Decoding motor performance from non-neural data

Our group developed a fast assessment of motor performance for Parkinson patients, the CopyDraw task. This paradigm enables us to record multiple repetitions of Parkinson patients performing the task with and without deep brain stimulation. We could later show, that we could decode the motor performance reliably from EEG data. Such experiments aim to develop markers which could be used to develop an adaptive closed-loop deep brain stimulation, which would provide the patients with an optimized stimulation setting, reducing side effects. During our measurement sessions, we also recorded a variety of non-neural recordings, such as heart rate, respiratory pressure and electromyogram (EMG).

This project investigates, how well motor performance can be decoded from these non-neural data sources. The target would be a one dimensional motor score, which should be predicted from non-neural data.

Skills required

  • Machine learning
  • Good programming skills in Python (familiarity with numpy, sklearn).
  • Good mathematical background and intuition.
Michael Tangermann
Michael Tangermann
Head of the Lab, Associate Professor, PI