Hidden Perils of a Learning Pill
Or, Don't Reopen that Critical Period!
A blog post about a future technology.
What it means to age, and in part to be human, is to become set in one's patterns and ways. But what if you could reawaken your ability to learn? Facts and events would suddenly stick in your mind again, the way young children are sponges for information. Imagine learning a second language at 60 as easily as a 5 year old. Perhaps there was a skill you wish you had begun at a younger age — piano, violin, soccer, coding. What would you learn?
For a few decades now the field of critical period plasticity has promised this redemptive vision. The promise is medicalized, as is usual: in cases of stroke, surgical removal of brain tissue, or traumatic brain injury, one may very much wish for the ability to re-learn an ability that has been lost. If you've lost the ability to speak from a stroke (as in Broca's aphasia), it can take years of difficult rehab to relearn to communicate your thoughts, and even then complex expressions may entirely evade you. Learning like an infant would seem a godsend.
There's no doubt that a boost to plasticity would be wonderful for people in need. However, I think there are three big reasons to be cautious about this capability, if we ever stumble upon it. Expanded below, these are 1) a potentially unavoidable downside to the technology itself, 2) its potential for societal transformation, and 3) an issue of medical ethics.
First, some background. Feel free to skip ahead.
Critical periods for learning
The "critical period" refers to a special phase in a developing brain when, put simply, learning is astonishingly fast and effective. If a young brain is damaged in this phase – a minor stroke, for example – the brain quickly repurposes non-damaged tissue for the lost functions. People can live decades before a doctor tells them they must have had a stroke, judging by an unusual brain scan. Individuals blind from birth repurpose their visual cortices for audition, action, and higher cognitive functions. At these young stages, one begins to think that brain tissue is just learning stuff capable of learning anything. (In reality a young brain won't learn "anything", though we still don't have any idea how to predict what sorts of things are in or out of bounds.)
The critical period ends at different times for different functions. After about 8 months in humans, if sight is restored in a blind eye the brain will nevertheless not use that eye for vision. This is why critical periods are called critical – it's critical that restorative therapies be delivered in this phase.
It's worth mentioning that there's debate about the critical period as a unified concept. The term arose for vision, for which certain things are simply not possible after a certain age. People may be able to learn other things to some degree at any age. Learning slows regardless. I use the term rather informally to refer to the general and striking slowdown in learning rate with age.
One is silver and the other gold
The critical period closes for a reason. A popular theory of why we evolved to slow our rate of learning relates to the so-called stability/plasticity tradeoff. This theory says that memories and abilities can be "overwritten" by new learning, so too much learning is dangerous. If we want to remember the old friends in our memory, we must limit how much new stuff we learn. In this framework, the reason you can't remember anything from your early years is because you learned so much afterwards.
If this story is true, reopening the critical period could be transformative in ways that are hard to predict. Wishing to simply regain one's sense of speech, one's memories of early life could fade. New skills could be gained, but old ones lost. In the worst-case scenario, a patient re-entering the womb of critical learning may show severe retrograde amnesia – for either procedural or declarative memory, or both.
There is a certain tragedy to this therapy. In Greek mythology, the fountain of Lethe brought total amnesia, and was to be drunk before passing the river into Hades. To forget was to die, but also to live on in a new realm. Reopening the critical period, if it unavoidably requires amnesia, is a Lethean therapy and worth good caution.
It is not at all clear that learning "too much" necessarily brings amnesia, however. The stability/plasticity tradeoff is a theory – and even if it's true in principle, it's not clear that the brain is up against the limit of how much can be learned. How much capacity for (consolidated) experience and memory does the brain actually have? Maybe we're far below the limit, and plasticity slows after the critical period for other reasons. Visual tasks could be learned "to completion" as the brain sees it, for example, and evolution had no reason to keep around sensory plasticity mechanisms after that point.
It is telling, perhaps, that neuroscience's friends in deep learning face the same tradeoff. There, researchers know that training artificial neural networks on a second task can completely destroy the ability to perform a previously learned task. This retroactive interference is dramatically termed "catastrophic forgetting". This relates to the way we train artificial neural networks, though, and not the networks themselves; the capacity of a single network to learn multiple tasks is massive. (Train on both tasks at the same time and the problem goes away). This story may or may not have relevance to neuroscience. What it does tell us is that preventing forgetting is, at the least, a hard engineering problem that AI research hasn't figured out how to solve. It tells us that maybe, just maybe, the critical period's brakes on plasticity are serious business.
To answer this question, it would be wonderful to know what protects memories in the first place. We certainly don't forget everything from young childhood. Skills and language grow over time – though you can forget languages if you move to a new country. Keeping memory while artificially increasing plasticity will likely require balancing whatever combination of continued practice and exposure, internal rehearsal, and molecular protections the brain naturally uses to retain memory. Until these questions are answered, it is very possible that a critical period therapy is, unavoidably, a Lethean bargain.
A future of lifelong rebirth
A fascinating thing happens when we imagine what a plasticity therapy would look like even in the best-case scenario: a learning pill with no forgetfulness. As with any technology, it's crucial we imagine early on what use would like beyond immediate (medical) application. Who would use such a pill? Would applications stick to regain-of-function or begin to stray?
In many ways this questions parallels the debate around brain-machine interfaces. Though initially funded as a cure to paralysis or other loss of "normal" bodily functions, already the developers of this technology imagine a world in which we communicate telepathically and transcend the limitations of our body-bound humanity. Sure, this is the hype of Elon Musk, but even the more down-to-earth developers at Facebook (I never thought I'd type that) are developing BMIs primarily for virtual reality control. When BMIs finally mature, I think it's clear that they'll be used primarily to gain new functions.
What sort of things would people do with a voluntary learning pill? It's hard to imagine. We'd see languages and professional skills, of course; new skills learned late. But it may also become possible to learn more of one thing than ever before. What sort of specialists will we produce when we unlock the secrets to plasticity? Fearing the worst, we may one day see individuals willing to repurpose massive amounts of their cortex to special functions. Olympic athletes and all those obsessed with greatness may go further than has ever been possible – potentially at the expense of other cognitive skills.
We might also see changes to our views of ourselves. What does it mean to be 50 years old? Or 90? We often think of late age as a settled, developed time. It is a "fact of life", as much as the impressionability of children is a fact of life. Late ages are marked by wisdom, but also by a deep sense of self-definition and identity. Yet if learning could be restored, the themes of the chapters of life may have to be edited. Perhaps these edits may not be more extreme than those of, say, antibiotics and a doubling of lifespan, but that's saying something. In a society with a cheap plasticity, would 30 be the new 10?
A mind by whose design?
Much of the way we structure society today takes for granted the adaptability of the young. Education is an obvious example. Or consider the criminal justice system, which frequently locks people away for decades on the premise that they cannot change and remain a threat to society. In this case, would extensible plasticity be a blessing or a curse?
It is not hard to imagine a criminal system that offers, instead of lifetime incarceration, a plasticity therapy. Take some pills, enter a learning program, and one's behavior and worldview may truly change. Childhood trauma may even fade. This is dicey ethical territory for the reason that no one can anticipate who they will become afterwards. This changes the notion of consent, as arguably one can only consent to things for which the outcome is understood. If offered to the incarcerated as an "alternative", it's not really consensual as the alternative is continued punishment. A medical treatment to "criminality" may seem kind, but how is a compulsive shift in personality, memory, and identity respectful of the dignity of a human being?
In the final scene of Kubrick's "One flew over the cuckoo's nest," (spoiler ahead! You don't want this spoiled, I assure you. Go watch it.), the empty, lost gaze of Randle (Jack Nicholson) stares past the camera. It's one of the most affecting scenes I've seen. To compel someone to undergo a lobotomy as a behavioral therapy is deeply unethical. This feels more extreme than a plasticity therapy, yet when you try it's fairly hard to draw any abstract lines of separation. Both are psychological alterations. Both potentially involve a loss of memory. They are compelled changes of the foundation of someone else's internal experience.
On the other hand, it's hard to separate this from standard education programs we look positively upon. Think literacy and job training, either in society at large or in the context of criminal justice. (Though there are certainly types of unethical education, like "brainwashing" or programs of cultural erasure and assimilation). Not all "psychological alterations" are unethical. The ethical line lies somewhere between your middle school experience and a lobotomy, and a plasticity therapy lies somewhere in there too.
It's probably premature to talk seriously about the ethics of this. A plasticity therapy could be so many things, with any variety of tradeoffs. Yet I find the ethical questions it brings up fascinating. By what principles should we judge the ethics of changing someone's mind? What right do we have to our own memories? If only Phillip K. Dick were still around to help us think through it.