D. PUDLES
We are so surrounded by gadgetry that it is sometimes hard to tell where devices end and people begin. From computers and scanners to multifarious mobile devices, an increasing number of humans spend much of their conscious lives interacting with the world through electronics, the only barrier between brain and machine being the senses — sight, sound and touch — through which humans and devices interface. But remove those senses from the equation, and electronic devices can become our eyes and ears and even our arms and legs, taking in the world around us and interacting with it through man-made software and hardware.
This is no future prediction; it is already happening. Brain–machine interfaces are clinically well established in restoring hearing perception through cochlear implants, for example. And patients with end-stage Parkinson's disease can be treated with deep brain stimulation (DBS) (see 'Human brain–machine applications'). Worldwide, more than 30,000 implants have reportedly been made to control the severe motor symptoms of this disease. Current experiments on neural prosthetics point to the enormous future potential of such devices, whether as retinal or brainstem implants for the blind or as brain-recording devices for controlling prostheses1.
Non-invasive brain–machine interfaces based on electroencephalogram recordings have restored communication skills of patients 'locked in' by paralysis2. Animal research and some human studies3 suggest that full control of artificial limbs in real time could further offer the paralysed an opportunity to grasp or even to stand and walk on brain-controlled, artificial legs, albeit likely through invasive means, with electrodes implanted directly in the brain.
Future advances in neurosciences together with miniaturization of microelectronic devices will make possible more widespread application of brain–machine interfaces. Melding brain and machine makes the latter an integral part of the individual. This could be seen to challenge our notions of personhood and moral agency. And the question will certainly loom that if functions can be restored for those in need, is it right to use these technologies to enhance the abilities of healthy individuals? It is essential that devices are safe to use and pose few risks to the individual. But the ethical problems that these technologies pose are not vastly different from those presented by existing therapies such as antidepressants. Although the technologies and situations that brain–machine interfacing devices present might seem new and unfamiliar, most of the ethical questions raised pose few new challenges.
Welcome to the machine
In brain-controlled prosthetic devices, signals from the brain are decoded by a computer that sits in the device. These signals are then used to predict what a user intends to do. Invariably, predictions will sometimes fail and this could lead to dangerous, or at the very least embarrassing, situations. Who is responsible for involuntary acts? Is it the fault of the computer or the user? Will a user need some kind of driver's licence and obligatory insurance to operate a prosthesis?
Fortunately, there are precedents for dealing with liability when biology and technology fail to work. Increasing knowledge of human genetics, for example, led to attempts to reject criminal responsibility that were based on the inappropriate belief that genes predetermine actions. These attempts failed, and neuroscientific pursuits seem similarly unlikely to overturn views on human free will and responsibility4. Moreover, humans are often in control of dangerous and unpredictable tools such as cars and guns. Brain–machine interfaces represent a highly sophisticated case of tool use, but they are still just that. In the eyes of the law, responsibility should not be much harder to disentangle.
But what if machines change the brain? Evidence from early brain-stimulation experiments done half a century ago suggests that sending a current into the brain may cause shifts in personality and alterations in behaviour. Many patients with Parkinson's disease who have motor complications that are no longer manageable through medication report significant benefits from DBS. Nevertheless, compared with the best drug therapy, DBS for Parkinson's disease has shown a greater incidence of serious adverse effects such as nervous system and psychiatric disorders5 and a higher suicide rate6. Case studies revealed hypomania and personality changes of which the patients were unaware, and which disrupted family relationships before the stimulation parameters were readjusted7.
Such examples illustrate the possible dramatic side effects of DBS, but subtler effects are also possible. Even without stimulation, mere recording devices such as brain-controlled motor prostheses may alter the patient's personality. Patients will need to be trained in generating the appropriate neural signals to direct the prosthetic limb. Doing so might have slight effects on mood or memory function or impair speech control.
Nevertheless, this does not illustrate a new ethical problem. Side effects are common in most medical interventions, including treatment with psychoactive drugs. In 2004, for example, the US Food and Drug Administration told drug manufacturers to print warnings on certain antidepressants about the short-term increased risk of suicide in adolescents using them, and required increased monitoring of young people as they started medication. In the case of neuroprostheses, such potential safety issues should be identified and dealt with as soon as possible. The classic approach of biomedical ethics is to weigh the benefits for the patient against the risk of the intervention and to respect the patient's autonomous decisions8. This should also hold for the proposed expansion of DBS to treat patients with psychiatric disorders9.
Bench, bedside and brain
The availability of such technologies has already begun to cause friction. For example, many in the deaf community have rejected cochlear implants. Such individuals do not regard deafness as a disability that needs to be corrected, instead holding that it is a part of their life and their cultural identity. To them, cochlear implants are regarded as an enhancement beyond normal functioning.
What is enhancement and what is treatment depends on defining normality and disease, and this is notoriously difficult. For example, Christopher Boorse, a philosopher at the University of Delaware in Newark, defines disease as a statistical deviation from "species-typical functioning"10. As deafness is measurably different from the norm, it is thus considered disease. The definition is influential and has been used as a criterion for allocation of medical resources11. From this perspective, the intended medical application of cochlear implants seems ethically unproblematic. Nevertheless, Anita Silvers, a philosopher at San Francisco State University in California and a disability scholar and activist, has described such treatments as a "tyranny of the normal"12, designed to adjust people who are deaf to a world designed by the hearing, ultimately implying the inferiority of deafness.
Electronic devices can become our eyes and ears and even our arms and legs.
Although many have expressed excitement at the expanded development and testing of brain–machine interface devices to enhance otherwise deficient abilities, Silvers suspects that prostheses could be used for a "policy of normalizing". We should take these concerns seriously, but they should not prevent further research on brain–machine interfaces.
Brain technologies should be presented as one option, but not the only solution, for paralysis or deafness. Still, whether brain-technological applications are a proper option remains dependent on technological developments and on addressing important safety issues.
One issue that is perhaps more pressing is how to ensure that risks are minimized during research. Animal experimentation will probably not address the full extent of psychological and neurological effects that implantable brain–machine interfaces could have. Research on human subjects will be needed, but testing neuronal motor prostheses in healthy people is ethically unjustifiable because of the risk of bleeding, swelling, inflammation and other, unknown, long-term effects.
People with paralysis, who might benefit most from this research, are also not the most appropriate research subjects. Because of the very limited medical possibilities and often severe disabilities, such individuals may be vulnerable to taking on undue risk. Most suitable for research into brain–machine interface devices are patients who already have an electrode implanted for other reasons, as is sometimes the case in presurgical diagnosis for epilepsy. Because they face the lowest additional risk of the research setting and will not rest their decision on false hopes, such patients should be the first to be considered for research13.
Brain–machine interfaces promise therapeutic benefit and should be pursued. Yes, the technologies pose ethical challenges, but these are conceptually similar to those that bioethicists have addressed for other realms of therapy. Ethics is well prepared to deal with the questions in parallel to and in cooperation with the neuroscientific research.
