Tremor dependant nonlinear interaction in deep brain local field potentials of Parkinson´s disease

Nonlinear interaction between neural oscillations may be a biomarker for closed-loop deep brain stimulation (DBS) in Parkinson´s disease. Bispectral analysis is an effective tool to detect quadratic phase coupling. In this paper we applied bispectral analysis to local field potentials (LFPs) recorded from Parkinsonism patients´ subthalamic nucleus (STN) during tremor and resting states. The significance of the coupling was statistically tested using bispectrum estimation with surrogate data approaches. Strong phase couplings were found in both states, whereas the coupling frequencies varied. There are bispectral peaks around 3-6Hz in both states and the peak amplitudes in tremor state are twice larger than those in resting state, while the bispectra in resting state shows peaks in narrow frequency bands of 7-9Hz, 10-12Hz and 15-17Hz. The measure of these nonlinear interactions between neural oscillations of local field potentials can be used to distinguish pathological states in Parkinson´s disease and could contribute to the development of intelligent deep brain stimulation strategies.