The Ted talk by MIT Prof Rosalind Picard centers on the interesting application of real-time measurement and analysis of physiological signals via a wearable device for the early detection of patterns that may be associated with generalized tonic-clonic seizures, and immediately alert caregivers. The ground-breaking device which is now approved by the FDA uses advances in machine learning to detect patterns in body conductance that may be associated with convulsive seizures(Picard, 2018).

According to Prof Picard, every year, more than 50,000 otherwise healthy people in the United States with epilepsy suddenly dies — a condition she describes as Sudden Unexpected Death in Epilepsy (SUDEP). In 100% of SUDEPs, the patient stops breathing usually minutes after a seizure ends, resulting mostly in healthy people dying without any sign of probable epileptic causes in their autopsy(Picard, 2018). These deaths could be prevented mainly if help was to arrive minutes after a seizure occurs. Hence, the researcher using cutting edge smatchwatch and advanced machine learning techniques can detect epileptic seizures before they happen and alert nearby family members or caregiver in time for help.

The innovative watch relies on the physiological signal associated with skin conductance for data collection, which is known to vary widely before and after seizures in epileptic patients, and using recent advances in machine learning algorithms predict with some level of precision the possible occurrence of seizure while alerting responsible caregivers’ minutes before an actual seizure. While one may be quick to recognize the innovative application of emerging machine learning algorithms in saving lives in this particular case, mention must be made of the equally innovative sensors developed during this research efforts with the combined ability to monitor movement patterns and physiological signals as quite promising and encouraging as well(Emphatica, 2019).

To achieve this phenomenon feat, the smartwatch device is packed with an Electro-Dermal Activity (EDA) sensor, a gyroscope, an accelerometer, and a peripheral temperature sensor. Electro-dermal activity is associated with the mammalian sweat glands on a microscopic level and is usually a response to emotional and other physical stimuli. In the case of convulsive seizures, large spikes in electro-dermal activity have shown similar correlations(Emphatica, 2019).

References

Emphatica (Producer). (2019, 01 29). Understanding the importance of electrodermal activity (EDA). Emphatica Support. Retrieved from https://support.empatica.com/hc/en-us/articles/115000198283-What-is-and-what-affects-Electrodermal-activity-EDA-

Picard, R. (Producer). (2018, 11 01). An AI smartwatch that detects seizures. Ted Talks. Retrieved from https://www.ted.com/talks/rosalind_picard_an_ai_smartwatch_that_detects_seizures_and_saves_lives