computing phenotype in movement disorders
learning invariant features, characterising variability
We group patients with movement disorders into core categories called phenotypes depending on the nature of dysfunctional movement. Many hundreds of different neural insults or diseases collapse down onto these limited number of phenotypes and we understand poorly why the sensorimotor system appears to fail in this set number of ways. In a major five-year project funded by the Wellcome Trust we will study patients with hyperkinetic disroders such as chorea, dystonia, myoclonus, tics and tremor. Using technology such as motion capture and wireless neurophysiology we will sample human movement disorders freely in naturalistic settings. Such high dimensional data is best probed with contempory data-led methods that lend alternative perspectives. By effectively capturing dynamic statistics we can define both invariant features and variability.