Evaluating task characteristics as predictors of continuous hand motion decoding performance
Year: 2025
Type: article
Abstract: Prosthetic hand control using biosignals requires robust decoding across diverse real-world conditions. We systematically evaluated how task characteristics influence muscular activity-based continuous hand motion decoding performance using the MyoKi database. Multiple linear regression with forward selection revealed that muscle fatigue modestly predicted task-wise decoding accuracy. Categorical task characteristics, such as grasp type, vertical task location, and force level, showed no significant effects on within-participant decoding performance in repeated measures ANOVAs, which demonstrates that the decoder maintains stable performance across diverse task characteristics. This is promising for prosthetic applications requiring robust performance across daily activities.
Source: Infinite Science GmbH
Authors: Andreas, Daniel, Beckerle, Philipp
Institution: Friedrich-Alexander-Universität Erlangen-Nürnberg
Language: English
Subfield: Biomedical Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7140918373
FWCI:
Cites:
Cited by:
Related to: