This paper presents a probabilistic-graphical model that can be used to infer characteristics of instantaneous brain activity by jointly analyzing spatial and temporal dependencies observed in electroencephalograms (EEG). Specifically, we describe a factor-graph-based model with customized factor-functions defined based on domain knowledge, to infer pathologic brain activity with the goal of identifying seizure-generating brain regions in epilepsy patients. We utilize an inference technique based on the graph-cut algorithm to exactly solve graph inference in polynomial time. We validate the model by using clinically collected intracranial EEG data from 29 epilepsy patients to show that the model correctly identifies seizure-generating brain regions. Our results indicate that our model outperforms two conventional approaches used for seizure-onset localization (5-7% better AUC: 0.72, 0.67, 0.65) and that the proposed inference technique provides 3-10% gain in AUC (0.72, 0.62, 0.69) compared to sampling-based alternatives.

This paper presents a probabilistic-graphical model that can be used to infer characteristics of instantaneous brain activity by jointly analyzing spatial and temporal dependencies observed in electroencephalograms (EEG). Specifically, we describe a factor-graph-based model with customized factor-functions defined based on domain knowledge, to infer pathologic brain activity with the goal of identifying seizure-generating brain regions in epilepsy patients. We utilize an inference technique based on the graph-cut algorithm to exactly solve graph inference in polynomial time. We validate the model by using clinically collected intracranial EEG data from 29 epilepsy patients to show that the model correctly identifies seizure-generating brain regions. Our results indicate that our model outperforms two conventional approaches used for seizure-onset localization (5-7% better AUC: 0.72, 0.67, 0.65) and that the proposed inference technique provides 3-10% gain in AUC (0.72, 0.62, 0.69) compared to sampling-based alternatives.

Introduction: Approximately 23 million or 30% of epilepsy patients worldwide suffer from drug-resistant epilepsy (DRE). The unpredictability of seizure occurrences, which causes safety issues as well as social concerns, restrict the lifestyles of DRE patients. Surgical solutions and EEG-based solutions are very expensive, unreliable, invasive or impractical. The goal of this research was to employ improved technologies and methods to epilepsy patient physiological data and predict seizures up to two hours before onset, enabling non-invasive, affordable seizure prediction for DRE patients. Methods: This research used a 1D Convolutional Neural Network-Based Bidirectional Long Short-Term Memory network that was trained on a diverse set of epileptic patient physiological data to predict seizures. Transfer learning was further utilized to personalize and optimize predictions for specific patients. Clinical data was retrospectively obtained for nine epilepsy patients via wearable devices over a period of about three to five days from a prospectively maintained database. The physiological data included 54 seizure occurrences and included heart rate, blood volume pulse, accelerometry, body temperature, and electrodermal activity. Results and Conclusion: A general deep-learning model trained on the physiological data with randomly sampled test data achieved an accuracy of 91.94%. However, such a generalized deep learning model had varied performances on data from unseen patients. When the general model was personalized (further trained) with patient-specific data, the personalized model achieved significantly improved performance with accuracies as high as 97%. This preliminary research shows that patient-specific personalization may be a viable approach to achieve affordable, non-invasive seizure prediction that can improve the quality of life for DRE patients.

An international study led by Monash University has done what could be the world's first demonstration of an AI model that can predict the optimal anti-seizure medication for newly diagnosed epilepsy patients. The research team has trained a deep-learning prediction model using clinical information from around 1,800 patients in five health care centres in Australia, Malaysia, China and the United Kingdom. The model is designed by the Monash Medical AI Group and is trained using Monash's MASSIVE computing facility. Findings from the study, which was published in the journal JAMA Neurology, showed that the AI model has a "modest" 65% accuracy in predicting the best anti-seizure medication. The research team is still improving the model by employing more complex information.

Locating the source of an epileptic seizure can be tricky. Even the most advanced MRI can't pinpoint lesions, scars or other abnormalities on the brain in one-quarter of epilepsy patients. Cleveland Clinic experts are turning to artificial intelligence to help bridge the gap. Neurologists and brain surgeons from the Clinic's Epilepsy Center are using AI and advanced medical imaging techniques to help locate the source of a patient's seizures. That gives surgeons a better chance of removing any brain tissue that's associated with those seizures, which could help the patient live seizure-free for years. The use of AI has already helped the Clinic improve the odds a surgery will result in a patient living without seizures, said Dr. Imad Najm, the director of the Epilepsy Center at the Cleveland Clinic Neurological Institute.

Researchers from Ben-Gurion University of the Negev in Israel have developed a wearable electroencephalogram (EEG) device they claim can predict epileptic seizures up to an hour before the onset. Epiness uses machine learning algorithms to analyze brain activity and detect potential seizures, and it can send a warning to a connected smartphone. Other devices on the market can detect seizures in real-time, but can't give advance warnings. However, researchers from the University of Louisiana at Lafayette last year unveiled an AI prediction model of their own. That was said to offer a similar level of prediction accuracy to Epiness, and it can also alert patients up to an hour in advance of a seizure taking hold.

MIT researchers have developed a wireless, private way to monitor a person's sleep postures -- whether snoozing on their back, stomach, or sides -- using reflected radio signals from a small device mounted on a bedroom wall. The device, called BodyCompass, is the first home-ready, radio-frequency-based system to provide accurate sleep data without cameras or sensors attached to the body, according to Shichao Yue, who will introduce the system in a presentation at the UbiComp 2020 conference on Sept. 15. The PhD student has used wireless sensing to study sleep stages and insomnia for several years. "We thought sleep posture could be another impactful application of our system" for medical monitoring, says Yue, who worked on the project under the supervision of Professor Dina Katabi in the MIT Computer Science and Artificial Intelligence Laboratory. Studies show that stomach sleeping increases the risk of sudden death in people with epilepsy, he notes, and sleep posture could also be used to measure the progression of Parkinson's disease as the condition robs a person of the ability to turn over in bed.

This paper presents a probabilistic-graphical model that can be used to infer characteristics of instantaneous brain activity by jointly analyzing spatial and temporal dependencies observed in electroencephalograms (EEG). Specifically, we describe a factor-graph-based model with customized factor-functions defined based on domain knowledge, to infer pathologic brain activity with the goal of identifying seizure-generating brain regions in epilepsy patients. We utilize an inference technique based on the graph-cut algorithm to exactly solve graph inference in polynomial time. We validate the model by using clinically collected intracranial EEG data from 29 epilepsy patients to show that the model correctly identifies seizure-generating brain regions. Our results indicate that our model outperforms two conventional approaches used for seizure-onset localization (5-7% better AUC: 0.72, 0.67, 0.65) and that the proposed inference technique provides 3-10% gain in AUC (0.72, 0.62, 0.69) compared to sampling-based alternatives. Papers published at the Neural Information Processing Systems Conference.

Recently, Hisham Daoud and Magdy Bayoumi of the University of Louisiana at Lafayette have introduced a completely new Artificial Intelligence (AI) system that predicts epilepsy seizures. According to the World Health Organization's reports, around 50 million people around the world are suffering from epilepsy and 70% of those patients can control the seizures through medications. The new AI technology shows 99.6% accurate results, and the best thing about it is that it predicts the attacks an hour before it happens. In this way, the patient can gear up for it and take medications that can prevent its occurrence. Having enough time to control the attack is what a patient needs.

Pioneering brain implants currently used for stroke or epilepsy patients could be used to track people's thoughts - and relay them back to governments or companies. That's the Orwellian warning from scientists at the Royal Society, who say reading chips pose a huge privacy issue. The society has urged the government to launch an inquiry and protect human rights against the commercial development of neural interface software. It comes after both Elon Musk and Mark Zuckerberg expressed plans to incorporate the technology into their businesses. This would allow users to interact with augmented reality environments using just their brain - no keyboards, touchscreens or hand gestures required.