Assessment of tardive dyskinesia by means of digital image processing

A wireless device for the assessment of tardive dyskinesia by means of digital image processing is presented. Four skin-cream dots placed around the subjects' mouth are recorded by a video camera. The image is passed to a framegrabber with a signal processor, where it is converted from analogue to digital. A fast spot-detecting algorithm implemented on the signal processor tracks the dots and passes the information to a personal computer where a Fourier transformation is performed to calculate the frequency spectrum of the movements. The device provided detailed information on the magnitude and on the frequencies of the movements. Data from a longitudinal investigation suggest a higher sensitivity and reliability than conventional rating scales to detect and evaluate abnormal perioral movements. The device might be useful for the early detection, for the longitudinal assessment (p.e. clinical trials) and in some cases for the differential diagnosis of tardive dyskinesia, thus providing a tool for both research and clinical purposes.