Objective: Our objective was to develop and test the concurrent validity of CMOR (Computational Motor Objective Rater; a computer vision-based software system) to quantify HT severity from conventional video recordings during a clinical examination of CD severity.

Background: Head tremor (HT) is a common characteristic of cervical dystonia (CD). HT severity is usually quantified by subjective observation using clinical rating scales. Technology developments offer alternatives for measuring HT severity that are objective and amenable to automation.

Method: We analyzed video recordings from 104 patients with HT enrolled by the Dystonia Coalition. Patients were recorded while they were instructed to do successive eye open and closing (“eyes” task) and let their head drift to its most comfortable dystonic position (“drift” task). Patients were screened for video recordings quality (camera instability and sufficient subject illumination), examination protocol omissions (tasks missing: ‘eyes’ or ‘drift’), and computer-vision related sources of artifactual head pose estimates. We used Spearman’s correlation to measure the agreement between CMOR’s HT severity measure based on spectral power of head movements and a movement disorders neurologist’s clinical tremor severity ratings.

Results: Nineteen patients were omitted because of recording quality (no tripod, insufficient illumination) and one was omitted because of missing tasks. An additional 15 patients were omitted because of insufficiently long periods of contiguous artifact-free data. For both the “eyes” and “drift” tasks, CMOR’s HT severity measure was positively and significantly related to the clinical HT severity ratings (Spearman’s rho = 0.67, p < 0.001 and rho = 0.71, p < 0.001, respectively).

Conclusion: CMOR exhibits concurrent validity with an expert movement disorder neurologist’s clinical ratings of HT severity. Thus we demonstrate a new method to measure HT severity in CD that is not only objective and unobtrusive but also requires only conventional video recordings that are readily acquired in the clinical setting.

[Previously presented at International Dystonia Symposium, Dublin, June 2020.]

« Back to MDS Virtual Congress 2020

MDS Abstracts - https://www.mdsabstracts.org/abstract/video-based-computer-vision-system-exhibits-concurrent-validity-with-clinical-ratings-of-head-tremor-severity-in-cervical-dystonia/