Save this PDF as:

Size: px
Start display at page:



1 1

2 SOME RECENT DEVELOPMENTS IN THE DESIGN OF BIOPOTENTIAL AMPLIFIERS FOR ENG RECORDING SYSTEMS John Taylor, Delia Masanotti, Vipin Seetohul and Shiying Hao Department of Electronic and Electrical Engineering University of Bath, UK 2

3 ACKNOWLEDGEMENTS Professor N Donaldson, Department of Medical Physics & Bioengineering University College, London, UK Dr P Langlois, Department of Electronic and Electrical Engineering, University College, London, UK Dr J Robbins, Department of Pharmacology, King College, London, UK Dr A Sapelkin, Department of Physics, Queen Mary University, London, UK Dr R Rieger, Department of Electrical Engineering, National Sun Yatsen University, Taiwan Dr D Pal, Department of Electronics & Communication Engineering, B.I.T. Mesra, India Dr M Schuettler, Department of Microsystems Engineering, University of Freiburg, Germany Dr N Rijkoff Centre for Sensory Motor Interaction (SMI), University of Aalborg, Denmark Engineering and Physical Sciences Research Council (UK) (EPSRC) 3

4 SUMMARY The importance of realtime recording of electroneurogram (ENG) signals The difficulties of achieving a stable interface between tissue and electronic devices Illustrations of current problems being studied at Bath University 1 Velocity selective recording (VSR) of ENG signals 2 In vitro recording of ENG from cloned neurons Some possible future directions 4

5 THE IMPORTANCE OF ENG RECORDING Although this area has been researched for some time, there is still much demand for improved systems for realtime ENG recording. Interest comes from eg: Neuroscientists requiring experimental data in fields such as neurophysiology and neuropharmacology Engineers requiring inputs for systems to control Functional Electrical Stimulation (FES) systems for a variety of rehabilitation applications such as neurogenic urinary incontinence by stimulation of the sacral roots There is a demand for recording methods with improved functionality, eg. Velocity/diameter selective recording (VSR) 5

6 NERVE CUFFS FOR ENG RECORDING nerve electrodes Avery and Wespic (1973) Avery (1973) cuff Naples et al Kallesoe (1996) (1988) Nerve cuff with tripolar electrode assembly Various types of cuff design 6

7 MULTIELECTRODE CUFF (MEC) cable ceramic adapter cuff Polyimide thinfilm technology Sputtered platinum electrodes Etched using oxygen plasma. The final MEC was 1.5 mm in diameter, 40 mm long and carried eleven 0.5 mm wide, ringshaped platinum electrodes M Schuettler,

8 RECORDING ENG USING NERVE CUFFS Nerve From tissue This type of cuff/amplifier connection is called a Quasi Tripole (QT). It Amplifier Output provides good suppression of EMG and other artifacts. It only requires one Electrode amplifier and is relatively simple to Cuff implement. It has been much used in practical ENG recording systems. To the central nervous system 8

9 10 CHANNEL ENG RECORDING SYSTEM nerve insulating cuff 1strank amplifiers AC coupling stage 2ndrank amplifiers Signal processing unit (SPUdigital) (N1) (N2) (N3) output for one matched velocity etc etc bandpass filter electrode (rings) subtractors (0) time delays adder digitisation 9

10 PREAMPLIFIER SCHEMATIC V DD M4 M5 M6 Q3 M7 V in Q1 Q2 V in M3 60k V out I bias M1 M2 34k 50pF M8 M9 M10 M11 M12 V ref V SS 10

11 10 CHANNEL ENG RECORDING SYSTEM Lownoise Preamps 2 nd rank Amps Channel Selection MUX 3mm AC Coupling Stage Die mounted in PGA package ASIC layout 11




15 MEASURED RESULTS IN FROG Stimulation intensity 0.13 µc Stimulation intensity 1.01 µc 15

16 MEASURED VSR DATA IN FROG Delay profiles corresponding to two different stimulation intensities: grey: 1.01 µc, white: 0.13 µc. The bars have a width of 25 µs (reciprocal value of the sampling frequency) 16

17 BIDIRECTIONAL INTERFACING OF ELECTRONICS AND CULTURED NEURONS This is a collaborative programme involving E&EE at Bath, KCL/Guy s Hospital (London), Dept of EE at University College (London) and the Department of Physics at Queen Mary University (London) The overall aim is to enhance our understanding of how mammalian nerve cells can be connected optimally to integrated electronic circuitry for neurobiological research and medical applications We intend to find an alternative method to traditional patch clamping which is nonintrusive, less labour intensive in use and is based on a simple, cheap, reproducible CMOS integrated circuit without any need for elaborate postprocessing 17

18 culture medium Si substrate NEURONELECTRODE INTERFACE: EQUIVALENT CIRCUIT sealing gap dendrite 5400µm nucleus cell body terminal axon cell membrane cleft sealing gap 0.220µm extracellular signal V out RNa Rk C gna gk ENa Ek V M Rseal Rspr Csh Cel Rel Rm CMOS circuitry RL gl EL V out metallic pad insulating layer 18

19 PSPICE REALISATION OF THE HODGKINHUXLEY MODEL OF A NEURON (1952) POTASSIUM CHANNEL EXP Beta(n) n PWR 4 gk IK Vmv GAIN = 1 Voltage Clamp V V1 = 0 V3 R10 V2 = 100 TD = 20m 1e12 TR = 1p TF = 1p PW = 20m PER = 30m 0 V V Vmv EXP PWRS1 Alpha(n) G10 G R12 C10 1e u VK SODIUM CHANNEL m EXP Beta(m) PWR 3 GAIN = 1 V G8 G R9 1e12 C8 1000u EXP PWRS1 Alpha(m) gna Vmv INa ITotal V GAIN = 1 G7 G R8 1e12 C7 1u V Cm h VNa V EXP PWRS1 Beta(h) GAIN = 1 G9 G R11 1e12 C9 1000u.0003 gl Vmv IL EXP 0.07 Alpha(h) 0 VL LEAKAGE CHANNEL 19

20 PADONLY CHIP LAYOUT Chip layout Packaged Die 20

21 GROWING NEURONAL CELLS ON CMOS NG10815 (neuroblastoma x glioma hybrid) cells, stained with methyl blue Grown on CMOS microchips precoated with a cationic polymer, (PE1), for 5 days 100μm 21

22 NEURONAL GROWTH ON POROUS SILICON Etched from polysilicon on a CMOS chip 22

23 ELECTRICAL IMPEDANCE SPECTROSCOPY (EIS) MEASUREMENTS ON PACKAGED CMOS CHIPS Z Frequency (Hz) theta Averaged EIS measurements Frequency (Hz) pin1_chip1_31may06.z pin2_chip1_31may06.z pin3_chip1_31may06.z pin4_chip1_31may06.z pin5_chip1_31may06.z Packaged die with insulated bond wires 23

24 FUTURE WORK Neural Prostheses: (i) complete the demonstration of velocity selectivity (start January 2007) (ii) consider new methods of neural recording, e.g. using optoelectronics Electroneural Interfacing: (i) Demonstrate capture of ENG from single excited neurons; design onchip signal processing for optimum SNR. (ii) Complete modelling work SMART Orthopaedic Sensors: (i) Develop an optimal implanted sensor for hip micromotion detection and verify with invitro experiments. (ii) Develop other possible applications for the use of SMART sensors and actuators in modern orthopaedic applications 24

25 SYSTEM SPECIFICATION Power supply Power consumption Circuit area Midband gain Bandwidth 1kHz Inputreferred voltage noise density 1 Hz, 1 khz Inputreferred current noise density 1 Hz, 1 khz Inputreferred rms voltage noise 1 Hz 5 khz Parameter Number of channels Residual DC input current Specification 10 ±2.5 V < 50 mw 10, Hz, 3.5 khz 100 db 20 nv/ Hz, 4 nv/ Hz 20 pa/ Hz, 2 pa/ Hz < 300 nv < 100 na Measured ±2.5 V 24 mw 12 mm 2 10, Hz, 3.3 khz 82 db 11.5 nv/ Hz, 3.8 nv/ Hz 17 nv/ Hz, 1.5 nv/ Hz 291 nv 15 na, 20 na,, inputs 25