Society has already seen multiple potential futures of neurotechnologies through the multifaceted lens of science fiction. We do not know which of those futures we will come to live, but the strong force in society's evolutionary process will likely continue to be the insight and oversight developed by the bioethics community, pushing and being pushed by the advances of technology, improvements in medical care, and market competitiveness.
Today's neurotechnologies are capable of only simple physical interactions with our brains, but it is easy to speculate that today's low-level devices will lead to significant advances in our understanding of neural function and our ability eventually to extract, synthesize, and inject high-level thought and perception processes through computer interfaces. In one possible future, computers themselves will manipulate knowledge and interact with each other as human brains do. How will the relationship between neurotechnology and humans then change? How will we deal with the societal aspects (rights, equality, free-will, ownership, passports, religion, morality) of cyborgs and robots? Although these questions are the domain of the future of neuroethics, can today's bioethics evolve and serve our multicultural society as the advances unfold? The current state of neurotechnologies and currently debated ethical issues can inform us in approaching the future we are designing for ourselves.
Neurotechnologies
Today's neurotechnologies can be divided into three main categories: neuroprostheses, internal direct brain interfaces (DBIs), and external direct brain interfaces. A derivative category, neurorobotics, which is a robotic extension of an internal or external direct brain interface, will be discussed separately.1
For comprehensive reviews, consult Wolpaw JR, Birbaumer N, McFarland DJ, Pfurtscheller G, Vaughan TM. Brain–computer interfaces for communication and control. Clinical Neurophysiology 2001;113:767–91; Donoghue JP. Connecting cortex to machines: Recent advances in brain interfaces. Nature Neuroscience Supplement 2002;5:1085–8; Millán J del R. Adaptive brain interfaces. Communications of the ACM 2003;46(3):5–80.
Neuroprostheses are devices that contain sensors and/or actuators, implanted in or communicating with the brain, to replace neural and sensory function. Cochlear and retinal implants, as well as hypocampus memory mediation chips in a rodent model, are recent examples.
Internal direct brain interfaces share with neuroprostheses the fact that they are implanted in the body and connect directly with biological neural structures. However, Internal DBIs do not replace neural function; they sense it through electrodes and potentially stimulate the brain to provide information to it. For example, the Cybernetics BraingateTM cortical implant is a 2-mm × 2-mm array of 100 electrodes that has been implanted in nonhuman primates and several human subjects in the motor cortex and several other cortical regions.2
Cyberkinetics Braingate neural implant. Available at: http://www.cyberkineticsinc.com/content/medicalproducts/braingate.jsp. Accessed November 28, 2005.
Shenoy KV, Meeker D, Cao S, Kureshi SA, Pesaran B, Mitra P, et al. Neural prosthetic control signals from plan activity. NeuroReport 2003;14:591–6.
Another type of internal DBI is the electrocorticogram (ECoG), used principally for epilepsy research and treatment. An ECoG electrode array is implanted on the surface of the cortex near the suspected focus of epilepsy seizures. Once found, the focus can then be surgically removed to alleviate symptoms.
Deep brain stimulation (DBS) electrodes are implanted in the basal ganglia to improve quality of movement of persons who have Parkinson's Disease. Electrode stimulation is controlled by a pacemaker-type device commonly placed subcutaneously under the collarbone. This device is also being studied for control of psychiatric conditions.
External direct brain interfaces rely on sensing through the skin of electrical activity related to nerve impulses. For example, electroencephalogram (EEG) electrode arrays can be mounted on caps to capture temporo-spatial patterns of electrical activity over the skull, but because the interelectrode spacing is approximately 1 cm, and because the electrodes are at least a skull's thickness from cortical activity, only the pattern from an entire region consisting of thousands or millions of neurons can be captured.4
Mason SG, Moore Jackson MM, Birch GE. A general framework for characterizing studies of brain interface technology. Annals of Biomedical Engineering 2005;33(11):1653–70.
A neurorobot is a system that typically involves an electromechanical device (robot) connected to a human or animal controller, or its converse, a biological system whose neural functions are partially taken over by a synthetic controller with sensor and/or motor connections. A human cyborg would be a neurorobot, as would a rat or flying insect whose brain receives signals from external vision sensors to steer its course. Although animal-level neurorobots are not fiction anymore, cyborgs are still in our future and will invite increasing bioethics scrutiny as the level of abstraction of neural processes increases and the level of interaction with humans evolves.
Bioethics Considerations
The ethical issues for the neuroprostheses of today are actively pursued by ongoing neuroethics research,5
Farah MJ, Wolpe PR. Monitoring and manipulating brain function: New neuroscience technologies and their ethical implications. Hastings Center Report 2004;34:35–45; Illes J, Blakemore C, Hansson MG, Hensch TK, Leshner A, Maestre G, et al. International perspectives on engaging the public in neuroethics. Nature Reviews Neuroscience 2005;6:977–84.
From the available literature on bioethics applied to various technologies, some of the main ethical questions that can be foreseen as particularly relevant to DBI–robotics are the following:
- Control: If a chip is implanted in a human's brain, will that person be in sole control of the inputs and outputs or may there be an outside agent, perhaps even a distributed system on the Internet, to determine information flow? Although we as a society have the benefit of science fiction to have been exposed to the specter of this issue already, the reality, when it arrives, will be more complex, with outcomes decided by the courts and the private sector, mediated by the available neurotechnology.
- Observability: Closely linked to the previous issue, how can the design and actions of the technology itself reveal its intentions in informing or controlling the human who uses it? In our society, we reveal our intentions to others through our behavior, gestures, and communication. When technology has the fundamental capacity to inject a layer of nonhuman reasoning to our thought processes, how can that be made apparent to the technology user, and, equally importantly, how can others affected by such bionic behavior be informed and protected?
- Functional improvement: The debate over the use of
assistive technology and pharmacology for people with disabilities as well
as for able-bodied persons to enhance “normal” function6is not specific to neurotechnologies, and the same arguments pertain. However, implicating brain function raises significantly the power of the technology to disrupt society through the introduction of cognitive superhumans or cyborgs.7
Daniels N. Normal functioning and the treatment-enhancement distinction. Cambridge Quarterly of Healthcare Ethics 2000;9(3):309–22; Bostrom N. In defence of posthuman dignity. Bioethics 2005;19(3):202–14; Hogle LF. Enhancement technologies and the body. Annual Review of Anthropology 2006;34:695–716.
Moreno JD. Mind Wars: Brain Research and National Defense. Washington, DC: Dana Press; 2006.
- Agency: Asimov's three (plus the zeroth) laws of
robotics8touch on the relationships between humans and machines sharing the planet. However, at what point does a robot's higher intelligence, if realized, require humans to give robots rights in society? How might artificial intelligence have to include artificial morality? Do we need to have “Kantian” as well as “utilitarian” interaction chips, and perhaps “Christian,” “Islam,” and “Hindu” morality overlays? Although this might appear to be more an issue for future humanoid robots, the same arguments will surface when direct brain interaction has the power to modify or control human behavior. What defines who is a human and who (or what) is not? Can we invent a “Moral Turing Test”9
Asimov I. I, Robot. USA: Gnome Press; 1950.
to guide us in making such a decision?Allen C, Varner G, Zinser J. Prolegomena to any future artificial moral agent. Journal of Experimental & Theoretical Artificial Intelligence 2000;12:251–61.
Although such questions may seem to become problems only far into the future, the exponential graphs of Ray Kurzweil10
Kurzweil R. The Singularity Is Near: When Humans Transcend Biology. New York: Penguin Group; 2005.
On a higher plane, as neuroprostheses and DBIs mature to levels of cognitive interaction with the brain that are still in the realm of science fiction today, bioethics will need to evolve as well to provide guidance for them. Nanoethics and roboethics,11
Wagner JJ, Cannon DM, Van der Loos HFM. Cross-cultural considerations in establishing roboethics for neuro-robot applications. In: Patton J, Weir R, eds. Proceedings IEEE International Conference on Rehabilitation Robotics, Chicago, 28 June–July 2005. Piscataway, NY: IEEE Press. Veruggio G (ed.). The Roboethics Roadmap. http://www.roboethics.org/site/modules/mydownloads/visit.php?cid=1&lid=37. Accessed March 5, 2007.
For example, Roco MC, Bainbridge WS, eds. Converging Technologies for Human Performance. NSF Report. Arlington, VA: National Science Foundation; 2002.
Conclusions
Even though the fields of philosophy and psychology have developed great theories over the past centuries that help explain how people think and interact, there is still no end to the debate about which of these theories is “right,” or even if there can be a “right” theory. Current neurotechnologies can only interact with the brain at very prosaic and well-understood levels such as simple movements (e.g., left, right) and single-modality sensing, and the ethical issues are mostly tractable. However, as the technologies are being developed to mimic higher levels of human thought, how can one reduce what are still theories to real practice as software and hardware, and how do we achieve any consensus on what might be the “right” approach? Bioethics of the future will have to grapple with this type of meta-challenge, as if the challenges of advanced neurotechnologies were not enough.
