Screen Guided Training Does Not Capture Goal-Oriented Behaviors: Learning Myoelectric Control Mappings From Scratch Using Context Informed Incremental Learning
Human-machine interfaces based on myoelectric signals typically use screen-guided training (SGT) for model calibration, but this approach fails to capture realistic user behaviors. This study evaluates a user-in-the-loop context-informed incremental learning (CIIL) framework, comparing SGT, SGT foll...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Transactions on Neural Systems and Rehabilitation Engineering |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10802919/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841546156390219776 |
|---|---|
| author | Evan Campbell Ethan Eddy Xavier Isabel Scott Bateman Benoit Gosselin Ulysse Cote-Allard Erik Scheme |
| author_facet | Evan Campbell Ethan Eddy Xavier Isabel Scott Bateman Benoit Gosselin Ulysse Cote-Allard Erik Scheme |
| author_sort | Evan Campbell |
| collection | DOAJ |
| description | Human-machine interfaces based on myoelectric signals typically use screen-guided training (SGT) for model calibration, but this approach fails to capture realistic user behaviors. This study evaluates a user-in-the-loop context-informed incremental learning (CIIL) framework, comparing SGT, SGT followed by CIIL adaptation (SGT-A), and a novel zero-shot adaptation (ZS-A) CIIL approach that begins adapting with no prior training. Sixteen participants completed a Fitts’ Law targeting task using these control schemes, with performance measured via online throughput and offline classification accuracy. Despite lower offline accuracy, the ZS-A model achieved the highest online throughput (<inline-formula> <tex-math notation="LaTeX">$1.47~\pm ~0.46$ </tex-math></inline-formula> bits/s), significantly outperforming the SGT baseline (<inline-formula> <tex-math notation="LaTeX">$1.15~\pm ~0.37$ </tex-math></inline-formula> bits/s) and reached competitive performance within 200 seconds. To further enhance control performance, a novel adaptive sigmoid-based proportional control mapping was introduced, dynamically adjusting control signals to allow precise control near neutral positions and rapid movements at higher activation levels, better aligning with natural user behaviors. These findings demonstrate that CIIL can surpass traditional SGT methods in online performance and emphasize the value of real-time user-in-the-loop data for developing adaptable and intuitive myoelectric interfaces, with implications for prosthetics, rehabilitation, and telerobotics. |
| format | Article |
| id | doaj-art-b9c663a3ff014a42a0b5c89c31621cbd |
| institution | Kabale University |
| issn | 1534-4320 1558-0210 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Transactions on Neural Systems and Rehabilitation Engineering |
| spelling | doaj-art-b9c663a3ff014a42a0b5c89c31621cbd2025-01-11T00:00:10ZengIEEEIEEE Transactions on Neural Systems and Rehabilitation Engineering1534-43201558-02102025-01-013333234210.1109/TNSRE.2024.351805910802919Screen Guided Training Does Not Capture Goal-Oriented Behaviors: Learning Myoelectric Control Mappings From Scratch Using Context Informed Incremental LearningEvan Campbell0https://orcid.org/0000-0001-5399-4318Ethan Eddy1https://orcid.org/0000-0002-8392-3729Xavier Isabel2https://orcid.org/0009-0007-1354-5578Scott Bateman3https://orcid.org/0000-0003-3592-2163Benoit Gosselin4https://orcid.org/0000-0003-1473-3451Ulysse Cote-Allard5https://orcid.org/0000-0003-3241-8404Erik Scheme6https://orcid.org/0000-0002-4421-1016Institute of Biomedical Engineering, University of New Brunswick, Fredericton, NB, CanadaInstitute of Biomedical Engineering, University of New Brunswick, Fredericton, NB, CanadaDepartment of Electrical and Computer Engineering, Université Laval, Québec City, QC, CanadaFaculty of Computer Science, University of New Brunswick, Fredericton, NB, CanadaDepartment of Electrical and Computer Engineering, Université Laval, Québec City, QC, CanadaDepartment of Technology Systems, University of Oslo, Oslo, NorwayInstitute of Biomedical Engineering, University of New Brunswick, Fredericton, NB, CanadaHuman-machine interfaces based on myoelectric signals typically use screen-guided training (SGT) for model calibration, but this approach fails to capture realistic user behaviors. This study evaluates a user-in-the-loop context-informed incremental learning (CIIL) framework, comparing SGT, SGT followed by CIIL adaptation (SGT-A), and a novel zero-shot adaptation (ZS-A) CIIL approach that begins adapting with no prior training. Sixteen participants completed a Fitts’ Law targeting task using these control schemes, with performance measured via online throughput and offline classification accuracy. Despite lower offline accuracy, the ZS-A model achieved the highest online throughput (<inline-formula> <tex-math notation="LaTeX">$1.47~\pm ~0.46$ </tex-math></inline-formula> bits/s), significantly outperforming the SGT baseline (<inline-formula> <tex-math notation="LaTeX">$1.15~\pm ~0.37$ </tex-math></inline-formula> bits/s) and reached competitive performance within 200 seconds. To further enhance control performance, a novel adaptive sigmoid-based proportional control mapping was introduced, dynamically adjusting control signals to allow precise control near neutral positions and rapid movements at higher activation levels, better aligning with natural user behaviors. These findings demonstrate that CIIL can surpass traditional SGT methods in online performance and emphasize the value of real-time user-in-the-loop data for developing adaptable and intuitive myoelectric interfaces, with implications for prosthetics, rehabilitation, and telerobotics.https://ieeexplore.ieee.org/document/10802919/Electromyography (EMG)deep learningincremental learninggesture recognitionzero-shot |
| spellingShingle | Evan Campbell Ethan Eddy Xavier Isabel Scott Bateman Benoit Gosselin Ulysse Cote-Allard Erik Scheme Screen Guided Training Does Not Capture Goal-Oriented Behaviors: Learning Myoelectric Control Mappings From Scratch Using Context Informed Incremental Learning IEEE Transactions on Neural Systems and Rehabilitation Engineering Electromyography (EMG) deep learning incremental learning gesture recognition zero-shot |
| title | Screen Guided Training Does Not Capture Goal-Oriented Behaviors: Learning Myoelectric Control Mappings From Scratch Using Context Informed Incremental Learning |
| title_full | Screen Guided Training Does Not Capture Goal-Oriented Behaviors: Learning Myoelectric Control Mappings From Scratch Using Context Informed Incremental Learning |
| title_fullStr | Screen Guided Training Does Not Capture Goal-Oriented Behaviors: Learning Myoelectric Control Mappings From Scratch Using Context Informed Incremental Learning |
| title_full_unstemmed | Screen Guided Training Does Not Capture Goal-Oriented Behaviors: Learning Myoelectric Control Mappings From Scratch Using Context Informed Incremental Learning |
| title_short | Screen Guided Training Does Not Capture Goal-Oriented Behaviors: Learning Myoelectric Control Mappings From Scratch Using Context Informed Incremental Learning |
| title_sort | screen guided training does not capture goal oriented behaviors learning myoelectric control mappings from scratch using context informed incremental learning |
| topic | Electromyography (EMG) deep learning incremental learning gesture recognition zero-shot |
| url | https://ieeexplore.ieee.org/document/10802919/ |
| work_keys_str_mv | AT evancampbell screenguidedtrainingdoesnotcapturegoalorientedbehaviorslearningmyoelectriccontrolmappingsfromscratchusingcontextinformedincrementallearning AT ethaneddy screenguidedtrainingdoesnotcapturegoalorientedbehaviorslearningmyoelectriccontrolmappingsfromscratchusingcontextinformedincrementallearning AT xavierisabel screenguidedtrainingdoesnotcapturegoalorientedbehaviorslearningmyoelectriccontrolmappingsfromscratchusingcontextinformedincrementallearning AT scottbateman screenguidedtrainingdoesnotcapturegoalorientedbehaviorslearningmyoelectriccontrolmappingsfromscratchusingcontextinformedincrementallearning AT benoitgosselin screenguidedtrainingdoesnotcapturegoalorientedbehaviorslearningmyoelectriccontrolmappingsfromscratchusingcontextinformedincrementallearning AT ulyssecoteallard screenguidedtrainingdoesnotcapturegoalorientedbehaviorslearningmyoelectriccontrolmappingsfromscratchusingcontextinformedincrementallearning AT erikscheme screenguidedtrainingdoesnotcapturegoalorientedbehaviorslearningmyoelectriccontrolmappingsfromscratchusingcontextinformedincrementallearning |