Preeti
Researchers at UHN’s KITE Research Institute have developed deep learning models to predict epileptic seizures, offering new hope for epilepsy research. Epilepsy affects over 50 million people globally and can lead to serious physical injury and death. Predicting seizures can reduce injuries and improve quality of life.
Dr. Shehroz Khan led a team that used deep learning models to analyze EEG data, which measures brain activity and helps understand seizure onset. By distinguishing pre-seizure EEG patterns, these models can help patients anticipate seizures and take preventive measures.
The researchers used supervised and unsupervised deep learning approaches to identify subtle changes in brain activity before seizures. Supervised learning uses labeled data, while unsupervised learning allows the model to learn from unlabeled data. Testing on two large seizure datasets showed promising results for both approaches, but variability in pre-seizure brain activity between individuals was noted.
While more work is needed, this research represents a significant advancement in epilepsy management. By harnessing the power of deep learning, personalized therapeutic interventions can be developed to save lives. The study was supported by generous donors to UHN Foundation, emphasizing the importance of collaboration in advancing healthcare.
Source link
Source link: https://uhnfoundation.ca/stories/deep-learning-for-epilepsy/
GIPHY App Key not set. Please check settings