A Novel Finger-Worn Sensor for Ambulatory Monitoring of Hand Use

Continuous remote monitoring of upper-limb movements in individuals with neurological conditions allows for assessing patients' motor behaviors outside of the clinic, with potential for improving the outcomes of rehabilitation interventions. This paper investigates the use of a miniaturized finger-worn sensor, combined with a wrist-worn sensor, to quantify the amount of hand use in the ambulatory setting. First, we introduce a new benchmark measure to quantify the amount of hand use based on a motion capture system. Then, we utilize a machine-learning based algorithm to estimate the aforementioned benchmark measure using wearable sensor data. The results herein presented show that (i) the benchmark measure derived using a motion capture system is reliable and responsive, and (ii) wearable sensors have great potential for remotely monitoring rehabilitation outcomes outside of the clinic.