Automatic assessment of the myoclonus severity from videos recorded according to standardized Unified Myoclonus Rating Scale protocol and using human pose and body movement analysis

Abstract

Myoclonus in progressive myoclonus epilepsy type 1 (EPM1) patients shows marked variability, which presents a substantial challenge in devising treatment and conducting clinical trials. Consequently, fast and objective myoclonus quantification methods are needed. Methods: Ten video-recorded unified myoclonus rating scale (UMRS) myoclonus with action tests were performed on EPM1 patients who were selected for the development and testing of the automatic myoclonus quantification method. Human pose and body movement analyses of the videos were used to identify body keypoints and further analyze movement smoothness and speed. The automatic myoclonus rating scale (ARMS) was developed. It included the jerk count during movement score and the log dimensionless jerk (LDLJ) score to evaluate changes in the smoothness of movement. Results: The scores obtained with the automatic analyses showed moderate to strong significant correlation with the UMRS myoclonus with action scores. The jerk count of the primary keypoints and the LDLJ scores were effective in the evaluation of the myoclonic jerks during hand movements. They also correlated moderately to strongly with the total UMRS test panel scores (r 2 = 0,77, P = 0,009 for the jerk count score and r 2 = 0,88, P = 0,001 for the LDLJ score). The automatic analyses was weaker in quantification of the neck, trunk, and leg myoclonus. Conclusion: Automatic quantification of myoclonic jerks using human pose and body movement analysis of patients' videos is feasible and was found to be quite consistent with the accepted clinical gold standard quantification method. Based on the results of this study, the automatic analytical method should be further developed and validated to improve myoclonus severity follow-up for EPM1 patients.

Computer vision for automated seizure detection and classification: A systematic review

“…Overall, the advantages of CNNs outweigh their disadvantages, as they have been used successfully to detect poses in videos of seizures. 26,27,30,51,62,63,66,67,70 PNNs PNNs are a subtype of FFNNs that estimate the probability of new data in the same category as previously learned data. 85 The probability is estimated by comparing the various parameters of the new data to the parameters of each previous case of training data to compute a Euclidean distance-a numeric value of similarity.…”

“…Several models have been developed to detect myoclonic seizures. 22,23,27,31 One model of nocturnal myoclonic seizures in children extracted and fed STIPs into an SVM, obtaining a sensitivity of 77% and a PPV of 87%. 23 An automated myoclonus detector developed to quantify the severity of myoclonic jerks in progressive myoclonus epilepsy type 1 27 tracked key point movement speed to create an automated myoclonus rating scale that was comparable to the gold standard Unified Myoclonus Rating Scale.…”

“…Myoclonus can also be benign, making false positive detections more common. Several models have been developed to detect myoclonic seizures 22,23,27,31 . One model of nocturnal myoclonic seizures in children extracted and fed STIPs into an SVM, obtaining a sensitivity of 77% and a PPV of 87% 23 .…”

“…22 OF captures motion patterns and movement periodicity of the body during seizures, providing an information-dense account of seizure semiology that forms an integral part of many seizure detection models. 14,23,24,25,26,27,28,29,30,31,32,33,34 Like FD, OF algorithms typically result in privacy preserving output, in this case by removing identifying facial information. Researchers often combine OF with other motion-based or appearance-based features to improve the overall performance of seizure detection algorithms.…”

“…CNNs excel on pattern recognition in images and have been widely used in seizure analysis. 13,26,27,30,44,51,59,60,62,64,66,68,69,70,71,72,73,81 In contrast to simple FFNNs, a CNN uses multiple layers of convolution operations, which pool information from local regions of image space (Figure 4). With sufficient layers and spatial extent of pooling, information is aggregated from the entire image.…”