Towards data-driven stroke rehabilitation via wearable sensors and deep learning

Aakash Kaku,Avinash Parnandi,Anita Venkatesan,Natasha Pandit,Heidi Schambra,Carlos Fernandez-Granda

DOI: https://doi.org/10.48550/arXiv.2004.08297

2020-06-19

Abstract:Recovery after stroke is often incomplete, but rehabilitation training may potentiate recovery by engaging endogenous neuroplasticity. In preclinical models of stroke, high doses of rehabilitation training are required to restore functional movement to the affected limbs of animals. In humans, however, the necessary dose of training to potentiate recovery is not known. This ignorance stems from the lack of objective, pragmatic approaches for measuring training doses in rehabilitation activities. Here, to develop a measurement approach, we took the critical first step of automatically identifying functional primitives, the basic building block of activities. Forty-eight individuals with chronic stroke performed a variety of rehabilitation activities while wearing inertial measurement units (IMUs) to capture upper body motion. Primitives were identified by human labelers, who labeled and segmented the associated IMU data. We performed automatic classification of these primitives using machine learning. We designed a convolutional neural network model that outperformed existing methods. The model includes an initial module to compute separate embeddings of different physical quantities in the sensor data. In addition, it replaces batch normalization (which performs normalization based on statistics computed from the training data) with instance normalization (which uses statistics computed from the test data). This increases robustness to possible distributional shifts when applying the method to new patients. With this approach, we attained an average classification accuracy of 70%. Thus, using a combination of IMU-based motion capture and deep learning, we were able to identify primitives automatically. This approach builds towards objectively-measured rehabilitation training, enabling the identification and counting of functional primitives that accrues to a training dose.

Signal Processing,Machine Learning,Quantitative Methods

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to objectively measure the training dose in stroke rehabilitation training. Specifically, the author points out that currently, in stroke rehabilitation training, there is no clear standard for the effective training dose (i.e., the number of functional repetitions), mainly because of the lack of accurate and practical tools to measure the training dose. Most rehabilitation studies use treatment time to approximate the dose, but this method is not accurate because the number of training actions varies greatly among different therapists and sessions. In addition, although manual counting can identify and count the movements of interest, it is very difficult to distinguish these movements in real - time due to the fluent and rapid functional movements, and video analysis is also very time - consuming. Therefore, the goal of the paper is to develop a method that combines wearable sensor data capture and deep learning to automatically identify and count the basic functional units in rehabilitation activities - functional primitives, so as to achieve an objective measurement of the rehabilitation training dose.