A new study from an international team of researchers – including one from Binghamton University – demonstrates that social media could be used to detect behaviors preceding sudden unexpected death in epilepsy (SUDEP), the leading cause of death in people with uncontrolled epileptic seizures.

The results, recently published in the journal Epilepsy and behavior, reveal that epilepsy patients’ social media activity increased prior to their sudden death. These digital behavior changes could be used as early warning signals to practice preventive interventions against SUDEP.

SUDEP occurs when a person with epilepsy dies suddenly and no cause of death is found. Although the physiological mechanisms underlying SUDEP are still a mystery, people with frequent attacks are known to be at higher risk. The best preventive strategy currently is to control seizures through medication, but reducing stress and controlling triggers are also key to reducing risk. However, measuring stress and other mood states can be difficult.

A study developed by researchers from Binghamton University, Indiana University and the Instituto Gulbenkian de Ciência (IGC) in Portugal explored the potential of using social media to identify behavioral signatures that could predict SUDEP.

“We know instantly when our best friend is not well,” said Rion Brattig Correia, study co-first author, researcher at IGC and visiting researcher in the Department of Systems at the Thomas J. Watson College of Engineering and Applied Science. Science and industrial engineering.

“They mumble, talk too much or maybe too little, eye contact is different, their tone is off – we just know that. Sometimes we know it on the phone, only after a few words. What if by detecting this sudden change in behavior, we could save a friend’s life?

Building on these insights, the study examined the Facebook timelines of six epilepsy patients who died of SUDEP, using various tools to decipher human emotion and any markers of stress hidden in their posts.

“The first thing we tried was just to answer the question of whether the amount of written text was increased in the platform just before they died. And that’s what we found,” Correia said. “For five subjects, the number of written words was significantly higher in their last days compared to the rest of their timeline.”

Additionally, the type of words used by the subjects changed, and there were drastic shifts in sentiment in their messages in the weeks leading up to their deaths.

“We found significant alterations in the patient’s digital behavior that could be picked up as a signal by our algorithms,” said Ian B. Wood of Indiana University and co-first author of the study.

These shifts in patient engagement on social media, as well as the sentiment behind their posts, can serve as early warning signals for SUDEP and guide preventive interventions.

“We thought that machine learning could be very useful in uncovering patient behaviors and outcomes from the wide range of unconventional data, such as social media,” said Luís M. Rocha, Professor George J. Klir of Systems Science at Binghamton and Research Director at IGC.

Rocha led the interuniversity group for the study, which was sponsored by the National Institutes of Health. Interdisciplinary work has involved computational/complex systems researchers, clinical/behavioral epilepsy scientists, and support from the Epilepsy Foundation of America.

“In general, SUDEP studies do not consider numerical behavioral data as we did here, but focus only on physiological and clinical data. As far as we know, this is the first time this type of data has been used in the SUDEP study,” said lead researcher Wendy Miller, an epilepsy specialist at IU’s School of Nursing. who also contributed to the study.

She acknowledged that the inclusion of this digital data could provide additional insight into patient behavior leading to SUDEP that is often missed during clinical consultations: “Any advancement in this area is likely to have a significant impact on life families affected. »

To validate the predictive power of these behavioral signals extracted from social networks, the researchers intend to set up clinical studies involving more people to collect more data. If patients’ digital behavior proves useful in predicting SUDEP, the analysis could be extended to platforms besides Facebook and possibly prevent unnecessary deaths.

“The method we employed could be applied to any digital behavioral data, such as text or chat exchanges, phone calls, and other means,” Wood said.

Sponsored by the NIH’s National Library of Medicine, the team is working on a personalized web-based service for epilepsy, myAura, which will include various clinical and non-clinical data, such as self-reported patient entries regarding seizures, adherence to medications and meetings with doctors. The web service will also include the ability for users to donate their social media timelines, making this data more easily accessible for larger studies.