Consciousness in Brain Organoids

We can begin by distinguishing between the kinds of brain organoids that raise concerns about moral status from those that do not. A clear moral line can be drawn at the onset of (phenomenal) consciousness — i.e., the point at which an entity can have experiences, or at which there is something it is like to be that entity.^{Reference Shepherd, Watt and Stone14} This is the first point at which brain organoids might become capable of suffering, and the first point at which they might develop some degree of moral status.

It is worth noting that because the brain itself does not have pain receptors, brain organoids might not be vulnerable to pain. However, it may be that meninges — which do have pain receptors — also develop. Moreover, we are discussing suffering generally, which is not limited to the experience of pain. For example, brain organoids might suffer from sensory deprivation. Given this potential for suffering, the onset of consciousness is the first point at which brain organ-oids are due moral consideration in their own right. Here, then, is an important moral threshold: we can treat brain organoids according to existing regulatory frameworks for stem cell research until the point at which organoids develop consciousness, but we should restrict the kinds of research that can take place beyond this point.

To enforce restrictions based on consciousness, we will need some means of deciding whether brain organoids are (or could plausibly be) conscious. It is not immediately obvious how consciousness could be detected in an in vitro model of the human brain, given that such an entity can neither communicate with us directly nor exhibit the kinds of behaviours that suggest conscious experience. How, then, should we determine whether a brain organoid might be conscious, and might therefore possess some degree of moral status?

One option is to look at structure. We could extrapolate from the threshold at which human fetuses begin to develop consciousness. We might have reasonable concerns about the consciousness of brain organoids that have reached an equivalent point of morphological development. Some estimates place the onset of consciousness at around 25 weeks' gestational age,^{Reference Lagercrantz15} and the beginning of pain perception at around 30 weeks,^{Reference Lee16} but there is arguably sufficient uncertainty about the exact threshold that we cannot rule out the possibility that sentience begins as early as 20 weeks' gestational age.^{Reference Brugger17} Assuming this is correct, we can be reasonably confident that a brain organoid lacks even a rudimentary form of consciousness until it resembles the brain of a fetus at 20 weeks' development. Beyond this point, we should treat brain organoids as if they could plausibly possess some degree of consciousness.

An alternative functional approach could involve measuring physical processes that suggest consciousness. For example, we could measure the brain's electroencephalographic responses to magnetic stimulation; the complexity of the brain's response might track the organoid's degree of consciousness.¹⁸ In principle, such metrics could provide a more fine-grained basis for research limits than extrapolating from structure. We endorse the development of such metrics. In the meantime, the extent to which brain organoids resemble the brains of human fetuses at the onset of consciousness could provide a useful, albeit coarse-grained, approach to determining whether consciousness is a realistic possibility.

Many philosophers argue — we think convincingly — that if we are unsure whether a particular being is conscious we should not treat them as if they lack moral status, but instead err on the side of generosity and treat them as if they have at least partial moral status.^{Reference Bradshaw, Knutsson, Munthe, Sebo, Koplin and Wilkinson19} We should not treat brain organoids as mere biological material if they could plausibly be conscious, even if we are not certain whether they possess consciousness. Beyond this point, experimentation with brain organoids is not merely experimentation with human tissue. It is experimentation with an entity that may have interests that matter for its own sake — i.e., an entity with moral status.

Research Limits Based on Consciousness

Would it be ethically permissible to experiment with brain organoids that are (or could potentially be) conscious — for example, to study disorders of the adult or ageing brain? Some people believe, on deontological grounds, that morality includes absolute prohibitions on harming or killing persons without their consent. Should we reject any experimentation with conscious brain organoids on these grounds? No. These absolute moral constraints are usually thought to apply only to persons. Personhood is generally thought to require one to have complex cognitive capacities such as autonomy, rationality, emotionality, moral agency, and/or sophisticated self-awareness.^{Reference Tooley, Kuhse and Singer20} Conscious brain organoids are unlikely to develop such capacities, especially in the absence of meaningful interactions with the outside world. Brain organoids would also lack other features that are sometimes taken to undergird personhood, such as membership in our human community or participation in the social contract.^{Reference Andrews21} Unlike human fetuses, human brain organoids would presumably lack even the potential to become persons at a later stage of development. In principle, some forms of experimentation with brain organoids might therefore be morally legitimate even if these organoids are conscious.

Indeed, many existing practices presuppose that consciousness per se confers neither an absolute right to life nor an absolute right against harmful treatment. Consider, for example, our treatment of nonhuman animals. Mice, rats, and other research animals possess consciousness, but we nonetheless allow many forms of animal research that harm these (and other) animals in order to promote human ends. Although many people — and arguably most philosophers writing on the subject — argue that research animals deserve greater protections, most also hold that it is at least sometimes justifiable to harm research animals in order to promote human ends.^{Reference DeGrazia and DeGrazia22} We can also draw a parallel with late term abortion. As we have seen, consciousness in human fetuses is generally thought to emerge between 20 and 30 weeks' gestational development. However, most jurisdictions permit abortions beyond this threshold. Even in the case of human fetuses, then, consciousness is not generally thought to confer an absolute right to life. Personhood — not consciousness — is the relevant moral yardstick.

We should nonetheless recognize some moral limits to research with conscious brain organoids. Brain organoids are not persons, but if they are conscious they may nonetheless have interests that could be set back by experimentation — for example, interests in achieving experiential well-being and avoiding suffering. These interests ought to be taken into account. Some principles for doing so can be found in the Three R's — Reduce, Refine, Replace — approach to animal welfare,^{Reference Russell and Burch23} which is already well entrenched in animal research guidelines. These three principles can be adapted as follows. First, we should reduce the number of conscious brain organoids used in research to the minimum number necessary to achieve the study's scientific aims. Second, we should refine experimental techniques to reduce possible harms — for example, by administering anaesthetic before experimentation in cases where there is some prospect for pain. Third, in line with the principle of replacement, researchers should use brains organoids only if the aims of the research cannot be realized using non-conscious organoids or other non-sentient material.

Despite its important historical legacy, Russell and Burch's three principles fall short of providing a comprehensive account of the ethics of animal research. A full account of the ethics of brain organoid research will need to draw on a more comprehensive set of moral principles. One such set of principles has recently been developed by Tom L. Beauchamp and David DeGrazia.^{Reference Beauchamp and DeGrazia24} Beauchamp and DeGrazia present six ethical principles of animal research, each of which is intended to be morally uncontroversial. Four of these principles are particularly useful for closing the gaps left by the Three R's. These are the Principle of Sufficient Value to Justify Harm (according to which the anticipated benefits of a research study must be sufficiently weighty to justify the harms to animal research subjects), the Principle of No Unnecessary Harm (according to which researchers should minimize harms to animal subjects), the Principle of Basic Needs (according to which researchers should meet animals' basic needs unless failure to do so is necessary for and morally justified by scientific purposes), and the Principle of Upper Limits to Harm (according to which animal subjects must not be made to endure severe long-term suffering except in rare cases of critically important research.)

These principles suggest three additional limits to organoid research. First, research with conscious (or potentially conscious) brain organoids should proceed only if the anticipated benefits of this research are significant enough to outweigh the expected harms. The greater the expected harm to brain organoids, the greater the expected benefits must be to justify the research. The reverse is also true: the more important the research, the greater the potential harms must be to justify preventing such research from going ahead.

Second, researchers should actively minimize the harms they might inflict through their research. This might involve not only refining experimental procedures to reduce harm, but also conducting research on whatever kind of brain organoids are likely to experience the least suffering. If (as seems plausible) a being's capacity for suffering is tied to the richness and complexity of its mental life, then we can limit the harm we inflict by using brain organoids with the lowest potential degree of consciousness compatible with achieving the goals of the research. In principle, it might be possible to further mitigate harms through gene editing. If brain organoids would develop conscious capabilities in excess of what is required for the research, gene editing could potentially be used to reduce their capacity for consciousness and/or to prevent brain organoids from experiencing pain or other aversive mental states.

Third, in line with the Principle of Upper Limits to Harm, conscious brain organoids should not be made to experience severe long-term suffering unless this is necessary to achieve some critically important scientific goal. As suggested above, it might be possible to mitigate severe suffering by conducting harmful forms of experimentation on brain organoids that possess the lowest possible degree of consciousness and/or have been genetically modified to ease aspects of consciousness that would otherwise contribute to severe suffering.

To summarize, we ought to introduce some limits to research with brain organoids that are or could plausibly be conscious. Such research should proceed only if (a) the research serves a sufficiently important purpose to outweigh the expected costs, including harms to the organoids themselves, (b) the research cannot be conducted using non-conscious organoids or other non-sentient material, (c) researchers use the minimum number of organoids than is required to answer the research question, (d) the organoids used do not have a higher potential capacity for suffering than is necessary to achieve the scientific objectives of the research, (e) the research is designed to minimize possible suffering, and (f) the research would not inflict severe long-term suffering, unless necessary to achieve some critically important purpose.

Research Limits Beyond Consciousness

So far, we have been assuming that brain organoids are unlikely to develop sophisticated forms of cognition. This assumption seems reasonable, at least insofar as these brain organoids would be unable to interact with their environment; the development of sophisticated brain networks is often thought to require both input (such as sensory input) and output (such as the ability to interact with surroundings). However, researchers are already exploring techniques that could eventually push these boundaries — for example, by wiring brain organoids to muscle tissue,^{Reference Giandomenico25} by connecting brain organoids to controllable robotic “bodies,”^{Reference Cohen26} or by implanting human brain organoids into non-human animals' brains.^{Reference Mansour27} Researchers have already created “photosensitive” brain organoids, which feature rudimentary eyes and display neural activity when light is shined on them.^{Reference Quadrato28} It is not implausible to think that sufficiently mature organoids could develop advanced cognitive capacities through interacting with the outside environment.

This possibility has two important implications. The first is that organoids with advanced cognitive capacities might have a wider range of morally relevant interests than organoids that are merely conscious. Cognitively sophisticated beings can develop a wide range of social, emotional, and cognitive needs, and they can suffer these needs go unmet. For example, humans and many non-human animals will suffer if they lack opportunities to socialize with conspecifics, or if they are treated in ways that cause lingering anxiety or fear. Self-consciousness (i.e., the capacity to think of oneself as oneself) has some especially weighty moral implications. Because self-conscious beings can have desires and ambitions for the future, they can be harmed by death even if they are killed painlessly.^{Reference Singer, McMahan, DeGrazia and Lurz29} To kill a self-conscious being therefore requires a weightier justification than to kill a being that lacks self-consciousness or possesses it only to a rudimentary degree. Self-consciousness might also have greater moral significance; indeed, some argue that it is never appropriate to use self-conscious beings for invasive research.^{Reference Hyun30} At a minimum, if brain organ-oids develop advanced cognitive capacities we ought to account for their full range of their welfare needs, not merely narrow interests in avoiding pain or other aversive sensations.

It might be difficult to predict the cognitive functioning or welfare requirements of such organoids. Consistent with existing ethical standards for part-human chimera research,^{Reference Hyun, Savulescu, Beauchamp and Frey31} we should screen human brain organoids for advanced cognitive capacities they could plausibly develop. If brain organoids do develop advanced cognitive capacities, researchers should take any associated welfare needs into account. This might involve providing environmental enrichment, and potentially even social opportunities, appropriate to the kinds of beings these brain organoids are.

There is another reason why it matters, morally, if brain organoids develop advanced cognitive capacities: because some cognitive capacities might affect brain organoids' moral status. Moral status is often (and we think plausibly) considered to be a matter of degree. There are two versions of this view.^{Reference DeGrazia32} On the first view, we might hold that the interests of all conscious beings deserve equal consideration and yet also hold that because different kinds of beings experience different forms of suffering to different degrees, some beings deserve greater protections than others. In line with this view, we might think that cognitively sophisticated brain organoids are susceptible to greater suffering than brain organoids that are merely conscious, and therefore deserve greater protection. On the second view, we might think that the degree to which a being possesses certain cognitive capacities — such as autonomy, sociality, and/or self-consciousness — should directly affect the degree of weight we attach to that being's interests. In line with this latter view, we might think that we ought to attach greater weight to the interests of cognitively sophisticated brain organoids than those that are merely conscious. On either view, research with cognitively sophisticated brain organoids requires a more powerful justification than research with those that lack advanced cognitive abilities.

We have argued that advanced cognitive capacities could give rise to new welfare requirements and potentially increase brain organoids' degree of moral status. Research with cognitively advanced brain organoids should therefore face a further set of research limits. Specifically, research with advanced brain organ-oids should proceed only if (a) they are screened for cognitive capacities they could plausibly develop, (b) any associated welfare requirements are taken into account, (c) brain organoids' cognitive capacities are not more sophisticated than is necessary to achieve the goals of the research, and (d) the research serves a sufficiently important purpose to outweigh the harms to the organoids themselves, taking into account these organoids' (potentially enhanced) degree of moral status.

In applying this framework, we should err on the side of generosity when resolving uncertainty regarding brain organoids' cognitive capacities and/or moral status. All else being equal, it would be worse to mis-takenly treat brain organoids as if they have advanced cognitive capacities (when in fact they lack them) than to mistakenly treat them as if they lack advanced cognitive capacities (and therefore fail to recognise important welfare requirements and/or underestimate their degree of moral status).