Neurobionics E-Book
134,99 €
Mehr erfahren.
- Herausgeber: John Wiley & Sons
- Kategorie: Wissenschaft und neue Technologien
- Sprache: Englisch
Technological advances have greatly increased the potential for, and practicability of, using medical neurotechnologies to revolutionize how a wide array of neurological and nervous system diseases and dysfunctions are treated. These technologies have the potential to help reduce the impact of symptoms in neurological disorders such as Parkinson’s Disease and depression as well as help regain lost function caused by spinal cord damage or nerve damage. Medical Neurobionics is a concise overview of the biological underpinnings of neurotechnologies, the development process for these technologies, and the practical application of these advances in clinical settings.
Medical Neurobionics is divided into three sections. The first section focuses specifically on providing a sound foundational understanding of the biological mechanisms that support the development of neurotechnologies. The second section looks at the efforts being carried out to develop new and exciting bioengineering advances. The book then closes with chapters that discuss practical clinical application and explore the ethical questions that surround neurobionics.
A timely work that provides readers with a useful introduction to the field, Medical Neurobionics will be an essential book for neuroscientists, neuroengineers, biomedical researchers, and industry personnel.
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 711
Veröffentlichungsjahr: 2016
Ähnliche
Table of Contents
Cover
Title Page
Copyright
Dedication
Contributors
Preface
References
Part I: Fundamentals of neural prostheses
Chapter 1: The Historical Foundations of Bionics
1.1 Bionics Past and Future
1.2 History in 1973
1.3 Anaesthesia
1.4 Aseptic Surgery
1.5 Clinical Observation and Experiments
1.6 Hermetic Packages
1.7 Encapsulation (Electrical Insulation)
1.8 Early Implanted Devices
1.9 Afterword
References
Chapter 2: Development of Stable Long-term Electrode Tissue Interfaces for Recording and Stimulation
2.1 Introduction
2.2 Tissue Responses in the Brain to an Implanted Foreign Body
2.3 Brain Computer Interfaces (BCI) – State-of-the-Art
2.4 Biocompatibility of BCI – on the Importance of Mechanical Compliance
2.5 Novel Electrode Constructs and Implantation Procedures
2.6 Concluding Remarks
Acknowledgements
References
Chapter 3: Electrochemical Principles of Safe Charge Injection
3.1 Introduction
3.2 Charge Injection Requirements
3.3 Electrode Materials
3.4 Factors Influencing Electrode Reversibility
3.5 Emerging Electrode Materials
3.6 Conclusion
References
Chapter 4: Principles of Recording from and Electrical Stimulation of Neural Tissue
4.1 Introduction
4.2 Anatomy and Physiology of Neural Tissue
4.3 Physiological Principles of Recording from Neural Tissue
4.4 Principles of Stimulation of Neural Tissue
4.5 Safety of Electrical Stimulation
4.6 Conclusion
References
Part II: Device Design and Development
Chapter 5: Wireless Neurotechnology for Neural Prostheses
5.1 Introduction
5.2 Rationale and Overview of Technical Challenges Associated with Wireless Neuroelectronic Interfaces
5.3 Wireless Brain Interfaces Require Specialized Microelectronics
5.4 Illustrative Microsystems for High Data Rate Wireless Brain Interfaces in Primates
5.5 Power Supply and Management for Wireless Neural Interfaces
5.6 Packaging and Challenges in Hermetic Sealing
5.7 Deployment of High Data Rate Wireless Recording in Freely Moving Large Animals
5.8 Summary and Prospects for High Data Rate Brain Interfaces for Neural Prostheses
Acknowledgements
References
Chapter 6: Preclinical testing of Neural Prostheses
6.1 Introduction
6.2 Biocompatibility Testing of Neural Implants
6.3 Testing for Mechanical and Electrical Integrity
6.4
In vitro
Accelerated Testing and Accelerated Aging of Neural Implants
6.5
In vivo
Testing of Neural Prostheses
6.6 Conclusion
References
Part III: Clinical Applications
Chapter 7: Auditory and Visual Neural Prostheses
7.1 Introduction
7.2 Auditory Prostheses
7.3 Visual Prostheses
7.4 Sensory Prostheses and Brain Plasticity
7.5 Conclusions
Acknowledgements
References
Chapter 8: Neurobionics: Treatments for Disorders of the Central Nervous System
8.1 Introduction
8.2 Psychiatric Conditions
8.3 Movement Disorders
8.4 Epilepsy
8.5 Pain
8.6 Future directions
Acknowledgements
References
Chapter 9: Brain Computer Interfaces
9.1 Introduction
9.2 Motor Physiology
9.3 The Clinical Population for Brain Computer Interfaces
9.4 BCI Modalities
9.5 BCI Decoding and Applications
9.6 Future Directions
9.7 Conclusion
References
Part IV: Commercial and Ethical considerations
Chapter 10: Taking a Device to Market: Regulatory and Commercial Issues
10.1 Introduction
10.2 Basic Research
10.3 Preclinical Development
10.4 Clinical Trials and Approval to Sell
10.5 Building a Business not a Product
10.6 Conclusions
References
Webliography
Chapter 11: Ethical Considerations in the Development of Neural Prostheses
11.1 Introduction
11.2 Individuals with Disabilities and Technology Development
11.3 Ethical Principles of Biomedical Research
11.4 Conclusions
References
Appendix: Examples of Companies Developing and/or Marketing Bionic Devices
Index
End User License Agreement
Pages
xv
xvi
xvii
xviii
xix
1
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
187
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
265
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
294
295
295
296
296
297
297
298
298
299
300
299
300
301
301
302
302
303
303
304
304
305
305
306
306
307
307
308
308
309
309
310
310
311
311
312
312
313
313
314
314
315
315
316
316
317
317
318
318
319
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
319
320
321
322
323
324
325
327
328
329
330
331
332
333
334
Guide
Cover
Table of Contents
Preface
Part I: Fundamentals of neural prostheses
Begin Reading
List of Illustrations
Chapter 1: The Historical Foundations of Bionics
Figure 1.1 Elmqvist-Senning pacemaker of 1958. It is powered by two nickel-cadmium cells (arrowhead) which can be recharged by induction. The two transistors are on the right (arrows). The encapsulant is epoxy resin. An external valve oscillator was used for recharging at a frequency of 150 kHz. Scale bar = 1 inch.
Figure Box 2 A non-scale drawing showing the main features of the most common type of implant during the last 40 years. The electronic components, and sometimes a battery are inside a metal enclosure which is “hermetic”, meaning that leak rate of moisture is low enough that the inside will remain dry for the required lifetime of the device. Conductors are brought out through annular “feed-throughs”, which comprise a metal pin, a glass or ceramic bush and a metal ring. On the outside of the enclosure, wires are joined to the pins of the feed-thoughs which may be part of the output cable (as shown) or may go to a surgical connector. The cables are usually either multi-strand wires (as shown) or helical single-strand wires, and there may be one or more wires in each cable. The polymer encapsulant is essential to insulate the exposed wires where they are joined to the pins of the feed-throughs: to be an effective insulator, the encapsulant must remain bonded to the enclosure and the feed-through.
Figure 1.2 The first modern pacemaker was made by Telectronics in 1971. The engineer David Cowdrey, who had been charged with developing a hermetic enclosure, selected Ti because of its light weight, strength, corrosion resistance and weldability. He used deep-drawn Ti half-cases that were welded together using TIG welding to keep the inside cool. He also developed ceramic/Ti alloy feed-throughs for the package. Interestingly, the technology was never patented because the company's patent attorney advised them that it “was obvious”! This picture shows a Telectronics ‘Slimline’ device from 1977.
Figure 1.3 Thermal expansion curves for some metals, some special alloys, some ceramics and a few glasses, taken from Strong (1938). The broken lines are the insulators.
Figure 1.5 Three types of package that were available in the 1970s, showing that several technologies were available that might be suitable for hermetic packages for implants.
Figure 1.4 High-reliability microelectronic module (aerospace) taken from Manfield (1969). The package has metal-in-glass feed-throughs and is sealed by electron beam welding.
Figure 1.6 Arrhenius plot for accelerated adhesive life tests, replotted from Donaldson (1982). Adhesive is Dow Corning 3140, an alcohol-evolving RTV. The substrates are alumina (*), tin-lead solder (), kovar (∇), copper-nickel alloy (○), oxidised titanium (Δ). Bonds to alumina are easily superior to all others.
Figure Box 4a The void on the surface of the glass bead of the feed-through fills with water that diffuses through the encapsulant. Soluble salts that are residues left after inadequate cleaning, or that are present in the encapsulant, dissolve in the water and increase its conductivity. Failures like this are often made visible because of corrosion products from the metals.
Figure 4b This diagram illustrates the disadvantage of using a high-modulus encapsulant if it will have any cause to shrink, for example due to cooling after cure. With this feed-through, which has a recessed glass bead, shrinkage sets up tension in the direction shown, tending to cause adhesion failure (a). This is exactly where adhesion is vital to prevent electrical leakage. Soft encapsulants are more likely to be able to relieve the tension by distortion at a free surface (b).
Figure 1.7 Dorsal column stimulator made by Medtronic from 1968–1978.
Figure 1.8 Intermediate design of pacemaker from Devices Implants Ltd. The hermetically-sealed package is on the right and the battery on the left, all encapsulated in epoxy resin (Kenny 1969). Figure with permission from the Annals of the New York Academy of Sciences.
Figure Box 6 Eighty-channel visual prosthesis, implanted in a blind volunteer in 1967. The extracranial part of the device had 80 radio frequency receivers, each connected to one of the electrodes on the intracranial silicone cap that was implanted over the right occipital cortex. Thirty-nine channels produced phosphenes after implantation. The receivers and the electrodes are visible in the X-ray image (Brindley and Lewin 1968). Each receiver in the implant was encapsulated in epoxy resin before being inserted into the silicone cap. Failures were probably due to wire breakage since failures were never partial. Biocompatibility had previously been tested in baboons.
Chapter 2: Development of Stable Long-term Electrode Tissue Interfaces for Recording and Stimulation
Figure 2.1 Schematic of the normal brain tissue response to the implantation of an electrode. From left to right: pre-implantation, acute to chronic phases after implantation. During the implantation, neurons, glia cells and blood vessels along and adjacent to the insertion track line are mechanically injured. As a consequence, local bleeding (illustrated in red) and edema are common. Within the next few days, an acute local tissue response to the injured cells and debris is initiated. This acute inflammatory tissue response includes activation and migration of microglia toward the implant (which starts within minutes after the injury) and proliferation of astrocytes in the tissue nearby the implant as well as neuronal death. During the subsequent weeks, an astrocytic scar that encapsulates the implanted electrode is formed, whereas many microglia cells undergo apoptosis. The number of neurons adjacent to the electrode is also significantly reduced.
Figure 2.2 Morphological alterations in neural morphology after 12 weeks implantation in the rat cortex cerebri depends on probe size and anchoring. Tethered (left) and untethered to the skull (right) with two different probe sizes (top 200 µm, bottom 50 µm). Neurons were stained for NeuN (from Thelin
et al
. 2010 [41] with permission). Note the flat shape of neurons adjacent to the tethered probes.
Figure 2.3 Comparison of glial reaction to rigid probes differing in density. The probes were not tethered to the cranium during the implantation time. The probes were explanted 6 weeks after implantation in rat cortex. (a–c) tissue reaction to low density probe, and (d–f) tissue reaction to high density probe. Scale bar: 200 um. Astrocytes were stained for GFAP and Microglia for ED1, cell nuclei was stained for DAPI. (g) GFAP staining plotted against distance from border of implant (in um). (h) Shape of probe used. The probes were 500 um in diameter and 3.5 mm long. Both low (tissue matched) and high density probes (density similar to platinum electrodes) had the same shape and were coated with Parylene C. Note the conspicuous difference in astrocyte proliferation between the two probes. While there was no difference in ED1 in the tissue sections, an examination of microglia attached to the probes after explantation showed a significantly larger number of microglia attached to the high density probes (not shown) (from Lind
et al
. 2013 [42] with permission).
Figure 2.4 A scanning electron microscopy (SEM) image of a novel polymer-based electrode array termed “u-foil electrode array” equipped with recording sites on protrusions [46]. The backbone of the electrode array is made of a thin film (<10 um) of a photo-structurable polyimide (Durimide 7505, FujiFilm Belgium) that serves to insulate the array of thin (<150 nm) gold leads. The electrode array is equipped with protrusions on which the individual recording sites are located at respective tips. The protrusions are directed backwards and serve the dual role to provide recording sites in less compromised tissue (i.e. at a distance from the major bulk of the array) and to anchor the electrode array in the tissue.
Figure 2.5 Gelatine-embedded electrode bundle inserted through a narrow opening and spread out in rat cortex to target a larger cortical volume. Each platinum lead has a diameter of 12 um and a 4 um thick layer of Parylene C insulation. A thick cortical section containing the entire implanted electrode bundle was clarified with methyl salicylate. In this process, lipids are extracted so that the brain tissue becomes transparent (from Lind
et al
. 2010 [50] with permission).
Figure 2.6 Gelatin-embedded bundle of electrodes for deep brain recording/stimulation. Platinum wires, 12 um in diameter and insulated with Parylene C. The distal part of the gelatin-embedded probe is shown enlarged to the right. Bar indicates 500 um.
Figure 2.7 Electrode array composed of ultra-thin electrodes that are individually flexible in three dimensions. Left, electrode array shown next to a British penny for comparison of size. Right, electrode array embedded and configurationally locked in hard gelatin for implantation.
Chapter 3: Electrochemical Principles of Safe Charge Injection
Figure 3.1 Damaging and non-damaging levels of stimulation are shown as a function of pulse charge density and charge/phase for platinum electrodes. The boundary between damaging and non-damaging stimulus intensity is delineated by the Shannon line [1] with k = 1.85. These data include acute animal studies of tissue damage thresholds and studies in humans for which no tissue damage was reported. The absence of reported damage does not exclude damage and the occurrence or absence of tissue damage can depend on a variety of factors including pulse frequency, duty cycle and the type of tissue being stimulated. ([1–25]).
Figure 3.2 Voltage transient response of an AIROF microelectrode in response to biphasic 100 μA, 400 µs current pulsing at 50 pps. The electrode was maintained at a positive bias of 0.4 V in the interpulse period. E
mc
maximum cathodal potential excursion (–0.42 V vs. Ag|AgCl); E
ma
maximum anodal potential excursion (0.6 V vs. Ag|AgCl), and V
a
access voltage (0.4 V vs. Ag|AgCl).
Figure 3.3 Voltage transient response of an over-pulsed AIROF microelectrode in response to biphasic 100 μA, 600 µs current pulsing at 50 pps. The electrode was maintained at a positive bias of 0.4 V in the interpulse period. E
mc
maximum cathodal potential excursion (–1.4 V vs. Ag|AgCl); E
ma
maximum anodal potential excursion (1.2 V vs. Ag|AgCl), and Va access voltage (0.4 V vs. Ag|AgCl).
Figure 3.4 Voltage transients for an SIROF electrode pulsed wih a balanced (150/150 mA) and two unbalanced biphasic current wveforms (125/150 μA and 120/150 μA). Polarization of the SIROF electrode to potentials negative of the water reduction potential occurs abruptedly as the imbalance is increased from 25 μA to 30 μA.
Figure 3.5 The effect of pulse frequency on the polarization for an SIROF electrode subjected to unbalanced biphasic pulsing. Each line represents the range of currents over which the polarization of the SIROF was observed to transition from potentials within the water window to potentials at which water electrolysis occurs. Data for both an anodal and cathodal imbalance are shown.
Figure 3.6 Voltage transient in response to a monophasic cathodal current pulse with charge-balance achieved by controlling the interpulse potential (E
bias
) of the electrode to 0.6 V (vs. Ag|AgCl) in the interpulse period.
Figure 3.7 SEM images of (A) bare Pt electrodes, (B) coated with PEDOT/pTS at 200× magnification, and (C) PEDOT/pTS at 1000× magnification (reproduced with permissions from [78]), shows the nodular structure of a typical CP film.
Figure 3.8 Cyclic voltammetry showing voltage response of Pt compared to PEDOT/pTS. Charge storage capacity is the integration of the current with respect to time. These curves were obtained in saline versus Ag/AgCl reference electrode with the voltage ramped at 150 mV/s.
Figure 3.9 Electrodes fabricated from CNTs. (A) A CNT microelectrode array (MEA), fan out wires and contact pads [109]. (B) Patch clamp recordings made from ganglion cells in an excised chick retina. Recordings were made following stimulation with the MEA shown in (A). The small arrow in (B) indicates the artifact coincident with the stimulation pulse. Following are a train of natural retinal ganglion depolarization spikes. (C) An SEM image of curly intertwined CNTs upon which rat neurons have been cultured [111]. (D) Higher magnification of the section indicated in (C) by the dashed lines. The fiber shown colored in blue and indicated by arrows is a neuronal process that has grown into and become intertwined with the nanotubes.
Figure 3.10 Proposed design of an all-diamond retinal prosthesis comprising a diamond capsule including a diamond back (b), a diamond feed-through and electrode array (c) with gold internal wiring. A dedicated stimulator chip (a) is encapsulated in the diamond case. (d) and (e) show rear and front views of the assembled device. The N-UNCD electrode array designed to sit directly on the retina is shown in (e).
Chapter 4: Principles of Recording from and Electrical Stimulation of Neural Tissue
Figure 4.1 Schematic of a myelinated and unmyelinated peripheral neuron demonstrating the basic structure and anatomy. Also demonstrated is the propagation of an action potential, and resulting accumulation of ions, along the axon of each of the neurons.
Figure 4.2 Example of changes in potassium and sodium conductances during an action potential as derived from the Hodgkin-Huxley equations (Hodgkin and Huxley 1952).
Figure 4.3 Amplitudes and bandwidth of several common biological signals.
Figure 4.4 Example of recording electrodes from patch-clamp to multi-electrode arrays. There are a number of variations of “patch-clamp” recordings that vary, depending on which part of the cell membrane is attached to the glass micropipette and the remainder of the cell. (A and B) Glass microelectrodes are typically made from quartz glass that has been pulled to a sharp tip and then filled with high molarity potassium chloride solution, while metal microelectrodes (C) are typically insulated with a thin covering of parylene-C. (D–F) Examples of multi-electrodes arrays from (D) Multichannel System, (E) NeuroNexus and (F) Blackrock Microsystems.
Figure 4.5 Examples of far-field recording techniques. Stylised ABR (A) and CAEP (B) recordings demonstrating the standard recording configuration of differential recording from scalp electrodes relative to a ground electrode (GND). The individual waves (I – V) in the ABR are generated by different sub-thalamic nuclei, as illustrated by the arrows in A. The waves in the CAEP (Na, Pa, Nb, P1, N1 and P2) represent different aspects of cortical processing. (C) Typical stages of amplification and signal conditioning.
Figure 4.6 (A) Schematic of a constant current, biphasic output pulse train for a typical neural stimulator. (B) Representation of monopolar and bipolar stimulation modes typically used by neural stimulators.
Figure 4.7 A point source of current in uniform tissue containing the axon of a neuron. Dotted circles with marked voltages are the spheres of constant voltage at various radii from a central point source of constant current. The approximate voltages appearing at five nodes of Ranvier for the section of neuron shown are indicated at the bottom of the picture.
Figure 4.8 A myelinated neuron with an idealised schematic of a the McNeal equivalent circuit shown underneath
Figure 4.9 (A) Typical strength-duration curve showing the threshold activating pulse width versus the stimulus current. The rheobase current (Ir) and the chronaxie pulse width (t
c
) are also illustrated. (B) Typical recruitment curve illustrating the threshold and maximum responses and the dynamic range.
Figure 4.10 Graph of the region where Shannon reported damage was observed (shaded). The region now considered safe in cochlear implants is indicated by dotted lines and arrows.
Chapter 5: Wireless Neurotechnology for Neural Prostheses
Figure 5.1 “Nested Doll” representation of a wireless neural implant as a system, highlighting multiple subsystems and their interfaces.
Figure 5.2 Schematic of electronic subsystems in a transcutaneous wireless link between an active neural implant, external peripherals and further downstream links.
Figure 5.3 One of the first headmounted wireless 32-channel recording devices (Miranda
et al
. 2010) for recording from electrode arrays implanted in rhesus macaque monkeys freely moving in their cages The active electronics of this system exploit selected features of a monolithic integrated neural interface (INI) microchip design which amplifies, digitizes and, in one implementation, transmits neural data across a ∼900 MHz wireless channel (Harrison 2010). On a single 2 A-hr battery pack, this system runs contiguously for several days. The total weight of this system including the batteries is 114 g, considerably heavier than human-wearable devices such as earpiece-mounted personal communication packages.
Figure 5.4 Three case examples of wireless high-data rate brain interface recording devices. (a) An implantable, polymer packaged device (Song
et al
. 2009) with an inductive coupled transcutaneous link for powering and near infrared signal transmitting microlaser. An analog preamplifier/multiplexer ASIC is flip-chip integrated with the intracortical microelectrode array; (b) an implantable 100-channel wireless neurosensor with RF telemetry (3–4 GHz link), housing all active microelectronics and a rechargeable Li-ion battery in a hermetically sealed Ti-enclosure weighing <50 g (Borton
et al
. 2013); (c) a 100-channel externally headmounted, battery-powered wireless neurosensor with RF telemetry up to 200 Mbits/sec (likewise with telemetry in the 3–4 GHz range and below 50 g in weight (Yin
et al
. 2014).
Figure 5.5 Distribution of the electronic and optoelectronic payloads of the implant of Figure 5.4a, with near infrared transcutaneous signal transmission. Power is supplied transcutaneously, either by continuous inductive coupling or by optically powered high efficiency custom multijunction tandem photocell
Figure 5.6 Implant placement into a NHP with the cortical front end module linked electrically by short flexible polymer ribbon cable to subcutaneous the back end module.
Figure 5.7 Block diagram of the RF telemetry scheme (and battery powering) of the wireless devices of Figure 5.4b and c; both using the same high efficiency, low-power custom RF transmitter ASIC.
Figure 5.8 Block diagram of a RF superheterodyne receiver, with neural data extracted by demodulation and passed through a dedicated neural signal processor to a laptop computer for decoding and storage.
Figure 5.9 Generic schematic of inductive coupling, here in the context of transcutaneous charging of a rechargeable battery through the transparent sapphire window of the Ti-enclosure.
Figure 5.10 Schematic of the circuits within a Ti-enclosure deployed in controlled charging of the on-board Li-ion battery (Yin
et al
. 2013a).
Figure 5.11 Two specific challenges in hermetic sealing. (a) sapphire window brazed to the Ti-enclosure; (b) arrays of metal-ceramic feed-throughs at the bottom of the Ti-can, to which wiring from a Kapton thin film fanout of the inputs from 100 wire bundle from neural sensor are wirebonded. The Kapton film is packaged by a PDMS silicone mold.
Figure 5.12 PDMS-based packaging of a neural implant. Upper image: profile of the implant with preamplifier/multiplexer ASIC is integrated with the neural sensor. Lower image illustrates the flexibility of a 7-wire ribbon cable connecting the cortical front end to the telemetry back end (Song
et al
. 2007).
Figure 5.13 A Yucatan minipig, implanted with the subcutaneous wireless neural sensor implant, engaged in foraging with a fully instrumented wireless feeder device, which synchronizes the animals intent and action to find food (from wells under sets of movable slats) with simultaneously recorded multichannel broadband neural data (Agha
et al
. 2013).
Figure 5.14 Process flow of the design for use of the Arduino microprocessor board and the application of the FPGA (Komar
et al
. 2014).
Figure 5.15 Wirelessly recorded data from an implanted neural interface in NHPs. (a) A selection of 15 of the 100 broadband recording channels richness of high-sample rate (20 ksps) data collection. Spikes were extracted (b) on all channels (c), showing single neuron activity on input channels. Raster plot (12 s) marking spike timestamps for all input channels is shown with behavior indicated by color overlay. In (e), the monkey reaches for food while neural data is recorded wirelessly. Spikes across all channels are reduced to a low-dimensional state space (f) through principal component analysis, where such neural trajectories produced during free movement of monkey scratching eye (blue), touching an apple (green) and turning head (purple) (Borton
et al
. 2013).
Figure 5.16 Sampling of
in vivo
recorded neural action potential from a rhesus macaque monkey sitting in chair.
Figure 5.17 (a) Overnight recording from an untethered NHP in home cage with a four antenna SIMO receiver system. (b) Wireless data represented as 3 hours of recorded power spectrograms below 50 Hz (top) and neuron firing rates histograms of a single neuron from two channels as field potential and firing rate patterns across sleep (grey) and awake (blue) cycles monitored by IR camera video recordings. (c) Raw neural data and firing rate histograms (1 s bin size) from the two channels over a 2-min sleep-to-wake transition. (d) Classification of neural states based on the action potential signals during the particular 2-min sleep-to-wake transition using PCA. (e) Comparison of average firing rates between asleep and awake states for 10 neurons with dominant contribution to the PCA data.
Figure 5.18 Futuristic concept of mobile, wireless neural prostheses, highlighting the functional roles of the subcutaneous wireless head implant together with the wearable external electronics for neural signal processing, decoding and wireless control of assistive neural prosthesis devices (the “BrainPhone” refers to an embedded microprocessor portable computing system, aimed at for real-time neural decoding by on-board algorithms). The wireless implant device is expected to also perform multichannel functional electrical stimulation.
Chapter 6: Preclinical testing of Neural Prostheses
Figure 6.1 Example of Arrhenius behavior of an implant. (A) Cumulative release of Dexamethasone from subcutaneous implants at five temperatures. The rate constant K is the slope of the plots after the initial fast release phase. (B) Arrhenius plot of Log (K) vs. 1/T, where T is the absolute temperature. The rate constant measured at 37°C (310
o
Kelvin) corresponds to the value obtained by extrapolation of the plot obtained at elevated temperatures (reprinted from Shen and Burgess 2012, with permission from Elsevier).
Figure 6.2 Examples of failure modes of an active implant caused by internal and external stresses, and by degradation of the electrical insulation in an aquaeous invironment
Figure 6.3 (A) Relation between charge density, charge per phase and neuronal injury after 8 hours of electrical stimulation applied through platinum electrodes implanted on cat cerebral cortex. Damaged neurons appeared dark, with shrunken somas, as shown in the inset (adapted from McCreery
et al
. 1990). (B) Numeric representation of the region of the graph shown in A, for which injury to neuron did or did not occur (“safe” region) (from Shannon 1992, with permission from IEEE).
Figure 6.4 A suggested protocol for monitoring the evolution of “biotic” and “abiotic” factors that contributes to the failure of a neural implant. The scheme incorporates metrics of the functional status of the implant, metrics linked to its functional status, and also metrics of the biological changes and status in the adjacent tissue that may affect the device's function (from Prasad
et al
. 2012, with permission from IOP Publishing).
Figure 6.5 (A) Light micrograph of a tissue section through the site of a microelectrode implanted chronically in cat cerebral cortex. The inset shows a photo of the microelectrode array and a scanning electron micrograph of one of the microelectrode tip sites. Neurons identified by the neuron-specific label NeuN are shown as red-brown profiles against the background of astrocyte processes stained with the astrocyte marker GFAP. (B) The density of cerebral cortex neurons surrounding 24 unstimulated microelectrode sites, and 21 sites from the same arrays stimulated for 240 h at 4 nC/phase, 50 Hz and ∼200 μC/cm
2
and with 100% duty cycle. (C) The density of neurons around 24 electrode sites stimulated with a 50% duty cycle, also at 50 pps, 4 nC/phase and 200 μC/cm
2
. (D) Neuron density around 8 microelectrodes pulsed with 100% duty cycle, 50 pps, 2 nC/phase and ∼100 μC/cm
2
(from McCreery 2010, with permission from IOP Publishing).
Figure 6.6 Injury to the myelinated axons of a peripheral nerve (cat sciatic nerve) induced by 8 h of continuous electrical stimulation. (A) The electrode array and nerve at the time of the implant surgery. (B) Light micrographs of cross-sections of a normal unstimulated sciatic nerve and of a stimulated nerve, showing the appearance of normal myelinated axons and damaged axons undergoing EAD. (C) Compound action potential recorded from a cat's sciatic nerve in response to electrical stimulation with the cuff electrode. The α component represents the activity in the largest and most rapidly conducting axons. (D) Plots of the relation between stimulus pulse amplitude and the % of axons undergoing EAD, with stimulus pulse rate as a parameter. The stimulus current is expressed as multiples of α, the stimulus amplitude required to excite all of the nerves' large myelinated axons (B,D from McCreery
et al
. 1995, with permission from IFMBE. C from Agnew
et al
. 1990b,
w
ith permission from Butterworth-Heinemann for BES).
Figure 6.7 Comparison of OCT image of retinal swelling with other indicators of cytotoxic damage. (A) OCT scan of retina region stimulated at 442 μC cm
2
shows swelling of the ILM and IPL under the electrode. (B) Low power light micrograph of the retinal surface after the stimulation electrode was removed, showing only a barely visible circular retinal opacity (arrow). (C) The same region as in (B), now stained with the vital dye PI. The fluorescence image reveals a disk-shaped region of damage at the location of the stimulating electrode, with bright specks at the retinal surface indicating that the local neurons were damaged. (D) Intensity profile of the florescence across the region shown in C (modified from Cohen
et al
. 2011, with permission from IOP Publishing).
Chapter 7: Auditory and Visual Neural Prostheses
Figure 7.1 Transverse section of the mammalian cochlea illustrating the three fluid-filled cochlear ducts (scala vestibule, scala media and scala tympani) spiralling around the modiolus containing the auditory nerve. The soma of the auditory nerve fibres, the spiral ganglion neurons (SGN), are located in a small bony canal (Rosenthal's canal) that also spirals from the cochlear base to its apex (arrowheads). There are ∼30 000 SGNs in a normal human cochlea. A cochlear implant electrode array is inserted into the scala tympani at the cochlear base either via the round window, or a membrane separating the scala tympani and the middle ear (not illustrated), or via a small hole drilled into the bone overlying (double arrowhead). The boxed area of the basal turn is illustrated in more detail in (b). (a) Schematic illustrating the afferent innervation pattern in the mammalian cochlea. SGN soma reside in Rosenthal's canal (arrowhead) and their peripheral processes (double arrowhead) project to the hair cells within the Organ of Corti. (b) Type I SGNs (I) make direct synaptic contact with a single inner hair cell (iH). In contrast, Type II SGNs (II) project across the floor of the basilar membrane (B) and make synaptic contact with numerous first, second and third row outer hair cells (oH). HA – Habenular perforata, OSL – osseous spiral lamina (adapted from Spoendlin 1984, reprinted with permission Elsevier Science).
Figure 7.2 Schematic diagram of a single cochlear turn illustrating the degenerative changes that occur following loss of the Organ of Corti. (a) Normal cochlea, illustrating the three fluid filled chambers (scala tympani, st; scala media, sm; scala vestibule, sv), the Organ of Corti containing the sensory hair cells (arrowhead), and Rosenthal's canal (rc) containing the spiral ganglion neuron (SGN) soma and their peripheral processes projecting to the Organ of Corti to synapse with the sensory hair cells. (b) The most common forms of SNHL target the hair cells of the Organ of Corti, eventually resulting in the degeneration of the Organ of Corti. Hair cell loss initially induces loss of SGN peripheral processes (arrow) that would normally innervate the Organ of Corti. (c) More gradual degenerative changes following hair cell loss includes extensive loss of peripheral processes and gradual loss of SGNs. This is typically the status of the cochleae receiving a cochlear implant. The typical position of the electrode array within the scala tympani is illustrated in this panel (adapted from Shepherd
et al.
2013b, with permission).
Figure 7.3 A modern cochlear implant system illustrating the (a) scala tympani electrode array containing platinum electrode contacts (arrow); (b) implant receiver-stimulator, the electrode array (ea); a monopolar return electrode (m), a hermetically sealed titanium case containing the stimulator electronics (arrow) and a magnet (mg) to align the external radio frequency (RF) coil (double arrowheads). The implant package is ∼15 mm across and 7 mm thick; (c) the external sound processor with the external RF coil and magnet; and (d) a schematic illustrating the surgical placement of the system together with the external speech processor and RF coil. Images courtesy of Cochlear Ltd.
Figure 7.4 Schematic illustrating the two most common electrode geometries used in cochlear implants. In a monopolar configuration, a scala tympani electrode proximal to the SGN population is stimulated against a remote large surface area electrode. In contrast, bipolar electrodes incorporate a single adjacent return electrode. Although monopolar stimulation is efficient, it also results in greater current spread compared with bipolar stimulation. Nevertheless, the majority of patients can identify electrodes via an orderly change in pitch using monopolar stimulation (images courtesy of Fallon
et al
. 2009, with permission).
Figure 7.5 Schematic diagram of the human auditory pathway. Five stimulation sites have been trialed clinically over the course of auditory prosthesis development (arrowheads) including: (i) the cochlea; (ii) the auditory nerve; (iii) the cochlear nucleus; (iv) the inferior colliculus; and (v) the primary auditory cortex. To date, only cochlear and cochlear nucleus based devices have received regulatory approval for clinical use (adapted from Purves
et al.
2001, with permission from Sinauer Associates, Inc.).
Figure 7.6 The anatomy of the eye in normal vision and following loss of photoreceptors. (a) Schematic of a normal eye. The boxed area of the retina is enlarged in panels b–d. (b) Schematic of a normal retina, choroid and sclera. The retina consists of several processing layers extending from the rods and cones of the outer retina through bipolar cells of the middle retina to the retinal ganglion cells (RGCs) that make up the inner retina. Axons of the RGCs project via the optic disc to form the optic nerve. (c) Schematic of a retina with widespread photoreceptor degeneration. This pathology also results in the remodelling of the neural architecture within the inner retina and shrinkage of the choroid. (d) Potential sites for the location of an electrode array for a retinal prosthesis including epiretinal, sub-retinal and suprachoroidal locations. Images courtesy of Bionic Vision Australia (image by C. Roce).
Figure 7.7 (a) Schematic diagram of a generic retinal prosthesis illustrating a receiver-stimulator unit implanted in the mastoid bone behind the ear (arrow), a leadwire assembly (arrowhead) connecting the output of the stimulator to an array of electrodes (e) implanted in the retina. The electrode array can be tacked to the retina (epiretinal); inserted between the choroid and the retina (subretinal); or inserted between the sclera and the choroid (suprachoroidal). (b) Overall schematic of a retinal prosthesis that includes a video camera incorporated onto glasses (arrow), an external vision processor (vp) that provides both data and power across the skin via a wireless link (w) to the implanted receiver-stimulator, leadwire and electrode array illustrated in (a). The camera continuously feeds video signals to the vision processor that contains the patient's phosphene map, visual processing algorithms and stimulation strategies. Each frame of the input video generates a sequence of commands at the vision processor that defines the electrodes and stimulus parameters required to generate a prosthetic image of the scene. (Images by Jack Parry; courtesy of the Bionics Institute).
Figure 7.8 Overview of the visual pathway from the retina to the primary visual cortex in the human. Visual prostheses can potentially target several sites along this pathway (arrowhead) including the retina, the optic nerve, lateral geniculate nucleus and the visual cortex. The retina, optic nerve and visual cortex have been trialed clinically over the course of visual prosthesis development. The majority of these structures have a well organised topographic map of the retina (adapted from Hubel 1988, with permission).
Chapter 8: Neurobionics: Treatments for Disorders of the Central Nervous System
Figure 8.1 A typical configuration of a DBS system: (A) sketch showing the main implanted components; (B) implantable pulse generator; (C) electrode array with four stimulation contacts
Figure 8.2 Sketch showing approximate positions of the principal anatomical targets used in DBS (A) part of sagittal section as indicated in the inset picture at upper left; (B) coronal section as indicated in the inset.
Figure 8.3 An image from a computed tomography scan showing DBS electrode arrays (center), leadwires, and connectors attached to extension cables (left).
Chapter 9: Brain Computer Interfaces
Figure 9.1
(
A) Example of neuron tuned for wrist movement. In the pioneering work done by Evarts, an electrode was placed into the M1 of an NHP trained to apply a torque to a bar. Spike rates for the recorded neuron increased and decreased as the wrist performed wrist movements (from
Principles of Neuroscience
, 5th edition, p. 848, with permission). (B) A NHP was trained to move a manipulandum to one of eight different positions. The firing rate of the neuron was modeled using a cosine function, relating the angular arm displacement to the firing rate
Figure 9.2 Block diagram of iBCI architecture
Figure 9.3 Schematic describing the conceptual steps involved in obtaining neural information. An electrode is placed into the cortex, which records voltage at a high sample rate. The voltage signals are band
-
pass filtered, and the spikes are identified where the filtered voltage exceeds some threshold. Multiple algorithms exist for spike sorting, that is, identifying specific waveforms of voltage signals corresponding to neuronal depolarizations
Figure 9.4 Conceptual drawing of a iBCI set-up. An implanted sensor (in this figure, Utah array) is implanted into the M1 of an individual with motor impairment. The neural signal is transmitted from the electrode to a pedestal fixated to the skull. Wires from the pedestal then descend to the data acquisition hardware located on the wheelchair, and subsequently to an external computer for additional processing. In a closed-loop setting, information from the decoded signal is interpreted and the real-time results are provided to the participant as feedback
Figure 9.5 iBCI-based neural control of a robotic arm in a person with LIS. In this photograph, the iBCI user is bringing herself a cup of coffee. This was the first time in nearly 15 years that she was able to take a drink solely of her own volition
Chapter 10: Taking a Device to Market: Regulatory and Commercial Issues
Figure 10.1 Overview of the commercialization process for medical devices. The four phases: (i) basic research, (ii) preclinical development, (iii) clinical trials, and (iv) regulatory approvals, are shown along with the major milestones which need to be accomplished in the phases. The value multiple is how much commercialization effort may be worth as a multiple of the investment and the time taken and a typical invested amount. The quality standards which operate over the three phases are also indicated (good laboratory practice (GLP), good clinical practice (GCP) and good manufacturing practice (GMP)).
Figure 10.2 The neuromodulation/neurobionic industry landscape. To the left of the Figure are the currently approved therapies and on the right the therapeutic targets currently being developed.
Figure 10.3 An example assessment of performance dimensions for a bionic eye. The dimensions (see text for explanation) are drawn in a spider diagram. Lines indicating the differences between non-treatment (shaded area), minimal performance required and the performance criteria required to expand the market. QALY, quality-adjusted life year; MTBF, Mean Time Before Failure (a measure of the reliability of the device); TCO (i.e. the sum of all costs for the therapy, including running and maintenance costs).
Figure 10.4 Basic research flow diagram illustrating the stages of development from identification of the therapeutic need through idea generation and then to market.
Figure 10.5 The number of patents in the neuromodulation industry over five year epochs from 1980 to 2010 categorised into product performance dimensions.
Figure 10.6 Spider diagram showing the number of patents in each technology area over five-year epochs from 1995 to 2010 and the total number of patents for each technology area from 1980 to 2010.
Figure 10.7 Application of Design Controls to Waterfall Design Process adapted from Figure 1 in http://www.fda.gov/medicaldevices/deviceregulationandguidance/guidancedocuments/ucm070627.htm
Figure 10.8 FDA classification of medical devices. Neurobionic devices are typically considered Class III devices and must pass through the FDA's PMA process in order to get into clinical trial.
Figure 10.9 Approval times and market access for medical devices in a number of jurisdictions. The Figure was redrawn from data in Basu and Hassenplug (2012). CMS refers to the Centers for Medicare and Medicaid Services, which is a US government health insurance program that covers 100 million people.
Figure 10.10 Funding sources, expected returns and amounts.
Chapter 11: Ethical Considerations in the Development of Neural Prostheses
Figure 11.1 ICF model and interaction of components.
Figure 11.2 The MPT model emphasizes evaluation of three distinct but interrelated components which have been found to influence the use of ATDs (See Figure 11.2).
List of Tables
Chapter 1: The Historical Foundations of Bionics
Table 1.1 Packaging Methods discussed at Stanford Meeting in 1979
Table 1.2 Implanted Devices reported from the MRC Neurological Prostheses Unit
Chapter 3: Electrochemical Principles of Safe Charge Injection
Table 3.1 Representative charge/phase (μC/ph) and charge density (μC/cm
2
) values for damaging and non-damaging levels of stimulation with planar platinum electrodes
Table 3.2 Water reduction and oxidation potentials on electrode materials at pH 7 vs. Ag|AgCl
Table 3.3 Comparison of saline and
in vivo
charge injection capacities (μC/cm
2
)
Chapter 9: Brain Computer Interfaces
Table 9.1 Sources of neural information for brain computer interfaces
Chapter 10: Taking a Device to Market: Regulatory and Commercial Issues
Table 10.1 The progression from performance categories to user requirements specifications verification and validation using a neuromodulation lead as an example of this process
Chapter 11: Ethical Considerations in the Development of Neural Prostheses
Table 11.1 The four basic principles in bioethics
NEUROBIONICS
The Biomedical Engineering of Neural Prostheses
Edited by
ROBERT K. SHEPHERD
Bionics Institute & The University of Melbourne, Australia
Copyright © 2016 by John Wiley & Sons, Inc. All rights reserved
Published by John Wiley & Sons, Inc., Hoboken, New Jersey
Published simultaneously in Canada
No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission.
Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.
For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002.
Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com.
Library of Congress Cataloging-in-Publication Data:
Names: Shepherd, Robert K., editor.
Title: Neurobionics : the biomedical engineering of neural prostheses / edited by Robert K. Shepherd.
Other titles: Neurobionics (Shepherd)
Description: Hoboken, New Jersey : John Wiley & Sons, Inc., [2016] | Includes bibliographical references and index.
Identifiers: LCCN 2016002807 (print) | LCCN 2016004110 (ebook) | ISBN 9781118814871 (cloth) | ISBN 9781118816141 (pdf) | ISBN 9781118816035 (epub)
Subjects: | MESH: Nervous System Physiological Phenomena | Bionics | Neural Prostheses | Brain-Computer Interfaces
Classification: LCC R857.M3 (print) | LCC R857.M3 (ebook) | NLM WL 102 | DDC 610.28–dc23
LC record available at http://lccn.loc.gov/2016002807
Cover image: Getty/Hemera Technologies
Inset image: Inset Photo used with permission of the Bionics Institute, East Melbourne, Australia
This book is dedicated to my wife, Ursula, for her wonderful support, encouragement and counsel over the last 40 years; to our children Damon and Anna; their partners Jo and Junior; and our grandchildren Harley, Michaela, Jordan and Heidi who enrich our lives daily.
Contributors
David Borton
Department of Engineering and Physics, Brown University, Providence, RI, USA
David M. Brandman
Department of Neuroscience, Brown University, Providence, RI, USA
Giles S. Brindley
(Retired) Implanted Devices Group, Department of Medical Physics & Bioengineering, University College London, London, UK
Paul M. Carter
Cochlear Ltd, Macquarie Park, NSW, Australia
Stuart F. Cogan
Department of Bioengineering, University of Texas at Dallas, Richardson, TX, USA
Nick Donaldson
Implanted Devices Group, Department of Medical Physics and Bioengineering, University College London, London, UK
James B. Fallon
Bionics Institute & Medical Bionics Department, University of Melbourne, East Melbourne, Victoria, Australia
David J. Garrett
Department of Physics, The University of Melbourne, Parkville and The Bionics Institute, East Melbourne, Victoria, Australia
Rylie A. Green
Graduate School of Biomedical Engineering, UNSW, Sydney, NSW, Australia
Leigh R. Hochberg
Department of Neuroscience, Brown University, Providence, RI, USA
Frank J. Lane
Illinois Institute of Technology Rehabilitation Psychology, Chicago, IL, USA
Kristian P. Nitsch
Department of Clinical and Rehabilitation Psychology Lewis College of Human Sciences Illinois Institute of Technology Rehabilitation Psychology Chicago IL, USA
Arto Nurmikko
Department of Engineering and Physics, Brown University, Providence, RI, USA
Douglas McCreery
Neural Engineering Program, Huntington Medical Research Institutes, Pasadena, CA, USA
Hugh McDermott
Bionics Institute & Medical Bionics Department, University of Melbourne, East Melbourne, Victoria, Australia
John L. Parker
Saluda Medical Pty Ltd, Artarmon, NSW, Australia
Marcia Scherer
University of Rochester Medical Center, Rochester, NY, USA
Jens Schouenborg
Neuronano Research Center, Experimental Medical Science and Nanometerconsortium, Lund University, Lund, Sweden
Peter M. Seligman
Bionics Institute & Medical Bionics Department, University of Melbourne, East Melbourne, Victoria, Australia
Robert K. Shepherd
Bionics Institute & Medical Bionics Department, University of Melbourne, East Melbourne, Victoria, Australia
Mohit N. Shivdasani
Bionics Institute & Medical Bionics Department, University of Melbourne, East Melbourne, Victoria, Australia
Ming Yin
Blackrock Microsystems, Salt Lake City, UT, USA
Preface
Neural prostheses are active implantable devices designed to: (i) provide therapeutic intervention, sensory feedback or motor function via electrical stimulation of nerves or muscles following trauma or disease; and/or (ii) record the electrical activity from nerve or muscle to detect disease states, enable the voluntary control of external devices such as prosthetic limbs, or to provide closed-loop feedback to modulate neural prostheses.
Since the introduction of the first commercial heart pacemakers in the late 1950s, there have been many devices approved for clinical use, resulting in a dramatic impact on the quality of life of millions of people around the world. Implantable heart pacemakers and defibrillators are a multi-billion dollar per annum industry. While the neural prosthesis industry is much younger, with an early wave of commercial devices appearing in the late 1970s, this is now a flourishing industry with impressive annual growth rates (Cavuoto et al. 2016). Four devices dominate this field: spinal cord stimulation for treatment of chronic pain; cochlear implants for stimulation of the auditory nerve in deafness; vagal nerve stimulation to treat epilepsy; and deep brain stimulation (DBS) to control motor disorders associated with Parkinson's disease and essential tremor.
Significantly, the development of neural prostheses is currently undergoing unprecedented expansion. There are a large number of devices in development or an early stage of commercialisation. These include visual prostheses for stimulation of the retina or visual cortex in blind patients; functional electrical stimulation to provide coordinated activation of nerve and muscle to assist with movement of the hand, arm and gait in stroke and spinal cord injury; DBS to treat pain, epilepsy or severe depression and related psychiatric disorders; vestibular prostheses to assist patients with balance disorders; and neural interfaces that record from the central or peripheral nervous system to monitor for the onset of seizures or to control external devices for amputees and severe spinal cord injured patients.
Recently neural prostheses have experienced an exciting new phase of innovation generated by the Obama Brain Initiative that encompasses the National Institutes of Health and the Defense Advanced Research Projects Agency, as well as GlaxoSmithKline's entry into the field to develop “electroceutical” techniques (Birmingham et al. [2014]). These initiatives call for greater multidisciplinary collaboration, including the development of detailed anatomical and physiological maps of neural circuits associated with disease and treatment combined with neural modelling to optimise the development of therapeutic stimulation strategies. While outside the scope of this book, we will watch with great interest as outcomes from these initiatives are delivered to the clinic over the next decade.
Given the multidisciplinary nature of neural prostheses, the field has adopted multiple terminologies that are reflected across the 11 chapters. “Bionics”, “medical bionics” or “neuroprosthesis” are used synonymously here with “neural prostheses”. We have used additional application-specific terms: “neuromodulation” refers to the stimulus-induced modulation of neural activity for therapeutic purposes – DBS for the control of motor symptoms associated with Parkinson's disease, or spinal cord stimulation to alleviate back pain are examples; “functional electrical stimulation” refers to stimulation of peripheral nerve and muscle to assist in the movement of limbs following paralysis; “sensory neural prostheses” refers to devices that operate under sensory control such as cochlear (auditory) and retinal (vision) implants; “neurobionics” refers to neural stimulation treatments for disorders of the central nervous system (e.g. DBS for the treatment of movement disorders, epilepsy and pain); and “closed-loop” describes a feedback mechanism, typically based on electrophysiological recordings, used to modify the electrical stimulation parameters delivered via a neural prosthesis for improved efficacy.
New developments in neural prostheses are built on advances in electronics, materials science, electrochemistry, battery technology, neuroscience, clinical and surgical practice, and rehabilitation techniques. This book provides a comprehensive historical overview of the field (Chapter 1); it covers the key sciences underpinning the technology including the electrode-tissue interface (Chapter 2); electrochemical principles of safe electrical stimulation (Chapter 3); principles of recording from and stimulating neural tissue (Chapter 4); wireless technology (Chapter 5); and preclinical device testing (Chapter 6). Subsequent chapters describe specific clinical applications, citing devices that are both commercially available and in development, including cochlear implants and vision prostheses (Chapter 7); neurobionics in the treatment of Parkinson's disease, severe depression, obsessive compulsive disorder, pain and epilepsy (Chapter 8); and brain machine interfaces for the control of external devices such as prosthetic limbs (Chapter 9). The final two chapters provide important insight into the process of regulatory approval and commercialisation – issues critical to the successful translation of research to the clinic (Chapter 10); and the key ethical considerations associated with the development of these devices (Chapter 11). Finally, the Appendix provides a list of companies and research organisations currently developing and/or manufacturing neural prostheses.
There are many individuals who have been instrumental in ensuring the successful completion of this book. I gratefully acknowledge the authors of all the chapters – it has been a privilege to work with such a professional and knowledgeable group of individuals without whose efforts and attention to detail this publication would not have existed. In acknowledging our authors I would like to highlight Professor Giles Brindley's contribution to the chapter on the historical foundations of bionics (Chapter 1). Professor Brindley is a pioneer of the field – developing the first visual prosthesis in the 1960's (Brindley and Lewin [1968]) – it is to his great credit that almost 50 years after this seminal work – and now in his 90th year – he continues to make important contributions to the advancement of neural prostheses. I am very grateful to Berenice Hale, Lyndal Borrell and Lauren Hill from the Bionics Institute for providing important administrative assistance; Justin Jeffryes, Stephanie Dollan and Allison McGinniss from Wiley for their endless advice and support for the project; and finally I acknowledge the staff and students of the Bionics Institute for providing such a stimulating environment in which to work.
Robert K. ShepherdMelbourne, Australia.
References
Birmingham, K., Gradinaru, V., Anikeeva, P., Grill, W.M., Pikov, V. et al. (2014) Bioelectronic medicines: a research roadmap.
Nat. Rev. Drug. Discov.
,
13
: 399–400.
Brindley, G.S. and Lewin, W.S. (1968) The sensations produced by electrical stimulation of the visual cortex.
J.Physiol
.,
196
: 479–493.
Cavuoto, J. (2016) The market for neurotechnology: 2016–2020,
Neurotech Reports
, 1–350.
Part IFundamentals of neural prostheses
Chapter 1The Historical Foundations of Bionics
Nick Donaldson and Giles S. Brindley
Implanted Devices Group, Department of Medical Physics & Bioengineering, University College London, London, UK
1.1 Bionics Past and Future
In 1973, Donaldson and Davis published a paper called “Microelectronic devices for surgical implantation” in which they listed neuroprostheses in use and under development: pacemakers for the heart (fixed-rate, atrial-triggered and demand), incontinence devices, visual prostheses, dorsal column stimulators and electromyogram (EMG)) telemeters1. The field of bionics was then very young, the idea of surgically implanting an electronic device was new and very few people had worked on the technical difficulties entailed. Only pacemakers were then commercial products and there were no regulations in force. Now, 40 years later, there are many more types of device, both in clinical use and under development. A number of these devices will be described in Chapters 7–9 and include implants for addressing sensory loss (e.g. hearing, sight, balance), disorders of the brain and the mind (e.g. epilepsy, migraine, chronic pain, depression), as well as brain-machine interfaces. Manufacturing these devices and going through the process of regulation is now a multi-billion dollar industry.
The year 2013 may be remembered as the year in which GlaxoSmithKline (GSK) announced that they were to invest in the development of neurobionic devices, which they call Electroceuticals or Bioelectronic Medicines2 (Famm et al. 2013; Birmingham et al
