This toolbox enables automated accelerometer-based assessment of UPDRS-tapping tasks, performed in clinical Parkinson-assessment routines.
The toolbox will provide 1) an automated prediction of UPDRS-subscore of a 10-second fingertapping task, and 2) provide detailed movement
features which can help the clinician or researcher to assess motor performance and bradykinesia.
Written by Jeroen Habets and Rachel Spooner as part of the ReTune-Consortium, a scientific collaboration between (among others) the HHU Düsseldorf and the Charité Berlin.
Work in Progress: data analysis ongoing.
environment installation: conda create --name updrsTapping python=3.9 jupyter pandas scipy numpy matplotlib statsmodels seaborn scikit-learn h5py
additional installed packages: pip install mne (for working with raw .poly5 files, e.g. via TMSI-amplifier) conda install openpyxl (for opening latest version Excel files) conda install pingouin (for calculating ICC)