We approach to develop the home-based clinical evaluation based on the quantitative functional performance and tremor metrics during the activities of daily living (ADLs) in the long term.
We have designed, developed, and assessed several computer-based standard tasks (i.e., Fitts’ law tasks, spiral and maze navigation tasks) along with their tremor metrics (dominant frequency and tremor power from the accelerometer data; (a) and (b) ). Although these studies were not designed to evaluate the functional performance of ADLs, we analyzed the performance of the computer-based tasks that projected on 2D computer screen. The performance metrics of computer-based tasks are highly correlated with the tremor metrics, particularity tremor power, which can represent clinical assessments.
Results of the Pre-clinical trial
We have developed a wireless wearable wrist device that collects a set of tremor and upper limb activities using a motion sensor unit and adapted a Fitts’ law based multi-directional tapping/clicking (MDT) task, and spiral navigation (SPN) and rectangular track navigation (RTN) tasks. We collected arm movement data from eleven participants with ET, ranging in age from 19 to 82 (six females and five males with a median age of 64) and three non-tremor participants ranging in age from 20 to 35 years old (two females and one male with a median age of 31). We first analyzed participants’ baseline tremor movements on the bean-transfer task using the three-axis accelerometer data and TETRAS scores by a non-medically trained professional researcher. To analyze the tremor metrics and their correlations with clinical scores, we defined the peak frequency of the power spectral density (PSD) as the tremor frequency and the integral of the PSD between 4 and 12 Hz as the tremor power. Using a linear regression analysis, we found a strong correlation between tremor power and TETRAS scores (R2 = 0.766).
We analyzed performance metrics such as completion time, outside area, path efficiency, throughput, and length of the actual traces for the three computer-based tasks and correlated them to the clinical evaluation (TETRAS) and tremor metrics (frequency and power of tremor). The performance metrics of the computer-based assessment tasks showed highly correlated linear regression models with tremor power (severity). With this information, we evaluated the feasibility of the new quantitative assessment method capable of accurately predicting metrics based on the characteristics of tremor movements.
Virtual-Reality based Evaluation
To expand in-home tremor assessment, we develop the VR based task setup that can extract the pathological characteristics while they perform some ADL tasks in VR environment.
Here are some tasks that can provide three categorized tremor movements: resting tremor, kinetic tremor, posture tremor.
