Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Update on Rehabilitation After Stroke: Global Changes and the Continued Importance of Therapy Intensity, Dose, and Timing.

Stroke·2026
Same author

A Sudden Decline in Global Stroke Rehabilitation Randomized Controlled Trials and Female Enrollment Associated With the COVID-19 Pandemic.

American journal of physical medicine & rehabilitation·2026
Same author

Evaluating the concurrent validity and test-retest reliability of markerless motion capture for static postural control assessment.

Journal of biomechanics·2026
Same author

RObotic WAlking for children who CAnnot WAlk (RoWaCaWa): feasibility and family impacts and perspectives of a family-led intervention.

Disability and rehabilitation·2026
Same author

Predictors of favorable modified Rankin Scale outcome during inpatient rehabilitation for subacute stroke: a secondary analysis of the Walk 'n Watch trial.

Disability and rehabilitation·2026
Same author

Shift happens: a fairness-oriented framework for medical classification under hidden bias.

International journal of computer assisted radiology and surgery·2026

Related Experiment Video

Updated: May 8, 2026

Enhancing Upper Limb Function and Motor Skills Post-Stroke Through an Upper Limb Rehabilitation Robot
04:49

Enhancing Upper Limb Function and Motor Skills Post-Stroke Through an Upper Limb Rehabilitation Robot

Published on: September 6, 2024

Reducing robotic upper-limb assessment time while maintaining precision: a time series foundation model approach.

Faranak Akbarifar1, Nooshin Maghsoodi2, Sean P Dukelow3

  • 1School of Computing, Queen's University, 25 Union St., Kingston, ON, K7L 2N8, Canada. f.akbarifar@queensu.ca.

Journal of Neuroengineering and Rehabilitation
|May 7, 2026
PubMed
Summary
This summary is machine-generated.

Foundation models can forecast missing trials in robotic assessments, significantly reducing testing time for stroke survivors. This method maintains accuracy, making evaluations more efficient and less burdensome.

Keywords:
Foundation modelsKinarm assessmentTime series forecastingVisually guided reaching

More Related Videos

A Structured Rehabilitation Protocol for Improved Multifunctional Prosthetic Control: A Case Study
06:58

A Structured Rehabilitation Protocol for Improved Multifunctional Prosthetic Control: A Case Study

Published on: November 6, 2015

Application of a Dual Upper Limb Task-Oriented Robotic System for the Functional Recovery of the Upper Limb in Stroke Patients
05:28

Application of a Dual Upper Limb Task-Oriented Robotic System for the Functional Recovery of the Upper Limb in Stroke Patients

Published on: October 11, 2024

Related Experiment Videos

Last Updated: May 8, 2026

Enhancing Upper Limb Function and Motor Skills Post-Stroke Through an Upper Limb Rehabilitation Robot
04:49

Enhancing Upper Limb Function and Motor Skills Post-Stroke Through an Upper Limb Rehabilitation Robot

Published on: September 6, 2024

A Structured Rehabilitation Protocol for Improved Multifunctional Prosthetic Control: A Case Study
06:58

A Structured Rehabilitation Protocol for Improved Multifunctional Prosthetic Control: A Case Study

Published on: November 6, 2015

Application of a Dual Upper Limb Task-Oriented Robotic System for the Functional Recovery of the Upper Limb in Stroke Patients
05:28

Application of a Dual Upper Limb Task-Oriented Robotic System for the Functional Recovery of the Upper Limb in Stroke Patients

Published on: October 11, 2024

Area of Science:

  • Robotics and Biomechanics
  • Artificial Intelligence in Healthcare
  • Neurorehabilitation

Background:

  • Visually Guided Reaching (VGR) assessments using the Kinarm robot provide valuable kinematic biomarkers for motor impairment.
  • Current VGR protocols require a high number of reaches (40-64), leading to extended assessment times and participant fatigue.
  • There is a need for more efficient methods to gather reliable kinematic data in VGR assessments.

Purpose of the Study:

  • To evaluate the efficacy of time series foundation models in forecasting unrecorded VGR trials.
  • To determine if forecasted trials can replace a significant portion of recorded trials while maintaining agreement with full-session Kinarm parameter estimates.
  • To assess the potential of foundation models to reduce the time and fatigue associated with VGR assessments.

Main Methods:

  • Analysis of VGR speed signals from 461 stroke and 599 control participants across 4- and 8-target protocols.
  • Utilizing ARIMA, MOMENT, and Chronos models, fine-tuned on a subset of participants, to forecast synthetic trials from an initial set of 8 or 16 recorded reaches.
  • Recomputing four key kinematic features (reaction time, movement time, posture speed, max speed) using combined recorded and forecasted trials, and comparing results to full-length assessments via ICC(2,1).

Main Results:

  • The Chronos model, when combining 8 recorded trials with forecasted trials, significantly increased agreement (ICC values) for all parameters, achieving results comparable to 24-28 recorded trials.
  • MOMENT models showed intermediate improvements, while ARIMA models provided minimal gains.
  • Synthetic trials successfully replaced a substantial number of reaches without compromising the reliability of the assessed kinematic features across different participant groups and protocols.

Conclusions:

  • Foundation model-based forecasting offers a substantial reduction in Kinarm VGR assessment duration.
  • For severely impaired stroke survivors, assessment time can be reduced from 4-5 minutes to approximately 1 minute while preserving the reliability of Kinarm parameter estimates.
  • This forecast-augmented approach presents a promising paradigm for efficient robotic evaluations of motor impairments post-stroke.