For this Research Topic on brain augmentation,
1
several authors discuss possibilities of brain stimulation (e.g., Duecker et al.,
2014), pharmacology (e.g., Lynch et al., 2014), and psychobiological training (e.g.,
Chapman and Mudar, 2014). According to a definition proposed by ethicists, such procedures
are human enhancement if and only if they are a “change in the biology or psychology
of a person which increases the chances of leading a good life in the relevant set
of circumstances” (Savulescu et al., 2011b, p. 7). Note how this definition describes
the individual as malleable and the circumstances as given. The authors continue to
explain that something counts as enhancement “so long as it tends to increase a person's
well-being” (Savulescu et al., 2011b). Similarly, Nagel emphasizes the notions of
happiness, well-being, and improvement in her discussion of the ethical challenges
of enhancement and discusses the possibilities and risks related to neuro-technology
and psychopharmacology (Nagel, 2014).
These and similar publications identify concepts like improvement or well-being as
foundational issues of the enhancement debate. This raises important questions, such
as who defines well-being and how to achieve it. In the three following sections,
I will discuss the conceptualization of well-being, the framing of enhancement, and
the translational promises given in the literature.
Whose well-being?
The majority of the experimental enhancement literature employs neuropsychological
test designs developed to measure the presence of psychological impairment in terms
of attention, learning, memory, and the like (for systematic reviews, see Repantis
et al., 2010; Smith and Farah, 2011; Bagot and Kaminer, 2014). Referring to this literature
in the human enhancement debate is problematic: That these tests can be used to inform
clinical decisions does not warrant their usefulness outside the clinics. Higher test
scores do not necessarily reflect a happier, more meaningful life in general. Yet,
clinical studies are often cited in ethical discussions to debate the benefits and
prospects of enhancement for the healthy. This carries the risk of a normative fallacy,
namely, the identification of clinical benefit with overall well-being.
This risk is often accompanied by another one, namely, that of a localizational fallacy.
It consists in only targeting individuals psychobiologically, not their circumstances.
In contrast, established measures such as the World Happiness Report which are provided
by United Nations institutions measure well-being macroscopically: GDP per capita,
social support, healthy life expectancy at birth, freedom to make life choices, generosity,
and perceptions of corruption together explain 75.5% of the international variance
of happiness rankings in 2012 (Helliwell et al., 2013). It goes without saying that
these indices are also based on norms, but not primarily driven by clinical needs,
instead broader in scope, and developed by institutions which are representing people
at large at least remotely.
An advanced recent proposal consists in the OECD Guidelines on Measuring Subjective
Well-being, operationalizing subjective well-being as consisting of life satisfaction,
affect, and eudaimonic well-being, which in turn consist of three subcategories each,
namely, income, health, and work satisfaction; anger, worry, and happiness; competence,
autonomy, and meaning and purpose (OECD, 2013). Based on these guidelines, people
can create their own Better Life Index, prioritizing 11 pre-defined dimensions (such
as housing, jobs, education, or safety), and more than 60,000 citizens from OECD countries
have so far participated
2
. Using such methods, the risk of a normative fallacy can be minimized, since people
can choose their own standards, although ideally they should be able to design the
methods, too. The results, including meaningful differences between countries, indicate
that human enhancement need not be localized in individual psychobiology, but can
also be achieved by socio-political reform.
It turned out, for example, that safety is valued most highly by participants from
Japan, income and housing by those in the United States, and education by those in
Finland. To assess the relevance of brain stimulation, pharmacology, and psychobiological
training for human enhancement, it would be informative to know to what extent these
methods can contribute to human well-being broadly understood. If it turned out that
the causal link is very remote and speculative, proponents of human enhancement could
conclude that socio-political reform is more promising a means than individual psychobiological
intervention. In the terms of the definition proposed by Savulescu and colleagues
above, this amounts to not changing the subject with respect to the circumstances,
but the circumstances with respect to the subject.
Framing and relevance
Cognitive enhancement has been framed as common by leading scholars in the field who
described it as a means “not to get high, but to get higher grades, to provide an
edge over their fellow students or to increase in some measurable way their capacity
for learning” (Greely et al., 2008, p. 702). Greely and colleagues subsequently stated
that almost 7% of students in the US already use stimulants like amphetamine or methylphenidate
for cognitive enhancement, with the prevalence reaching 25% on some campuses. In a
comment gathering some anecdotal evidence, I pointed out that such framings occur
regularly in the ethical literature (Schleim, 2010). This impression is shared by
Lucke et al. (2011) who also carried out a media analysis of newspaper articles and
found that 94% of the reports mentioning the prevalence of psychopharmacological enhancement
described it as common, increasing, or both (Partridge et al., 2011). Actually, 66%
of these reports referred to the academic literature as evidence. It goes without
saying that this framing of the practice as common and/or increasing lends the topic
high urgency.
In the systematic review of prevalence studies in student samples by Smith and Farah,
the most comprehensive I know of, the authors conclude that “[a]mong college students,
estimates of use vary widely but, taken together, suggest that the practice is commonplace”
(Smith and Farah, 2011, p. 717). Referring to this review, Nagel even claims that
the usage is increasing (Nagel, 2014). Both claims are difficult to justify, though,
with respect to cognitive enhancement: First of all, it is in the eye of the beholder
what to count as common. The decision is complicated by the variance in findings,
ranging from 1.7 to 34% in studies with more than thousand students (N = 12; mean
= 9.5%, median = 6.7%). Sometimes the reported figures reflect past month prevalence
(N = 2; mean = 4.6), sometimes they refer to last year (N = 6; mean = 6.7) or even
lifetime usage (N = 4; mean = 16.1). Secondly, their authors often investigated non-medical
use, which allows many different motives for stimulant consumption that do not indicate
cognitive enhancement, such as feeling high or losing weight. Smith and Farah summarize
that in those surveys addressing motives, study-related answers were dominant but
regularly accompanied by recreational/lifestyle choices (Smith and Farah, 2011). However,
detailed interviews with consumers at an elite university in the United States suggest
that emotional rather than cognitive motives drive non-medical use even for improving
studying, since people report feeling better and overcoming motivational problems
with stimulants (Vrecko, 2013).
For the time being, framing the relevance as common and non-medical use as cognitive
enhancement is therefore, in my view, in contrast to the best available evidence.
It is even more problematic to claim that the practice is increasing, because this
would require repeated cross-sectional studies of comparable samples under standardized
conditions. Yet, even within research groups definitions of inclusion criteria and
ways of sampling data often differ. Nevertheless, what has been increasing steeply
during the last decades was the production of stimulants like amphetamine and methylphenidate,
particularly in the United States, and publications on enhancement (see Figure 1).
That the former increase is not reflected in the prevalence studies previously mentioned
is most likely due to the concept of non-medical use. Both drugs are controlled prescription
stimulants and most epidemiologists as well as ethicists strictly distinguish medical
use as treatment from non-medical use as either drug abuse or enhancement.
Figure 1
Stimulant production and enhancement papers increased strongly. Lines show a steep
increase in publications on cognitive enhancement (blue) and neuroenhancement (yellow),
but only modestly on mood enhancement (orange). Publication numbers are based on a
Web of Science topic search. Bars show a strong increase in production quotas for
amphetamine (red) and methylphenidate (green). In the shown 10-year period from 2004
to 2013, the former increased 5.5-fold, the latter 3.4-fold, after quotas had already
been increasing in the 1990s (not shown, but see Rasmussen, 2008). Figures based on
US Drug Enforcement Agency, October 2, 2013, http://www.deadiversion.usdoj.gov/quotas/quota_history.pdf
(accessed May 30, 2014), accumulating amphetamine produced for sale and conversion.
This framing has wider ramifications for the scientific community: Without the treatment/enhancement
distinction, the consumption of stimulants can and has been analyzed by medical sociologists
under labels such as medicalization or pharmaceuticalization (Abraham, 2010; Bell
and Figert, 2012); and without the claim that enhancement is common or even increasing,
the problem appears much less urgent. By framing stimulant consumption as enhancement
and common, though, neuroethicists generated a new ethical problem, new prospects
and risks, that they subsequently could manage (see also Conrad and De Vries, 2011;
Littlefield and Johnson, 2012). Indeed, the steep increase in publications on enhancement
topics coincides with the inception of instutionalized neuroethics (Marcus, 2002;
Farah, 2012; Figure 1). It thus becomes apparent that both, medical sociologists and
neuroethicists, have a conflict of interest in framing stimulant consumption in the
competition for research funds and high-impact publications.
Promises
The abundant literature on enhancement suggests the possibility to increase learning,
to feel better, and to become more intelligent by means of brain stimulation, pharmacology,
or psychobiological learning (Savulescu et al., 2011a; Farah, 2012; Hildt and Franke,
2013; Nagel, 2014). However, it is also noted that there is much that is not known
about the working of stimulants, for example, and that funding of empirical research
is difficult because it is not about treatment and therefore outside the purview of
disease-oriented schemes and it is too applied for funders of basic science (Smith
and Farah, 2011). As mentioned in the section on well-being above, it is furthermore
not clear what the goal of the intervention is and whether changing the individual
in its circumstances is actually more promising than changing the circumstances for
the individual.
However, by analogy with biological psychiatry it is possible to at least engage in
informed speculation on what the situation might be like had there been more agreement
on the research goals and more funding of enhancement research. When psychiatric researchers
started to prepare the fifth edition of the Diagnostics and Statistical Manual of
Mental Disorders (DSM) they set the aim to include biomarkers, particularly based
on genetic and neuroimaging research, to improve diagnosis and treatment (Hyman, 2007).
Note that the previous fourth edition of the DSM listed more than 300 disorders and
their respective symptoms guiding clinical diagnosis (APA, 2000). It is now widely
acknowledged that this attempt for the fifth edition was unsuccessful, though views
on why this happened and what to do about it differ (Hyman, 2010; Kapur et al., 2012;
Walter, 2013; Kirmayer and Crafa, 2014). Certainly, with more than one billion dollars
annually spent on research at the National Institutes of Mental Health alone, lack
of funding was not the problem
3
. In the light of decisions by pharmaceutical companies to close their psychiatric
laboratories because of negative prospects (Amara et al., 2011; Van Gerven and Cohen,
2011) and reports that prescription stimulants do not even seem to have a lasting
positive effect on individuals diagnosed with Attention Deficit/Hyperactivity Disorder
(Currie et al., 2013; Sharpe, 2014), the frequently promised translational possibilities
of enhancement research may be unrealistic (Schleim, 2014). Perhaps we need to minimize
risks of committing a translational fallacy, too.
When Quednow speaks of a “phantom debate” (Quednow, 2010) or Lucke and colleagues
want to deflate the “neuroenhancement bubble” (Lucke et al., 2011), they appear to
have good reasons for doing so. We should also not forget that people in many countries
are already quite happy and that in those where they are not, the difference in happiness
is probably not due to limited access to enhancement technology. Clinical research
for those suffering from a disorder should keep the priority over enhancement. It
could even be the case that too much focus on increasing well-being and happiness,
on how things might yet be better than they presently are, might make more people
unhappy in the first place; or, in Schopenhauer's words:
“We then recognize that the best, which the world has to offer, is a painless, calm,
bearable existence and we confine our claims to these in order to accomplish them
better. Because not to become very unhappy, it is the best means that one may not
demand to be very happy.” (Schopenhauer, 1874, p. 434; author's translation).
Conflict of interest statement
The author declares that the research was conducted in the absence of any commercial
or financial relationships that could be construed as a potential conflict of interest.