Congratulations to Dr. Hisham Daoud for receiving the Biomedical Circuits and Systems best paper award from the IEEE Circuits and Sytems Society for his paper "Efficient Epileptic Seizure Prediction Based on Deep Learning" co-authored with Dr. Magdy A. Bayoumi.
In this research, an AI algorithm that predicts seizures up to one hour in advance with 99.6 percent accuracy has been developed. Anticipating seizures could greatly improve the quality of life for patients. “Due to unexpected seizure times, epilepsy has a strong psychological and social effect on patients,” explains Dr. Daoud. In the new approach, Dr. Daoud and Dr. Bayoumi developed a deep-learning algorithm that analyzes the spatial-temporal activity of a patient’s brain from the perspective of different electrodes. And to speed up the prediction, they applied an extra algorithm to identify the most prognostic channels of electrical activity. The researchers developed and tested their approach using long-term EEG data from 22 patients at Boston Children’s Hospital. Their model showed a low tendency for false positives, at 0.004 false alarms per hour. However, it did require some setup, as the model had to be trained on each patient. With the software complete, Daoud says the next step is to develop a customized computer chip to process the algorithms so that this training could be completed in the comfort of a patient’s home.