Identifying six core motor learning variables for neurorehabilitation: a comprehensive review with natural language...

Motor learning (MoL) plays a key role in restorative therapies after neurological injury. However, its integration into clinical protocols is not yet fully systematized. This study aims to identify and organize essential MoL variables (practice, task, feedback, environment, individual, and therapist), contributing to the development of a practical checklist to support clarity, coherence, and reproducibility in neurorehabilitation.

A two-level approach was employed to identify and structure MoL variables relevant to neurorehabilitation. In the first level, three expert raters independently reviewed chapters from authoritative MoL manuals to extract and hierarchically organize variables into core, intermediate, and peripheral categories. In the second level, a comprehensive literature review was conducted following PRISMA guidelines. Natural language processing (NLP) techniques, including latent dirichlet allocation (LDA) and N-gram analysis, were used to validate and expand the initial taxonomy, assess terminology frequency, and examine semantic relationships among variables. This combined expert-driven and data-driven process ensured both theoretical consistency and empirical relevance. The final outcome of this methodology was the development of a practical and structured checklist to support the clear and reproducible application of MoL variables in neurorehabilitation protocols.

The analysis led to the identification of six core variables Practice, Task, Feedback, Environment, Individual, and Therapist consistently recognized across expert manuals and validated through NLP-based literature analysis. While four variables (Practice, Task, Feedback, Individual) were confirmed by both methods, Environment and Therapist emerged exclusively from manual analysis, highlighting their conceptual relevance despite lower recurrence in literature. Semantic analysis confirmed approximately 80% of related intermediate and peripheral variables, supporting the robustness of the taxonomy. Based on this framework, two structured checklists were developed: one to characterize each task before its delivery, and one to document key variables during execution. These tools aim to translate the taxonomy into a usable format for both clinical and research contexts.

This work introduces practical, standardized MoL checklists, bridging research and clinical practice. It supports reproducibility, optimizes rehabilitation protocols, and lays a foundation for future neurorehabilitation research and innovation.