Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Calibration-Free Electromyography Motor Intent Decoding Using Large-Scale Supervised Pretraining

Olsson, Alexander E. LU ; Maleševic, Nebojša LU ; Björkman, Anders and Antfolk, Christian LU orcid (2026) In Advanced Intelligent Systems
Abstract

Machine learning algorithms for myoelectric pattern recognition require substantial user-specific training data, limiting broader applications of electromyography (EMG) in human–computer interfacing. Here, we present a framework for EMG-mediated motor intent decoding designed to function for new users without collecting user-specific training data. We introduce a Transformer-based architecture, termed the Spatially Aware Feature-learning Transformer (SAFT), which processes EMG time windows with variable numbers of channels from arbitrary spatial electrode configurations by combining channel-wise temporal feature extraction with learned spatial encoding of electrode positions and attention across channels. This enables training of a... (More)

Machine learning algorithms for myoelectric pattern recognition require substantial user-specific training data, limiting broader applications of electromyography (EMG) in human–computer interfacing. Here, we present a framework for EMG-mediated motor intent decoding designed to function for new users without collecting user-specific training data. We introduce a Transformer-based architecture, termed the Spatially Aware Feature-learning Transformer (SAFT), which processes EMG time windows with variable numbers of channels from arbitrary spatial electrode configurations by combining channel-wise temporal feature extraction with learned spatial encoding of electrode positions and attention across channels. This enables training of a single model across heterogeneous EMG datasets. In the present study, large-scale supervised pretraining refers to pretraining on a pooled corpus of 29 public EMG databases comprising 506 subjects, 108 movement classes, and ≈9.9 million nonrest EMG windows after preprocessing. A pretrained SAFT model was fine-tuned on a held-out database and evaluated for cross-user performance. On the 3DC benchmark, the pretrained-only model achieved 28.7% balanced accuracy (vs. 10% chance), while pretrained and fine-tuned cross-user SAFT models achieved 81.8% balanced accuracy, comparable to conventional user-specific linear discriminant analysis (LDA) models (82.9%). These findings indicate the feasibility of EMG intent decoding models that work “out of the box” without end-user calibration.

(Less)
Please use this url to cite or link to this publication:
author
; ; and
organization
publishing date
type
Contribution to journal
publication status
epub
subject
keywords
electromyography, human–computer interfacing, machine learning, myoelectric control, zero-shot learning
in
Advanced Intelligent Systems
publisher
John Wiley & Sons Inc.
external identifiers
  • scopus:105035475599
ISSN
2640-4567
DOI
10.1002/aisy.202500791
language
English
LU publication?
yes
id
52891a08-4e38-4077-9b2c-39c1ac3ff718
date added to LUP
2026-06-23 15:28:50
date last changed
2026-06-23 15:29:04
@article{52891a08-4e38-4077-9b2c-39c1ac3ff718,
  abstract     = {{<p>Machine learning algorithms for myoelectric pattern recognition require substantial user-specific training data, limiting broader applications of electromyography (EMG) in human–computer interfacing. Here, we present a framework for EMG-mediated motor intent decoding designed to function for new users without collecting user-specific training data. We introduce a Transformer-based architecture, termed the Spatially Aware Feature-learning Transformer (SAFT), which processes EMG time windows with variable numbers of channels from arbitrary spatial electrode configurations by combining channel-wise temporal feature extraction with learned spatial encoding of electrode positions and attention across channels. This enables training of a single model across heterogeneous EMG datasets. In the present study, large-scale supervised pretraining refers to pretraining on a pooled corpus of 29 public EMG databases comprising 506 subjects, 108 movement classes, and ≈9.9 million nonrest EMG windows after preprocessing. A pretrained SAFT model was fine-tuned on a held-out database and evaluated for cross-user performance. On the 3DC benchmark, the pretrained-only model achieved 28.7% balanced accuracy (vs. 10% chance), while pretrained and fine-tuned cross-user SAFT models achieved 81.8% balanced accuracy, comparable to conventional user-specific linear discriminant analysis (LDA) models (82.9%). These findings indicate the feasibility of EMG intent decoding models that work “out of the box” without end-user calibration.</p>}},
  author       = {{Olsson, Alexander E. and Maleševic, Nebojša and Björkman, Anders and Antfolk, Christian}},
  issn         = {{2640-4567}},
  keywords     = {{electromyography; human–computer interfacing; machine learning; myoelectric control; zero-shot learning}},
  language     = {{eng}},
  publisher    = {{John Wiley & Sons Inc.}},
  series       = {{Advanced Intelligent Systems}},
  title        = {{Calibration-Free Electromyography Motor Intent Decoding Using Large-Scale Supervised Pretraining}},
  url          = {{http://dx.doi.org/10.1002/aisy.202500791}},
  doi          = {{10.1002/aisy.202500791}},
  year         = {{2026}},
}