100% found this document useful (1 vote)
15 views67 pages

Circuit Design Considerations For Implantable Devices 1st Edition Peng Cong

The document is a promotional overview for the book 'Circuit Design Considerations for Implantable Devices' edited by Peng Cong, which is part of the River Publishers Series in Electronic Materials and Devices. It covers various topics related to the design and application of electronic materials and devices, particularly in the context of biomedical electronics. The book aims to provide insights into the latest research and developments for professionals and researchers in the field.

Uploaded by

atarifobah
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
100% found this document useful (1 vote)
15 views67 pages

Circuit Design Considerations For Implantable Devices 1st Edition Peng Cong

The document is a promotional overview for the book 'Circuit Design Considerations for Implantable Devices' edited by Peng Cong, which is part of the River Publishers Series in Electronic Materials and Devices. It covers various topics related to the design and application of electronic materials and devices, particularly in the context of biomedical electronics. The book aims to provide insights into the latest research and developments for professionals and researchers in the field.

Uploaded by

atarifobah
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 67

Read Anytime Anywhere Easy Ebook Downloads at ebookmeta.

com

Circuit Design Considerations for Implantable


Devices 1st Edition Peng Cong

https://ebookmeta.com/product/circuit-design-considerations-
for-implantable-devices-1st-edition-peng-cong/

OR CLICK HERE

DOWLOAD EBOOK

Visit and Get More Ebook Downloads Instantly at https://ebookmeta.com


River Publishers Series in Electronic Materials and Devices

Circuit Design Considerations for


Implantable Devices

Peng Cong (Editor)

B
Circuit Design Considerations for
Implantable Devices
RIVER PUBLISHERS SERIES IN ELECTRONIC
MATERIALS AND DEVICES

Series Editors

Edoardo Charbon Mikael Östling Albert Wang


EPFL KTH Stockholm University of California
Switzerland Sweden Riverside, USA

Indexing: All books published in this series are submitted to the Web of Science Book
Citation Index (BkCI), CrossRef and to Google Scholar.

The “River Publishers Series in Electronic Materials and Devices” is a series of com-
prehensive academic and professional books which focus on the theory and applications
of advanced electronic materials and devices. The series focuses on topics ranging from
the theory, modeling, devices, performance and reliability of electron and ion integrated
circuit devices and interconnects, insulators, metals, organic materials, micro-plasmas,
semiconductors, quantum-effect structures, vacuum devices, and emerging materials.
Applications of devices in biomedical electronics, computation, communications, dis-
plays, MEMS, imaging, micro-actuators, nanoelectronics, optoelectronics, photovoltaics,
power ICs and micro-sensors are also covered.
Books published in the series include research monographs, edited volumes, hand-
books and textbooks. The books provide professionals, researchers, educators, and
advanced students in the field with an invaluable insight into the latest research and
developments.
Topics covered in the series include, but are by no means restricted to the following:
• Integrated circuit devices
• Interconnects
• Insulators
• Organic materials
• Semiconductors
• Quantum-effect structures
• Vacuum devices
• Biomedical electronics
• Displays and imaging
• MEMS
• Sensors and actuators
• Nanoelectronics
• Optoelectronics
• Photovoltaics
• Power ICs
For a list of other books in this series, visit www.riverpublishers.com
Circuit Design Considerations for
Implantable Devices

Editor

Peng Cong
Verily Life Science, US

River Publishers
Published, sold and distributed by:
River Publishers
Alsbjergvej 10
9260 Gistrup
Denmark

River Publishers
Lange Geer 44
2611 PW Delft
The Netherlands

Tel.: +45369953197
www.riverpublishers.com

ISBN: 978-87-93519-86-2 (Hardback)


978-87-93519-85-5 (Ebook)

c 2018 River Publishers

All rights reserved. No part of this publication may be reproduced, stored in


a retrieval system, or transmitted in any form or by any means, mechanical,
photocopying, recording or otherwise, without prior written permission of
the publishers.
Contents

Preface ix

List of Contributors xi

List of Figures xiii

List of Tables xxiii

List of Abbreviations xxv

1 Introduction 1
Peng Cong
1.1 Implantable Devices . . . . . . . . . . . . . . . . . . . . . 1
1.2 Neural Interface . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Circuit Design Considerations for Neural Sensing . . . . . 5
1.4 Circuit Design Considerations for Stimulation . . . . . . . 7
1.5 Closed-Loop System . . . . . . . . . . . . . . . . . . . . . 11
1.6 RF Powering and Data Telemetry for Implantable Devices . 12
References . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2 Neural Stimulation and Interface 17


Kunal Paralikar
2.1 Neuronal Element . . . . . . . . . . . . . . . . . . . . . . 17
2.1.1 Neuron Structure . . . . . . . . . . . . . . . . . . 17
2.1.2 Neuron: Electrical Characteristics . . . . . . . . . 19
2.1.2.1 Membrane model and resting membrane
potential . . . . . . . . . . . . . . . . . 20
2.1.2.2 Action potential . . . . . . . . . . . . . 22
2.1.3 Neuron: Cable Equation . . . . . . . . . . . . . . 24

v
vi Contents

2.2 Tissue Element . . . . . . . . . . . . . . . . . . . . . . . . 27


2.3 Electrode–Tissue Element . . . . . . . . . . . . . . . . . . 28
2.3.1 Mechanism of Charge Transfer . . . . . . . . . . . 29
2.3.1.1 Charge injection . . . . . . . . . . . . . 30
2.3.2 Electrode Properties and Materials . . . . . . . . . 31
2.3.3 Limits to Electrical Stimulation . . . . . . . . . . . 32
2.4 Discussion: Bringing it All Together . . . . . . . . . . . . . 34
References . . . . . . . . . . . . . . . . . . . . . . . . . . 35

3 Front-End Sensing Amplifiers 37


You Zou
3.1 System Level Architecture . . . . . . . . . . . . . . . . . . 37
3.2 Current Mirror-Based OTA . . . . . . . . . . . . . . . . . 39
3.3 Folded-Cascode OTA . . . . . . . . . . . . . . . . . . . . 40
3.4 Complementary-Input OTA . . . . . . . . . . . . . . . . . 44
3.5 Current Reuse OTA . . . . . . . . . . . . . . . . . . . . . 48
3.6 Reconfigurable OTA . . . . . . . . . . . . . . . . . . . . . 49
References . . . . . . . . . . . . . . . . . . . . . . . . . . 52

4 Circuits for Neural Stimulation 53


Peng Cong
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.2 Stimulation Waveform . . . . . . . . . . . . . . . . . . . . 54
4.3 Circuit Design for Stimulation . . . . . . . . . . . . . . . . 57
4.3.1 Circuit Design . . . . . . . . . . . . . . . . . . . . 57
4.3.2 Safety Consideration . . . . . . . . . . . . . . . . 62
4.3.2.1 Charge balance . . . . . . . . . . . . . 62
4.3.2.2 Electrode safety consideration . . . . . 64
4.3.2.3 Neuron safety consideration . . . . . . . 67
4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
References . . . . . . . . . . . . . . . . . . . . . . . . . . 68

5 Embedded Signal Analysis 71


Z. Wang and N. Verma
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.2 Challenges in Medical-Signal Analysis . . . . . . . . . . . 73
5.3 Data-Driven Medical-Signal Analysis . . . . . . . . . . . . 75
Contents vii

5.4 Overview of Inference Systems . . . . . . . . . . . . . . . 77


5.4.1 Feature Extraction . . . . . . . . . . . . . . . . . 77
5.4.2 Inference . . . . . . . . . . . . . . . . . . . . . . 80
5.5 Specialized Inference Algorithms and Processors . . . . . . 82
5.5.1 Rationale . . . . . . . . . . . . . . . . . . . . . . 82
5.5.2 Classifier Configurability . . . . . . . . . . . . . . 84
5.5.2.1 Inference models . . . . . . . . . . . . 85
5.5.2.2 Model parameters . . . . . . . . . . . . 86
5.5.2.3 Classification meta-algorithms . . . . . 88
5.5.2.4 Active learning . . . . . . . . . . . . . 90
5.6 Training for Low-Energy Systems . . . . . . . . . . . . . . 92
5.6.1 Model Bit Precision . . . . . . . . . . . . . . . . . 93
5.6.2 Dataset Imbalance . . . . . . . . . . . . . . . . . 97
5.6.3 Embedded Training . . . . . . . . . . . . . . . . . 97
5.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
References . . . . . . . . . . . . . . . . . . . . . . . . . . 101

6 Wireless Power Transmission to mm-Sized Free-Floating


Distributed Implants 105
S. Abdollah Mirbozorgi and Maysam Ghovanloo
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 105
6.1.1 Free-Floating Implants (FFIs) . . . . . . . . . . . 105
6.1.2 WPT Measurement Accuracy . . . . . . . . . . . 107
6.2 Modeling and Analysis of Inductive Links . . . . . . . . . 108
6.3 Design Procedure . . . . . . . . . . . . . . . . . . . . . . . 111
6.3.1 System Configuration and HFSS Simulation . . . . 112
6.3.2 Inductive Link Design Rules . . . . . . . . . . . . 113
6.3.3 Coil Segmentation . . . . . . . . . . . . . . . . . 118
6.3.4 SAR Calculations . . . . . . . . . . . . . . . . . . 119
6.4 Experimental Results . . . . . . . . . . . . . . . . . . . . . 120
6.4.1 Implementation and Setup . . . . . . . . . . . . . 120
6.4.2 Wireless Link Measurement Methods . . . . . . . 124
6.4.3 Measurement Results . . . . . . . . . . . . . . . . 127
6.5 Practical Aspects of the FFIs . . . . . . . . . . . . . . . . 131
6.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 134
References . . . . . . . . . . . . . . . . . . . . . . . . . . 135
viii Contents

7 Next Generation Neural Interface Electronics 141


IanWilliams, Lieuwe Leene and Timothy G. Constandinou
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 141
7.1.1 Today’s Neural Interfaces . . . . . . . . . . . . . . 141
7.1.2 The Neural Interface of the Near Future . . . . . . 142
7.2 Sensing Neural Activity . . . . . . . . . . . . . . . . . . . 144
7.2.1 Power Density Target for High Channel Count
Recording . . . . . . . . . . . . . . . . . . . . . . 144
7.2.2 Emerging Technologies . . . . . . . . . . . . . . 147
7.2.2.1 Advanced CMOS technologies . . . . . 147
7.2.2.2 Chemical sensing . . . . . . . . . . . . 148
7.2.2.3 Optical sensing . . . . . . . . . . . . . 149
7.3 Signal Processing . . . . . . . . . . . . . . . . . . . . . . . 149
7.3.1 Single Neuron Processing . . . . . . . . . . . . . . 149
7.3.2 Population Level Processing . . . . . . . . . . . . 153
7.3.3 Signal Processing Summary . . . . . . . . . . . . 154
7.4 Neuromodulation . . . . . . . . . . . . . . . . . . . . . . . 154
7.4.1 Existing Approaches: Issues and Challenges . . . . 157
7.4.1.1 Current controlled stimulators . . . . . . 157
7.4.1.2 Voltage controlled stimulators . . . . . . 158
7.4.1.3 Charge controlled stimulators . . . . . . 158
7.4.2 Trends and Limits . . . . . . . . . . . . . . . . . . 158
7.4.2.1 Channel count and size trends . . . . . . 158
7.4.2.2 Power consumption trends . . . . . . . 159
7.4.2.3 Charge balancing . . . . . . . . . . . . 161
7.4.3 System on Chip . . . . . . . . . . . . . . . . . . . 162
7.4.4 Alternative Approaches and Modalities . . . . . . 163
7.4.4.1 Optogenetics . . . . . . . . . . . . . . . 164
7.4.4.2 Acoustic/Mechanical . . . . . . . . . . 164
7.4.4.3 Thermal . . . . . . . . . . . . . . . . . 165
7.4.5 Neuromodulation Summary . . . . . . . . . . . . 165
7.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . 166
References . . . . . . . . . . . . . . . . . . . . . . . . . . 167

Index 179

About the Editor 183


Preface

To my beloved wife, Rui

This book is based on a tutorial lecture, “Circuit Design Considerations for


Implantable Devices,” I gave at the IEEE International Solid-State Circuits
(ISSCC) in San Francisco, CA, February 2016. Recently, there are increasing
interests in implantable medical devices, especially implantable neuromod-
ulation devices. Implantable devices are unique area for circuit designers.
Comprehensive understanding of design trade-offs at system level is impor-
tant to ensure device success. To expand the scope of the tutorial, many world-
class experts in the field contribute much to this book, Dr. Kunal Paralikar
(Medtronic Inc, US), Dr. You Zou (Verily Life Science, US), Dr. Zhou Wang
(Princeton University, US), Dr. Naveen Verma (Princeton University, US),
Dr. Abdollah Mirbozorgi (Georgia Institute of Technology, US), Dr. Maysam
Ghovanloo (Georgia Institute of Technology, US), Dr. Ian Williams (Imperial
College London, UK), Dr. Lieuwe Leene (Imperial College London, UK) and
Dr. Timothy G. Constandinou (Imperial College London, UK). The goal of
this book is to provide knowledge to circuit designers with limited biomedical
background to understand design challenge and trade-offs for implantable
devices. This book will also be welcomed by those readers with limited circuit
design knowledge who need to understand the essentials of design challenge
and trade-off on circuit design considerations.

Peng Cong

ix
List of Contributors

Peng Cong, Verily Life Science, US

Kunal Paralikar, Medtronic Inc., USA

You Zou, Verily Life Science, US

Zhou Wang, Princeton University, US

Naveen Verma, Princeton University, US

Abdollah Mirbozorgi, Georgia Institute of Technology, US

Maysam Ghovanloo, Georgia Institute of Technology, US

Ian Williams, Imperial College London, UK

Lieuwe Leene, Imperial College London, UK

Timothy G. Constandinou, Imperial College London, UK

xi
List of Figures

Figure 1.1 Example of implantable medical devices:


(a) a pacemaker, (b) spinal cord stimulator, and
(c) cochlear implant. . . . . . . . . . . . . . . . . 2
Figure 1.2 A diagram of a typical implantable medical
device. . . . . . . . . . . . . . . . . . . . . . . . . 2
Figure 1.3 Examples of neuron communication
and a bi-directional neural interface. . . . . . . . . 4
Figure 1.4 An electrode interface model for sensing. . . . . . 6
Figure 1.5 Circuit design considerations for a sensing neural
interface. . . . . . . . . . . . . . . . . . . . . . . . 6
Figure 1.6 A typical stimulation waveform. . . . . . . . . . . 8
Figure 1.7 An electrode interface model for stimulation. . . . 9
Figure 1.8 The Shannon criteria for stimulation safety. . . . . 10
Figure 1.9 Series of steps of classification. . . . . . . . . . . . 11
Figure 1.10 Closed-loop architecture for (a) pacemakers
and (b) brain interface devices. . . . . . . . . . . . 13
Figure 1.11 Equivalent circuit of wireless powering with
inductive link. . . . . . . . . . . . . . . . . . . . . 14
Figure 2.1 Neuron. Inputs are received at the cell body and
dendrites. Action potentials initiate at the axon
hillock and terminate at the axon terminal.
Their propagation down the axon is aided by
insulating myelin sheath secreted by the Schwann
cells. Nodes of Ranvier enable regeneration of the
action potential during its propagation. . . . . . . . 18
Figure 2.2 Communication of electrical impulse/action
potential down the axon results in release of
neurotransmitters in the synapse which are
absorbed by the receptors located on dendrites
and cell bodies of down-stream neurons. . . . . . . 19

xiii
xiv List of Figures

Figure 2.3 Electrical equivalent representation of a cell


membrane. gx represents variable conductance to
ions X and leakage currents, Ex represents their
Nernst Potential and Cm is the membrane
capacitance. . . . . . . . . . . . . . . . . . . . . . 20
Figure 2.4 Schematic for recording extracellular action
potential. . . . . . . . . . . . . . . . . . . . . . . 23
Figure 2.5 Intracellular current flow between nodes n and n + 1
with voltage Vn and V n+1 , respectively, and
resistance per unit length of ri . . . . . . . . . . . . 25
Figure 2.6 Ladder model of a neuron in unmyelinated
location. . . . . . . . . . . . . . . . . . . . . . . . 26
Figure 2.7 Potential distribution due to bipolar (a) and tripolar
(b) stimulation configurations for electrodes used in
spinal cord stimulation. . . . . . . . . . . . . . . . 28
Figure 2.8 Two electrode–tissue interfaces are formed during
bipolar stimulation between two electrodes with the
tissue in between. Each component of the interface
(Z F , Cdl , and Z Tissue ) can have non-linear proper-
ties that may be dependent on the stimulation condi-
tions. Capacitive processes are denoted by Cdl and
the faradaic processes (reversible and irreversible)
are represented by Z F . RS represents the impedance
of the wire connecting the stimulation source like an
implantable neurostimulator and the electrode.
RS , Z F , and Cdl can be modified by appropriate
selection of electrode material and surface
topography. . . . . . . . . . . . . . . . . . . . . . 30
Figure 2.9 Graph of charge and charge density per phase for
safe stimulation. . . . . . . . . . . . . . . . . . . . 33
Figure 2.10 Strength–duration curve shows the required
stimulation intensity at a given stimulus duration
to elicit action potential. Rheobase is the current
intensity when using an infinitely long pulse-width.
Chronaxie is the pulse-width duration required
when using a current intensity that is twice the
rheobase. . . . . . . . . . . . . . . . . . . . . . . 34
Figure 3.1 Capacitive-coupled neural amplifier
architecture. . . . . . . . . . . . . . . . . . . . . . 38
List of Figures xv

Figure 3.2 Current mirror-based OTA. . . . . . . . . . . . . . 39


Figure 3.3 Folded-cascode OTA. . . . . . . . . . . . . . . . . 41
Figure 3.4 Folded-cascode OTA with current-splitting
technique. . . . . . . . . . . . . . . . . . . . . . . 43
Figure 3.5 Single-ended complementary input OTA. . . . . . . 44
Figure 3.6 Fully-differential OTA with complementary input
stage. . . . . . . . . . . . . . . . . . . . . . . . . 45
Figure 3.7 Hybrid OTA. . . . . . . . . . . . . . . . . . . . . 47
Figure 3.8 Current-reuse OTA. . . . . . . . . . . . . . . . . . 48
Figure 3.9 Mid-rail current source. . . . . . . . . . . . . . . . 49
Figure 3.10 Tunable capacitor array for reconfigurable
midband gain. . . . . . . . . . . . . . . . . . . . . 50
Figure 3.11 Tunable feedback resistance for lower-frequency
pole tuning. . . . . . . . . . . . . . . . . . . . . . 51
Figure 3.12 Tunable bias current. . . . . . . . . . . . . . . . . 51
Figure 4.1 Typical stimulation waveform for
neuromodulation. . . . . . . . . . . . . . . . . . . 54
Figure 4.2 Simplified electrode–tissue interface. . . . . . . . . 54
Figure 4.3 Simplified schematic of circuit for stimulation
therapy. . . . . . . . . . . . . . . . . . . . . . . . 55
Figure 4.4 Bipolar and unipolar stimulation waveform. . . . . 56
Figure 4.5 Simplified stimulation circuit. . . . . . . . . . . . . 57
Figure 4.6 Illustration of power efficiency drop with
stimulation amplitude and electrode impedance
decrease. . . . . . . . . . . . . . . . . . . . . . . 58
Figure 4.7 Power efficiency improvement with dynamic high
voltage adjustment. . . . . . . . . . . . . . . . . . 58
Figure 4.8 Charge pump design example with gain of 3x. . . . 59
Figure 4.9 Example of stimulation circuit with improved
current matching. . . . . . . . . . . . . . . . . . . 60
Figure 4.10 R–2R-based current reference DAC. . . . . . . . . 61
Figure 4.11 An implementation using stimulation circuit
for optogenetic application. . . . . . . . . . . . . . 62
Figure 4.12 Example of an active charge balancing design. . . . 63
Figure 4.13 Combination of active and passive charge
balancing design. . . . . . . . . . . . . . . . . . . 64
Figure 4.14 Stimulation waveform in constant current mode
with charge balance. . . . . . . . . . . . . . . . . . 64
xvi List of Figures

Figure 4.15 Voltage over double-layer capacitor during


stimulation. . . . . . . . . . . . . . . . . . . . . . 65
Figure 4.16 Electrochemical processes and potential waveforms
during charge-balanced stimulation: (a) capacitive
charging only; (b) reversible hydrogen plating;
(c) irreversible hydrogen evolution. . . . . . . . . . 66
Figure 4.17 Issues with charge balance stimulation
for irreversible stimulation. . . . . . . . . . . . . . 67
Figure 5.1 Distinguishing seizure onset from normal activity
in EEG signals necessitates high-order models. . . 73
Figure 5.2 Substantial patient-to-patient variability is seen
in EEG expressions of physiologic/pathophysiologic
processes. . . . . . . . . . . . . . . . . . . . . . . 74
Figure 5.3 Changes expressed in ECG following a myocardial
infraction. . . . . . . . . . . . . . . . . . . . . . . 75
Figure 5.4 Data-driven methods for ECG-based cardiac
arrhythmia detection. . . . . . . . . . . . . . . . . 76
Figure 5.5 Generic structure of a data-driven inference
system. . . . . . . . . . . . . . . . . . . . . . . . 77
Figure 5.6 Seizure detection system with EEG spectral energy
distribution as features. . . . . . . . . . . . . . . . 78
Figure 5.7 Comparison of two different feature sets for
EEG-based seizure detection. . . . . . . . . . . . . 79
Figure 5.8 Performance and energy of cardiac-arrhythmia
detector with respect to complexity of the
data-driven inference model. . . . . . . . . . . . . 83
Figure 5.9 Die photo and measurement results of a
heterogeneous microprocessor, demonstrating
substantial energy savings thanks to hardware
specialization. . . . . . . . . . . . . . . . . . . . . 84
Figure 5.10 Processor architecture maps high-level
inference computations to structured hardware. . . 85
Figure 5.11 Microarchitecture of the specialized Data
Computation Unit (DCU). . . . . . . . . . . . . . 87
Figure 5.12 Ensemble learners construct a strong classifier
with multiple weak classifiers. . . . . . . . . . . . 89
Figure 5.13 10-class classification through all-versus-all voting
on 45 binary classifier outputs. . . . . . . . . . . . 90
List of Figures xvii

Figure 5.14 Prototypical patient-adaptive detector, illustrated


through an ECG-based cardiac-arrhythmia
detector. . . . . . . . . . . . . . . . . . . . . . . . 91
Figure 5.15 Batch data (a) with only marginal-distance metric,
leading to concentrated selections, and (b) with both
marginal-distance and diversity metrics, leading to
coverage over large regions of the feature space. . . 92
Figure 5.16 Classification error rate in image-detection system,
where training is done with high precision, but
derived model is then quantized. . . . . . . . . . . 93
Figure 5.17 Profile of parameter values, tending to result in
severe quantization errors, shown for (a) a linear
classifier in arrhythmia detector, and (b) an
FIR-filter feature-extractor in seizure detector. . . . 94
Figure 5.18 Co-linearity, (a) causing unreasonably large-valued
model parameters, can be addressed through
(b) principal-component regression, leading to a
more robust distribution of model parameters. . . . 95
Figure 5.19 Comparison of quantization error for fixed-point
(6 bit) and floating-point (3-bit mantissa, 3-bit
exponent) representation of FIR filter
coefficients. . . . . . . . . . . . . . . . . . . . . . 96
Figure 5.20 Quantization error minimization approach through
parameter value scaling. . . . . . . . . . . . . . . . 96
Figure 5.21 Compared to AdaBoost, the training algorithm in
Chazal et al. [25] reduces training dataset size
by employing a new training dataset for each
iteration. . . . . . . . . . . . . . . . . . . . . . . . 99
Figure 5.22 Illustration of tradeoff between training complexity
and classification complexity. . . . . . . . . . . . . 100
Figure 6.1 The circuit models for (a) two-coil and (b) three-coil
inductive links. . . . . . . . . . . . . . . . . . . . 108
Figure 6.2 The equivalent circuit models of (a) two-coil and
(b) three-coil inductive links. . . . . . . . . . . . . 109
Figure 6.3 Conceptual rendering of the proposed WPT link
for distributed free-floating implants (FFI)
encompassed by an implanted high-Q resonator.
Inset: simplified equivalent circuit model of the
proposed inductive link. . . . . . . . . . . . . . . . 111
xviii List of Figures

Figure 6.4 (a) The presented 3-coil link configuration for FFI
using a high-Q resonator (doutRes : diameter of
the resonator) around a doutR = 1 mm Rx coil;
(b) Cross-section of the transcutaneous WPT link
including tissue layers, location of various
components, and coil insulation, and (c) HFSS
simulations comparing the Poynting vector
without (left) and with (right) a passive high-Q
resonator. . . . . . . . . . . . . . . . . . . . . . . 113
Figure 6.5 The flowchart of algorithm presented for PTE
optimization of the FFI 3-coil link considering
body layers and SAR limits. . . . . . . . . . . . . 114
Figure 6.6 (a) The HFSS simulation of Rx efficiency (Rx-PSR)
as a function of frequency for different Rx turns,
with the presence of body layers and RL,ORIG of
500 Ω, and (b) the HFSS simulation of Rx efficiency
(Rx-PSR) as a function of frequency and actual load
resistor (RL,ORIG ). . . . . . . . . . . . . . . . . . 115
Figure 6.7 HFSS simulation results of (a) the resonator’s
Q-factor as a function of frequency and width
of the resonator, and b) the Tx’s Q-factor
as a function of frequency and width of the
Tx coil. . . . . . . . . . . . . . . . . . . . . . . . 116
Figure 6.8 HFSS simulation results for coupling coefficient
between L1 and L2 (k12 ) as a function of frequency
and (a) width of L2 , and (b) width of L1 . . . . . . 117
Figure 6.9 (a) HFSS simulation of the PTE for the presented
link as a function of frequency and resonator width.
(b) Relative permittivity of human head issue layers
for skin, fat, bone, dura, and brain vs. frequency. . . 118
Figure 6.10 (a) HFSS simulation of the E-field, comparing the
complete loop and segmented coils, showing
uniformity of the current distribution in the case
of three-coil link, (b) HFSS simulation results for
the SAR simulation presenting the peak of average
SAR values for tissue layers in conventional
two-coil link, three-coil link with loop coils, and
three-coil link with segmented coils, while deliver-
ing identical power of 1.3 mW to the receiver. . . . 119
List of Figures xix

Figure 6.11 The measured quality factor of (a) 1-mm 4-turns


square Rx coil, (b) resonator and (c) Tx coil as a
function of frequency. . . . . . . . . . . . . . . . . 121
Figure 6.12 De-embedding technique was utilized reduce the
effects of EM-field on interconnects from a 1-mm
Rx coil to the VNA for PTE measurements and
characterization of the 3-coil wireless link. . . . . . 122
Figure 6.13 The test setup including a sheep brain and skull,
VNA, and 3-coil link. Insets clockwise: Rx coil
close up view, resonator placement within the
cranial space, S21 measurements with/without
resonator. . . . . . . . . . . . . . . . . . . . . . . 123
Figure 6.14 Transmission coefficient (S21 ) of the prototype
three-coil link before and after implantation
showing the resulting shift in resonance frequency
as a result of the parasitic effects of the surrounding
tissue. . . . . . . . . . . . . . . . . . . . . . . . . 124
Figure 6.15 Theory behind the presented PTE and PDL
measurement. . . . . . . . . . . . . . . . . . . . . 125
Figure 6.16 Measurement setup configuration including NA and
the additional resistors for all the possible source
and load conditions. . . . . . . . . . . . . . . . . . 126
Figure 6.17 (a) The HFSS model and (b) implemented prototype
of a three-coil link to verify the presented PTE and
PDL measurement method. . . . . . . . . . . . . . 127
Figure 6.18 Simulated and measured PTE (a) as a function of
RL including S21 and S21R vs. frequency in the
inset, and (b) as a function of RS . . . . . . . . . . . 128
Figure 6.19 Measurement test setup using SA for measuring
PDL, and the measured PDL as a function
of frequency. . . . . . . . . . . . . . . . . . . . . . 129
Figure 6.20 (a) Measured uniformity of the PTE when sweeping
the Rx coil across the resonator area, indicating the
significant improvement in the PTE in powering
the FFIs on the brain at 16 mm depth using the
proposed three-coil link, compared to the case
without a resonator, (b) PTE measurement while
sweeping the Rx coil depth along z-axis at the center
of the Tx and resonator coils. . . . . . . . . . . . . 130
xx List of Figures

Figure 6.21 HFSS simulation and measurement results of PTE


and PDL as a function of Rx rotation, while the Rx
is located at the center of the Tx and resonator coils
at a depth of 16 mm. Error bars show maximum
and minimum of six consecutive measurements. . . 132
Figure 6.22 (a) The number of Rx devices as a function of
center-to-center spacing between devices that are
uniformly distributed inside the area enclosed
by the resonator. Red dots: Rx devices, Yellow:
The area surrounding each Rx device cleared
of other devices. (b) HFSS model for seven
Rx devices within the resonator. . . . . . . . . . . 133
Figure 6.23 (a) Simulation and measurement of coupling
coefficient between two adjacent Rx coils vs.
their center-to-center spacing, and the impact on
PTE when they are located at the center of the
resonator. Rx coil dimensions are given in
Table 6.1. The effects of surrounding tissue
is also considered at 60 MHz. (b) PDL of an Rx
coil, when it is located at the center of the resonator,
and the overall PTE of all Rx coils as a function of
the number of FFIs within the resonator and the Rx
coils’ center-to-center spacing. . . . . . . . . . . . 134
Figure 7.1 Primary factors influencing development of leading
edge neural interfaces. . . . . . . . . . . . . . . . 143
Figure 7.2 System power density with respect to sensing area
for state of the art recording systems. . . . . . . . . 145
Figure 7.3 (a) Diagrammatic representation of different spike
shapes recorded around a neuron depending on
electrode Paralikar neuron distance and relative
orientation. (b) The recorded neural signal is a
composite of high SNR signals from nearby neurons
(1), low-SNR multi-unit activity from slightly more
distant neurons (2) and a local field potential (LFP)
resulting from the sum activity of many distant
neurons (3). (c) The recorded neural signal
List of Figures xxi

broken down into the 3 components


described in (b). . . . . . . . . . . . . . . . . . . . 150
Figure 7.4 (a) The raw recorded signal, (b) after filtering,
(c) the spikes all peak aligned (d) the spikes after
clustering. . . . . . . . . . . . . . . . . . . . . . . 150
Figure 7.5 The NGNI v1 headstage. . . . . . . . . . . . . . . 151
Figure 7.6 The NGNI v1 hardware showing the headstage
with its connections either to a microcontroller or
via a USB link to a computer. . . . . . . . . . . . . 152
Figure 7.7 The two-phase spike sorting process. In the first
phase the neural signal recording is processed
offline to generate templates, in the second phase
the neural data is passed to an FPGA
and the templates are used to perform real time
spike sorting. . . . . . . . . . . . . . . . . . . . . 153
Figure 7.8 Diagrammatic representations of output stages for
neural stimulators. . . . . . . . . . . . . . . . . . . 157
Figure 7.9 Trend in channel number per IC stimulator.
Green and red devices not included for trend line
calculation – green dot indicates an ineffective
device, red is an in vitro device included to indicate
limits of currently achievable integration. . . . . . 159
Figure 7.10 (a) A biphasic current pulse. (b) Dotted lines
indicate typical stimulator power consumption,
solid line indicates power necessary to drive the
current through the electrodes, the grey area
indicates power waste. (c) An ideal current
controlled stimulator would continuously
modulate the voltage across the current
controlling transistors to minimise power waste
(although some minimal headroom is likely
necessary to keep the transistors in saturation).
(d) Typical solutions in the literature have made
use of discrete positive voltage levels leading
to power consumption. . . . . . . . . . . . . . . . 160
xxii List of Figures

Figure 7.11 Charge balancing techniques. (a) A standard


biphasic current pulse generator – mismatch
between the reference currents and mirrors leads
to charge imbalance. (b) An H-bridge enables the
use of a single reference current – greatly enhancing
charge balance. (c) A current-copying calibration
approach, during calibration just switches S1 are
closed, copying the PMOS sourced current, then
during stimulation switches S1 are open and
switches S2 and S3 are utilized. . . . . . . . . . . . 161
List of Tables

Table 5.1 Power comparison for EEG-based seizure-detection


systems with and without local processing . . . . . . 72
Table 5.2 Tradeoffs for seizure detection with support vector
machine . . . . . . . . . . . . . . . . . . . . . . . . 88
Table 6.1 Specifications of the inductive link at 60 MHz
with/without the Surrounding Tissue . . . . . . . . . 120
Table 6.2 Benchmarking wireless power transmission to
small implants . . . . . . . . . . . . . . . . . . . . 132
Table 7.1 NGNI v1 performance measures . . . . . . . . . . . 153

xxiii
List of Abbreviations

ADC Analogue to Digital Converter


ASIC Application Specific Integrated Circuit
AVA All-vs.-All
BMI Brain Machine Interface
CDS Correlated Double Sampling
CMOS Complementary Metal-Oxide Semiconductor
CMRR Common-Mode Rejection Ratio
CNS Central Nervous System
DAC Digital to Analogue Converter
DBS Deep Brain Stimulation
ECG Electrocardiography
ECOG Electrocorticography
EEG Electroencephalography
EKG Electrocardiography
FDA Food and Drug Administration
FSMs Finite-state machines
IC Integrated Circuit
IR Infra-Red
LED Light Emitting Diode
LFP Local Field Potential
MEA Micro Electrode Array
NEF Noise Efficiency Factor
NGNI Next Generation Neural Interfaces
OOK On-Off Keying
PCB Printed Circuit Board
PEDOT Polyethylenedioxythiophene
PNS Peripheral Nervous System
SAR Specific Absorption Rate
SD Secure Digital

xxv
xxvi List of Abbreviations

SNR Signal to Noise Ratio


SOC System on Chip
SRAM Static Random Access Memory
SVM Support Vector Machine
VCO Voltage Controlled Oscillator
1
Introduction

Peng Cong
Verily Life Science, US

1.1 Implantable Devices


Implantable medical devices, such as pacemakers, spinal cord stimulators,
and cochlear implants, play increasingly important role in saving and improv-
ing the quality of people’s lives. Circuit design is one of the important
enabling technologies for those devices. It is helpful for the circuit designers
to understand the uniqueness of implantable medical devices in order to make
big contribution.
There are many commercially available implantable devices. A pace-
maker, as shown in Figure 1.1(a), is a medical device to regulate heartbeat. It
monitors electrocardiography (ECG or EKG) signal using sensing electronics
and delivers an electrical stimulation through electrodes using stimulation
circuit when missing heartbeat was detected. Due to accident or neurological
diseases, some patients experience chronic pain with no explainable reasons
and no conservative therapy works. In some cases, they can benefit from
spinal cord stimulators as shown in Figure 1.1(b). A spinal cord stimulator
is a device [1], which delivers pulsed electrical stimulation to the spinal cord
to minimize chronic pain using similar stimulation circuit as a pacemaker,
although the amplitude and stimulation frequency are much higher. Cochlear
implants, as shown in Figure 1.1(c), may help provide hearing through
stimulation of hearing nerve within cochlear in patients who are deaf because
of damage to sensory hair cells.
In spite of different device functionalities, the electronic system of those
devices has much in common. As shown in Figure 1.2, the system may
include energy source with power management, data telemetry, sensing block,
signal processing for algorithm, system control, and therapeutic stimulation
sub-block. For disease intervention, the therapeutic stimulation is the most

1
2 Introduction

Figure 1.1 Example of implantable medical devices: (a) a pacemaker, (b) spinal cord
stimulator, and (c) cochlear implant.

Figure 1.2 A diagram of a typical implantable medical device.

important sub-blocks, which interacts with nervous system through therapeu-


tic electrical stimulation. Sensing is another key block, with signal processing
unit, to determine disease state or abnormal nervous system behavior in
order to delivery effective therapy, e.g., detecting a missing heartbeat for
a pacemaker. Bi-directional neural interfacing with both sensing and stim-
ulation capability is the key challenge and the uniqueness associated with
implantable medical devices. As an isolated device implanted in human sub-
jects, energy source is one critical sub-block in the system. The energy source
could be a primary battery, a rechargeable battery, or continuous wireless
1.2 Neural Interface 3

powering coupled through electromagnetic field. The choice of the energy


source highly depends on application desirability and technical feasibility.
For example, primary battery is typically selected for pacemaker due to
its low power consumption (stimulation frequency is typically smaller than
5 Hz and stimulation voltage is typically less than a couple volts) and highly
reliability requirement. Rechargeable battery is typically selected for spinal
cord stimulator due to it high-power consumption (stimulation frequency
could be more than 100 Hz and stimulation voltage could be larger than 10 V).

1.2 Neural Interface


Understanding the challenges of the neural interface-associated circuit design
considerations is critical for successful medical devices. Human nerve system
was typically divided into central nervous system and peripheral nervous
system. The central nervous includes brain and spinal cord. The brain is the
center of the nervous system with more than 86 billion neurons (compared to
5.5B transistors in the latest commercially available CPU as of 2015). Spinal
cord connects brain to peripheral nervous system, which then controls limbs
and organs, and sends sensational signal back to brain or spinal cord. The
nervous system is mainly formed by neurons. One basic function of neuron
is to transfer information through both electrical and chemical activities.
Figure 1.3 depicts two neurons communicating with each other. When an
electrical signal, typically called action potential or spike, is generated, it
flows down the axon of the cell on the left, with the help of ions such as
sodium, potassium, and calcium to transmit the electrical message. When
it reaches its terminal gap with the next neuron, a neurotransmitter, called
synapse, will be released and passed along to the next cell. The neurotransmit-
ter will help generate another spike in the next neuron. Using drug to regulate
neurotransmitters is one way to do intervention; however, for an implantable
device, it typically uses stimulation through electrodes [2].
If there is an ideal electrode, which can be very small, close to neuron
membrane even inside the neuron, and have a zero impedance between
electrode and the medium around, as shown in Figure 1.3, a spike can be
measured with resting potential about –70 mV, active potential level roughly
+30 to 40 mV and the pulse width ∼1 ms. Action potentials are generated
by voltage-gated ion channels in a cell’s membrane. These channels are
shut when the membrane potential is near the resting potential of the cell,
but they rapidly begin to open if the membrane potential increases to a
4 Introduction

Figure 1.3 Examples of neuron communication and a bi-directional neural interface.

precisely defined threshold value. When the channels open, they allow an
inward flow of sodium ions, which change the electrochemical gradient, and
in turn produce a further rise in the membrane potential. The process proceeds
explosively until all of the available ion channels are open, resulting in a
large upswing in the membrane potential. The rapid influx of sodium ions
causes the polarity of the plasma membrane to reverse, and the ion channels
then rapidly inactivate. As the sodium channels close, sodium ions can no
longer enter the neuron, and then they are actively transported back out of
the plasma membrane. Potassium channels are then activated returning the
electrochemical gradient to the resting state. These action potentials can be
engaged by using therapeutic stimulation in neuromodulation.
Fundamentally, for neural interface, electrodes are used to listen to neural
communication and to apply stimulation to control neural communication.
Most of the readers, who are interested in this book, must know how to
use an oscilloscope to measure a voltage on a PCB. Oscilloscope probes are
typically used to interface with the test points on the PCB, where it makes a
metal-to-metal ohm contact; therefore, what displayed on the scope matches
the voltage at the test-point. It would be a different case for electrode–tissue
interface, which would be discussed in details in Chapter 2.
1.3 Circuit Design Considerations for Neural Sensing 5

1.3 Circuit Design Considerations for Neural Sensing


Typically, there are two different types of signal, which neural interface
circuit works with, spike and local field potential (LFP). Spike is the funda-
mental signal within neural system. Some microelectrodes can pickup neural
spike; however, many electrodes are not ideal in reality. They are not small
enough to pick up signal from only one neuron, not close enough to pick full
scale of the signal, and have finite impedance resulting in signal distortion.
Therefore, in an implantable system, this signal is rarely seen. Instead, the
sensed signal it typically much small in amplitude with slightly different
characteristics. The signal bandwidth is typically in kHz range. The other
typical signal is LFP, which is an electrophysiological signal generated by
the summed electrical signal from 100 to 1000’s of nearby neurons within
a small volume of nervous tissue. The signal bandwidth is much reduced
up to a few hundreds of Hertz due to the averaging effect and presentation
of the signal is oscillation in nature instead of spike. However, this signal
may contain more high-level biological meaning compared to spikes. Spikes
can typically be measured by using micro wire electrode or microelectrodes
array and LFP can be accessed by more “traditional” electrodes as shown in
Figure 1.3, and the choice of spike or LPF strongly depends on applications
and this can have a big impact on the architectural choices.
Electrodes are used to detect neural signals; however, electrodes are
not ideal. When an electrode, which is usually metal, is placed inside a
physiological medium (e.g., tissue or body fluid, which contain electrolytes),
electrochemical activity will occur, and an interface will form between the
two phases. In the metal electrode phase, charges are carried by electrons;
in tissue (or, more generally, in the electrolyte phase), charges are carried by
ions. At zero biasing, a layer of oriented water molecules with two layers of
charges can be found in the interface. This behaves as a capacitor with a small
signal and is called a double-layer capacitor (CD ), as illustrated in Figure 1.4.
There is also a resistive component associated with the electrolyte, RS .
The values of both CD and RS are a function of electrode geometry, mate-
rial, and surface profile. As the size decreases from traditional electrodes
to microelectrodes, the value and mismatch tend to increase. As shown in
Figure 1.5, with the input impedance of the amplifier RIn , the two electrodes
can be replaced by the electrode model RS and CD . At the same time, the
neural circuit can be replaced by an ideal signal source. A few circuit design
considerations can be determined by the circuit representation. The first is the
mismatch between electrodes (RS1 , CD1 and RS2 , CD2 ), and the second is the
6 Introduction

Figure 1.4 An electrode interface model for sensing.

Figure 1.5 Circuit design considerations for a sensing neural interface.

high electrode impedance, especially when electrodes move from traditional


ones to microelectrodes. A large mismatch results in high signal susceptibility
to common-mode interference, and large electrode impedance results in large
signal attenuation and, as a consequence, a more pronounced mismatch.
1.4 Circuit Design Considerations for Stimulation 7

The electrode performances, for both matching and impedance values,


are typically beyond the control of circuit designers. Generally, improving
input impedance and common-mode rejection ratio (CMRR) helps for both
cases. Another important consideration for electrodes is polarization, which
results in a DC offset between two electrodes. This polarization could be
caused by material and tissue mismatch or stimulation effect. The DC offset
can typically be managed by using an AC-coupled amplifier design.
Based on the nature of neural signal characteristics and the electrode
model, a few circuit design considerations can be extrapolated:
• DC offset: an AC-coupled or AC-coupled equivalent (e.g., with DC
servo to remove DC offset) amplifier
• Electrode impedance mismatch: large circuit input impedance and high
CMRR
• Large electrode impedance, especially for microelectrodes: large circuit
input impedance.
In addition, the following factors should be considered. For spike detection,
there is generally a need for low-power, low-noise (i.e., thermal noise), small
size, and high-input impedance. As the signal band is typically 300 Hz to
10 kHz, 1/f noise may not be very critical. However, high-input impedance
and small size are critical, as these are often required when working with
microelectrodes and large electrode counts. For LFP, requirements typically
include low power, low noise (i.e., thermal noise and 1/f noise), and high-
input impedance. As the bandwidth of LPF is from sub-Hz to 300 Hz,
minimizing 1/f noise is important. The impedance requirement is not as
rigid as for the spike amplifier, but a reasonable value will help system
performance.
Circuit design considerations and circuit architecture for different sensing
amplifiers would be presented in Chapter 3.

1.4 Circuit Design Considerations for Stimulation


Typically, two electrodes are needed for stimulation, where one as reference
and the other as stimulation electrode. The goal of stimulation is to delivery
charge into tissue to influence the potential outside of neuron membrane
so that to initiate spikes/action potential. A typical stimulation waveform is
shown in Figure 1.6 with the x-axis time and the y-axis current. The positive
and negative indicate the direction of the charge flow, in or out of tissue.
8 Introduction

Figure 1.6 A typical stimulation waveform.

The waveform typically includes stimulation phase followed by recovery


phase. It could be an active recharge by using the same or scaled current with
reversed direction, or passive recovery by shorting the electrodes as shown in
the dash-line, which follows RC discharge curve.
In the case of sensing, the electrode and tissue interface can be consid-
ered as a pure capacitor, where there is no charge going across the double
layer structure. However, in the case of stimulation (typical stimulation
voltage could vary from 1 to 10 V), there could be real charge transfer with
electrochemical reaction. The leakage through the double layer capacitor is
typically called Faradeic impedance. The model of the electrode interface
for stimulation is shown in Figure 1.7. Note that the value of the Faradeic
impedance may vary with stimulation amplitude and other factors.
The most important consideration for stimulation circuit design is safety,
which largely depends on whether the electoral-chemical reactions are
reversible or irreversible. If the products do not diffuse away during the
stimulation phase, it could be recovered during discharge phase. In this case,
although electron transfer occurs, in terms of the electrical model, it is better
modeled as a capacitor. So it was named pseudocapacitance, one example is
Pt. During the stimulation phase, the reaction product of electron, hydrogen
1.4 Circuit Design Considerations for Stimulation 9

Figure 1.7 An electrode interface model for stimulation.

ions, and platinum is Pt–hydrogen bond, and these products only stay at
the electrode surface (called hydrogen plating) thanks to the characteristic
of Pt. If the total charge injected into tissue during stimulation phase is
totally recovered during recharge phase (charge-balanced stimulation), all
the production will be recovered too. This is the reason why Pt is used
widely in implantable devices. In short, for a reversible case, charge-balanced
stimulation is critical to ensure the safety.
If the products diffuse away from the electrode, the charge cannot be
recovered upon reversing the direction of current. One example is water
reduction, where hydrogen will be formed. Therefore, control voltage cross
double layer capacitor within a limitation is essential in this case with the
control of total DC current under a limit.
Beside the electrode safety considerations, neuron safety is an even bigger
topic and so far people still do not have a complete understanding to deter-
mine neuron safety under stimulation. A wide accepted standard is called
Shannon criteria illustrated in Figure 1.8, proposed by professor Shannon
in 1992. Note the criteria is pure experimental by plotting previous results
against charge per phase (x-axis) and charge densities (the charge divided by
the area of electrode) per phase (y-axis). The line in Figure 1.8 is defined by
the equation below:
10 Introduction

Figure 1.8 The Shannon criteria for stimulation safety.

k = log(Q) + log(Q/A),
where, Q is charge per phase and A is the area of electrode. If K is greater
than 1.5–1.8, it would not be safe. Shannon criteria have been evolved in
recent years for microelectrodes.
Below is the summary of circuit considerations for stimulation:
1. Work with electrode designer and understand electrode model (careful!
Low signal vs. large signal model).
2. Small DC leakage current from electrode to tissue (<1 µA or 10 nA).
3. Charge balanced stimulation for reversible reaction.
4. Monitor and control voltage across double layer capacitor within “good”
reaction window if irreversible reaction might happen.
5. Limit maximum charge and charge density for neuron safety.
6. Constant current versus constant voltage stimulation.
7. Power efficient design (Stimulation therapy circuit typically is the major
power consumption source in implantable devices).
8. “Good” performance (frequency, stimulation amplitude, and stimulation
pulse width accuracy).
9. Pay more attention to micro-electrodes!
The detailed safety consideration and circuit design of stimulation would be
presented in Chapter 2 and Chapter 4, respectively.
1.5 Closed-Loop System 11

1.5 Closed-Loop System


As shown in Figure 1.2, the other important module of the system architecture
is system control unit processing unit, which often runs signal process-
ing and algorithm control. One example of how the module involved in
the system can be found in cardiac devices, e.g., pacemakers. Pacemakers
monitor ECG or EKG signal using sensing electronics. The EKG signal
is analyzed by on-device algorithm unit. An electrical stimulation through
electrodes is then applied if a missing heartbeat was detected. As captured in
Figure 1.9, the pathway to classification generally involves a series of steps:
from raw biosignal recording to initial feature extraction, e.g., identifying the
peak of an EKG; to detection of some feature of relevance, e.g., the peak
of the R-wave and the rhythm associated with ventricular contraction in the
heart; finally to the classification and to the estimation of the potential state
of the patient, e.g., atrial fibrillation if the spacing between R-wave feature
peaks becomes erratic and/or extraordinarily fast. The process from signal
extraction to classification requires knowledge of the desired information
that is relevant for the clinician and/or an algorithm, as well as a method
of signal processing for deriving that classification. Identifying the desired
classification is often collaborated with biomedical engineers; but the signal
processing is domain knowledge relevant to good circuit design.
The classification of EKG can be achieved by using time domain analysis
with “small” power consumption and “low” computation capability. One

Figure 1.9 Series of steps of classification.


12 Introduction

example of system architecture can be found in Figure 1.10 (a). However,


when researches move into central nerve system, e.g., brain, the analysis gets
more involved, which is sometimes involved in frequency domain analysis as
illustrated in Figure 1.10(b). In addition, instead of single channel analysis as
in pacemaker, brain interface device need multiple channels up to hundreds
to thousands, which post-significant challenge to embedded system design
with significant power constrain within implantable devices. Chapter 5 will
present latest research in this area.

1.6 RF Powering and Data Telemetry for Implantable


Devices
One of the most difficult challenges for implanted device technologies to
overcome is in providing implants with sufficient power. There has been
great progress in understanding how to design miniature low-power circuits
for biological applications, which enables devices, such as pacemakers, via
primary batteries with reasonable size. However, batteries are difficult to
miniaturize and remain the size-limiting component of many implants. In
addition, the lifetime of batteries limits the useful life of potential implants.
Battery replacement for implantable devices often requires an additional
surgery and can cause many complications. Alternatively, rechargeable bat-
teries allow for longer useful lifetimes but need an additional means of
delivering power to recharge, such as RF approaches, which suffer from
low-efficiency power transfer and require relatively large, aligned antennas.
Other non-RF methods to wirelessly power implanted devices have been
proposed but are only in very early stages of development and will require
many advances before they are practical. WiTricity, which uses magnetic
resonance coupling, allows for highly efficient energy transfer but requires
large coils [3]. Ultrasound energy can be used to deliver power to implanted
devices, but the efficiency of power delivery is very small [4]. Energy scav-
enging [5] and optical energy [6] have also started to be investigated but
currently produce too little energy to reliably power implantable devices.
Currently, wireless powering through inductive link is still the main
method for implant with rechargeable battery. A typical model of such a
system can be found in Figure 1.11(a). L1 is the coil of external charger and
L2 is the coil of implant. The mutual inductance of the two coils services
the link of powering. An analysis of this configuration, where two circuits
are coupled together through a mutual inductance, is inconvenient to perform
Figure 1.10 Closed-loop architecture for (a) pacemakers and (b) brain interface devices.
1.6 RF Powering and Data Telemetry for Implantable Devices
13
Exploring the Variety of Random
Documents with Different Content
Aus den Privataufzeichnungen des
Sekundaners Heppenheimer.
Erstes Bruchstück.
… Am 24. Februar 18**
omit soll ich zunächst anfangen? – Es klingt eigentümlich, aber
es ist nichtsdestoweniger wahr: jeder Anfang hat für mich
etwas Peinliches. Bei meinen deutschen Aufsätzen hocke ich
oft stundenlang und kaue an der Feder, ohne zu wissen, wie ich dem
Dinge beikommen soll. Gewöhnlich helfe ich mir dann dadurch, daß
ich mich nach einem geeigneten Zitat umsehe und dasselbe als
Motto oben rechts in die Ecke schreibe. Hieran läßt sich dann
gewöhnlich in ungezwungener Weise anknüpfen, indem man etwa
fortfährt wie nachstehend:
»Der große Dichter, dem wir diese Worte entlehnen, hat ohne
Zweifel dabei die hochwichtige Frage im Auge gehabt, deren
Behandlung mir heute von Amts wegen obliegt.«
Es ist mir bis jetzt noch stets gelungen, den erforderlichen
Nachweis zu liefern, zumal wenn das Zitat von Schiller war, dessen
Aussprüche das Angenehme haben, daß sie für alle Verhältnisse des
Lebens gleich brauchbar sind. Ich will also dieser meiner
angestammten Gewohnheit auch heute nicht untreu werden und
mein Tagebuch mit den herrlichen Worten aus Schillers Glocke
einleiten:
Von der Stirne heiß
Rinnen muß der Schweiß.

Dies ist nämlich die Ansicht meines Mitschülers Leopold Hutzler,


der in der Nähe des Fensters sitzt und durchaus nicht leiden kann,
wenn man auch nur ein kleines Quadratchen öffnet, um frische Luft
hereinzulassen. Wie zu Eingang notiert, ist es noch Februar, und die
Witterung läßt manches zu wünschen übrig. Wir besitzen nun einen
Lehrer, der zum Schlagfluß neigt und vor Kongestionen fast
umkommt, wenn alles geschlossen ist. Kaum tritt er ins Zimmer, so
ruft er mit seiner dröhnenden Baßstimme: »Schwarz, machen Sie
mal 's Fenster auf!«, und Schwarz tut, wie geheißen. Bis zum 20.
Februar ging die Sache auch ihren stillen, friedlichen Gang. An
diesem Tage aber gelangten die exercitia pro loco zur Verteilung,
und Hutzler, der ein Feind aller Zugluft ist, kam in die Nähe des
Fensters zu sitzen …
Doktor Perner ließ wie gewöhnlich oben die Klappe öffnen und
wollte eben seinen Vortrag beginnen, als der stramme Hutzler sich
von seinem Platze erhob und mit aufgestelltem Rockkragen und
frostschauernder Stimme in die geflügelten Worte ausbrach:
»Herr Doktor, es zieht so!«
Doktor Perner wird nun jedesmal nervös, wenn jemand behauptet,
es ziehe. Er sagt, das sei Einbildung, und wenn die Bewegung der
atmosphärischen Luft die Gesundheit schädige, so könne kein
Mensch mehr über die Straße gehen, ohne eine
Rippenfellentzündung oder die Diphtheritis zu bekommen.
»So, es zieht Ihnen?« erwiderte er in wegwerfendem Tone. »Wie
alt sind Sie eigentlich?«
»Im nächsten Januar werde ich siebzehn!« entgegnete Hutzler mit
Würde.
»Und demungeachtet zieht es Ihnen? – Nun, dann ist es die
höchste Zeit, daß Sie endlich einmal dieses Vorurteil ablegen. Setzen
Sie sich, das Fenster bleibt auf!«
Hutzler zog den Rockkragen noch höher, setzte sich nicht und
sagte mit männlicher Festigkeit:
»Herr Doktor, der Arzt hat es mir dringend verboten, mich der
Zugluft auch nur auf wenige Minuten auszusetzen. Ich bitte um die
Erlaubnis, meinen Platz wechseln zu dürfen!«
»Meinetwegen«, sagte der Doktor Perner mit einem
geringschätzigen Achselzucken. – »Hanau, wechseln Sie einmal mit
dem Hutzler den Platz!«
»Herr Doktor,« sagte Hanau, »ich bin erst gestern
wiedergekommen und neige sehr zum Katarrh. Es wäre vielleicht
doch besser, wenn wir das Fenster zumachten.«
»Seien Sie still! Gildemeister, setzen Sie sich dort an das Fenster!«
Gildemeister hustete dumpf, und es klang wie ein Bierfaß.
»Wenn Sie erlauben,« sagte er mit heiserer Stimme, »so möchte
auch ich hier auf meinem Platze bleiben. Ich habe jetzt schon acht
Senfpflaster verbraucht, um meinen Luftröhrenkatarrh los zu
werden, und bin immer noch nicht damit zustande gekommen.«
»Gut,« sagte Doktor Perner jetzt stirnrunzelnd, »so bleiben Sie,
wo Sie sind. Wenn es dem Hutzler zieht, so mag er seinen Paletot
umhängen.«
»Wenn ich meinen Paletot umhänge, so wird mir's zu warm, und
dann erkälte ich mich erst recht.«
»Meinetwegen erkälten Sie sich sechsmal.«
»Nun, Sie werden ja sehen, was Sie anrichten, Herr Doktor«,
sagte Hutzler gekränkt. – »Ich merke jetzt schon einen
eigentümlichen Kitzel im Halse, und so fängt es bei mir jedesmal
an.«
Mit diesen Worten begann er zu hüsteln.
Ich muß nun an dieser Stelle bemerken, daß Hutzler einer unserer
gesundesten Schüler ist. Wie oft hat der Direktor Samuel Heinzerling
ihm die vernichtenden Worte zugerufen: »Schwächläch! Sä,
schwächläch? Non, hären Sä änmal, Hutzler, äch wollte, jäder
Mänsch onter der Sonne wäre so schwächläch wä Sä! Faul sänd Sä,
aber nächt schwächläch!« Es handelt sich bei der ganzen Opposition
Hutzlers lediglich um das, was man eine parlamentarische
Unterbrechung nennt. Er will in die Monotonie der Lehrstunden eine
gewisse Frische und Abwechslung bringen. Aus diesem
Gesichtspunkte wird uns auch die Weigerung der beiden erwähnten
Mitschüler begreiflich.
Der Lehrer begann nun den Unterricht, und Hutzler, das Haupt
trotzig in die Hand gestützt, bereitete sich zur Fortsetzung seiner
planvoll erwogenen Störung vor.
Als es ein Viertel schlug, hustete er dreimal tief auf und stöhnte
dann, als ob sich ihm die Luftröhre krampfhaft zusammenschnüre.
Fünf Minuten später hatte sein Husten einen so dröhnenden
Charakter angenommen, daß es Herrn Doktor Perner unmöglich war,
den Unterricht fortzusetzen.
Er hielt einen Augenblick inne.
»Sind Sie nun bald fertig?« rief er, die Augen rollend, während er
das Buch heftig wider die Platte des Katheders stieß.
Hutzler hustete noch lauter, und so natürlich, daß ich noch heute
nicht begreife, wie er diese gewaltigen Erschütterungen seines
Kehlkopfes zuwege bringen konnte, ohne ernstlich Schaden zu
nehmen.
»Hutzler!« schrie Doktor Perner außer sich.
Jetzt trat in dem trefflich erkünstelten Anfalle eine Pause ein.
Hutzler erhob sich.
»Herr Doktor, darf ich nun vielleicht das Fenster da zumachen?«
»Das Fenster bleibt auf! Sie sollten sich schämen, auf so
pöbelhafte Weise etwas erzwingen zu wollen, was ich Ihnen
grundsätzlich verweigern muß.«
Kaum hatten diese Worte Hutzlers Trommelfell erreicht, als er
sofort wieder zu husten begann, und zwar so krachend und klirrend,
daß ich jeden Augenblick meinte, die Brust müsse ihm zerspringen.
»Ich lasse Sie sofort nach dem Karzer führen!« rief Doktor Perner,
außer sich vor Zorn. »Wenn Sie so empfindlich sind gegen jede
erbärmliche Kleinigkeit, so wickeln Sie sich in Watte! Ich meinesteils
dulde nicht, daß man in meinen Lehrstunden solche Komödien
aufführt.«
»Komödien?« hustete Hutzler. »Wenn ich erkältet bin, werde ich
doch wohl noch husten dürfen? … Hätten Sie beizeiten das Fenster
geschlossen …«
»Sie sind einer der frechsten Gesellen, die mir noch jemals
vorgekommen. Gehen Sie nach Hause und ziehen Sie sich wärmer
an! Ich bin es müde, mich fortwährend mit Ihnen herumzuzanken!«
»Recht gern«, hustete Hutzler. »Hätte ich gewußt, daß es hier so
ziehen würde, so wäre ich von Anfang an in einem wärmeren
Kostüm erschienen.«
Hutzler wohnt nur drei Schritte vom Gymnasium entfernt. Er ging,
und Doktor Perner setzte seinen Vortrag fort. Es dauerte ungefähr
zehn Minuten. Dann erschien der treffliche Leopold wieder in der
Tür, und mit einem Male herrschte in Sekunda ein Leben, dessen
reizende, überschwängliche Ausgelassenheit sich nicht in Worte
kleiden läßt. Zuerst erscholl ein dreisalviges Gelächter; dann ein
dumpfes Geheul, wie es die Rothäute bei ihren Angriffen auf die
Weißen auszustoßen pflegen; dann ein Klatschen, Pfeifen, Scharren,
Trappeln und Rütteln, daß mir selbst, der ich doch an das
Schlimmste gewöhnt bin, fast Hören und Sehen verging. Hanau und
ich hoben in der allgemeinen Verwirrung unseren Tisch ungefähr
drei Zoll hoch über den Boden und ließen ihn dann mit aller Wucht
aufdonnern, so daß der Staub wie Opferrauch nach der Decke stieg.
Es war dies nur eine verdiente Huldigung an die Adresse unseres
liebenswürdigen Kameraden Hutzler.
Ahnst Du, o Genius meines Tagebuches, wie Hutzler im
Schulzimmer erschien? Hinten auf den Rücken und vorn vor den
Bauch hatte er sich vermittelst roter Schnüre zwei Federkissen
gebunden. An den Füßen trug er die großen Reisepelzstiefel seines
Vaters; zwei Müffe, die seinen beiden Schwestern angehörten,
dienten ihm als Pulswärmer, und um den Hals trug er in unzähligen
Windungen einen halbzölligen Hanfstrick, wie ihn die Packer beim
Aufwinden der Warenballen benutzen.
Doktor Perner stand wie versteinert, während Hutzler sich ganz
gelassen anschickte, seinen Platz einzunehmen.
»Halt!« schrie der Professor. »Keinen Schritt weiter! Denken Sie
etwa, Sie befinden sich hier in einer Bierkneipe?«
»Gewiß nicht, Herr Doktor!« entgegnete Hutzler ehrerbietig. »So
viel ich weiß, befinde ich mich in Sekunda.«
»Schweigen Sie! Ihr Zynismus übersteigt alle Begriffe. Sofort
entledigen Sie sich dieses Unrats und verfügen sich nach dem
Karzer!«
»Welchen Unrates, Herr Professor?«
Doktor Perner war außer sich, er trat auf den Schüler zu und faßte
ihn an dem Strick, der um seinen Hals lag.
»Hier, dieses nichtswürdigen Tandes!« schrie er, daß uns allen die
Ohren gellten.
»Ach so,« sagte Hutzler, »es ist ja wahr, da wollte ich Sie noch
ganz ergebenst um Entschuldigung bitten. Meine beiden wollenen
Halstücher sind in der Wäsche, und Mutter wollte mir das ihrige
nicht hergeben. Da meinte der Vater, so ein Strick sei auch nicht zu
verachten, und es käme ja nicht darauf an, wie es aussehe, wenn es
nur warm hielte. Aber wenn Sie meinen, es wäre unziemlich, so bin
ich gern bereit, ihn wieder abzulegen.«
Mit diesen Worten begann er, das halbzöllige Seil von seinem
Halse loszuwickeln.
In immer größeren Kreisen fegte der hanfgeflochtene Radius um
Hutzlers Kopf, und jetzt fehlte nicht viel, und die Spitze hätte den
Professor ernstlich in seiner Integrität verletzt. Ich gebrauche hier
Integrität als Euphemismus, da es mir nicht wohl ansteht, diejenigen
Teile des Doktor Perner namhaft zu machen, die von dem wuchtigen
Strick Hutzlers zunächst bedroht wurden.
»Mensch!« rief Doktor Perner wutschnaubend, indem er das Ende
des Strickes ergriff und daran zerrte. »Das sollen Sie mir büßen!«
Hutzler bemühte sich, die Augen zu verdrehen und zwischen den
Lippen die Zunge sichtbar werden zu lassen.
»Herr Doktor!« stöhnte er mit verlöschender Stimme, nach rechts
und links mit den Armen in die Luft greifend. »Ich ersticke! Ich
ersticke!«
Doktor Perner ließ los. Hutzler reckte seinen Hals und begann, ihn
vollends auszuwickeln.
»Nun, was zögern Sie noch?« rief der Lehrer, indem er mit der
rechten Hand nach der Tür deutete. »Losgeschnallt! sage ich, und
dann hinauf! Vor nächstem Montag kommen Sie mir nicht wieder
herunter!«
»Also weil ich mich Ihrer ausdrücklichen Anordnung entsprechend
etwas wärmer gekleidet habe, belegen Sie mich mit Karzerstrafe?«
sagte Hutzler, gleich darauf in einen erneuten Hustenanfall
ausbrechend. »Erlauben Sie, Herr Doktor, ist der Karzer geheizt?«
»Der Pedell wird das Nötige besorgen«, entgegnete Doktor Perner.
»Schwarz, bestellen Sie einmal, daß Nummer fünf geheizt wird. Und
jetzt bringen Sie sich unverzüglich in eine anständige Verfassung.
Diese schamlosen Wülste hier dulde ich nicht!«
»Aber, Herr Doktor, es sind ja zwei Kopfkissen! Meine Mutter hält
so viel auf ihr Weißzeug, die würde sich schön wundern, wenn sie
erführe …«
»Kein Wort mehr, oder ich vergesse mich!«
Hutzler schwieg und begann, seine Kissen loszuschnallen. Mit
einemmal stieß er einen Schrei des Entsetzens aus. Die Naht war
geplatzt, und eine Flut von Federn ergoß sich in das Schulzimmer.
Wir andern eilten sofort hinzu, um die kostbaren Daunen
aufzulesen, und bald wirbelte es rings wie Schneeflocken.
Doktor Perner machte vergebliche Anstrengungen, die Ordnung
wiederherzustellen. Gelles Geheul derjenigen, die in dem Tumult
umgestoßen wurden und auf den Boden zu liegen kamen, mischte
sich unter die Wehklagen Hutzlers, der sich nicht genug tun konnte
in elegischen Ausrufen und Seufzern.
»Ach, was wird meine Mutter sagen! Das kommt davon, daß ich
das so aufschnallen mußte. Hätte ich die Kissen ruhig anbehalten,
dann wären sie nicht entzwei gegangen. Ach, und jetzt zieht es
wieder, und die Federn kommen mir in den Hals. Herr Doktor,
erlauben Sie mir, nach Hause zu gehen. Ich fühle mich sehr, sehr
unwohl!«
»Gehen Sie!« rief Doktor Perner in ohnmächtigem Zorn.
Hutzler ließ sich die Sache nicht zweimal sagen. Hastig raffte er
seine Kissen zusammen, ergriff den Strick an einem Ende und schritt
majestätisch zur Tür hinaus, sein improvisiertes Halstuch wie eine
Schlange hinter sich herziehend. Wir übrigen vergnügten uns den
Rest der Stunde hindurch damit, daß wir die Flaumfedern, die rings
durch das Zimmer wirbelten, geschickt in die Höhe bliesen, wodurch
ein ununterbrochenes Schneegestöber entstand, das mir viel Spaß
machte. Doktor Perner ließ in das Tagebuch einschreiben, es habe
dem Hutzler in kindischer Weise gezogen; auch sei derselbe mit
einem Strick um den Hals im Lehrzimmer erschienen und habe sich
die frechsten Störungen erlaubt. Der Eintrag schloß mit dem
Vermerk einer zweitägigen Karzerstrafe. Hutzler war indes klug
genug, die nächsten acht Tage wegen Unwohlseins zu fehlen. Erst
am folgenden Sonnabend kam er wieder, und am Dienstag darauf
fragte ihn Doktor Perner, ob er die Strafe abgesessen habe.
»Ich?« sagte Hutzler erstaunt.
»Ja, Sie! Antworten Sie mir!«
»Aber, Herr Doktor, Sie haben mir die zwei Tage ja geschenkt!«
entgegnete Hutzler, tief beleidigt.
»So? Davon weiß ich kein Wort!«
»Doch, Herr Doktor!« wagte ich jetzt schüchtern zu bemerken.
»Ja, Herr Doktor!« riefen zwei, drei Stimmen aus dem
Hintergrunde. »Am vorigen Sonnabend haben Sie gesagt: Nun, für
diesmal will ich es Ihnen noch erlassen, aber in Zukunft geht es
Ihnen schlecht, das gebe ich Ihnen schriftlich!«
Doktor Perner begann bei dieser bestimmten Formulierung
unserer Lüge stutzig zu werden. Vielleicht auch schien es ihm das
geratenste, mit Hutzler Frieden zu schließen. Daher sagte er
geringschätzig:
»Nun, es mag gut sein. Wenn ich es denn einmal gesagt habe, so
will ich nicht weiter darauf bestehen. Aber das sage ich Ihnen:
kommt mir wieder einmal etwas Ähnliches vor, so ist es alle mit uns.
Überhaupt konstatiere ich seit einiger Zeit das Überhandnehmen
eines Geistes, der den Zwecken dieser Anstalt schnurstracks
zuwiderläuft. Es herrscht hier statt des wissenschaftlichen Ernstes
eine läppische Puerilität, die mich anwidert. Ich bin auch einmal jung
gewesen und habe mich meines Lebens gefreut, aber ich würde
noch jetzt schamrot werden, wenn ich mir jemals eine so kindische
Haltung hätte vorwerfen müssen, wie sie Ihnen zur zweiten Natur
geworden ist. Bessern Sie sich, ich rate es Ihnen im guten. Ostern
steht vor der Tür, und die Versetzungen sind noch lange nicht
entschieden. Es könnte manchem passieren, daß er sich grimmig
verrechnete. Zum Aufrücken in eine höhere Klasse ist nicht nur eine
gewisse Summe von Kenntnissen erforderlich, sondern vor allen
Dingen ein würdiges Betragen. Die Primaner werden sich bedanken,
ihren Lehrsaal mit einer Gesellschaft von Kindsköpfen teilen zu
sollen, wie sie hier auf den Bänken von Obersekunda sitzen. Was
lachen Sie, Hutzler? Weinen sollten Sie und in sich gehen, ehe die
verderblichen Wege, auf denen Sie sich befinden, vollends zum
Abgrund geführt haben. Ein Mensch von Ihren Gaben! Es ist
himmelschreiend!«
Hutzler erhob sich.
»Herr Doktor, ich habe nur ein freundliches Gesicht gemacht«,
versetzte er mit unerschütterlicher Ruhe.
»Machen Sie Ihre freundlichen Gesichter, wenn Sie zu Hause sind!
Hier ist Haltung und Ernst erforderlich, und Sie hätten am
allerwenigsten Ursache, ihrem Übermut freien Lauf zu lassen.«
Hiermit schloß das Hutzlersche Intermezzo.
Aus den Privataufzeichnungen des
Sekundaners Heppenheimer.
Zweites Bruchstück.
Am 5. Mai 18**
… Von mir selber zu reden, verbietet mir eigentlich die mir
innewohnende Bescheidenheit. Indes neulich habe ich unserem
Religionslehrer einen so köstlichen Streich gespielt, daß ich nicht
umhin kann, diese wohlgelungene »Störung« hier aufzuzeichnen.
Und eine »Störung« war es in des Wortes tiefgehendster Bedeutung,
insofern sie nämlich nicht allein den regelmäßigen Verlauf des
Lektionsplanes, sondern mehr noch das gesamte seelische
Gleichgewicht unseres trefflichen Lehrers »störte«. Aber ich konnte
ihm nicht helfen. Einmal hat er's durchaus nicht um mich verdient,
daß ich den Regungen des Mitleids Audienz gebe, da er mir in der
Religionsstunde schon dreimal »wegen hartnäckigen
Widersprechens« Ordnungsstrafen erteilte. Und dann hätte ich
Nerven besitzen müssen von der Dicke und der Dauerhaftigkeit jenes
Strickes, den Hutzler im Februar dieses Jahres um seinen Hals
wickelte, wenn ich die qualvolle Monotonie, die sich in der religiösen
Gesinnung des Herrn Pastors geltend machte, länger hätte ertragen
sollen.
Der Herr Pastor …! Wir nennen ihn so, weil er in früheren Zeiten
eine Predigerstelle an einem benachbarten Dorfe versah. Er ist indes
schon seit geraumer Zeit an unserem Gymnasium in bester Form
angestellt und gibt in allen Klassen Glaubenslehre und
Kirchengeschichte. Ich unterlasse es, hier auf seine persönliche
Beschreibung einzugehen, da es mir doch nicht möglich wäre, seinen
wohlwollenden Gesichtszügen und dem sanften Behagen, das um
seine schmalen, blutlosen Lippen spielt, stilistisch gerecht zu
werden.
Der Herr Pastor ist nämlich ein sehr frommer Mann, was ich ihm
durchaus nicht verüble, denn wenn jemand Pastor ist, so versteht es
sich von selbst, daß er gewisse Gesinnungen hegt. Wohl aber
verüble ich dem Herrn Pastor im höchsten Grade, daß er seiner
Frömmigkeit seit so und so viel Jahren in sämtlichen Klassen stets
denselben Ausdruck verleiht und so zum Beispiel an jedem Morgen,
den Gott werden läßt, aus dem Klassengebetbuch dasselbe Gebet
abliest. Wir alle kennen es längst auswendig: aber der Herr Pastor
scheint nun einmal die Überzeugung zu hegen, dieses Gebet sei
besonders wirksam und gottwohlgefällig. Es beginnt mit den Worten:
»So treten wir denn wiederum vereint vor die Stufen Deines
Thrones, o Allmächtiger, und flehen zu Dir mit kindlichem Herzen um
die Gnade Deines Beistandes …«
Es ist mehr als zwei Seiten lang und enthält unter anderm die sehr
richtige Bemerkung:
»Schritt für Schritt wandeln wir dem Ende zu und sind ihm jeden
Morgen näher gebracht.«
Es hat mir nun fast den Anschein, als ob der Herr Pastor geglaubt
habe, durch die fortwährende Betonung dieser unleugbaren
Tatsache das immer näher rückende Ende weiter hinausrücken zu
können. Denn nur so vermag ich mir zu erklären, wie er immer und
immer wieder dieselben Phrasen zum besten gab, ohne zu
bedenken, daß jedes unverkünstelte Menschengehirn bei solchem
Geklapper aus dem Leim gehen muß. Er scheint eine ganz eigen
organisierte Natur zu besitzen. Wir bekamen das trostlose Gebet
doch nur jeden Dienstag und Freitag zu hören: er aber trug es seit
Menschengedenken auch Montags, Mittwochs, Donnerstags und
Sonnabends vor (in Prima und Tertia nämlich), ohne daß es ihm bis
jetzt irgend geschadet hätte.
Nun, es heißt schon in Goethes Faust: »Die Kirche hat einen guten
Magen«. Da ich aber in keiner Beziehung zur Kirche gehöre und
mich überhaupt von der sogenannten kirchlichen Richtung prinzipiell
fern halte – mein Vater ist Freimaurer –, so erscheint es begreiflich,
daß ich infolge dieser ununterbrochenen Gebetsidentität nahezu
krank wurde und eine wahre Wut gegen den wiederkäuenden Lehrer
faßte.
Da kam mir ein köstlicher Gedanke, den ich um so bereitwilliger
durchführte, als ich mir sagen mußte, das Faktum werde, ganz
abgesehen von der Befriedigung meiner Gebetwünsche, auch einen
reizvollen Zwischenfall absetzen, wie ein lebensfroher Gymnasiast
ihn stets brauchen kann. So zögerte ich denn nicht länger und
»vollendete das Werk dieser Woche«. (Ein schönes Zitat! Es ist dem
Schlußgebet entnommen, das wir jeden Sonnabend um zwölf mit
anhören müssen.)
Das Klassengebetbuch liegt in der Regel auf dem Katheder, damit
es dem Lehrer gleich zur Hand sei, sobald er das Bedürfnis fühlt,
sich mit Gott zu unterhalten. Wie Möros in der Schillerschen Ballade,
schlich ich mich eines Morgens in aller Frühe – ich war eigens eine
halbe Stunde vor Beginn der Lehrstunde erschienen, um der erste zu
sein – zum Katheder, öffnete das Buch mit dem schwarzen,
unheimlichen Einband, der so oft in den Händen meines Peinigers
geruht hatte, und suchte mit fiebernder Hast nach dem
verhängnisvollen Kapitel.
»Aha!« sagte ich mit diabolischer Wollust, als ich den Gegenstand
meines Hasses entdeckt hatte, – »da steht es:
›So treten wir denn wiederum vereint vor die Stufen Deines
Thrones, o Allmächtiger, und flehen zu Dir mit kindlichem Herzen um
die Gnade Deines Beistandes …‹
Du sollst keinen mehr kränken!« Und mit keckem Griffe riß ich die
beiden Blätter, auf denen das Leibgebet des Herrn Pastors
verzeichnet stand, aus dem Buche, zerpflückte sie in hundert
mikroskopische Stückchen und trug sie kaltblütig, als ob nichts
geschehen wäre, nach dem Hofe, wo ich sie dem Spiel der
Frühlingswinde überantwortete.
In dem beseligenden Gefühle, ein gutes Werk vollbracht zu haben,
setzte ich mich auf meinen Platz und wartete der Dinge, die da
kommen sollten.
Es schlug sieben. Die Tür öffnete sich, und herein wandelte
wuchtigen Schrittes unser gottwohlgefälliger Religionslehrer. Er
setzte sich auf den Kathederstuhl, schneuzte sich zweimal und legte
dann sein Gesicht in jene frommen Gebetsfalten, die ich so oft mit
Schrecken an ihm bemerkt hatte. Das war immer die Introduktion;
eine halbe Minute später ging's los: »So treten wir denn wiederum
vereint vor die Stufen Deines Thrones, o Allmächtiger …«
Der Herr Pastor erhob sich und ergriff mit einem verschleierten
Blick gen oben das Gebetbuch. Auch die Schüler standen ehrerbietig
von ihren Plätzen auf und falteten schweigend die Hände.
Lautlose Stille.
Der Herr Pastor schlug das Gebetbuch auf und spitzte die Lippen.
Merkwürdigerweise war das so oft vorgetragene Gebet heute nicht
auf den ersten Griff zu finden.
Der Herr Pastor blätterte. Und blätterte wiederum. Und blätterte
abermals.
Es war ganz unbegreiflich!
Er beschaute das Buch von außen, als wolle er sich überzeugen,
ob es noch das alte, stille, traute Gebetbuch von ehedem sei, mit
dessen Hilfe er so manches Mal vereint vor die Stufen des Thrones
getreten war.
In der Tat, der Einband hatte sich nicht im geringsten verändert.
Nun suchte der Herr Pastor, immer befremdlicher dreinschauend,
im Register.
Richtig, da stand es, Seite 50!
Der Herr Pastor suchte demgemäß Seite 50, aber siehe da! der
schöne Spruch: Suchet, so werdet Ihr finden, wurde diesmal
zuschanden! … Jetzt erst ging dem unglücklichen Lehrer ein Licht
auf. Er stieß einen unartikulierten Ton der Entrüstung aus und sagte
dann mit einer Stimme, die an die Posaune des Jüngsten Gerichts
gemahnte:
»Da hat mir ein miserabler Junge die Blätter herausgerissen, auf
denen unser Morgengebet stand! Ich will nicht untersuchen, wer
sich diesen sakrilegischen Akt erlaubt hat: ich überlasse den
Betreffenden dem Gefühl seiner eigenen Schande. Aber traurig ist es
doch, daß so etwas in einer Klasse von gesitteten jungen Leuten
vorkommen kann! Pfui!«
Ich mußte mir auf die Lippen beißen, um nicht in helles Gelächter
auszuplatzen. Der Herr Pastor bemerkte es.
»Was lachen Sie?« sagte er mit einem vernichtenden Blick auf
mein unschuldiges Gesicht. »Gehen Sie hinaus: Sie sind in dieser
Stimmung nicht würdig, der Schulandacht beizuwohnen.«
»Aber Herr Pastor« … sagte ich demütig.
»Sie verlassen das Zimmer!« wiederholte er schneidig. »Wenn Sie
bei einem solchen Anlaß überhaupt lachen können, so läßt das tief
blicken.«
Ich verließ also das Zimmer, stand aber nahe genug an der Tür,
um zu hören, daß der Herr Pastor noch eine längere Rede hielt.
Hierauf öffnete er wieder das Gebetbuch und las ein anderes Kapitel
vor, das halb so lang war, als die »Stufen des Thrones«, und schon
um seiner Neuheit willen die gebührende Aufmerksamkeit fand.
»Amen!« sagte der Herr Pfarrer wuchtig und salbungsvoll, und
gleich darauf fügte er hinzu: »Sie können den Heppenheimer jetzt
wieder hereinholen.«
War das nicht ein köstlicher Streich? Aber nicht genug! Das
Gerücht von dem herausgerissenen Gebet verbreitete sich wie ein
Lauffeuer durch alle Klassen, und ehe der folgende Morgen graute,
waren die »Stufen des Thrones« aus sämtlichen Gebetbüchern
entfernt! So ging dem Herrn Pastor denn ein für allemal die
Möglichkeit verloren, seiner langjährigen Leidenschaft fürder zu
frönen!
Er verhängte jetzt eine umfassende Untersuchung, die jedoch
ohne Resultat blieb. Wenn ich der Eventualität einer Karzerstrafe so
gleichmütig gegenüberstünde, wie Schwarz oder Rumpf, die beide
einen unbeschreiblichen Stoizismus besitzen, so würde ich dem
Herrn Pastor ohne weiteres entgegentreten und ihm zurufen: »Ja,
Verehrtester, c'est moi qui l'ai fait! Und wissen Sie, warum ich's
getan habe? Weil es ein Sprüchlein gibt, das da lautet: So Ihr betet,
sollt Ihr nicht plappern wie die Heiden, die da meinen, sie würden
erhöret, wenn sie viele Worte machen! Weil ferner sich der Betende
in sein stilles Kämmerlein einschließen soll! Und weil schließlich
geschrieben steht, daß der Buchstabe tot macht, um wieviel mehr
ein ellenlanges Geleier, das aus mehreren tausend Buchstaben
besteht!«
Da ich heute gerade bei der Persönlichkeit des Herrn Pastors bin,
so will ich noch die Geschichte von den Bescherungen meines
Freundes Boxer notieren. Die Sache ist zwar sehr einfach, aber sie
hat mich königlich amüsiert.
Vor einiger Zeit nämlich, es mögen jetzt vielleicht acht oder zehn
Wochen her sein, hatten wir es eingeführt, dem Herrn Pastor
jedesmal morgens bei dem Beginn der Lehrstunde eine Bescherung
auf den Katheder zu setzen. Mein Freund Boxer war in dieser
Beziehung der Haupt-Entrepreneur und ging dabei ebenso malerisch
als humoristisch zu Werke. Der Herr Pastor entdeckte beim Eintritt in
das Lehrzimmer folgende Gegenstände in zierlicher Gruppierung auf
der Platte seines Lehrpultes: in der Mitte eine große, halb
zerbrochene Blumenscherbe, mit zwei alten Schreibärmeln
behangen, wie eine Totenurne; oben darauf eine Schneelawine (es
war damals noch Winter und hatte tüchtig geschneit), recht fest
geknetet und von der Form eines menschlichen Kopfes. Die Augen
hatten wir aus gekautem, blauem Papier gefertigt; desgleichen die
Nase; den Mund stellte ein hereingedrücktes Bleistiftchen dar. Unser
Mitschüler Schwarz hatte seine brandrote Mütze hergeben müssen,
auf daß sie diesem Schneekopfe zur Bedeckung diene. Rechts und
links prangten zwei stattliche Haufen frischgeschlagener
Chausseesteine und ein Paar alte Stiefel, die Boxer sich von dem
Hausknecht im »Goldenen Pfau« hatte schenken lassen. Die Lücken
unseres künstlichen Aufbaues waren mit Äpfeln und Eierschalen,
rohen Kartoffeln, Besenreisern und ähnlichen Gegenständen
dergestalt ausgefüllt, daß man eine Barrikade en miniature vor sich
zu sehen glaubte, wie denn Boxer überhaupt viel Talent zum
Kommunismus verrät. Der Herr Pastor näherte sich dieser
Bescherung jedesmal mit einem Blick, als könne das Ding
explodieren, bestieg den Katheder und legte dann seinen Arm links
auf die Platte des Pultes. Mit einem einzigen, gewaltigen Ruck fegte
er die Barrikade herunter, daß wir jedesmal Angst hatten, der
Wandschrank, der rechts vom Katheder steht, möge durch den
Anprall so mannigfacher Objekte zertrümmert werden.
Die Klasse aber brach in ein diabolisches Jauchzen aus, und die
Vordersten liefen herzu, um die Brocken der Schneelawine aus dem
Fenster zu werfen, und besonders um die alten Stiefel in Sicherheit
zu bringen, die bei jeder Bescherung aufs neue verwendet wurden.
Der Herr Pastor beobachtete bei solchen Vorgängen stets das
Prinzip des unbedingten Ignorierens. Kein Wort kam über seine
Lippen: er strafte uns mit stiller Verachtung.
Zuweilen fiel ihm das recht schwer, denn einmal hatten wir ihm
einen alten, zerrissenen Familienschirm an die Lehrtafel genagelt,
dessen Fischbeindrähte dergestalt über den Kathedersessel
hinausragten, daß es nur bei einer eigentümlichen Haltung des
Kopfes möglich war, ihnen auszuweichen. Der Herr Pastor ertrug
diese Tortur mit einer bewundernswerten Geduld etwa fünf Minuten
hindurch; dann sagte er dem Katheder Valet und tat, als ob er nur
zur Abwechslung einmal einen anderen Standpunkt einnehme,
während er bisher niemals seinen gewöhnlichen Platz hinter dem
Pult verlassen hatte.
Endlich aber zerriß auch diesem Gerechten der Faden der Geduld.
Boxer hatte ihm nämlich, des ewigen Hinabwerfens müde, einen
großen Eimer voll Wasser auf die Platte gestellt, dessen Fall eine
wahre Sündflut herbeigeführt haben würde.
Der Herr Pastor kam, beschaute sich das Ding mit rollenden
Blicken, blähte die Nüstern, stieg wieder vom Katheder herunter,
schritt zornig im Zimmer auf und ab, bestieg den Katheder von
neuem und sagte endlich mit Donnerstimme:
»Boxer, schaffen Sie das augenblicklich hinweg!«
Ich bewunderte hier den Instinkt des scheinbar so stillen Mannes,
der sofort wußte, wo er den Feind seiner Ruhe zu suchen hatte.
Boxer erhob sich.
»Ich?« sagte er indigniert. »Weshalb denn gerade ich?«
»Tun Sie, was ich Ihnen sage! Augenblicklich schaffen Sie mir das
Ding da fort!«
»Wenn ich den Eimer dahin gesetzt hätte,« entgegnete Boxer,
»mit Vergnügen! Aber so sehe ich in der Tat nicht ein …«
»Augenblicklich!« wiederholte der Herr Pastor, indem er den Arm
ausstreckte und die Spitze seines Zeigefingers auf den Boden
richtete.
»Gut!« sagte Boxer, »ich bin Schüler und muß gehorchen. Aber ich
will mich doch einmal bei dem Herrn Direktor erkundigen, ob ich
Ihnen die Eimer ausleeren muß.«
Mit diesen Worten trat er aus den Bänken heraus und schritt
langsam und würdevoll dem Katheder zu. Stirnrunzelnd ergriff er das
in diesen Räumen sehr ungewöhnliche Gefäß und wußte es so
einzurichten, daß er beim Herabtreten vom Katheder stolperte und
langwegs ins Zimmer fiel.
Ein Hallo sondergleichen durchbrauste die Räume Sekundas. Ich
versichere bei allen Göttern, es hat ganz über alle Maßen schön
geklatscht, und das Wasser floß bis in den fernsten Winkel des
Saales. Die Verwirrung wurde noch dadurch gesteigert, daß einige
von uns riefen, sie könnten es in einem so feuchten Zimmer nicht
aushalten, sie hätten sich neulich erst erkältet, als der Pedell so
unsinnig aufgewaschen, und sie bäten um ihre Entlassung. Vier oder
fünf wurden in der Tat beurlaubt; dann aber wandte sich der Herr
Pastor zu Boxer und sagte:
»Boxer, ich mache Sie von jetzt an für alles verantwortlich, was in
diesen Räumen geschieht. Ist morgen wieder etwas auf die
Kathederplatte gestellt, so werden Sie die Folgen zu tragen haben!«
Boxer trocknete sich inzwischen die Beinkleider und erwiderte in
vorwurfsvollem Tone:
»Also wenn der Schwarz ein Paar alte Stiefel auf den Katheder
legt, dann bin ich dafür verantwortlich?«
»Was?« rief Schwarz, »ich hätte ein Paar alte Stiefel auf den
Katheder gelegt? Herr Pastor, das muß ich mir doch sehr von dem
Boxer verbitten.«
»Ich sage ja nur: wenn«, erwiderte Boxer.
»Still,« gebot der Herr Pastor; »was ich gesagt habe, dabei bleibt
es! Und nun rufen Sie den Pedell, daß er hier aufwischt!«
Die Situation Boxers war offenbar eine kritische. Gerade für den
folgenden Tag hatten wir uns so etwas Hübsches ausgedacht! Wir
wollten dem Herrn Pastor nämlich zwei Eimer auf den Katheder
stellen, und Schwarz hatte zu diesem Behufe bereits seiner Tante,
bei der er zur Miete wohnte, einen gestohlen. Und nun sollte unser
guter Freund Boxer für alle diese Streiche verantwortlich gemacht
und vielleicht mit einer mehrtägigen Karzerstrafe belegt werden! Es
war nicht zu verlangen, daß wir unser reizendes Amüsement
aufgaben; aber ebensowenig durften wir Boxer zumuten, um
unseres Gaudiums willen die Räume des Karzers zu beziehen.
Wir überlegten hin und her, ohne zu einem Resultat zu gelangen.
Plötzlich sagte Boxer: »Laßt mich nur machen!«
Am folgenden Tage arrangierten wir die Bescherung, wie
verabredet. Boxer war der emsigste, und in der Tat, es war keine
kleine Arbeit, denn wir mußten die Eimer glasweise füllen, weil der
Pedell sonst Lunte gemerkt hätte. Einer von uns stand Wache, um
das Herannahen des Herrn Pastors rechtzeitig anzukündigen.
»Er kommt!« rief es jählings aus dem Munde unseres Warners.
Sofort ergriff Boxer seine Mütze, nahm einen möglichst starken
Stoß Bücher unter den Arm und verließ das Lehrzimmer, um sich auf
dem Korridor hinter einen der größten Schränke zu stellen.
Zwei Minuten später erschien der Lehrer, und unmittelbar hinter
ihm her tappte keuchend und atemlos unser trefflicher Freund Boxer,
die Bescherung auf dem Katheder genau ebenso verblüfft
anstarrend, wie der Herr Pastor.
– »Boxer! Wo ist der Boxer?« zürnte der entrüstete Schulmann.
»Hier!« rief es hinter ihm. »Entschuldigen Sie, daß ich mich heute
verspätet habe.«
Der Herr Pastor biß sich ärgerlich auf die Lippe. Der Tatsache
gegenüber, daß Boxer fast mit ihm zugleich ins Zimmer getreten war,
konnte er seine Drohung von gestern nicht wohl verwirklichen.
»Heute brauche ich die Eimer wohl nicht auszuleeren?« fragte
Boxer triumphierend.
»Schwarz,« sagte der Lehrer, »rufen Sie einmal den Pedell.«
Der Pedell erschien.
»Quaddler,« begann der Herr Pastor in ernstem Tone, »nehmen
Sie einmal diese Gefäße da hinweg. Ich mache Sie von jetzt ab dafür
verantwortlich, daß solche Ungehörigkeiten sich nicht wiederholen.
Stellen Sie sich an den Brunnen, damit kein Wasser geschöpft
werden kann, oder lassen Sie Ihre Frau Wache halten. Wozu sind Sie
verheiratet?«
»Entschuldigen Sie gütigst,« stammelte Quaddler in höchster
Verwirrung, »aber meine Frau war gerade damit beschäftigt, sich
anzukleiden, und da mußte ich Kaffee brennen.«
»Brennen Sie Ihren Kaffee von sieben bis acht! Kommen Sie
vorher Ihren Pedellpflichten nach! Was ist das für eine Art! Die Eimer
da können doch nur von Ihnen herrühren! Warum geben Sie nicht
besser auf Ihre Küchengerätschaften acht?«
Quaddler näherte sich dem Katheder.
»Ich bitte um Verzeihung«, sagte er. »Das sind allerdings
sozusagen zwei Eimer, aber zwischen Eimer und Eimer ist ergebenst
ein Unterschied. Meine Eimer sind ganz anders. Und wenn die
Herren Sekundaner hinter meinem Rücken solche Sachen da
mitbringen …«
»Seien Sie still und machen Sie, daß Sie so schnell wie möglich
hinauskommen. Die Eimer können Sie behalten, denn wenn
dieselben auch nicht Ihrer Küche entstammen, so bezweifle ich
doch, daß die wahren Besitzer sich melden werden.«
Und Quaddler nahm die Eimer und verschwand in den Gängen des
Schulgebäudes. Der Herr Pastor irrte sich indes, wenn er glaubte,
über fremdes Eigentum so ohne weiteres verfügen zu dürfen. Die
Tante meines Freundes Schwarz machte bei der Polizei die Anzeige,
es sei ihr am Nachmittage des Dreizehnten ein Eimer gestohlen
worden. Die Behörden stellten umfassende Recherchen an; Quaddler
mußte zu wiederholten Malen auf das Gericht, und selbst der Herr
Pastor ward eidlich vernommen. Schwarz brauchte nicht zu
schwören, denn er ist erst fünfzehn Jahre alt, und so lief die Sache
höchst günstig ab. Fortes fortuna juvat.
Doktor Veit.
Eine Porträt-Skizze.
u den reinsten Genüssen meiner Schuljahre zähle ich den
Unterricht des originellen Mathematiklehrers Doktor Veit, der
uns von Quarta bis Obersekunda den Pfad zur Vollendung führte.
Man verbindet gewöhnlich mit dem Begriffe des Mathematikers die
Vorstellung eines regelrecht konstruierten Behälters öder, lebloser
Formeln: nur im äußersten Winkel dieses Behälters lauert die
Individualität eines menschlich fühlenden Wesens. Unser lieber,
trefflicher Veit war die verkörperte Widerlegung dieser einseitigen
Theorie. Kein Zug seiner charaktervollen Erscheinung erinnerte an
die Schablone. Seine innere und äußere Physiognomie entfernte sich
himmelweit von jenen typischen Schreckbildern, die alles auf
Gleichungen zurückführen, die selbst in der Liebe von positiven und
negativen Größen phantasieren und die Erkorene mit dem
Parallelogramm der Kräfte an die hochklopfende Brust ziehen.
Doktor Veit war vielmehr ein erquickliches Original, in derber
Holzschnittmanier ausgearbeitet, aber nirgends geometrisch korrekt.
In seinen Lehrstunden herrschte durchaus nicht der Ton der reinen
Mathematik. Sein Vortrag war reich an subjektiven Streiflichtern, an
reizvollen Impromptus, an ergötzlichen Zwischenfällen …
Erst in Sekunda lernten wir diesen Mann nach dem ganzen
Umfange seiner Vorzüge schätzen.
Wenn es elf schlug, und der Xenophon-Interpret, »froh der
bestandenen Gefahr,« von hinnen geeilt war, dann erschien in der
Tür eine mittelgroße Gestalt mit rötlich angestrahltem Gesicht, den
Hut ein wenig im Nacken, um die Lippen ein freundliches Lächeln.
Rasch warf er die Kopfbedeckung auf den Tisch neben dem Eingang
und schritt beweglich nach dem Katheder, ohne nach pädagogischer
Würde zu haschen, ohne in professorenhafter Manier den Rock
zuzuknöpfen, ohne an der Brille zu rücken. Mit prüfendem Blick
musterte er die lebhaft plaudernde Versammlung und nahm dann
halb wie im Traum den Schwamm, der neben der Kreide lag, um ihn
mit Wasser zu sättigen. Allerlei groteske Figuren beschreibend,
wischte er die große Schultafel ab; dann kehrte er sein freundliches
Antlitz von neuem dem schwatzhaften Publikum zu, trommelte mit
den Fingern auf die Holzfläche des Katheders und nickte still vor sich
hin …
Das währte so drei, vier Minuten. Plötzlich schien er sich zu
besinnen, weshalb er hierher gekommen. Mit der Faust auf die
Fläche des Lehrpultes schlagend wie ein Tambour, der die Kolonne
zum Sturme führt, rief er mit Donnerstimme:
»Allez! vite! vite! vite! wacker! wacker! wacker! Boxer, komme Se
mal raus an die Tawel!«
Diese Wendung kehrte in ähnlicher Form als Einleitung zu jeder
Lehrstunde wieder. Doktor Veit vermochte sich nämlich gewisser
dialektischer Eigentümlichkeiten nie zu entschlagen, namentlich in
erregter Stimmung; wenn er im Eifer des Dienstes erglühte, wenn
der Zorn ihm die Nerven schüttelte, stets verfiel er dem Banne der
Mundart, und seine Mundart war nicht die gefeilteste. In
grammatikalischer wie in lexikographischer Hinsicht borgte er bei
dem Volke. So ergoß sich ein Hauch echter Urwüchsigkeit und
reinen, gediegenen Menschentums über die exklusive Klassizität
unseres Gymnasialkatheders.
Boxer stand auf und trat an die »Tawel«, die rechts vom Katheder
auf den Holzzapfen der Staffelei ruhte …
»Ich bitt' mer jetzt aus, daß Ihr Ruh' halt't!« rief Doktor Veit
kategorisch den hinteren Bänken zu. »Na, Boxer! Allez! vite! vite!
Hier ist die Kreide! Schreibe Se emal folgende Gleichung!«

You might also like