Fast, reliable and automatic assessment of the severity of myoclonic jerks from video footage is now possible, thanks to an algorithm using deep convolutional neural network architecture and pretrained models that identify and track keypoints in the human body.
