Imagine having a stroke or Parkinson’s disease and how difficult it could be to get to a doctor’s appointment. Or being in a remote part of the world where medical facilities are few and far between. Having a video clip of your movements being reviewed and diagnosed by a computer might be the next best thing to being there.
According to a study by Dr. Hardeep Ryait and associates at CCBN-University of Lethbridge in Alberta, Canada, this is a possibility. Video images of neurological patients could be assessed automatically, allowing quantification of behavior as part of a check-up or to assess the effects of drug treatments.
In their study performed on rats in which a stroke had been induced, human experts assessed the movements of the rats and scored their degree of impairment when reaching for food. This data was input into an artificial neural network (ANN) and, using a deep learning program, the ANN was able to learn to assess the movements.
According to the study, the ANN was then able to review a new set of rats’ movements following strokes and provided an analysis of their impairments with similar human-like accuracy.
Artificial Neural Networks are currently used to drive cars, interpret video surveillance and regulate traffic. They do so by using deep learning programs.
Deep learning is a type of machine learning that trains a computer to perform human-like tasks, such as recognizing speech, identifying images or making predictions. Instead of organizing data to run through predefined equations, deep learning sets up basic parameters about the data and trains the computer to learn on its own by recognizing patterns using many layers of processing.
Interestingly, the results revealed that this network can use a wider range of information than that included by experts in a behavioral scoring system. This method of analysis could aid standardization of the diagnosis and monitoring of neurological disorders, and in the future could be used by patients at home for monitoring daily symptoms.