYSkill learning is deemed successful if it can generalize to untrained conditions. Despite the potential clinical relevance of interlimb skill generalization for motor rehabilitation, such as in case of unilateral stroke, the underlying mechanisms that can influence such generalization are not fully understood. Specifically, whether interlimb skill memory generalization can be impacted by another memory system is unknown. Here, we show that engaging in a working memory task following short skill training is detrimental for interlimb skill generalization, and such interference caused by working memory is not evident when the skill training is longer. These behavioral findings bridge the link between traditionally distinct memory systems, i.e., procedural skill memory and working memory.