Home AI News Remote Motor Function Evaluation: MIT’s Solution to Assessing Patients from Afar

Remote Motor Function Evaluation: MIT’s Solution to Assessing Patients from Afar

Remote Motor Function Evaluation: MIT’s Solution to Assessing Patients from Afar

Mitigating Stress with Remote Evaluation of Patients’ Motor Function

It can be challenging and stressful for parents of children with motor disorders like cerebral palsy to take them for regular evaluations at the doctor’s office. These evaluations often require an hour of in-person examination and can be expensive and emotionally taxing. However, MIT engineers have come up with a solution to alleviate some of this stress. They have developed a new method that remotely evaluates patients’ motor function using computer vision and machine-learning techniques.

By analyzing videos of patients in real-time, the method detects certain patterns of poses and computes a clinical score of motor function. The researchers tested the method on over 1,000 children with cerebral palsy and found that it accurately matched the clinical scores assigned by clinicians during in-person visits. The video analysis can be done on various mobile devices, allowing patients to be evaluated at home. Patients can simply record a video of themselves moving around and upload it for analysis. The video and clinical score can then be sent to a doctor for review.

The MIT team is now working on tailoring the approach to evaluate children with metachromatic leukodystrophy and patients who have experienced a stroke. They believe that this technology can be used to remotely evaluate any condition that affects motor behavior, reducing the need for hospital visits for evaluations.

The method incorporates real-time skeleton pose data obtained from videos of children with cerebral palsy. They utilized pose estimation algorithms to generate skeleton pose data and then trained a machine-learning algorithm to classify the data into different clinical motor function levels. The pretrained algorithm learned to extract features from a sequence of human poses and accurately classify children’s mobility levels. The researchers tested the method on various mobile devices and found that it could generate clinical scores from videos in close to real-time.

The MIT team plans to develop an app that parents and patients can use to analyze videos of patients taken at home. The results can then be sent to a doctor for further evaluation. They also aim to adapt the method for other neurological disorders like stroke or Parkinson’s disease.

This new method could significantly reduce the stress and cost associated with in-person evaluations, improve patient care, and increase compliance with interventions. It also has the potential to predict how patients would respond to interventions sooner by evaluating them more frequently.

This research was supported by Takeda Development Center Americas, Inc.

Source link


Please enter your comment!
Please enter your name here