Action tremor kinematics exhibit features of stabilization dysfunction in essential tremor patients.

BACKGROUND

Action tremor that begins with voluntary movement is a common manifestation of essential tremor (ET). We investigated a novel approach to detect reach direction and tremor direction across naturalistic reaches, elucidating this relationship.

METHODS

ET patients undergoing awake DBS implantation surgery performed a center-out reaching task, which elicited multi-joint, naturalistic reaching movements made to evenly spaced radial targets. This project utilized a deep-learning based markerless motion tracking system to monitor movements in 3-dimensions. For analysis, to test novel hypotheses, we compared the direction of the tremor vectors (during hold) to the direction of the reach vector, as well as to the other tremor direction vectors for each subject.

RESULTS

We analyzed 13 limb recordings from 9 subjects. Tremor directions were not significantly dependent on preceding reach directions in any recording (p > 0.05). The distribution of observed tremor directions was significantly different from the expected distribution of random vectors for all 11 analyzed recordings (likelihood of observed tremor directions vs. randomly generated vectors, p < 0.01).

COMPARISON WITH EXISTING

METHODS

Our markerless capture method for tremor measurement does not rely on accelerometry or electromyography and does not require fiducials for tracking and combines two open-source python toolkits to leverage computer vision for three-dimensional pose estimation. Further we test a novel solution for measuring the relationship between tremor direction and movement.

CONCLUSIONS

Here, we describe a novel method of tremor detection. Our findings suggest that an action tremor is not generated as an oscillation of a function involved in reach, such as braking, but is consistent with a function involved in stabilization to maintain a posture.