Wireless neural-interfacing microsystems based on CMOS technology have been widely used for brain monitoring, diagnostic, therapeutic, and prosthetic applications, such as deep-brain stimulation, epileptic seizure detection and intervention, and brain-computer interface (BCI). However, even recent neural interfacing devices still rely on bulky external components such as antennas, capacitors, or inductors for system functionality and/or efficiency improvements, limiting their utility when ultra-miniaturized integration is required. To overcome this challenge, we took the opportunity to lead a team of graduate researchers to develop the first prototype of wireless neural-interface-on-chip that incorporates all functionality on a single chip: antenna, adiabatic stimulator, analog front-end, ADC, and wireless power and data telemetry, requiring no external wires, substrates, batteries, or any other external components. This system is one of the smallest neural-interfacing microsystems with a volume of 3×3×0.3 mm3, small enough to be placed amongst the folds and curves of the cortical surface, and to be implanted through minimally invasive surgical procedures inserted through small skull fissures. We validated each component of the system on the electrical bench-top and have proceeded with first in vivo tests [Ha et al., VLSI 2015].
We are pursuing three main research directions in this area. First, we have been further refining the development of neural-interface-on-chip platforms targeting neurological applications. In doing so we have been addressing challenges in aspects of EMI interference, biocompatible encapsulation, interrogator, multi-user communication, etc. This work will truly enable a chronic implanted BCI system that makes enormous impact on neuroscience research and clinical treatments for numerous patients with neurological disorders and paralyzed limbs.
Second, We are pursuing fundamental limits of performance in core sub-components for wireless power and data telemetry, stimulation, recording, on-chip signal processing, and sensor networking. For example, the power and data telemetry technology we developed broke through the conventional performance barrier of single-link inductive telemetry, achieving the highest data rate normalized to RF carrier frequency. With this novel technique, we have demonstrated data rates up to 10 Mbps, greater than half the 13.56 MHz carrier frequency, with simultaneous delivery of more than 10 mW power over the same inductive link [Ha et al., JSSC 2016]. This technology is applicable not only to neural recording systems, but also various RFID-based sensor systems and near-field communication between hand-held devices. Likewise, we anticipate further innovation on sub-components through our hands-on experience in various developments of ADCs, stimulators, analog front-ends, PLL and RF transmitters will greatly impact further advances of biomedical instrumentation.
The third research direction is non-amperometric in vitro neural/cellular interfacing as a more advantageous tool for investigating cellular-level functions and interactions than in vivo settings. By applying the technique we developed for non-amperometric (capacitive) recording and stimulation, we are planning to develop a non-contact MEA-based neural interface platform that minimally interferes with the targeted neurons and cells, tailoring to new emerging research opportunities.