Introduction

Cognitive neuroenhancement, the improvement of intellectual function through medical means without therapeutical intentions, has been the topic of public and academic debate for several years. For example, up to 25% of American students use psychostimulants1 and 5% of the working population in Germany use pharmaceutical drugs to enhance their cognitive functions.2 Furthermore, the excessive rise in prescriptions of some pharmaceuticals such as Ritalin confirms widespread off-label use.3 In addition, 80% of students in Germany would use neuroenhancers as long as there were no adverse effects.4 The question regarding the ethical assessment of this phenomenon thus arises.

While many authors strongly oppose neuroenhancement, others cannot identify any sustainable objections to it and believe that the advantages outweigh the disadvantages. This liberal argument can be found, among others, in the memorandum of a German team of authors5 and in a well known article by an American–British group of authors.1 The advocates of neuroenhancement do not claim that today's means are unproblematic or even preferred. On the contrary, they rightly stress the fact that these pharmaceuticals have serious adverse effects and do not even show the desired efficacy. The substances in use at the present time such as methylphenidate, modafinil, or selective serotonin reuptake inhibitors are not the neuroenhancers their positive judgements refer to. Therefore, the liberal argument is hypothetical and exhibits an if–then structure. If there were substances that verifiably improved certain mental abilities without adverse effects or with an acceptable risk–benefit profile, then no convincing arguments against the use of these substances or procedures could be made. The conditional portion of this argument (the ‘if…’) is not meant in a purely speculative or hypothetical way. In fact, the authors assume that such pharmaceuticals could be available in the foreseeable future.

While several participants in the discussion criticise the second part of the liberal argument, the first part (the ‘if…’) will be examined as follows. How probable is it that substances without any addiction potential will be available in the foreseeable future?

However, the second part of the if–then argument, that is the actual ethical assessment of neuroenhancement, comes back into play when the question arises as to which research strategy to follow. The ethical argument requires a certain plausibility of the if assumption, which will ultimately be empirically tested, but the evaluation of empirical studies also depends on the general ethical assessment of neuroenhancement.

Risk of addiction to neuroenhancers: facts and implicated mechanisms

The proponents of neuroenhancement base their considerations in part on what we believe to be mistaken assumptions and tend to overlook significant differences between neuroenhancing drugs and food or human practices to enhance performance.

Firstly, several liberal proponents often work with an out-of-date differentiation in regard to addiction medicine when they assume that ‘physical dependence’ would be a cogent argument against the use of neuroenhancers, while ‘psychological dependence’, defined as an irrational desire for an object and the extreme discomfort caused by the absence of it, could indeed arise as a result of the use of neuroenhancers; but this is a widespread phenomenon and cannot be an argument against neuroenhancement.5 A German group of authors claims that such a ‘psychological dependence’ could also be an innocuous phenomenon like ‘dependence’ on mobile phones (see Galert et al5). However, the distinction between physical and psychological dependence is scientifically outdated and assumes precisely the functional dualistic distinction between psychological and physical phenomena that the authors dismissed elsewhere as untenable (Galert, et al5).

Modern addiction medicine no longer distinguishes between psychological and physical phenomena,6 but now traces the particular symptoms back to their different neurobiological correlates. Development of tolerance and withdrawal symptoms normally occur due to neuroadaptive processes in inhibitory neurotransmitter systems, where a sudden discontinuation of substance consumption can lead to excitation and observable ‘physical’ withdrawal symptoms such as seizures or vegetative symptoms. However, symptoms, classified by some authors as a ‘psychological’ dependence such as the desire for a substance and the loss of control when exposed to it, are regarded as disorders associated with dysfunctions in motivational systems and their dopaminergic modulation.7–9 This is also true for behavioural addictions, such as gambling, with a strong desire for the activity and a diminished control over its practice: imaging studies among patients with gambling addiction revealed impaired functional activation in the ventral striatum during the anticipation and arrival of reward, an observation that was also reported in patients with alcohol addiction.10

Secondly, the dopaminergic neurotransmitter system plays a key role in modern addiction theories. In fact, all of the known substances with addiction potential release dopamine in the so-called reward system and therefore boost further consumption. This occurs through dopaminergic effects on learning processes, which ultimately lead to drugs and drug-related stimuli becoming particularly desirable, while so-called naturally rewarding experiences such as food intake and social interaction lose their appeal.11–13 Therefore, the effects of neuroenhancers on the dopamine system are of particular interest when discussing their addiction potential.

In this context, several liberal proponents suggest that there are several ‘natural’ activities as well as food and drink, such as coffee, that affect the brain and therefore can influence neurotransmitter systems such as the dopaminergic system.1 5 However, there is a substantial difference between dopamine release caused by a particular experience or food without addiction potential and one resulting from psychotropic substances used for the purpose of neuroenhancement with respect to the extent and the habituation of dopamine release. Sensory environmental stimuli never have a direct effect on the brain, but are always mediated through sense organs, which transform the stimulus constellations into neuronal codes. Dopamine concentrations, which are triggered by food, sex and human communication, increase by approximately 50% to 100%, while drugs that directly affect dopamine transporter function (eg, amphetamine, cocaine and the psychostimulants methylphenidate and modafinil) induce dopamine releases ranging from 175% to 1000%.6 14–16 As a result, subjects learn to crave for the ‘more effective’ drugs of misuse and lose interest in non-drug-associated stimuli.

Similarly, an obvious difference can be seen in habituation: the repeated blockade of dopamine transporters by psychotropic substances and/or the repeated use of drugs with addiction potential consistently lead to an increase in the release of dopamine due to direct pharmacological effects, whereas dopamine release triggered by environmental stimuli rapidly habituates.7 11 The repeated use of such pharmaceuticals or drugs induces counter-regulatory neuroadaptive processes (eg, a reduction of dopamine D2 receptors in the ventral striatum), which reduce response to non-drug-associated reinforcers and impairs goal-directed behaviour driven by such reinforcers.7 8 17 Repeated use of drugs that strongly stimulate dopamine release thus increases the desire for further drug consumption and shifts behavioural control away from goal-directed behaviour towards habitised drug intake.9 12 As a result, chronic intake of drugs of misuse severely impairs behavioural flexibility and voluntary behavioural control and reduces motivation and ability to learn from non-drug-related feedback.12 17

Thirdly, modafinil, presently the only pharmacological substance that seems to effectively lead to some genuine cognitive improvement, also has strong effects on the dopaminergic system. Nevertheless, some liberal proponents of neuroenhancement distinguish between psychostimulants such as amphetamines, which are ‘not suitable as neuroenhancers due to their addiction potential and serious side effects’, and modafinil as a potentially effective neuroenhancer.5 However, is the categorical separation of modafinil and amphetamines empirically justified? Unlike methylphenidate, modafinil is unstable when heated and thus cannot be injected intravenously. Still, a study shows that the oral administration of modafinil causes substantial dopamine release in the range observed with other psychostimulants.18 Particularly alarming is the fact that dopamine release in the ventral striatum, a core area of the reward system and therefore highly relevant for potentially addictive effects of a pharmaceutical agent, was much greater than in other areas of the striatum, which generally contribute to psychomotor activation. In addition, the European Medicines Agency19 listed 485 reports of modafinil misuse, tolerance development and addiction, and case reports describe severe side effects such as death from polysubstance misuse including modafinil, pathological gambling that started with modafinil intake and terminated when modafinil medication was stopped, as well as modafinil-induced psychosis, which may all be triggered by the dopaminergic effects of this psychostimulant.20–23 Therefore, there is an evidence-based concern that more widespread and uncontrolled use of for example, modafinil for NE rather than for treatment could place many more subjects at the risk of developing addictive disorders, and recent pharmacological regulations for example, in Germany explicitly warn against using modafinil in subjects at risk of developing addiction.19

Fourthly, coffee plays a prominent role in the discussion of neuroenhancers. Even here, proponents of neuroenhancement neglect important distinctions in addiction medicine when they claim that coffee consumption has a similar effect to that of the pharmaceuticals being discussed.1 However, it should be noted that coffee, unlike psychostimulants and other drugs, apparently only triggers a dopamine release in the prefrontal cortex, whereas a dopamine release in the ventral striatum does not occur.24 25 Therefore, there is a neurobiologically definable difference that is highly relevant to addiction medicine between coffee consumption and the use of drugs such as modafinil, which indicates the addictive potential of the latter. Coffee and neuroenhancers are not the same.

Regarding the addiction potential of neuroenhancers, we suggest that it is not a matter of the contingent nature of today’s drugs, but rather a matter of the fundamental properties of such pharmacological agents aimed at cognitive performance, because such substances necessarily modulate systems relevant for learning and memory also implicated in drug addiction.7 Three arguments support this assumption:

1. Modulating learning and memory performance necessarily affect dopamine neurotransmission: basic learning mechanisms such as Pavlovian and operant conditioning are strongly driven by reward and punishment, which is neuronally encoded by monoaminergic systems including dopamine.7 26 27

2. The cognitive capacity for flexible behaviour adjustment and problem solving, fluid IQ, is directly associated with dopamine function in the ventral striatum, a core area of the brain reward system.28

3. Drugs of misuse induce addiction precisely because of their strong interaction with the same monoaminergic systems, particularly dopamine neurotransmission, thus inducing learning mechanisms that promote further drug misuse, for example, via drug craving elicited by Pavlovian conditioned drug cues or via operant reinforcement of drug seeking.8 11 12

We thus suggest that precisely the same neurotransmitter systems implicated in drug addiction have to be influenced by neuroenhancement, which aims to improve learning and fluid IQ. Addictive behaviour is manifested when drugs of misuse exert strong effects on dopamine and related neurotransmitter systems and thus induce counter-regulative adaptive processes, for example, downregulation of dopamine receptors; as a consequence, natural reinforcers lose impact and subjects crave for the non-physiologically strong effects of drugs of misuse.6–11 Indeed, persistence of such alterations in dopamine synthesis and D2 receptors during abstinence is directly correlated with drug craving and the risk of relapse.29 30 In principle, neuroenhancing drugs can have rather mild effects on dopamine neurotransmission. However, if the size of these neurobiological effects is similar to the physiological effects of natural reinforcers, then there is no advantage of neuroenhancement compared to traditional methods such as practice and study; if they exceed physiological effects and induce counteradaptive neuroadaptation, then neurobiologal research6–11 and clinical evidence21 29 30 suggest that there is a substantial risk of inducing addictive behaviour. Addiction severely impairs voluntary behavioural control and learning from non-drug-related outcomes,12 17 and we suggest that any risk associated with NE (no matter how low or eventually treatable) results in a particularly unfavourable risk–benefit ratio for the use of such drugs.

In summary, there is currently no substance fulfilling the requirements of the liberal proponents for approval of real cognitive neuroenhancement, that is, leading effectively to some genuine cognitive improvement. Furthermore, due to neurological findings and theoretical considerations, there is reason to doubt that such agents will be available in the future. Addiction is fundamentally more linked to pharmacological neuroenhancement than the liberal authors admit. Thus, the central precondition for their positive evaluation of neuroenhancement is more speculative than they acknowledge.

Normative conclusions

Even though the above-mentioned reasons for the assumption of an unavoidable addiction potential of today's and future neuroenhancers have an empirical basis, they are theoretical in nature. If there shall be one or more substances in the future without the feared adverse effects, the assumption would be falsified. But how could this assumption be falsified? It is only possible by doing research. Questions concerning the risks and adverse effects of neuroenhancement can ultimately only be answered empirically.31 Here, we do not consider research on neuroenhancement in clinical disorders with severe cognitive impairments but among healthy volunteers aimed to optimise cognitive capacity. Relevant studies would have to test the efficacy of such drugs and their safety, because otherwise there would be the danger of systematically overestimating the effect. Principally, clinical research is justified, provided that the following conditions, among others, are fulfilled: a reasonable goal must be investigated, study participants must give informed consent and an acceptable risk–benefit profile in the investigation as well as a high scientific quality must be guaranteed.32

Secure empirical data on the effects and side effects of neuroenhancers are indeed preferable, but are the necessary elaborate investigations therefore justified? The criterion of the acceptable risk–benefit profile argues against it:

1. If, as suggested above, such interventions are subject to serious risks and involve a significant addiction potential, one cannot assume that such research is relatively unproblematic.

2. The legitimacy of risky research is growing proportionately to the expected benefit. High benefit exists if serious problems can be eliminated or reduced, that is, new therapies for serious illnesses, whereas the investigation of neuroenhancers is without doubt a ‘luxury problem’. All statements by national commissions or ethics commissions in regard to the prioritisation of healthcare give the improvement of human qualities as either low or no priority whatsoever.33 34 Even if this empirical fact is not a strong argument against the legitimacy of such research, it reveals at least a widespread moral intuition.

Therefore, clinical research on neuroenhancers in healthy volunteers is subject to serious risks and is of low priority.

What about research specifically targeted to assess addictive side effects of neuroenhancers? To perform research on the addictive effects of neuroenhancers, larger groups of previously healthy individuals with varying vulnerabilities for addiction will have to be exposed to a potentially addictive drug. The risk of developing addiction may be substantially lower when application of neuroenhancers is limited to clinical research, that is, when subjects with an increased risk of developing addiction are screened for and excluded. However, to date no single variable has been identified that reliably predicts addiction to neuroenhancers. To identify such predictive factors, a substantial number of subjects would have to be exposed to potentially addictive drugs in prospective studies. Since these healthy volunteers do not have a disease, the risk–benefit ratio for such exposures remains rather unfavourable.

Defenders of neuroenhancement suggest that cognitive enhancement could lead to significant advances in productivity and research and point to the role of IQ in productivity and social success.35 36 High IQ is indeed associated with strong work performance and low levels of unemployment and poverty; however, such associations are rather weak: IQ explains about 16% of job performance37 and IQ, parental social status and age together explain no more than 10.4% of the variance when assessing poverty,37 1.7% to 3.1% when assessing unemployment,37 and 1.5% of self-reported crime.37 Moreover, in industrialised countries, average cognitive performance has risen by more than a SD (equivalent to 15 IQ points) since World War II38 without a correlated decrease in these social problems. Therefore, we feel that risking damage to the health of some to improve performance of already healthy individuals cannot be justified ethically by such potential social benefits.

The social consequences of neuroenhancement

But that is not all. Even if one accepts effective and low-risk neuroenhancers, there are important arguments against neuroenhancers—and therefore against clinical studies using them. Some of these arguments refer to the expected social consequences of neuroenhancement. The chances are that neuroenhancement will create a considerable amount of social pressure to use these agents. People who refuse to use neuroenhancers will be forced into accepting occupational disadvantages or have to resolve their convictions and resort to using such agents.

Galert et al would like to invalidate this argument by stating that our society ‘already puts us at risk and under tremendous pressure to conform’.5 However, the fact that there is already a social problem does not justify its expansion to other areas. Above all, neuroenhancement distinguishes itself from other often quoted technological innovations such as cell phones or computers, which also end up forcing one to conform, in that it is a direct intervention into the neurobiological basis of one's own personality.

Furthermore, Galert et al demand a thorough ethical reflection in order to gauge which risk is reasonable and ‘socially adequate’. That is correct. However, if they demand ‘empirical findings’ to be included in the reflection, caution is the key here. Because if that means requiring a widespread social practice of neuroenhancement in order to investigate empirically the mechanisms of social pressure, precisely those social conditions will be regarded as a prerequisite for a sound judgement, the acceptability of which is the issue at the moment. That is to say, one would have to facilitate social conditions that, depending on the outcome of the research, may have to be prevented. Such social processes cannot be sociotechnically undone at will. In addition, the phenomenon of social pressure caused by the use of pharmacological enhancers is widely known and has been adequately studied due to doping in competitive sports.

In short, provided that one does not regard the social consequences of a societal practice of neuroenhancement as morally preferable, even in the event of unproblematic research results with respect to addiction medicine, further arguments arise against the research of these substances. At this point, it becomes clear how the fundamental ethical arguments for and against neuroenhancement are interconnected with the question of the ethical legitimacy of the empirical research needed to develop it.

Further arguments against neuroenhancement include questions of self-efficacy, authenticity and self-coherence. We do not discuss these topics here, although we feel that they are highly relevant for a thorough assessment of the ethical aspects of neuroenhancement.39

In conclusion, even though we do not know for sure what kind of long-term, adverse effects neuroenhancers will have, based on our knowledge of neuropharmacology, we can assume that there is a serious addiction potential even of future pharmaceuticals of this kind; an assumption that could only be falsified through detailed, empirical research. However, there are numerous objections to such research that are based on research ethics, basic morals and ethics of good life. Despite their cumulative weight, these arguments do not justify an absolute ban on the research of neuroenhancers, but argue against the use of scarce resources to carry out this research. We therefore strongly recommend refraining from such research.