#Dystonia study: visual deep learning reveals head movement dynamics. #Neurology

Head movement dynamics in dystonia: a multi-centre retrospective study using visual perceptive deep learning

The study sourced clinical video data from two prospective cohort studies on the therapeutic effect of pallidal deep brain stimulation (DBS) on cervical and generalised/segmental dystonia, as well as a retrospective investigation using advanced neuroimaging techniques. The data was split into two datasets based on dystonia subtype, and included videos from pre-operative and post-operative phases. A total of 232 videos from 116 individual patients were analyzed, along with a cohort of 22 healthy controls. Computer vision models were developed to assess dystonia phenotype and severity, combining facial landmark tracking and movement state extraction. Features were engineered to capture kinematic observations in dystonia, such as movement overflow, tremor, and complexity. The models demonstrated high accuracy in training and testing. The study evaluated the performance of the visual perceptive framework in assessing dystonic head deviation and tremor severity. Statistical analyses were conducted to identify differences in kinematic features between stimulation conditions in cervical dystonia and between cervical and generalised dystonia. Correlation analyses were also performed to investigate relationships between head angle excursions and annotated scores. The study provides a comprehensive framework for automated kinematic evaluation of dystonia from video data, with potential applications for clinical assessment and monitoring of dystonia patients.

Source link

Source link:

What do you think?

Leave a Reply

GIPHY App Key not set. Please check settings

WWDC: Here's Why Elon Musk Is Furious Over Apple AI!

#ElonMusk Furious Over Apple AI at WWDC Event. #Competition

The Best Productivity Apps for 2024 - PCMag AU South Africa’s Top 10 AI Tools for 2024 #AItools