Machine learning detects changes in early upward behavior in mouse models of Parkinson's disease

Detecting early elevated Parkinson's disease (PD) symptoms may improve treatment outcomes by enabling early intervention. It's new Eneuro A paper and colleague of Daniil Berezhnoi, a former member of Georgetown University, used machine learning techniques to detect subtle, early behavioral changes in mouse models of PD. The researchers also evaluated whether levodopa, the primary approved treatment for PD, could effectively treat these symptoms.

Berezhnoi et al. We evaluated the movement of various mouse models of PD using a previously developed motion sequencing platform. The main advantage of this machine learning platform is that it can automatically detect subsectan posture changes from animal 3D videos. Researchers found that faster, higher velocity movements were the first influenced behavior in the early stages of PD. Levodopa improved movement speed during fine times but did not improve other attributes of these movements.

Speaking about the meaning of this work, Beletznoy says.Perhaps applying the same machine learning approach we used might help identify early biomarkers of Parkinson's disease in people. ”