Timon Merk: Decoding movement from invasive cortical vs. subthalamic for clinical brain computer interfaces for Parkinson’s disease
BCCN Berlin / TU Berlin
Subthalamic (STN) Deep Brain Stimulation (DBS) is an effective treatment for patients suffering from Idiopathic Parkinson Syndrom (IPS). Currently DBS targets are stimulated continuously with high frequency (130 Hz). First studies however show significant advantages of demand actuated adaptive stimulation over current in practice continuous DBS. Therefore invasively measured feedback signals based on neural activity show high research relevance. Neuron population activity in form of local field potentials (LFP) from the basalganglia and electrocorticographical (ECOG) signals from the cortex are measured. For a clinical “Brain-Computer-Interface” those signals could adapt DBS stimulation parameters movement dependently. Movement decoding of this kind had been postulated mostly based on time consuming patient individual training. In this master thesis real-time movement decoding is presented through LFP –and ECOG signals, which had been implemented without patient individual training. Therefore STN-LFP and Cortex-ECOG signals are transformed frequency band specific, the respective electrode positions are localized and interpolated on a cortex – and STN spanning grid including 94 grid points in the Montreal Neurological Institute (MNI) standard space. The results show that statistical significant movement decoding is possible across patients. The movement detection rate is improved by use of neuronal networks or ensemble classifiers like Random Forests over linear regression. The results also show similar high movement detection rates compared to publications using patient individual training. This could lead to a better, demand actuated DBS for patients with Idiopathic Parkinson syndrome. Overall the clinical potential of machine learning in invasive neuromodulation is presented.
Master Thesis Defense
If you would like to join the online talk, please contact graduateprograms(at)bccn-berlin.de
Wolf-Julian Neumann / Lisa Velenosi
Location: Zoom (Online)