Edited by: Martin J. Herrmann, University of Würzburg, Germany
Reviewed by: Chao Liu, Beijing Normal University, China; Elena Rusconi, University College London, UK
*Correspondence: R. Ryan Darby
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Biomedical enhancement refers to the use of biomedical interventions to improve capacities beyond normal, rather than to treat deficiencies due to diseases. Enhancement can target physical or cognitive capacities, but also complex human behaviors such as morality. However, the complexity of normal moral behavior makes it unlikely that morality is a single capacity that can be deficient or enhanced. Instead, our central hypothesis will be that moral behavior results from multiple, interacting cognitive-affective networks in the brain. First, we will test this hypothesis by reviewing evidence for modulation of moral behavior using non-invasive brain stimulation. Next, we will discuss how this evidence affects ethical issues related to the use of moral enhancement. We end with the conclusion that while brain stimulation has the potential to alter moral behavior, such alteration is unlikely to improve moral behavior in all situations, and may even lead to less morally desirable behavior in some instances.
Biomedical enhancement refers to biomedical interventions used to improve certain capacities beyond normal, rather than to restore capacities deficient as a result of a disease (Chatterjee,
Moral enhancement refers to improving moral or social behavior, rather than cognition or physical attributes (Harris and Savulescu,
Moral enhancement raises a number of unique ethical questions that are different than ethical issues in the use of cognitive enhancement. For example, cognitive enhancement benefits the individual but can be harmful to society in terms of distributive justice and fairness. However, the opposite argument could be made for moral enhancement, which is more likely to benefit society but not necessarily the individual (Douglas,
A critical point is that discussions about cognitive enhancement often focus on improving cognitive capacities as a means to improve cognition, while discussions in moral enhancement have focused on the ultimate ends of moral enhancement. However, by doing so, the practical means of how altering specific cognitive-affective capacities can lead to changes in moral behavior is neglected. Given the intense and conflicting ethical opinions about moral enhancement, it is especially important to consider whether morality can be enhanced, and, if so, what the likely consequences of such modification will be Douglas (
In this article, we review the existing literature on the use of non-invasive brain stimulation (NIBS) to modulate moral behavior. Rather than improving or impairing a single “moral” process, we will show that non-invasive brain stimulation alters specific neuropsychological processes contributing to normal moral behavior. Such alterations can be viewed as moral “enhancement” in certain situations, but may lead to immoral behavior in other situations. Finally, we will discuss how the evidence from non-invasive brain stimulation affects the ethical debate on moral enhancement.
We searched Pubmed (from indexing through 1/2016) for all articles involving non-invasive bran stimulation (using the terms “transcranial magnetic stimulation” OR “transcranial direct current stimulation” OR TMS OR tDCS OR “non-invasive brain stimulation” OR “theta burst stimulation” OR TBS) and moral behavior (using the terms moral OR morality OR altruism OR cooperation OR fairness OR unfairness OR empathy OR social OR aggression). We additionally searched references in review articles for additional cases. This search resulted in 470 articles. From these, 48 specifically involved the use of non-invasive brain stimulation on morally relevant behaviors. We classified articles based on the methodology (neurophysiological response to moral stimuli using NIBS vs. modulation of moral behavior using NIBS), NIBS parameters (TMS vs. tDCS, stimulation, frequency, duration, location, use of neuronavigation), psychological tests used, and significant findings.
A small but growing number of studies have examined the role of non-invasive brain stimulation in moral behavior and decision-making in normal persons (Table
Sellaro et al., |
Anodal/Cathodal/Sham | R. TPJ | Anodal stimulation of R. TPJ led to diminished moral blame for accidental harms |
Intensity of 1 mA | Target: CP6, 35 cm2 | ||
Duration of 20 min | Ref. L. SO, 35 cm2 | ||
Testing post-stimulation | |||
Sellaro et al., |
Anodal/Cathodal/Sham | mPFC | Anodal stimulation of mPFC reduced racial bias on IAT for reaction time and errors (Cohen's |
Intensity of 1 mA | Target: CPz, 35 cm2 | ||
Duration of 20 min | Ref. Oz, 35 cm2 | ||
Testing during last 10 min | |||
Kuehne et al., |
Anodal/Cathodal | L. DLPFC | Anodal stimulation of L. DLPFC led to LESS utilitarian responses to hard personal dilemmas |
Intensity of 1 mA | Target: F3, 35 cm2 | ||
Duration of 20 min | Ref. P4, 35 cm2 | ||
Testing during last 10 min | |||
Kelley et al., |
Anodal/Cathodal/Sham | L. or R. DLPFC | Anodal stimulation of L. DLPFC stimulation increased jealousy ratings relative to right and sham after social exclusion |
Intensity of 2 mA | Target: F3 or F4, 35 cm2 | ||
Duration 15 min | Ref: F3 or F4, 35 cm2 | ||
Testing post-stimulation | |||
Nihonsugi et al., |
Anodal/Sham | R. DLPFC | Anodal stimulation of R. DLPFC increases trust/cooperation |
Intensity 2 mA | Target: MNI (44, 34, 22), 35 cm2 | ||
Unknown Duration | |||
Testing after 5 min | Ref: Oz, 35 cm2 | ||
Dambacher et al., |
Anodal/Sham | R. DLPFC | Anodal stimulation of R. DLPFC reduced proactive aggression in men only (Cohen's |
Intensity of 2 mA | Target: F4, 35 cm2 | ||
Duration 12.5 min | Ref: L. SO, 35 cm2 | ||
Testing during stimulation | |||
Wang et al., |
Anodal/Cathodal/Sham | L. DLPFC | Anodal stimulation of L. DLPFC increased empathic pain ratings |
Intensity of 2 mA | Target: F3, 35 cm2 | ||
Duration of 5 min | Ref: FP4, 25 cm2 | ||
Testing post-stimulation | |||
Riva et al., |
Cathodal/Sham | R. VLPFC, | Cathodal stimulation to R. VLPFC increased negative feelings associated with exclusion (Cohen's |
Intensity of 1.5 mA Duration of 20 min | Target: F6, 25 cm2 | ||
Testing during last 15 min | Ref: L. SO, 35 cm2 | ||
Riva et al., |
Anodal/Sham | R. VLPFC, | Anodal stimulation of R. VLPFC decreased aggression after social exclusion (Cohen's |
Intensity of 1.5 mA Duration of 20 min | Target: F6, 25 cm2 | ||
Testing during last 15 min | Ref: L. SO, 35 cm2 | ||
Riva et al., |
Anodal/Sham | R. VLPFC, | Anodal stimulation of R. VLPFC decreased hurt feelings (Cohen's |
Intensity of 1.5 mA Duration of 20 min | Target: F8, 25 cm2 | ||
Testing during last 15 min | Ref: L. SO, 35 cm2 | ||
Ruff et al., |
Anodal/Cathodal/Sham Intensity of 1 mA | R. VLPFC | 1. Anodal stimulation of R. VLPFC increases (+33.5%), while cathodal decreases (–22%), giving in ultimatum game, |
Unknown Duration | Target: MNI (52, 28, 14), 35 cm2 | ||
Testing during stimulation | Ref. Cz, 35 cm2 | ||
Mameli et al., |
Anodal/Sham | Bilateral DLPFC Target: F3/F4, 32 cm2 each | Anodal stimulation to bilateral DLPFC decreased reaction times to making lies about general knowledge (Cohen's |
Intensity of 2 mA | |||
Duration of 15 min | Ref. R. deltoid, 64 cm2 | ||
Testing post-stimulation | |||
Fumagalli et al., |
Anodal/Cathodal | Bilateral mPFC | Anodal stimulation of the bilateral mPFC increased, and cathodal decreased, utilitarian judgments in females only |
Intensity of 2 mA | Target: “above eyes,” 54 cm2 | ||
Duration 15 min | Ref. R. deltoid, 64 cm2 | ||
Testing post-stimulation | |||
Karim et al., |
Anodal/Cathodal/Sham Intensity of 1 mA | Anterior PFC | Cathodal stimulation of R. DLPFC improved lying score, reduced reaction time, and reduced guilt of lying |
Duration of 18 min | Target: FP2, 24 cm2 | ||
Testing after 3 min | Ref: PO3, 24 cm2 | ||
Hortensius et al., |
Anodal/Cathodal/Sham Intensity of 2 mA | R. or L. DLPFC | Anodal Stimulation of L. DLPFC/Cathodal stimulation of R. DLPFC increased aggressive responses |
Duration 15 min | Target: F3 or F4, 35 cm2 | ||
Testing post-stimulation | Ref. F3 or F4, 35 cm2 | ||
Priori et al., |
Anodal/Cathodal/Sham | Bilateral DLPFC | Anodal stimulation of bilateral DLPFC prolonged reaction times to lies (Cohen's |
Intensity of 1.5 mA | Target: F3/F4, 32 cm2 each | ||
Duration of 10 min | Ref. Deltoid, 64 cm2 | ||
Testing post-stimulation | |||
Knoch et al., |
Cathodal/Sham | R. DLPFC | Cathodal stimulation of R. DLPFC increased acceptance of unfair offers (+21.2%), without changing judgments of fairness, in ultimatum game |
Intensity of 1.5 mA | Target: F4, 35 cm2 | ||
Duration of 10 min | Ref. L. SO, 100 cm2 | ||
Testing after 3 min | |||
Civai et al., |
Cathodal//Sham | Bilateral mPFC | Cathodal stimulation of mPFC increases rejection of unfair offers in ultimatum game when playing for a third party |
Intensity of 2 mA | Target: MNI (–2, 58, 8), 35 cm2 | ||
Duration of 20 min | Ref: R. arm, 35 cm2 | ||
Testing after 2 min | |||
Fecteau et al., |
Anodal/Cathodal/Sham | R. and L. DLPFC | Both stimulation conditions increased lie generation compared with sham stimulation |
Intensity of 2 mA | Target: F3 or F4, 35 cm2 | ||
Duration of 20 min | Ref: F3 or F4, 35 cm2 | ||
Testing post-stimulation | |||
Boggio et al., |
Anodal/Sham | L. DLPFC, M1, V1 | L. DLPFC anodal stimulation reduced unpleasantness (−5.8%) and emotional discomfort (−8.9%) to viewing pain in others |
Intensity of 2 mA | Target: F3, C3, or Oz, 35 cm2 | ||
Duration of 5 min | Ref: R. SO, 35 cm2 | ||
Testing during session | |||
Rêgo et al., |
Anodal/Cathodal/Sham | L. or R. DLPFC | L. DLPFC cathodal/R. DLPFC anodal reduced emotional valence and arousal ratings to viewing pain in others |
Intensity of 2 mA | Target: F3 or F4, 35 cm2 | ||
Duration of 15 min | Ref: F3 or F4, 35 cm2 | ||
Testing after 5 min | |||
Dambacher et al., |
Anodal/Cathodal/Sham | L. or R. DLPFC | No effects of either stimulation paradigm on Taylor aggression paradigm |
Intensity of 1.5 mA | Target: F7 or F8, 35 cm2 | ||
Duration of 22 min | Ref: F7 or F8, 35 cm2 | ||
Testing post-stimulation | |||
Wang et al., |
Anodal/Sham | R. Orbitofrontal | Anodal stimulation to R. orbitofrontal cortex increased trust/giving in trust game (+15%) |
Intensity of 2 mA | Target: FP2, 9 cm2 | ||
Duration of 15 min | Ref: F4, 9 cm2 | ||
Testing post-stimulation | |||
Ye et al., |
Anodal/Cathodal/Sham | R. or L. TPJ | 1. Anodal L. TPJ/ cathodal R. TPJ reduced blame for attempted but unsuccessful harms. |
Intensity of 2 mA | Target: CP5 or CP6, 35 cm2 | ||
Duration of 20 min | Ref: CP5 or CP6, 35 cm2 | ||
Testing after 15 min | |||
Sowden et al., |
Anodal/Active Sham | R. TPJ vs. Occipital | Anodal R. TPJ improved detection of lying in others (6.8%) |
Intensity of 1 mA | Target: CP6 or Oz, 35 cm2 | ||
Duration of 20 min | Ref: Vertex, 35 cm2 | ||
Testing post-stimulation | |||
Jeurissen et al., |
3-pulses, 150 ms apart | R. DLPFC, R. TPJ Target: Talaraich (39, 47, 7) or (60, −40, 19) | 1.3-pulse inhibition at 2.5 s of the R. DLPFC DECREASED utilitarian decisions to personal dilemmas. |
Intensity of 70% machine output | |||
Stimulation 1.5, 2, 2.5, or 3 s into decision | |||
Balconi and Canavesio, |
TMS 10 Hz vs. Sham Intensity of 120% RMT Duration of 1 s per trial for total of 80 trials | L. DLPFC | 10 Hz TMS of L. DLPFC increased decision to help in all scenarios except neutral |
Target: Talairach (–1, 45, 15) | |||
Baumgartner et al., |
rTMS 1 Hz vs. Sham Intensity of 110% RMT Duration of 20 min, Testing post-stimulation | L. TPJ or R. TPJ | 1 Hz TMS of R. TPJ decreased punishment of outgroup persons |
Target: MNI (−45, −60, 21) or (57, −60, 30) | |||
Tassy et al., |
rTMS 1 Hz vs. Sham Intensity of 54% stimulator output | R. DLPFC | 1 Hz TMS of R. DLPFC decreased utilitarian responses to high conflict personal moral dilemmas (OR 0.248) |
Duration of 15 min | Target: Talairach (45, 36, 24) | ||
Testing post-stimulation | |||
Young et al., |
1. rTMS 1 Hz vs. control Intensity of 70% machine output | R. TPJ | 1.1 Hz TMS of R. TPJ increased permissibility for attempted but unsuccessful harms. |
Duration of 25 min | Target: MNI (60, −54, 34) | ||
Testing post-stimulation | Control stimulation 5 cm posterior to this region | ||
2. TMS 10 Hz vs. control Intensity of 60% machine output | |||
Duration of 500 ms at beginning of each judgment | |||
Knoch et al., |
rTMS 1 Hz vs. Sham | R. or L. DLPFC, Target: Talaraich (±39, 37, 22) | 1 Hz rTMS of R. DLPFC increased acceptance of unfair offers in ultimatum game (+35.4%) |
Intensity of 54% machine output | |||
Duration of 15 min | |||
Testing post-stimulation | |||
Buckholtz et al., |
rTMS 1 Hz vs. Sham Intensity of 30% machine output | R. or L DLPFC, Target: Talairach (±39, 37, 22) | 1 Hz TMS of either R. or L. DLPFC reduced punishment for responsible moral violators, without changing judgments of blameworthiness, or responsibility (Cohen's |
Duration of 30 min | |||
Testing post-stimulation | |||
Perach-Barzilay et al., |
cTBS (3-pulse 50 Hz, delivered at rate of 5 Hz) vs sham, intensity 100% aMT, duration 1 min, testing post-stimulation | R or L. DLPFC | R. DLPFC cTBS reduces both reactive and proactive aggression (Cohen's |
Target: 5 cm anterior to motor hot-spot | |||
Strang et al., |
rTMS 1 Hz vs. Sham | R. or L DLPFC, Target: Talairach (±39, 37, 22) | 1.1 Hz rTMS to R. DLPFC reduced giving in both dictator and ultimatum. |
Intensity of 110% RMT | |||
Duration of 15 min | |||
Testing post-stimulation |
One important and well-studied factor in moral behavior is the aversion to violating the personal rights of other persons, such as causing them physical harm. One method of experimentally testing aversion to harming others has been to use personal and impersonal moral dilemmas, where a subject must chose whether to harm one person in order to save many others (Greene et al.,
Complicating matters further, anodal stimulation to the left DLPFC led to decreased utilitarian responses (or an increased sensitivity to moral harms) (Kuehne et al.,
Non-invasive brain stimulation can also influence altruism, trust, cooperation, and other prosocial behaviors. One approach to studying prosocial behavior uses economics games where one person (player A) is given a sum of money and has the option to give a portion of that money to another person (player B). In the dictator game, there is no threat of retribution for unfair offers, while in the ultimatum game, player B can reject an unfair offer, in which case neither player will receive any money (in a sense, paying to punish the unfair offer of player A). Anodal tDCS to the right DLPFC led to lower offers without the threat of punishment but higher offers with the threat of punishment, while cathodal tDCS had the opposite effect (Ruff et al.,
A related issue is whether to punish others who are unfair or violate other social norms. Both low frequency rTMS (Knoch et al.,
Finally, several studies have investigated modulation of the right temporal-parietal junction (right TPJ), an area important in mediating the effects of intentions and mental states on moral judgments, a process referred to as theory of mind (Young et al.,
Based on these studies, tentative proposals can be mad for the regionally specific effects of brain stimulation on moral behavior.
Inhibition of the right DLPFC reduces the influence of harm on decision-making, both when deciding to perform a harmful action oneself, and when punishing harmful or unfair behaviors in others. This reduces decisions to punish or disapprove of potentially harmful personal moral violations in others. However, it also increased the likelihood of performing a harmful action oneself, such as lying or reacting aggressively. In contrast, excitatory brain stimulation to the right DLPFC reduced aggression and increased adherence to social norms. Taken together, these results suggest that the right DLPFC is important for representing aversion to harm, leading individuals to avoid harming others, but also to punish those who do cause harm.
This role may not be specific to the right hemisphere, as stimulation of the left DLPFC may also results in increased cooperation and increased consideration of harm violations in moral dilemmas.
Finally, the right TPJ presumably exerts a more fundamental role in modulating the role of beliefs and intentions in moral judgments. This may influence the degree to which a harmful action is perceived as moral (e.g., intended) vs. non-moral (e.g., an accident). It also more generally shifts focus toward the viewpoint of other persons, leading to improved lie detection, for example Sowden et al. (
The previous section provides evidence to support our hypothesis that non-invasive brain stimulation can modulate certain cognitive-affective capacities involved in moral cognition. In the following section, we will discuss how modulating capacities contributing to moral behavior, rather than assuming enhancement of a single moral capacity, alters key ethical considerations in moral enhancement.
Autonomy refers to the capacity of a person to define his or her preferences, desires, values, and ideals (Dworkin,
Certain proponents have argued that moral enhancement should be required, or obligatory (Persson and Savulescu,
Again, such an argument assumes that one intervention can be used to enhance all moral decisions, without considering the mechanism by which such enhancement must occur. The evidence presented from non-invasive brain stimulation instead shows that only specific cognitive-affective capacities contributing to moral behavior are likely to be enhanced, affecting some but not all behaviors, and not always in the desired direction. For example, improving empathy may lead to improved moral behavior toward one's in-group, but more harsh retaliation against those outside of one's group. Conversely, increasing cognitive processes contributing to utilitarian reasoning to consider the interests of many groups of persons might come at the expense of decreasing one's aversion to harming persons in general. Given this reality, it is unlikely that moral biomedical enhancement will alter behavior to the extent necessary to requiring individuals to use moral enhancing technologies. Proponents of moral enhancement agree that such technologies are unlikely in the near future but should be possible eventually (Persson and Savulescu,
Harris (
This argument is perhaps similar to the authenticity debate in cognitive enhancement, where some have expressed concern that enhancement would taint the authentic, or true, self (Parens,
Part of the difficulty in understanding moral enhancement is that many moral challenges do not have clear solutions. So while there is agreement that improving memory, multi-tasking, and executive functions are indicators of cognitive enhancement, it is not clear how moral enhancement is supposed to affect our views on controversial topics like abortion, distributed justice, or choosing between the welfare of different groups of persons. A more reasonable goal would be to enhance certain moral motivations that are universally agreed upon, which research in non-invasive brain stimulation has shown to be theoretically possible. However, enhancing a single capacity is unlikely to lead to what one would consider a superior moral decision in every instance.
That being said, it is reasonable that enhancing certain cognitive-affective processes would lead to improved moral behavior on average (Douglas,
In conclusion, limited but growing evidence thus far suggests that brain stimulation can modulate specific cognitive-affective processes involved in moral behavior, making moral enhancement possible. However, rather than improving one single moral capacity, brain stimulation alters specific neuropsychological processes contributing to moral behavior. Enhancement of these processes can lead to morally enhanced behavior in some situations, but less morally desirable behavior in other circumstances. This influences the ethical debate regarding moral enhancement, showing that technologies will be unlikely to change moral behavior to the extent required to make moral enhancement obligatory, or to raise concern regarding our freedom to act immorally. However, the more modest goal of improving our tendencies to act in accordance with our moral motivations is likely possible, and may be desirable for large numbers of people. Given the very real threats of weapons of mass destruction and other technologies noted by past and current scholars, there is a clear need to research moral interventions in greater depth. Such research is needed to determine the implications of enhancing certain capacities contributing to moral behavior so that informed, rational debates regarding the use of moral enhancement are possible.
RD for study concept and design and for writing the manuscript, AP for critical revisions of the manuscript.
This work was supported by funding from the Sidney R. Baer, Jr. Foundation (RD, AP), the NIH (R01HD069776, R01NS073601, R21 NS082870, R21 MH099196, R21 NS085491, R21 HD07616 to AP), the Football Players Health Study at Harvard University (AP), and Harvard Catalyst. The Harvard Clinical and Translational Science Center (NCRR and the NCATS NIH, UL1 RR025758 to AP).
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.