Deep brain stimulation of the subthalamic nucleus: model-based analysis of the effects of electrode capacitance on the volume of activation

Deep brain stimulation (DBS) has rapidly emerged as an effective clinical treatment for movement disorders. However, our understanding of the neural effects of DBS is limited, and significant opportunities exist to optimize electrode design to enhance therapeutic effectiveness. To address these issues, we have developed computational tools to predict the neural response to stimulation. For decades the electrostatic approximation has been applied in neural stimulation modeling, treating the electrode as a perfect current source and the neural tissue as a purely conductive medium. However, clinical DBS electrodes are voltage controlled, utilize an asymmetrical biphasic stimulus waveform, and are surrounded by a 3D anisotropic, inhomogeneous tissue medium. To more accurately model DBS in the human, we have developed finite element models (FEM) of the electrode and tissue medium that incorporate a Fourier FEM solver to determine the potential distribution in the tissue in time and space simultaneously. The field data is then coupled to multi-compartment neuron models to predict neural activation. Our results show that electrostatic models overestimate the volume of activation (VOA) by -30% compared to voltage-controlled stimulation for typical therapeutic stimulation parameter settings. The error is directly related to the electrode capacitance and the stimulation pulse width. These results illustrate the need for detailed models of neural stimulation to accurately predict the effects of DBS