Wireless multichannel biopotential recording using an integrated FM telemetry circuit

Abstract

This paper reports on the design, implementation, and testing of wireless multichannel recording microsystems featuring on-chip AC amplification, DC input stabilization, time-division-multiplexing, and wireless FM reconstruction of input biopotentials with frequency contents from 0.05-6 kHz with measured I/O correlation coefficients in the range of 70-94% per channel for spike train input amplitudes of 0.2-2 mV(p-p) while dissipating only 2.2 mW from 3 V. The 4.84 mm(2) IC is fabricated using AMI 1.5 microm 2P2M CMOS process, and is successfully interfaced with a micromachined silicon probe for simultaneous multichannel wireless in vitro recording of simulated neural spikes at 98 MHz with measured I/O correlation coefficients of >80%.

Embedded Neural Recording With TinyOS-Based Wireless-Enabled Processor Modules

“…The high-pass corner frequency of this stage is (1) where is the resistance of the subthreshold MOS resistors. The resistance of a subthreshold MOS resistor is governed by (2) where , , , , and are the device width, length, carrier mobility, subthreshold-slope factor, and gate capacitance, respectively [16]. In addition, is the thermal voltage, is the gate-source voltage (controllable by the 12-bit mote DAC), is the drain-source voltage, and is the carrier Fermi energy [17].…”

Using Pulse Width Modulation for Wireless Transmission of Neural Signals in a Multichannel Neural Recording System

“…Matching between the receiver antenna and front-end RF LNA results in transferred noise power of (12) where is the receiver RF front-end bandwidth. At room temperature, (12) in dBm becomes (13) Every stage in Fig. 2 has a thermal noise characterized by its noise factor, , and noise figure .…”

“…The advantage of this method is its simplicity and low power consumption. However, analog samples are susceptible to noise, and the transitions from one channel to another in short sampling periods can result in significant crosstalk among adjacent channels on the receiver side [13], [14]. To get a sense of how much information needs to be transferred across the wireless link, we should note that the neural signal spectrum spans from 0.1 Hz to 10 kHz.…”

The case for a generic implant processor

“…FiniteState-Machine (FSM)-based systems [18], [19], [20] to more generic and software-based (µP /µC-based) ones [21], [22], [23]. This trend has been well-studied [16] and is depicted in Fig.1.…”