Motor task event detection using Subthalamic Nucleus Local Field Potentials

Abstract

Deep Brain Stimulation (DBS) provides significant therapeutic benefit for movement disorders such as Parkinson's disease. Current DBS devices lack real-time feedback (thus are open loop) and stimulation parameters are adjusted during scheduled visits with a clinician. A closed-loop DBS system may reduce power consumption and DBS side effects. In such systems, DBS parameters are adjusted based on patient's behavior, which means that behavior detection is a major step in designing such systems. Various physiological signals can be used to recognize the behaviors. Subthalamic Nucleus (STN) Local Field Potential (LFP) is a great candidate signal for the neural feedback, because it can be recorded from the stimulation lead and does not require additional sensors. A practical behavior detection method should be able to detect behaviors asynchronously meaning that it should not use any prior knowledge of behavior onsets. In this paper, we introduce a behavior detection method that is able to asynchronously detect the finger movements of Parkinson patients. As a result of this study, we learned that there is a motor-modulated inter-hemispheric connectivity between LFP signals recorded bilaterally from STN. We used non-linear regression method to measure this connectivity and use it to detect the finger movements. Performance of this method is evaluated using Receiver Operating Characteristic (ROC).

A Multiple Kernel Learning approach for human behavioral task classification using STN-LFP signal

“…The quantitative results of our method are compared with a recently proposed single kernel SVM-based method [15]. We also assess the effect of three different kernel functions on the SVM classifier separately: .…”

“…To obtain the spectrogram of a continuous signal, we apply the continuous wavelet transform (CWT) on the bilaterally re-referenced LFPs. We employ the complex Morlet (C-Morlet) mother wavelet which is proven to be a suitable choice for biomedical signal processing [1,15]: …”

“…Many algorithms have been developed based on the processing of electroencephalography (EEG) and electrocorticography (ECoG), e.g., P300 speller paradigm [7,8], seizure detection [9], brain-switch based on motor imaginary [10]. Neural Networks [11,12] and SVM-based classifiers together with wavelet-domain feature extraction have been successful in detecting human behavior [5,13,14]. Most recently, STNLFPs have been considered as another useful neural feedback signal to recognize human behavioral activities.…”

“…Timefrequency analysis of motor cortex recorded with ECoG and STN-LFPs has shown that PD exhibits oscillatory behavior modulated by motor activities, which results in suppression of β (13-35 Hz) frequency spectral power during motor tasks [1]. Taking this point into account, an SVM-based classifier was proposed in [15]. A motor task detection method using LFPs and non-linear regression has been addressed in [14].…”

“…Taking this point into account, an SVM-based classifier was proposed in [15]. A motor task detection method using LFPs and non-linear regression has been addressed in [14]. An adaptive learning approach based on LFP signal was suggested in [16], where a hybrid model for combining SVM and hidden Markov model was used for human behavioral clustering.…”