Neurosurg Focus 28 (5):E25, 2010

Brain-computer interfaces: military, neurosurgical, and ethical perspective

*Iv a n S. Ko t c h e t k o v , B.A., Br i a n Y. Hw a n g , B.A., Ge o f f r e y Ap p e l b o o m , M.D., Ch r is t o p h e r P. Ke l l n e r , M.D., a n d E. Sa n d e r Co n n o l l y Jr., M.D. Department of Neurological Surgery, Columbia University College of Physicians and Surgeons, New York, New York

Brain-computer interfaces (BCIs) are devices that acquire and transform neural signals into actions intended by the user. These devices have been a rapidly developing area of research over the past 2 decades, and the military has made significant contributions to these efforts. Presently, BCIs can provide humans with rudimentary control over computer systems and robotic devices. Continued advances in BCI technology are especially pertinent in the military setting, given the potential for therapeutic applications to restore function after combat injury, and for the evolving use of BCI devices in military operations and performance enhancement. Neurosurgeons will play a central role in the further development and implementation of BCIs, but they will also have to navigate important ethical questions in the translation of this highly promising technology. In the following commentary the authors discuss realistic ex- pectations for BCI use in the military and underscore the intersection of the neurosurgeon’s civic and clinical duty to care for those who serve their country. (DOI: 10.3171/2010.2.FOCUS1027)

Ke y Wo r ds • brain-computer interface • military neurosurgeon • ethics • brain-machine interface

r a i n -c o m p u t e r interfaces, also called brain-ma- tential applications of BCIs have greatly exceeded the chine interfaces or neural interface systems, rep- current state of the technology. Nevertheless, practical resent a direct communication pathway between and clinically useful BCIs are increasingly becoming a theB brain and an external device.18,32,59 The devices used reality, and this has important implications for neuro- for the BCI acquire brain signals such as an EEG rhythm surgeons practicing and conducting research in military or electrophysiological recordings of neuronal firing settings. Our aim in this commentary is to use an un- and translate them into commands intended by the user. derstanding of current achievements in the field of BCIs Brain-computer interfaces accomplish this through novel to discuss realistic expectations for future adaptation of output pathways that do not use the normal conduits of BCI systems in military settings and to highlight the the .32 complex role of military neurosurgeons in the further During the past 40 years, BCIs have rapidly pro- expansion of this technology. The implications of neu- gressed from mere neuroscientific theory into a rudimen- rosurgical BCI implantation for restoring function to in- tary yet highly promising technology. Increasing levels jured soldiers as well as the potential to enhance military of brain-derived control have been attained in nonhuman training and operations are considered from an ethical primates and in humans.13,21,32,48,50,51,54,55 Moreover, rapid perspective. In the context of our national duty to support progress of supporting technologies from the fields of those who serve our country, we reinforce the importance computational , biomaterial engineering, of physician beneficence and nonmaleficence in any BCI and computer processing have significantly contributed application, alongside responsible and just distribution of to the ongoing development of BCIs. Importantly, an the technology. ever-increasing understanding of motor, cognitive, and sensory functions through cortical mapping has led to improvements in device designs as well as a wider gamut State of BCI Technology at Present of possible BCI applications.32,42,43 For BCIs to translate the user’s intentions accurately Media hype and popular imagination regarding po- into actions, “learning” must take place on both ends of the interface: the user must modulate his or her brain sig- nals to improve performance of the BCI, while the device Abbreviations used in this paper: BCI = brain-computer interface; DARPA = Defense Advanced Research Projects Agency; ECoG = must identify, interpret, and adapt to the neural signals ; EEG = . that are most predictive of the desired output. This is * Mr. Kotchetkov and Mr. Hwang contributed equally to this achieved through feedback and fine-tuning mechanisms work. that are similar to those used when learning a new mo-

Neurosurg Focus / Volume 28 / May 2010 1

Unauthenticated | Downloaded 10/05/21 02:43 PM UTC I. S. Kotchetkov et al.

Fig. 1. Three classes of BCIs: their anatomical locations, advantages, and limitations. tor task.32 Some BCI designs rely on a training phase in at the cortical surface placed either above or below the which the subject performs a designated task and a com- dura mater.10,32,40,48 Platforms that belong to the latter 2 putational algorithm is employed to select the neuronal classes require neurosurgical implantation. signals that best correlate with execution of that task. A Using an EEG-based system, humans with motor de- code is generated for each command that can subsequent- bilities, including those that result from injury ly be used for control of an external device.31,39,40 Alter- or amyotrophic lateral sclerosis, have been able to control natively, a real-time adaptive algorithm can be employed a computer cursor in 2 dimensions.29,60 This technology during the learning phase to concurrently select for the has also been used by motor-intact individuals to com- signals that are most predictive of the user’s intentions mand robots to manipulate objects, and has the potential by continuously refining them based on comparisons of to be applied in operating limb prosthetics.3 The EEG de- past and intended trajectories.40,54,60 Recordings from vices, however, are fundamentally limited by their signal larger populations of , or neuronal ensembles, are content, which does not convey information about com- generally the preferred source for extracting useful and ponents of movement such as position and velocity, and relevant information to guide appropriate activity.2,6,40 Al- recordings are prone to interference from the electromyo- though input from a single can result in success- graphic activity of cranial musculature.10,32 ful BCI control,7,22 averaging neuronal signals over many Invasive BCIs, on the other hand, can acquire more trial sessions is often necessary for predicting behavior,46 informative signals that enable higher performance lim- and therefore synthesizing the electrical firing of neuronal its.50 For instance, human patients with locked-in syn- ensembles can remove the variability associated with us- drome are able to move cursors on a 2D keyboard to ing the input of a single neuron.2,40 As the user learns to communicate using typed messages after undergoing im- operate the BCI, neuronal plasticity leads to a tuning of plantation of electrodes that attract growth of myelinated ensemble signals such that activity in more discrete popu- nerve fibers.24,25 With the aid of a 96-microelectrode ar- lations of neurons becomes the best determinant of action ray that records signals from primary , tet- commands.8,30 raplegic patients have been able to move a 2D cursor as Depending on the source from which they derive well as to execute basic control over robotic devices, such their neural signals, BCIs can be classified into those that as opening and closing a prosthetic hand, years after their use noninvasive, invasive, and partially invasive platforms initial spinal cord injury.21 Subsequent reports on tetraple- (Fig. 1). Electroencephalography, which obtains electri- gic patients who were enrolled in a pilot clinical trial of cal signals from the scalp, has been the dominant method BCIs have demonstrated modest improvements in cursor of recording used for noninvasive BCIs due to its rela- control, thereby achieving greater functionality for practi- tive safety and practical technical requirements. Invasive cal tasks.13,26 BCIs retrieve signals from single-neuron recordings via More recently, ECoG has proven to be a useful tool in microelectrodes implanted in the cortical layers. Partially detecting input signals for BCIs.33,51 Unlike EEG, ECoG invasive BCIs use ECoG readings that come from sensors can detect high-frequency gamma wave activity that is

2 Neurosurg Focus / Volume 28 / May 2010

Unauthenticated | Downloaded 10/05/21 02:43 PM UTC Brain-computer interfaces: military, , and ethics the product of smaller cortical ensembles and correlates multijointed robotic arm to exert variable grip strength with discharge of action potentials from cortical neu- and perform 3D movements to feed themselves.55 These rons.19,32 Because they are not embedded in brain paren- results have been achieved with motor prostheses, which chyma, ECoG electrodes inflict less damage to the cortex continuously process cortical signals to guide movement and also experience less signal deterioration than invasive and speed.48 Their counterparts, communication prosthe- electrodes.32,34 In patients with intractable epilepsy who ses, which extract information from higher cortical ar- required invasive monitoring, ECoG signals have been eas about intended goals, have also been successful and used for 2D movement control at a level of performance promise to enhance BCI performance as it might be ap- similar to that achieved with invasive BCIs.51 Although plied to humans executing more complicated goal-orient- this approach has not been tested in patients with motor ed tasks.17,37,48,50 The end point of the ongoing advances impairment, it can be applied more safely than invasive in BCIs lies in the translation of such functionality from electrodes, and produces greater information content than preclinical studies to humans, and thus BCI research de- EEG systems.32 serves continued academic and clinical as well Although invasive and partially invasive BCIs hold as government funding to help those who risk their lives great potential for functional recovery, the current risks for their country. and limitations associated with device implantation pre- Combat casualties from the wars in Afghanistan and vent BCIs from widespread use. One of the major short- Iraq have a marked propensity to present with the types of comings of the current technology is associated with the injuries that may be addressed with new developments in loss of signal reliability over time. In response to damage BCIs. Possibly in connection with increased use of body incurred by microelectrode penetration of the cortex, mi- armor,15,28,44 the number of combat-related extremity inju- croglia and astrocytes begin a reactive process that en- ries has been on the rise, representing 54% of all combat sheaths the prosthesis and disrupts its initial impedance injuries and 63% of primary diagnoses for admission.41 properties.5,16,45,58 Neural and vascular damage at the site In a recent analysis, Masini and colleagues35 concluded of insertion can also lead to development of infections.5,23 that extremity injuries are the leading cause of disability As a result, the length of time over which an implant- among soldiers not returning to duty, and have the great- able electrode is able to produce signals is measured in est projected disability benefit costs. In addition, spinal months, and typically does not make it over 1 year with- cord injuries represent a significant proportion of casu- out significantly losing quality.48,53 Moreover, motion be- alties,56 occurring in nearly 10% of the entire caseload tween the implanted electrode and brain parenchyma, as of closed or penetrating head trauma seen by military is often caused by changes in brain volume or physical neurosurgeons in the US.4 Neuroprosthetic use of BCIs activity, can influence signal production.49 can directly address the major clinical problems affecting active soldiers, improve quality of life for those injured in The Future of BCI Therapeutics: combat, and ultimately reduce the individual, fiscal, and Restoring Function After Combat Injury social impact of disability. In the face of new developments in the field of BCIs, Although BCI technology at present has had a narrow whether to take on the risks of surgery and implantation scope of application in a small group of severely impaired must remain a question of utmost clinical and ethical im- patients, robust research efforts are underway to over- portance for a military neurosurgeon. Up to the present come current limitations and to boost the effectiveness of moment, BCI devices have been able to benefit only the invasive electrode platforms. For instance, the design of most severely injured patients, such as those with locked- biocompatible microelectrode coatings27 and algorithms in syndrome or tetraplegia.21,24 It is only with improve- that can adapt to vigorous movement9 promises to endow ments in the safety, longevity, and effectiveness of BCIs BCIs with the durability needed to withstand common- that neurosurgeons can begin to consider routine use of place disturbances encountered outside the confines of a such devices, even for significantly disabled and willing controlled laboratory environment. Additionally, develop- soldiers. In any future discussion with a potential patient ment of ECoG-based platforms constitutes an important regarding BCI implantation, expectation management investigative avenue, given their lower risk profile when must be a central element of the conversation, because the compared with single-neuron electrodes and their higher clinical trial stage will still consider BCIs to be an exper- information content than EEG studies.32 These proper- imental therapy rather than a treatment with confirmed ties would allow ECoG BCIs to be implanted in patients benefits. Furthermore, in light of the increasing incidence harboring less severe injury, and may make this the mo- of posttraumatic stress disorder and psychiatric issues ex- dality of choice in the future when restoring function to perienced by veterans of recent wars,36 it will be crucial wounded soldiers. for military neurosurgeons to take into account the influ- The comparatively greater success of BCIs in animal ence of mental health on decision making as well as any studies further substantiates the potential for these de- preconceived assumptions held by the patient about BCI vices to become functional prostheses. Rhesus monkeys treatments. Irrespective of the situation, any application implanted with electrodes recording from as few as 18 of BCIs in humans must be conducted in accordance with cortical neurons are able to move a cursor in 3 dimen- the guiding principles of patient autonomy and informed sions while also receiving visual feedback from their consent as well as physician beneficence and nonmalefi- brain-controlled environment.54 In follow-up studies, cence as espoused by the Hippocratic Oath, the Belmont implanted monkeys have also been trained to control a Report, and the Declaration of Helsinki.14,38,61

Neurosurg Focus / Volume 28 / May 2010 3

Unauthenticated | Downloaded 10/05/21 02:43 PM UTC I. S. Kotchetkov et al.

Military Enhancement With BCIs teration of personal identity. Unlike the use of BCIs for therapeutic interventions, cognitive, physical, and psy- Alongside therapeutic interventions, rapid advances chological enhancement of healthy individuals does not in BCI technologies will also create opportunities for fall under the principle of physician beneficence that ob- neurosurgeons to participate in improving military train- ligates doctors to restore health to normal levels through ing and operations, particularly through combat perfor- the treatment and prevention of disease. Nevertheless, mance modification and optimization. In fact, the use the ability of BCI devices to expand human capacities of neuroscientific approaches for achieving these goals must also be viewed in light of the advantage they grant is already an evolving area of research. During the last soldiers to perform and succeed in combat missions. In decade, the Pentagon’s DARPA launched the “Advanced this context, development of BCIs can be seen as mak- Speech Encoding Program” to develop nonacoustic sen- ing a paramount contribution to the national security, sors for speech encoding in acoustically hostile environ- which citizens, including physicians, have a social duty ments, such as inside of a military vehicle or an urban en- 12 to support. Equally important is the distribution of this vironment. The DARPA division is currently involved technology: in the hands of a responsible military, BCIs in a program called “Silent Talk” that aims to develop can protect national interests and the population at large, user-to-user communication on the battlefield through but if obtained by rogue groups, they can promote ter- EEG signals of “intended speech,” thereby eliminating rorism and instability. On a further level, if an existing the need for any vocalization or body gestures.11 Such BCI application had the potential for significant benefit capabilities will be of particular benefit in reconnais- in a much wider population, such as through therapeutic sance and special operations settings, and successful ap- uses, it would be ethically questionable to sequester its plications of silent speech interfaces have already been use without justly distributing it to the society at large. reported.12 All these and other considerations make the role of a neu- Enhancements of soldiers’ and control of rosurgeon in the development of BCIs particularly com- vehicles or heavy machinery with BCIs are also within plex. Nevertheless, ethical considerations must be fore- the realm of possibilities. A recent DARPA proposal for most applied to the most realistic expectations, such as a “Cognitive Technology Threat Warning System” in- BCI therapeutics, and deferred in those that are presently cludes a requirement for operator-trained high-resolution more speculative. BCI binoculars that can quickly respond to a subcon- sciously detected target or a threat. Such biological vision devices can have detection ranges of up to 10 km against Conclusions dismounts and vehicles, and can expand soldiers’ field of Brain-computer interfaces and their potential appli- view to 120°.11,57 Thus, future generations of auditory and cations engender great excitement. However, it must be visual neuroprostheses may allow soldiers to perform stressed that in their present state, it remains to be seen better during combat situations through automated de- how far, and in what direction, applications for BCIs will tection and interpretation capabilities.52 The concept of develop. In the near future, guided by a responsibility to telepresence, in which a soldier is physically present at the patient and civic duty, the military neurosurgeon can a base or concealed location, but has the ability to sense meet the clinical and ethical challenges presented by the and interact in a removed, real-world location through a field of BCIs to make optimal decisions for those who put mobile BCI device, is also being actively investigated and themselves in harm’s way to serve their country. has even been projected to be available in limited appli- cations by 2015.1 These expectations are substantiated by Disclosure recent advances in the operation of robots using EEG sig- nals,3 such that control of cargo-loading machines, demo- Brian Y. Hwang was supported by a grant from the Doris Duke Charitable Foundation. The authors report that no financial, lition robots, or unmanned aerial vehicles, as enabled by personal, or professional conflicts of interest pertaining to this BCIs, is more than a progressive goal; it is also a realistic manuscript exist. expectation.1,20,47 In its earliest stages, this type of BCI Author contributions to the study and manuscript prepara- could be used in manned vehicles, vessels, and aircraft to tion include the following. Conception and design: CP Kellner, IS make their operation more efficient by reducing the need Kotchetkov, BY Hwang, G Appelboom, ES Connolly. Drafting for manual input of key functions as required by today’s the article: IS Kotchetkov, BY Hwang, G Appelboom. Critically navigation and weapons deployment protocols. revising the article: CP Kellner, IS Kotchetkov, BY Hwang, G Appelboom, ES Connolly. Reviewed final version of the manuscript The ethical considerations of employing BCIs for and approved it for submission: CP Kellner, ES Connolly. Study performance modification depend largely on the type of supervision: ES Connolly. intervention required to implement the device. Because a majority of unclassified DARPA projects are based on References noninvasive BCIs, use of such platforms will not be asso- ciated with additional risks and can be viewed in a simi- 1. Aliberti K, Bruen TL: Telepresence: harnessing the human- lar manner as the use of night-vision goggles or radiofre- computer-machine interface. Army Logistician: Profes- sional Bulletin of United States Army Logistics 38: 2006 quency signals. However, the application of invasive or (http://www.almc.army.mil/alog/issues/NovDec06/browse. partially invasive BCIs in soldiers presents an ethically html) [Accessed March 1, 2010] challenging scenario that raises concerns of surgical risk 2. Averbeck BB, Lee D: Coding and transmission of information as well as issues of neurocognitive enhancement and al- by neural ensembles. Trends Neurosci 27:225–230, 2004

4 Neurosurg Focus / Volume 28 / May 2010

Unauthenticated | Downloaded 10/05/21 02:43 PM UTC Brain-computer interfaces: military, neurosurgery, and ethics

3. Bell CJ, Shenoy P, Chalodhorn R, Rao RP: Control of a hu- 24. Kennedy PR, Bakay RA: Restoration of neural output from a manoid robot by a noninvasive brain-computer interface in paralyzed patient by a direct brain connection. Neuroreport humans. J Neural Eng 5:214–220, 2008 9:1707–1711, 1998 4. Bell RS, Vo AH, Neal CJ, Tigno J, Roberts R, Mossop C, et al: 25. Kennedy PR, Bakay RA, Moore MM, Adams K, Goldwaithe Military traumatic brain and spinal column injury: a 5-year J: Direct control of a computer from the human central ner- study of the impact blast and other military grade weaponry vous system. IEEE Trans Rehabil Eng 8:198–202, 2000 on the . J Trauma 66 (4 Suppl):S104– 26. Kim SP, Simeral JD, Hochberg LR, Donoghue JP, Black MJ: S111, 2009 Neural control of computer cursor velocity by decoding motor 5. Bjornsson CS, Oh SJ, Al-Kofahi YA, Lim YJ, Smith KL, Turn- cortical spiking activity in humans with tetraplegia. J Neural er JN, et al: Effects of insertion conditions on tissue strain and Eng 5:455–476, 2008 vascular damage during neuroprosthetic device insertion. J 27. Kipke DR, Shain W, Buzsáki G, Fetz E, Henderson JM, Hetke Neural Eng 3:196–207, 2006 JF, et al: Advanced for chronic neural in- 6. Breakspear M, Stam CJ: Dynamics of a neural system with a terfaces: new horizons and clinical opportunities. J Neurosci multiscale architecture. Philos Trans R Soc Lond B Biol Sci 28:11830–11838, 2008 360:1051–1074, 2005 28. Kosashvili Y, Hiss J, Davidovic N, Lin G, Kalmovic B, 7. Brecht M, Schneider M, Sakmann B, Margrie TW: Whisker Melamed E, et al: Influence of personal armor on distribution movements evoked by stimulation of single pyramidal cells in of entry wounds: lessons learned from urban-setting warfare rat motor cortex. Nature 427:704–710, 2004 fatalities. J Trauma 58:1236–1240, 2005 8. Carmena JM, Lebedev MA, Crist RE, O’Doherty JE, Santucci 29. Kübler A, Nijboer F, Mellinger J, Vaughan TM, Pawelzik DM, Dimitrov DF, et al: Learning to control a brain-machine H, Schalk G, et al: Patients with ALS can use sensorimotor interface for reaching and grasping by primates. PLoS Biol rhythms to operate a brain-computer interface. 64: 1(2):E42, 2003 1775–1777, 2005 9. Chestek CA, Gilja V, Nuyujukian P, Kier RJ, Solzbacher F, 30. Lebedev MA, Carmena JM, O’Doherty JE, Zacksenhouse M, Ryu SI, et al: HermesC: low-power wireless neural recording Henriquez CS, Principe JC, et al: Cortical ensemble adapta- system for freely moving primates. IEEE Trans Neural Syst tion to represent velocity of an artificial actuator controlled Rehabil Eng 17:330–338, 2009 by a brain-machine interface. J Neurosci 25:4681–4693, 10. Daly JJ, Wolpaw JR: Brain-computer interfaces in neurologi- 2005 cal rehabilitation. Lancet Neurol 7:1032–1043, 2008 31. Lebedev MA, Nicolelis MA: Brain-machine interfaces: past, 11. Defense Advanced Research Projects Agency: Department of present and future. Trends Neurosci 29:536–546, 2006 Defense Fiscal Year 2010 Budget Estimates. Washington, 32. Leuthardt EC, Schalk G, Roland J, Rouse A, Moran DW: Evo- DC: Department of Defense, 2009 lution of brain-computer interfaces: going beyond classic mo- 12. Denby B, Schultz T, Honda K, Hueber T, Gilbert JM, Brum- tor physiology. Neurosurg Focus 27(1):E4, 2009 berg JS: Silent speech interfaces. Speech Communication 33. Leuthardt EC, Schalk G, Wolpaw JR, Ojemann JG, Moran 52:270–287, 2009 DW: A brain-computer interface using electrocorticographic 13. Donoghue JP, Nurmikko A, Black M, Hochberg LR: Assistive signals in humans. J Neural Eng 1:63–71, 2004 technology and robotic control using motor cortex ensemble- 34. Margalit E, Weiland JD, Clatterbuck RE, Fujii GY, Maia M, based neural interface systems in humans with tetraplegia. J Tameesh M, et al: Visual and electrical evoked response re- Physiol 579:603–611, 2007 corded from subdural electrodes implanted above the visual 14. Edelstein L: The Hippocratic Oath: text, translation and inter- cortex in normal dogs under two methods of anesthesia. J pretation, in Temkin O, Temkin CL (eds): Ancient Medicine: Neurosci Methods 123:129–137, 2003 Selected Papers of Ludwig Edelstein. Baltimore: Johns 35. Masini BD, Waterman SM, Wenke JC, Owens BD, Hsu JR, Hopkins University Press, 1967, pp 3–64 Ficke JR: Resource utilization and disability outcome assess- 15. Greer MA, Miklos-Essenberg ME, Harrison-Weaver S: A ment of combat casualties from Operation Iraqi Freedom and review of 41 upper extremity war injuries and the protective Operation Enduring Freedom. J Orthop Trauma 23:261– gear worn during Operation Enduring Freedom and Opera- 266, 2009 tion Iraqi Freedom. Mil Med 171:595–597, 2006 36. Milliken CS, Auchterlonie JL, Hoge CW: Longitudinal as- 16. Griffith RW, Humphrey DR: Long-term gliosis around chroni- sessment of mental health problems among active and re- cally implanted platinum electrodes in the Rhesus macaque serve component soldiers returning from the Iraq war. JAMA motor cortex. Neurosci Lett 406:81–86, 2006 298:2141–2148, 2007 17. Hatsopoulos N, Joshi J, O’Leary JG: Decoding continuous and 37. Musallam S, Corneil BD, Greger B, Scherberger H, Andersen discrete motor behaviors using motor and premotor cortical RA: Cognitive control signals for neural prosthetics. Science ensembles. J Neurophysiol 92:1165–1174, 2004 305:258–262, 2004 18. Hatsopoulos NG, Donoghue JP: The science of neural inter- 38. National Commission for the Protection of Human Subjects face systems. Annu Rev Neurosci 32:249–266, 2009 of Biomedical and Behavioral Research: The Belmont Re- 19. Heldman DA, Wang W, Chan SS, Moran DW: Local field po- port: Ethical Principles and Guidelines for the Protection tential spectral tuning in motor cortex during reaching. IEEE of Human Subjects of Research. Washington, DC: Depart- Trans Neural Syst Rehabil Eng 14:180–183, 2006 ment of Health, Education, and Welfare, OPRR Reports, 1979, 20. Hoag H: Remote control. Nature 423:796–798, 2003 (http://www.hhs.gov/ohrp/humansubjects/guidance/belmont. 21. Hochberg LR, Serruya MD, Friehs GM, Mukand JA, Saleh htm) [Accessed March 5, 2010] M, Caplan AH, et al: Neuronal ensemble control of prosthet­ 39. Nicolelis MA: Actions from thoughts. Nature 409:403–407, ic devices by a human with tetraplegia. Nature 442:164–171, 2001 2006 40. Nicolelis MA, Lebedev MA: Principles of neural ensemble 22. Houweling AR, Brecht M: Behavioural report of single neu- physiology underlying the operation of brain-machine inter- ron stimulation in somatosensory cortex. Nature 451:65–68, faces. Nat Rev Neurosci 10:530–540, 2009 2008 41. Owens BD, Kragh JF Jr, Wenke JC, Macaitis J, Wade CE, Hol- 23. Johnston JM Jr, Mangano FT, Ojemann JG, Park TS, Treva- comb JB: Combat wounds in Operation Iraqi Freedom and than E, Smyth MD: Complications of invasive subdural elec- Operation Enduring Freedom. J Trauma 64:295–299, 2008 trode monitoring at St. Louis Children’s Hospital, 1994-2005. 42. Patil PG, Carmena JM, Nicolelis MA, Turner DA: Ensemble J Neurosurg 105 (5 Suppl):343–347, 2006 recordings of human subcortical neurons as a source of motor

Neurosurg Focus / Volume 28 / May 2010 5

Unauthenticated | Downloaded 10/05/21 02:43 PM UTC I. S. Kotchetkov et al.

control signals for a brain-machine interface. Neurosurgery mary motor cortex. IEEE Trans Neural Syst Rehabil Eng 55:27–38, 2004 13:524–541, 2005 43. Patil PG, Turner DA: The development of brain-machine in- 54. Taylor DM, Tillery SI, Schwartz AB: Direct cortical control of terface neuroprosthetic devices. Neurotherapeutics 5:137– 3D neuroprosthetic devices. Science 296:1829–1832, 2002 146, 2008 55. Velliste M, Perel S, Spalding MC, Whitford AS, Schwartz AB: 44. Peleg K, Rivkind A, Aharonson-Daniel L: Does body armor Cortical control of a prosthetic arm for self-feeding. Nature protect from firearm injuries?J Am Coll Surg 202:643–648, 453:1098–1101, 2008 2006 56. Weaver FM, Burns SP, Evans CT, Rapacki LM, Goldstein B, 45. Polikov VS, Tresco PA, Reichert WM: Response of brain tis- Hammond MC: Provider perspectives on soldiers with new sue to chronically implanted neural electrodes. J Neurosci spinal cord injuries returning from Iraq and Afghanistan. Methods 148:1–18, 2005 Arch Phys Med Rehabil 90:517–521, 2009 46. Quian Quiroga R, Panzeri S: Extracting information from 57. Williams E, JASON Defense Advisory Panel Reports: Human neuronal populations: information theory and decoding ap- Performance. Washington, DC: Office of Defense Research proaches. Nat Rev Neurosci 10:173–185, 2009 and Engineering, JSR-07-625, 2008 (http://www.fas.org/irp/ 47. Rudolph A: Military: brain machine could benefit millions. agency/dod/jason/human.pdf) [Accessed March 5, 2010] Nature 424:369, 2003 (Letter) 58. Williams JC, Hippensteel JA, Dilgen J, Shain W, Kipke DR: 48. Ryu SI, Shenoy KV: Human cortical prostheses: lost in trans- Complex impedance spectroscopy for monitoring tissue re- lation? Neurosurg Focus 27(1):E5, 2009 sponses to inserted neural implants. J Neural Eng 4:410–423, 49. Santhanam G, Linderman MD, Gilja V, Afshar A, Ryu SI, 2007 Meng TH, et al: Hermes B: a continuous neural recording sys- tem for freely behaving primates. IEEE Trans Biomed Eng 59. Wolpaw JR: Brain-computer interfaces as new brain output 54:2037–2050, 2007 pathways. J Physiol 579:613–619, 2007 50. Santhanam G, Ryu SI, Yu BM, Afshar A, Shenoy KV: A high- 60. Wolpaw JR, McFarland DJ: Control of a two-dimensional performance brain-computer interface. Nature 442:195–198, movement signal by a noninvasive brain-computer interface in 2006 humans. Proc Natl Acad Sci U S A 101:17849–17854, 2004 51. Schalk G, Miller KJ, Anderson NR, Wilson JA, Smyth MD, 61. World Medical Association: Declaration of Helsinki: Ethi- Ojemann JG, et al: Two-dimensional movement control us- cal Principles for Medical Research Involving Human ing electrocorticographic signals in humans. J Neural Eng Subjects, 2008 (http://www.wma.net/en/30publications/10 5:75–84, 2008 policies/b3/index.html) [Accessed March 5, 2010] 52. Schmorrow DD, Estabrooke IV, Grootjen M (eds): Founda- tions of Augmented : Neuroergonomics and Op- erational Neuroscience. 5th International Conference, FAC Manuscript submitted January 18, 2010. 2009, held as part of HCI International 2009 San Diego, Cali- Accepted February 16, 2010. fornia, July 19–24. New York: Springer, 2009 Address correspondence to: Christopher P. Kellner, M.D., Co­­ 53. Suner S, Fellows MR, Vargas-Irwin C, Nakata GK, Dono- lumbia University College of Physicians and Surgeons, The Neu­ ghue JP: Reliability of signals from a chronically implanted, rological Institute of New York, 710 West 168th Street, Room 404, silicon-based electrode array in non-human primate pri- New York, New York 10032. email: [email protected].

6 Neurosurg Focus / Volume 28 / May 2010

Unauthenticated | Downloaded 10/05/21 02:43 PM UTC