Artificial intelligence tools for engagement prediction in neuromotor disorder patients during rehabilitation

Abstract Background Robot-Assisted Gait Rehabilitation (RAGR) is an established clinical practice to encourage neuroplasticity in patients with neuromotor disorders. Nevertheless, tasks repetition imposed by robots may induce boredom, affecting clinical outcomes. Thus, quantitative assessment of eng...

Full description

Saved in:
Bibliographic Details
Main Authors: Simone Costantini, Anna Falivene, Mattia Chiappini, Giorgia Malerba, Carla Dei, Silvia Bellazzecca, Fabio A. Storm, Giuseppe Andreoni, Emilia Ambrosini, Emilia Biffi
Format: Article
Language:English
Published: BMC 2024-12-01
Series:Journal of NeuroEngineering and Rehabilitation
Subjects:
Online Access:https://doi.org/10.1186/s12984-024-01519-2
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850101804954550272
author Simone Costantini
Anna Falivene
Mattia Chiappini
Giorgia Malerba
Carla Dei
Silvia Bellazzecca
Fabio A. Storm
Giuseppe Andreoni
Emilia Ambrosini
Emilia Biffi
author_facet Simone Costantini
Anna Falivene
Mattia Chiappini
Giorgia Malerba
Carla Dei
Silvia Bellazzecca
Fabio A. Storm
Giuseppe Andreoni
Emilia Ambrosini
Emilia Biffi
author_sort Simone Costantini
collection DOAJ
description Abstract Background Robot-Assisted Gait Rehabilitation (RAGR) is an established clinical practice to encourage neuroplasticity in patients with neuromotor disorders. Nevertheless, tasks repetition imposed by robots may induce boredom, affecting clinical outcomes. Thus, quantitative assessment of engagement towards rehabilitation using physiological data and subjective evaluations is increasingly becoming vital. This study aimed at methodologically exploring the performance of artificial intelligence (AI) algorithms applied to structured datasets made of heart rate variability (HRV) and electrodermal activity (EDA) features to predict the level of patient engagement during RAGR. Methods The study recruited 46 subjects (38 underage, 10.3 ± 4.0 years old; 8 adults, 43.0 ± 19.0 years old) with neuromotor impairments, who underwent 15 to 20 RAGR sessions with Lokomat. During 2 or 3 of these sessions, ad hoc questionnaires were administered to both patients and therapists to investigate their perception of a patient’s engagement state. Their outcomes were used to build two engagement classification targets: self-perceived and therapist-perceived, both composed of three levels: “Underchallenged”, “Minimally Challenged”, and “Challenged”. Patient’s HRV and EDA physiological signals were processed from raw data collected with the Empatica E4 wristband, and 33 features were extracted from the conditioned signals. Performance outcomes of five different AI classifiers were compared for both classification targets. Nested k-fold cross-validation was used to deal with model selection and optimization. Finally, the effects on classifiers performance of three dataset preparation techniques, such as unimodal or bimodal approach, feature reduction, and data augmentation, were also tested. Results The study found that combining HRV and EDA features into a comprehensive dataset improved the synergistic representation of engagement compared to unimodal datasets. Additionally, feature reduction did not yield any advantages, while data augmentation consistently enhanced classifiers performance. Support Vector Machine and Extreme Gradient Boosting models were found to be the most effective architectures for predicting self-perceived engagement and therapist-perceived engagement, with a macro-averaged F1 score of 95.6% and 95.4%, respectively. Conclusion The study displayed the effectiveness of psychophysiology-based AI models in predicting rehabilitation engagement, thus promoting their practical application for personalized care and improved clinical health outcomes.
format Article
id doaj-art-5d656a4d9433480eb2848a10b07a86df
institution DOAJ
issn 1743-0003
language English
publishDate 2024-12-01
publisher BMC
record_format Article
series Journal of NeuroEngineering and Rehabilitation
spelling doaj-art-5d656a4d9433480eb2848a10b07a86df2025-08-20T02:39:55ZengBMCJournal of NeuroEngineering and Rehabilitation1743-00032024-12-0121112110.1186/s12984-024-01519-2Artificial intelligence tools for engagement prediction in neuromotor disorder patients during rehabilitationSimone Costantini0Anna Falivene1Mattia Chiappini2Giorgia Malerba3Carla Dei4Silvia Bellazzecca5Fabio A. Storm6Giuseppe Andreoni7Emilia Ambrosini8Emilia Biffi9Department of Electronics Information and Bioengineering, Politecnico di MilanoScientific Institute, IRCCS “E. Medea”Scientific Institute, IRCCS “E. Medea”Scientific Institute, IRCCS “E. Medea”Scientific Institute, IRCCS “E. Medea”Scientific Institute, IRCCS “E. Medea”Scientific Institute, IRCCS “E. Medea”Scientific Institute, IRCCS “E. Medea”Department of Electronics Information and Bioengineering, Politecnico di MilanoScientific Institute, IRCCS “E. Medea”Abstract Background Robot-Assisted Gait Rehabilitation (RAGR) is an established clinical practice to encourage neuroplasticity in patients with neuromotor disorders. Nevertheless, tasks repetition imposed by robots may induce boredom, affecting clinical outcomes. Thus, quantitative assessment of engagement towards rehabilitation using physiological data and subjective evaluations is increasingly becoming vital. This study aimed at methodologically exploring the performance of artificial intelligence (AI) algorithms applied to structured datasets made of heart rate variability (HRV) and electrodermal activity (EDA) features to predict the level of patient engagement during RAGR. Methods The study recruited 46 subjects (38 underage, 10.3 ± 4.0 years old; 8 adults, 43.0 ± 19.0 years old) with neuromotor impairments, who underwent 15 to 20 RAGR sessions with Lokomat. During 2 or 3 of these sessions, ad hoc questionnaires were administered to both patients and therapists to investigate their perception of a patient’s engagement state. Their outcomes were used to build two engagement classification targets: self-perceived and therapist-perceived, both composed of three levels: “Underchallenged”, “Minimally Challenged”, and “Challenged”. Patient’s HRV and EDA physiological signals were processed from raw data collected with the Empatica E4 wristband, and 33 features were extracted from the conditioned signals. Performance outcomes of five different AI classifiers were compared for both classification targets. Nested k-fold cross-validation was used to deal with model selection and optimization. Finally, the effects on classifiers performance of three dataset preparation techniques, such as unimodal or bimodal approach, feature reduction, and data augmentation, were also tested. Results The study found that combining HRV and EDA features into a comprehensive dataset improved the synergistic representation of engagement compared to unimodal datasets. Additionally, feature reduction did not yield any advantages, while data augmentation consistently enhanced classifiers performance. Support Vector Machine and Extreme Gradient Boosting models were found to be the most effective architectures for predicting self-perceived engagement and therapist-perceived engagement, with a macro-averaged F1 score of 95.6% and 95.4%, respectively. Conclusion The study displayed the effectiveness of psychophysiology-based AI models in predicting rehabilitation engagement, thus promoting their practical application for personalized care and improved clinical health outcomes.https://doi.org/10.1186/s12984-024-01519-2Robot-Assisted Gait RehabilitationEngagementPsychophysiological signalsClassificationK-Nearest NeighborsRandom forest
spellingShingle Simone Costantini
Anna Falivene
Mattia Chiappini
Giorgia Malerba
Carla Dei
Silvia Bellazzecca
Fabio A. Storm
Giuseppe Andreoni
Emilia Ambrosini
Emilia Biffi
Artificial intelligence tools for engagement prediction in neuromotor disorder patients during rehabilitation
Journal of NeuroEngineering and Rehabilitation
Robot-Assisted Gait Rehabilitation
Engagement
Psychophysiological signals
Classification
K-Nearest Neighbors
Random forest
title Artificial intelligence tools for engagement prediction in neuromotor disorder patients during rehabilitation
title_full Artificial intelligence tools for engagement prediction in neuromotor disorder patients during rehabilitation
title_fullStr Artificial intelligence tools for engagement prediction in neuromotor disorder patients during rehabilitation
title_full_unstemmed Artificial intelligence tools for engagement prediction in neuromotor disorder patients during rehabilitation
title_short Artificial intelligence tools for engagement prediction in neuromotor disorder patients during rehabilitation
title_sort artificial intelligence tools for engagement prediction in neuromotor disorder patients during rehabilitation
topic Robot-Assisted Gait Rehabilitation
Engagement
Psychophysiological signals
Classification
K-Nearest Neighbors
Random forest
url https://doi.org/10.1186/s12984-024-01519-2
work_keys_str_mv AT simonecostantini artificialintelligencetoolsforengagementpredictioninneuromotordisorderpatientsduringrehabilitation
AT annafalivene artificialintelligencetoolsforengagementpredictioninneuromotordisorderpatientsduringrehabilitation
AT mattiachiappini artificialintelligencetoolsforengagementpredictioninneuromotordisorderpatientsduringrehabilitation
AT giorgiamalerba artificialintelligencetoolsforengagementpredictioninneuromotordisorderpatientsduringrehabilitation
AT carladei artificialintelligencetoolsforengagementpredictioninneuromotordisorderpatientsduringrehabilitation
AT silviabellazzecca artificialintelligencetoolsforengagementpredictioninneuromotordisorderpatientsduringrehabilitation
AT fabioastorm artificialintelligencetoolsforengagementpredictioninneuromotordisorderpatientsduringrehabilitation
AT giuseppeandreoni artificialintelligencetoolsforengagementpredictioninneuromotordisorderpatientsduringrehabilitation
AT emiliaambrosini artificialintelligencetoolsforengagementpredictioninneuromotordisorderpatientsduringrehabilitation
AT emiliabiffi artificialintelligencetoolsforengagementpredictioninneuromotordisorderpatientsduringrehabilitation