In January 2015, a joint American-Australian research team won an American Epilepsy Society’s completion to detect seizures. The researchers developed an algorithm, which accurately predicts seizures 82% of the time. Previously, some health professionals believed that seizures could not be detected. “Until recently,” says Dr Francis Collins of the National Institute of Health, USA, “the best algorithm was hardly better than flipping a coin”.
Epilepsy, which usually presents at the end of the first or second decade, is a chronic condition consisting of more than 40 clinical syndromes affecting about 50 million people worldwide. Its cause is unknown, but may stem from birth trauma, perinatal infection, anoxia, infectious diseases, ingestion of toxins, brain tumors, inherited disorders or degenerative disease, head injury, metabolic disorders, cerebrovascular accident, and alcohol withdrawal. Treatment is through medication or surgery, and the prognosis is variable.
The most common form of the condition is temporal lobe epilepsy (TLE), which is characterized by recurrent, unprovoked seizures. About 13% of patients receiving medication for TLE have inadequate seizure control. The prognosis for such patients includes a higher risk of memory loss, mood challenges, quality of life impairment, and, in some cases, death.