Rieger, Robert; (2004) Implantable analogue circuits for improved methods of nerve signal recording. Doctoral thesis (Ph.D.), University College London (United Kingdom).

Text Implantable_analogue_circuits_.pdf Download (8MB)

Abstract

Every year thousands of persons suffer injuries to the central nervous system with the permanent loss of sensation and voluntary motor functions. Functional electrical stimulation is a useful technique to restore some of these motor functions. However, the challenge remains to improve control of the stimulated muscle through provision of a feedback path. Biopotentials recorded from nerves (Electroneurogram, ENG) by means of cuff electrodes can provide the required feedback in a fully implantable system. The information and signal-to-noise ratio (SNR) obtainable from ENG recording can be increased by the use of multi-electrode cuffs (MECs). A MEC contains several electrode pairs embedded in the inside wall. The signal and noise performance of the MEC are discussed in this thesis. Typical signals recorded with this method are extremely small, on the order of 1μV and require amplification by a very low-noise, high-gain system providing multiple recording channels at very low power consumption. This thesis is concerned with the design of a fully integrated analogue multichannel ENG recording system, which due to its small size can be combined with the recording cuff to form a 'smart electrode', i.e. an electrode joined with active circuitry to provide additional functionality and higher signal quality. The specification of the system improves on previously reported designs for single tripole recordings making it suitable for MEC applications. Special attention is given to low-noise optimisation. The relevant noise equations are derived in this thesis, which leads to an optimum design strategy for very low-noise amplifiers. Due to the high gain of the system, AC-coupling of the gain stages is essential to preserve the available dynamic range. A suitable passive high-pass filter is discussed, which removes DC offsets without degrading the noise performance. Furthermore, a micropower, continuous-time integrator with a very large time-constant is proposed. Such an integrator is a main building block in an ENG adaptive amplifier, which was proposed to reduce the effect of residual muscle signal interference on the recorded ENG. A recording system incorporating this integrator has the potential to further improve recording quality in a future design. Measured results for all the proposed circuit blocks are presented as well as results for an 11-electrode ASIC combining the proposed circuits. Furthermore, in vitro tests with frog nerve show the feasibility of extracting parameters such as compound action potential propagation velocity and direction from the recorded multi-channel ENG.

Download activity - last month

Export as CSVXMLJSON

Download activity - last 12 months

Export as CSVXMLJSON

Downloads by country - last 12 months

Export as JSONCSVXML

Archive Staff Only

View Item