Stent with Piezoelectric Transducers for High Spatial Resolution Ultrasound Neuromodulation - a Finite Element Analysis

Abstract

DEEP brain stimulation is currently the only technique used in the clinical setting, capable of stimulating the deep brain nuclei. Recently, low intensity focused ultrasound has been shown to reversibly modulate brain activity. At present, ultrasound neuromodulation is performed through the transcranial pathway. This requires the ultrasound waves to pass through the skull, thus attenuating the ultrasound waves. To reduce the effect of attenuation, lower frequencies are used of around 0.5 MHz, however, with decreasing frequency the focal region in which the stimulation occurs also increases. Therefore, a stent with an array of ultrasound transducers operating at high frequencies of around 5 MHz has been proposed to bypass the skull, thus significantly
reducing the attenuation. Moreover, such a device would require a minimally invasive procedure removing the need for open brain surgery as currently performed for deep brain stimulation. This research was performed to evaluate the feasibility of using ultrasound arrays within the brain vasculature to stimulate deep brain nuclei using the finite elementmethod.

Initially, a literature review was conducted to identify the optimal ultrasound parameters for achieving neuromodulation. For stimulation, a PRF ranging from 500 to 1500 Hz and a DC of around 50 to 70% were found to be optimal. Whereas for suppression, a PRF of<100 Hz and a DC of around >=10% were found to be ideal. The intensity for both suppression and stimulation was found to greatly depends on the target region and individual as well as anaesthesia for stimulation, nonetheless, for 5 MHz an intensity of around 1W/cm2 was required to achieve neural excitation.

Next, the brain nuclei were studied and the subthalamic nucleus, ventral intermediate nucleus and globus pallidus internus were identified to be the most commonly targeted nuclei for deep brain stimulation. Brain vasculature was then examined in terms of the blood vessel inner diameter, shape, orientation with respect to the nuclei and the distance to the nuclei. Basal vein of Rosenthal, internal cerebral vein and internal carotid arterywere found to be most suitable to target the subthalamic nucleus, ventral intermediate nucleus and globus pallidus internus respectively.

Afterwards, two-dimensional simulationswere performed usingCOMSOLMultiphysics to identify the minimum required array width and length for stimulating each nucleus from its respective blood vessel. However, when compared to 3D counterparts it was found that 2D significantly overestimated the size of the focal region.

Finally, 3D simulations of 100 x 100 x 262 ¹m transducers in a 104 (2D 26 x 4) and 105 (Checkered [3 x 2] x 21) arrays within the BVR were analysed targeting the STN. These arrays were found to produce a focal region of full width half intensity within the subthalamic nucleus model boundaries while reaching intensities and pressures of 0.558 W/cm2 or 0.140 MPa and 0.311 W/cm2 or 0.101 MPa with a 1 V terminal for 2D and checkered array respectively. Therefore, these arrays are capable of reaching the neural excitation threshold of 1 W/cm2 with terminal input of less than 2 V . Additionally, a maximum input of 9.2 V is required to reach 1 MPa pressure threshold that has been reported by other studies. Nonetheless, these values do not take into account the fully unused length of the stents, as the arrays can be designed to be longer. In conclusion, finite element method result are promising for deep brain stimulation via a vascular stent with ultrasound transducers for high-spatial neuromodulation. Nonetheless, further work is required to develop an initial prototype as well as initially test the prototype and array within a phantom.