Micro-miniaturized Neural Recording Motes using Mid-field wireless Energy Transfer for Implantable Brain Machine Interface

Abstract

Invasive Brain-Machine Interface(BMI) has revealed the high potential to assist motor or communication functions of patients with paralysis or neurological impairment. However, most of the system requires maintained transcutaneous wiring between recording electrodes and external devices, which raises the risks of infection and limits the range of behavior of subjects. Thus, many research teams have tried to develop wireless neural implant with an internal battery or external power transmission strategies. Nevertheless, a number of safety problems still remain unsolved due to a large footprint of implant, heat generation and limited power supply. Herein, to address those challenges, we developed the micro-miniaturized neural recording sensors by combining Mid-field Wireless Power Transfer and RFID(Radio-Frequency Identification) technique. Mid-field Wireless Power Transfer is a recently suggested method that realizes optimal current density to create energy-focused region deep in biological tissue allowing the receiving coil to be made extremely small. By using an electromagnetic structural simulator, we designed Mid-field WPT antenna that can approximate optimal current density with single-phase source. This antenna also operates at a dual frequency (2.46 GHz and 4.92 GHz) realizing high efficiency both in wireless powering and backscattered signal readout. On the top of that, we designed neural implant as passive RFID sensor consisted by commercial electronic components, such as Schottky diode and RF bipolar junction transistor(BJT). The Schottky diode generates harmonic component of received source power and RF BJT performs highly nonlinear mixing of RF carrier and amplified neural potentials. All components are compactly bonded together small enough to be inserted inside 2mm diameter coil and the overall implant was encapsulated in epoxy resin. Performance test of the powering system and the neural implant was conducted at multi-layer agar phantom simulating relative permittivity and electrical conductance of skin, skull, dura and gray matter, white matter. In this measurement, our wireless powering system achieves a transmission efficiency of 3.3×10-3 for 2.46 GHz and 4.6×10-4 for 4.92 GHz when the miniature coil is placed in target separation, 17.5mm. This powering system allows the recording sensor to have a remarkably small footprint (3 mm × 5 mm), which is only 6% of a comparable system. With this micro-miniaturized footprint, our implant is theoretically able to monitor the extracellular potential as low as 47 μVpp from a high impedance electrode (20 Kohm). Our system demonstrated its ability to extract neural signals while minimizing potential trauma or physiological interference from the implant. Further improvement of this study will pave the way for a broad range of clinical application of wireless BMI from hospital to daily life.