matching considerations for wireless and batteryless brain ...introduction to deep brain...
TRANSCRIPT
Katrina Guido and Asimina KiourtiElectroScience Laboratory
Dept. of Electrical and Computer EngineeringThe Ohio State University
[email protected], [email protected]
Matching Considerations for Wireless and Batteryless Brain Implants
Abstract
Having previously proposed a novel wireless, batterylessneurosensing system able to detect deep brain signals down to 20 μVpp during in vitro testing, we now seek to maintain this high sensitivity under in vivo conditions as well. The latter implies recording of deep brain activity via clinical neural electrodes which have inherently high impedances. In this paper, we report a link budget analysis that illuminates the significance of matching the brain implant to its associated electrode. Potentiostaticmeasurements of clinical microelectrodes are also provided, showing impedances in the tens of kΩ to MΩ range. In the future, these studies can be leveraged to maintain the wireless, passive, batteryless operation while matching implant impedance to the high impedance of clinical electrodes.
Index Terms: brain implants, impedance matching, implantable antennas.
Short Bio of Asimina KiourtiAsimina Kiourti is an Assistant Professor in the Department ofElectrical and Computer Engineering at The Ohio State Universitywhere she leads the Wearable and Implantable Technologiesgroup. From 2013 to 2016 she served as a Post-DoctoralResearcher and then a Senior Research Associate at Ohio State’sElectroScience Laboratory. Prior to that, she received the Ph.D.degree in Electrical and Computer Engineering from the NationalTechnical University of Athens, Greece (2013) and the M.Sc.degree from University College London, UK (2009). Prof. Kiourti’sresearch interests lie in bio-electromagnetics, wearable andimplantable antennas, sensors for body area applications, andflexible textile-based electronics. Her publication record includes9 book chapters, 10 patents (3 awarded, 7 pending), 50 journalpapers, and over 80 conference papers. Her work has beenrecognized with over 40 scholarly awards, including the 2018URSI Young Scientist Award, the 2014 IEEE EMB-S BRAIN YoungInvestigator Award, the 2013 Chorafas Foundation OutstandingPh.D. Dissertation Award, the 2012 IEEE MTT-S GraduateFellowship for Medical Applications, and the 2011 IEEE AP-SDoctoral Research Award. Prof. Kiourti is active in IEEE andUSNC-URSI, where she serves in elected and appointed roles. Hercontributions have gained international reputation and have beenfeatured by TechCrunch, Gizmodo, Times of India, AustraliaNetwork News, and the ALN Magazine, among others.
Topics
• Introduction to Deep Brain Neurosensing
• Wireless and Batteryless Brain Implants
• Detecting Low-Voltage neural signals– Challenge 1: Lack of implanted amplifiers
– Challenge 2: Matching to high-impedance electrodes
• Next Steps
Existing Brain Implants are Restrained by Several Safety Concerns
Brain implants can improve medical care and treatment.Example applications: epilepsy, tremor, Parkinson’s, Alzheimer’s, prosthetics control, mental disorders, etc.
Existing implants are restrained by:
1) Wired connections to/from theimplant
2) Heat generated by excessiveelectronics
3) Batteries requiring replacement /recharging
Brain Signals of Interest
Brain Signals Waveform Voltage Range Frequency Range
Extracellular action potentials (spikes) 100–2000 μVpp 100 Hz – 5 kHz
ElectroCorticoGraphic signals (ECoG) 20–10000 μVpp 0.5 – 500 Hz
Local Field Potentials (LFPs) 20–2000 μVpp < 200 Hz
LFPs are low frequency and low amplitude = difficult to detect
Vision: Wireless and Batteryless Acquisition of Deep Brain Signals
• Wireless neurosensors foreveryday use
• Fully passive
• Very simple electronics
• Tiny footprint tominimize trauma
• Acquisition of extremelyweak signals
Overview of Wireless and Batteryless Neurosensing
Signal generator
Demodulator
2.4 GHz
4.8 GHz
4.8 GHz ± fneuro
electrodes implanted antenna
mixer
Batteryless Implant
ExteriorInterrogator
very low-power neurosignals
(fneuro)
2.4 GHz
4.8 GHz ± fneuro
4.8 GHz ± fneuro
fneuro
2.4 GHz
C. Lee, A. Kiourti, et al., “A High–Sensitivity Fully–PassiveNeurosensing System for Wireless Brain SignalMonitoring,” IEEE Trans. Microw. Theory Techn., 2015.
CHALLENGE #1: Lack of Implanted Amplifier
Detector
MixerMixer
LNA
Batteryless Implant Interrogator
2xSig. Gen.
Freq. Doublerfneuro
2.4 GHz
4.8 GHz ± fneuro
2.4 GHz 4.8 GHz
Electrodes Antenna Antenna
Neurosignal power = -90dBm (20μVpp)
Minimum Detectable Signal = -120dBm
System Loss (Mixer + Wireless + Mismatch)
must be < 30 dB
IS THIS POSSIBLE TO ACHIEVE?
𝐟𝐟𝐧𝐧𝐧𝐧𝐧𝐧𝐧𝐧𝐧𝐧
CHALLENGE #1: Lack of Implanted Amplifier
Brain Implant Design Summary
Impl. Ant. Mixer CircuitImpl. Ant. Mixer Circuit
Ground
Antenna Circuit
Approximatelyfold into two
Antenna
Circuit
Ground
Via
Sideview16 mm
Antenna
Circuit
Ground
Via
Sideview10 mm
1. Higher εr2. Meandering3. Cap. & ind.
replaced by trans. lines
145 mm
ExternalAntenna
Impl. Ant. Mixer Circuit
8.7 mm × 10 mmMDS = 20 μVpp
39 mm × 15 mmMDS = 200 μVpp
16 mm × 15 mmMDS = 63 μVpp
18.8 mm
10.1 mm
2015 2016 2017
Ground
Antenna Circuit
Sideview39 mm
ImplantedDevice
Ohio State’s Work vs. Previously Reported Wireless Brain Implants
Ref. Type Footprint Power consumption Min. detectable signal
Yin, 2014 Exterior 52 x 44 mm2 17 mA from a 1.2 Ah battery to run for 48 hours N/A
Szuts, 2011 Exterior N/A 645 mW 10.2 μVpp (rat)Rizk, 2007 Exterior 50 x 40 mm2 100 mW N/A
Miranda, 2010 Exterior 38 x 38 mm2 142 mW 14.2 μVpp (non-human primate)
Yin, 2010 Exterior N/A 5.6 mW 13.9 μVpp (rat)Sodagar, 2009 Exterior 14 x 16 mm2 14.4 mW 25.2 μVpp (guinea)
Borton, 2013 Implanted 56 x 42 mm2 90.6 mW 24.3 μVpp (non-human primate)
Rizk, 2009 Implanted 50 x 40 mm2 2000 mW 20 μVpp (sheep)Sodagar, 2007 Implanted 14 x 15.5 mm2 14.4 mW 23 μVpp (guinea)Moradi, 2014 Implanted N/A N/A, yet >0 mW N/A
Schwerdt, 2011 Implanted 12 x 4 mm2 Fully passive 3.4 mVpp (in-vitro)
Lee, 2015 Implanted 39 x 15 mm2 Fully passive 200 μVpp (in-vitro)
Kiourti, 2016 Implanted 16 x 15 mm2 Fully passive 63 μVpp (in-vitro)
Lee, 2017 Implanted 8.7 x 10 mm2 Fully passive 20 μVpp (in-vitro)
A. Kiourti et al., “A Wireless Fully-Passive Neural Recording Device for Unobtrusive Neuropotential Monitoring,” IEEE Trans. Biomed. Eng., 2016.C.W.L Lee et al., “A high–sensitivity fully–passive neurosensing system for wireless brain signal monitoring,” IEEE Trans. Microw. Theory Techn., vol. 63, pp. 2060–2068, Jun. 2015.
C.W.L. Lee et al., “Miniaturized Fully-Passive Brain Implant for Wireless Neuropotential Acquisition," IEEE Antennas Wireless Propag. Lett., 2017.
CHALLENGE #2: Multi-Channel Sensing Needed in PracticeCHALLENGE #2: Electrodes Have High Impedance
Our previous devices can detect down to 20 μVpp generated by a (50 Ω) function generator, but real electrodes have much higher impedances
(kΩ to MΩ range)
To improve system sensitivity, we need:
Higher implant impedance
or
Lower electrode impedanceFHC, Inc
Electrode Impedance Spectroscopy
10 0 10 2 10 4
Frequency (Hz)
10 1
10 2
10 3
10 4
10 5
|Z| (
)
-55
-45
-35
-25
-15
-5
Phas
e (d
egre
es)
Clinical electrode
impedance is7.4 kOhms at 20 Hz
Previous implants
matched to 50 Ohms
Clinical Electrode ImpedanceClinical Neural
Electrode
Ag/AgClReference Electrode
Pt CounterElectrode
CMOS chips use a pre-amplifier on the first stage to account for high impedance, but no pre-amplifiers are allowed in our design (batteryless!).
Impedance-Matched Implant Layout
Since the BJT pair has much larger input impedance than the RF diode alone, it can boost the overall input impedance of the implant.
Impedance-Matched Implant Layout
SchottkyDiode
SchottkyDiode
• DC Mode – Diode = Rectifier
• DC current self-biases theBJT
• RF Mode – Diode = Mixer
• Harmonic mixing generatesthird-order products
Performance of our 2019 Impedance-Matched Implant
• Neural signals ≥ 100 µVppcan be retrieved for a 33 kΩelectrode impedance
• Sensitivity of our previoussystem was 50 times lower
Recorded Time Domain Signal Improvement
0
1000
2000
3000
4000
5000
2011 2013 2015 2017 2019
Sens
itivi
ty (μ
Vpp
)
Year
Sensitivity Improvement by Year Using50 and 33 kΩ Electrodes
50 OhmElectrode
33 kOhmElectrode
Work-in-Progress
Improve Wireless Interface Robustness
Transmission loss at 2.4 GHz
Transmission loss at 4.8 GHz
Work-in-Progress
Clinical Neural
Recording System
Splitter
Fully PassiveNeurosensor
Macro
Micro
Macroelectrode
Microelectrode
Sterile Environment
References
1) G. Schalk and E.C. Leuthardt, “Brain-computer interfaces using electrocorticographic signals,”IEEE Rev. Biomed. Eng., vol. 4, pp.104-154, 2011.
2) K. D. Wise, “Microelectrodes, microelectronics, and implantable neural microsystems,” Proc. IEEE,vol. 96, no. 7, pp. 1184–1202, Jul. 2008.
3) J. P. Blountetal, “Advances in intracranial monitoring,” Neurosurg. Focus, vol. 25, no. 3, pp. 1–8,Sep. 2008.
4) A. Waziri, C. A. Schevon, J. Cappell, R. G. Emerson G. M. McKhann, and R. R. Goodman, “Initialsurgical experience with a dense cortical microarray in epileptic patients undergoing craniotomyfor subdural electrode implantation,” Neurosurgery, vol. 64, pp. 540-545, 2009.
5) E. Moradi, T. Bjorninen, L. Sydanheimo, J. M. Carmena, J. M. Rabaey, and L. Ukkonen,“Measurement of wireless link for brain–machine interface systems using human-head equivalentliquid,” IEEE Antennas Wireless Propag. Lett., vol. 12, pp. 1307–1310, 2013.
6) S. Kim, P. Thathireddy, R. A. Normann, and F. Solzbacher, “Thermal impact of an active 3-Dmicroelectrode array implantedin the brain,” IEEE Trans. Neural Syst. Rehabil. Eng., vol. 15, pp.493-501, 2007.
7) C. W. L. Lee, A. Kiourti, and J. L. Volakis, “Miniaturized fully passive brain implant for wirelessneuropotential acquisition,” IEEE Antennas Wireless Propag. Lett., vol. 16, pp. 645–648, 2017.
8) A. Kiourti, C. W. L. Lee, J. Chae, and J. L. Volakis, “ A high-sensitivity fully-passive neuralrecording device for unobtrusive neuropotentional monitoring,” IEEE Trans. Biomed. Eng., vol. 63,pp. 131-137, 2016.
9) J. Ruiz-Amaya, A. Rodriguez-Perez, and M. Delgado-Restituto, “A low noise amplifier for neuralspike recording interfaces,” Sensors, vol. 15, pp. 25313–25335, 2015.
Questions?
Katrina Guido and Asimina KiourtiElectroScience Laboratory
Dept. of Electrical and Computer EngineeringThe Ohio State University