Prediction of vision recovery rate after stroke based brain graph network and deep neural networks

This multidisciplinary project draws from the fields of neurology, informatics and medical engineering research to develop a new method for the prediction and diagnostics of visual dysfunctions after visual system damage. The final goal is to find methods to improve vision after optic nerve damage, for example after glaucoma or optic neuropathy, and for stroke. About ¹/₃ of all stroke patients’ su er posterior artery territory damage which leads to visual impairments (hemianopia) which decreases of life quality. Less is known about the mechanism of how brain works with the neurons which managed to survive and how the brain could recover and which kinds of treatments are useful. According to the "residual vision activation theory”, visual functions can in part be activated and restored because some residual structures are usually spared after damage. EEG is an electrophysiological monitoring method to record electrical activity of the brain. Brain stimulation was a typically noninvasive common method to treat the brain injuries for lot of clinical applications, here 24 patients were assigned into three groups and accepted the brain stimulation therapy for ten days, resting state EEG data was recorded while patients kept eyes closed in a no task condition, the data was preprocessed and resourced into a 3D brain model, brain connectivity were analyzed on power and phase as well as the correlation with HRP data, the di erent areas will be marked for next step machine learning. Deep neural network (deep learning) can allow us to gain lots of insight based on its high performance with undefined features. Therefore, we combine the deep learning technology and brain graph network to make prediction how the brain recovers following brain stimulation treatment. Generally, this topic would be highlighted by the integrated technologies such brain imaging and deep learning, the result could be referred as an alternatively way to help the stroke patients in their daily life.