Objective, computerized video-based rating of blepharospasm severity

Abstract

Objective: To compare clinical rating scales of blepharospasm severity with involuntary eye closures measured automatically from patient videos with contemporary facial expression software.

Methods: We evaluated video recordings of a standardized clinical examination from 50 patients with blepharospasm in the Dystonia Coalition's Natural History and Biorepository study. Eye closures were measured on a frame-by-frame basis with software known as the Computer Expression Recognition Toolbox (CERT). The proportion of eye closure time was compared with 3 commonly used clinical rating scales: the Burke-Fahn-Marsden Dystonia Rating Scale, Global Dystonia Rating Scale, and Jankovic Rating Scale.

Results: CERT was reliably able to find the face, and its eye closure measure was correlated with all of the clinical severity ratings (Spearman ρ = 0.56, 0.52, and 0.56 for the Burke-Fahn-Marsden Dystonia Rating Scale, Global Dystonia Rating Scale, and Jankovic Rating Scale, respectively, all p < 0.0001).

Conclusions: The results demonstrate that CERT has convergent validity with conventional clinical rating scales and can be used with video recordings to measure blepharospasm symptom severity automatically and objectively. Unlike EMG and kinematics, CERT requires only conventional video recordings and can therefore be more easily adopted for use in the clinic.

Figure 1. Computer Expression Recognition Toolbox video processing pipeline

Video segments are analyzed on a frame-by-frame basis. Within the frame image, faces are located with the use of standard facial landmarks. After the face is found, Gabor features are extracted and used in a high-dimensional support vector machine (SVM) large-margin linear classifier, and the distance of the patient's metrics from the hyperplane determines the normalized eye closure magnitude, in arbitrary units (modified from Littlewort et al., figure 2, with permission).

Figure 2. Clinical severity ratings: video vs live ratings

Clinical severity measures based on video vs live ratings for (A) Burke-Fahn-Marsden Dystonia Rating Scale (BFM) and (B) Global Dystonia Rating Scale (GDRS), including linear regression lines and 95% confidence intervals (shaded).

Figure 3. Eye closure thresholding

Frame-by-frame continuous-valued Computer Expression Recognition Toolbox measures of eye closure during clinical examination, corresponding distributions over time, and bimodality of the distributions. Two sample patients: (A) one with and (B) one without a clear bimodal distribution in eye closure. (C) Bimodality of patients' eye closure distribution (sorted by bimodality measure). Horizontal line represents threshold above which distribution is conventionally considered bimodal (Sarle b, see appendix e-1). BFM = Burke-Fahn-Marsden Dystonia Rating Scale; GDRS = Global Dystonia Rating Scale; JRS = Jankovic Rating Scale. AU = arbitrary units.

Figure 4. Convergent validity between Computer Expression Recognition Toolbox (CERT) and clinical severity ratings

Convergent validity between percent time of eye closure based on CERT and clinical severity ratings, separately for video (A–C) and live (D and E) clinical rating modalities. Lines are linear regressions, and shaded regions show the 95% confidence intervals.