Self-Organizing Maps to Assess Rehabilitation Progress of Post-Stroke Patients

Abstract

The scarcity of adequate rehabilitation and treatment centers for post-stroke patients, a relatively common disease among the Egyptian populace, and the lack of awareness and trained physiotherapists, causes many patients to forgo treatment until they are transported to the hospital. Even then, the high cost of treatment will impede most rehabilitation attempts to those who survive. Thankfully, rehabilitation robotics can be used to replace the need for trained physiotherapists. This paper uses the Myo armband as a rehabilitation assessment device, tracking the progress of Post-Stroke patients and comparing them with healthy subjects. By taking a total of 60 samples from 3 healthy subjects and using self-organizing maps, a clustering system that can differentiate between regular and irregular motions using kinematic data with less than 10% error was produced.