The advent of high-resolution imaging has provided healthcare providers and scientists with a greater understanding of the brain circuit malfunctions seen in epileptics, but less is known about how epilepsy affects behavior. A new study has used state-of-the-art AI on mice to catch epilepsy-related behavior that can be missed by the human eye.
Epilepsy is the most common chronic brain disease, affecting millions of people worldwide. It can affect people of any age, and, for some, treatment not only produces nasty side effects but does not prevent seizures from occurring.
A traditional approach to epilepsy diagnosis and treatment assessment involves the use of continuous video-electroencephalogram (EEG) monitoring over days or weeks. But it can be a fairly blunt tool, given the complexity and diversity of the condition and the fact that some seizures do not appear on EEG. In addition, it is both labor-intensive and subjective. A healthcare professional must view and analyze hours of video-EEG recordings, and rely on their ability to notice often slight behavioral changes.
Now, researchers have used AI technology called MoSeq (or Motion Sequencing) to analyze the behavior of epileptic mice, identifying the behavioral “fingerprints’ that can go unnoticed by the human eye.
MoSeq is a machine-learning technology that trains an unsupervised machine to identify repeated patterns of behavior. After identifying the behaviors, MoSeq offers a set of visualization tools and statistical tests to help scientists understand those behaviors and compare them to a range of experimental conditions.
Using MoSeq to analyze 3D videos of freely moving mice, the researchers were able to locate, track and quantify the behavior of the mice. They found that the technology could better distinguish between epileptic and non-epileptic mice, outperforming trained human observers. Moreover, it required only one hour of video recording and did not need a seizure to occur before offering its analysis, unlike traditional methods.
Researchers were able to use the AI to differentiate between patterns of behavior in the mice after they were given one of three anti-epileptic medications.
The successful use of machine-learning technology demonstrates its potential for use in humans to provide a faster, less labor-intensive, less costly and more objective way to diagnose epilepsy and test the efficacy of anti-epileptic medications.
Source: newatlas.com, Pau; McClure