Small, Wearable Sensor Captures Progression, Limb Dysfunction in Multiple Sclerosis

woman scratching her arm
Application of technology developed in the gaming and computer control industry offers the opportunity to create new metrics with greater sensitivity and reproducibility.
The following article is part of live conference coverage from the 2018 ACTRIMS Forum in San Diego, California. Neurology Advisor’s staff will be reporting breaking news associated with research conducted by leading experts in neurology. Check back for the latest news from ACTRIMS 2018.

A small, multisensor wearable device showed intertest reliability and was effective at distinguishing cases from controls and distinguishing strong correlations with disability outcomes in multiple sclerosis (MS). These findings were presented at ACTRIMS Forum 2018 in San Diego, California.

“Gold standard MS outcome metrics are often insensitive to disability changes over the short term and are prone to inter-rater variability. Application of technology developed in the gaming and computer control industry offers the opportunity to create new metrics with greater sensitivity and reproducibility,” wrote Jennifer Graves, PhD, MD, from the University of California, San Francisco, and colleagues.

To determine whether a small, wearable sensor could capture limb dysfunction and subtle MS progression, researchers conducted a study in which consecutive clinic patients who met published MS criteria and healthy controls were offered enrollment. In total, 88 patients with MS and 23 healthy controls were enrolled. Participants completed finger and foot taps wearing the device, which had 3 sensors: accelerometer, gyroscope, and surface electromyogram. Raw signals from the devices were processed to reduce noise and artifacts. Researchers extracted time and waveform-based textural features from the sensor data.

Analyses included intraclass correlation coefficients to assess intertest reproducibility, logistic regression adjusting for age and sex to compare case vs control differences, and Spearman correlation coefficients and multivariable regression methods to compare disability metrics with extracted features.

Patient-specific models of change over time were created for participants who returned for follow-up, and handedness was addressed in all analyses.

The researchers found that intraclass correlation coefficients for temporal signal features ranged from 0.81 to 0.84 for all limbs. This distinguished cases from controls, even for those with Expanded Disability Status Scale <2.0 (odds ratio 2.0/sec; 95% CI, 1.2-3.3, P =.01). The most informative combination of waveform textural features from the 3 sensors strongly correlated with Expanded Disability Status Scale score (r2=0.77-0.82).

Related Articles

In longitudinal analyses (n=56), a composite measurement of irregularity in the waveforms of all 3 sensor signals over time distinguished participants with progressive disease from those with relapsing disease over the 1-year observation period (sensitivity 87.5%, specificity 76.9%). In one participant, researchers detected a relapse event by the composite metric before clinical detection.

The researchers concluded that this multisensor device showed intertest reliability, distinguished cases from controls, and showed strong correlations with disability outcomes. “Ideal for future generalizability in the clinic, the assessments were brief (a few minutes) and wearing of the small device was nondisruptive to a standard MS physical examination,” the researchers wrote.

For more coverage of ACTRIMS Forum 2018, click here


Akhbardeh A, Arjona J, Gourraud P-A. Wearable multi-sensor device captures limb dysfunction and MS progression. Presented at: ACTRIMS Forum 2018; February 1-3, 2018; San Diego, CA. Abstract P018.