Clerestory

Serotonin and Learning

June 15, 2020

I’m still thinking about the creative/editorial cycles involved in writing and in filters more generally.

Kristijan Ivancic pointed out that this cycle of exploration and consolidation relates to the explore-exploit trade-off, and to the trade-offs between going “a mile wide and inch deep” versus going “an inch wide and a mile deep.” In short, another way to look at the two modes is in the trade-off between generalists and specialists.1 The generalists seek new horizons (explore), switching frequently, whereas the specialists stay put and double down on what they’re doing (exploit). This is loosely the same opposition I discussed between the avant-garde and its critics.

I have a hunch that these two modes, namely exploration and exploitation, may be mediated by the neurotransmitters serotonin and dopamine, respectively.

The backstory

I’ve been thinking about these two transmitters for a few years, since reading Michael Pollan’s How to Change Your Mind, and Robin Carhart-Harris’ phenomenal paper The Entropic Brain, which the book discusses at length. I wrote about both works here in 2018.

Since then, one question that’s stuck with me is this: What possible purpose could the neurotransmitter receptors to which psychedelics bind have had over evolutionary time? Carhart-Harris argues that the effect of psychedelics depends entirely on how potent of a partial agonist of a specific serotonin receptor a substance is (the receptor is 5-HT2A, one of 14 different serotonin Receptors found in mammals). LSD is extremely potent, possibly because of its similarity in shape to the shape of serotonin itself, which can leave the drug stuck in a receptor for a long time.

Obviously these brain receptors had already evolved by the time LSD was discovered in 1943, and they must also have existed before psilocybe mushrooms, which act by the same mechanism, could have had any effect when eaten. Even if you accept McKenna’s highly controversial “stoned ape” theory, which not only dates mushroom consumption to around 100,000 years ago, but credits them for much of subsequent human development, we would still have needed to have some response to psychedelic compounds before that. And of course, the fact that scientists experiment on rodents who also have these receptors in their brains suggests that these receptors are ancient (though of course their purposes may have changed).

People have long noted the similarities between the psychedelic experience and religious experience (e.g., Alan Watts in 1968). Religions as we typically think of them today arose only a few thousand years ago, but even if you include much older animistic religions, religion itself must also have evolved (culturally speaking) to co-opt existing systems in the brain which had evolved for other purposes.

Carhart-Harris has also noted the similarity with near-death experiences. But it seems unlikely to me that there could have been sufficiently strong selection pressure for producing a pleasurable or trippy death, doesn’t it?

What do these neurotransmitters do again?

For a refresher, see Healthline. The tl;dr is that serotonin seems to maybe have something to do with happiness, though it’s not as simple as “more serotonin means more happiness.” Dopamine pretty clearly has something to do with rewards/pleasure/craving, and therefore with addiction. SSRIs, drugs which increase the amount of serotonin in the brain by preventing its reuptake, do help some people with depression, though the mechanism is not straightforward .

But what if serotonin’s primary use is to facilitate learning?

  • There’s at least some evidence that agonists at this receptor are “cognition-enhancing” and that they’re involved in learning and memory consolidation. This paper gives evidence that 5-HT2AR is involved in “acquisition of fear memory extinction,” which is a nice way to say that scientists terrified mice then taught them not to be terrified and in the process found that this receptor was involved in learning not to be afraid again.
  • Antidepressants seem to work better when administered in conjunction with therapy or other interventions, suggesting that these drugs make changes easier rather than cause changes directly.
  • There’s also evidence that psychedelics increase the personality trait of openness, even in adults. Personality traits don’t change much by adulthood, so it’s surprising that psychedelics can do this, and that the change is in the direction of openness to new learning. The changes were found to persist for a year.
  • Serotonin levels seem to peak in infancy if I’ve read this paper correctly, and to decline with age.
  • Psychedelics seem to attenuate the activity of the Default-Mode Network, or DMN, which among other things seems to act as a sort of “top-down routing table” for the brain. Interfering with this network seems to improve neuroplasticity (as well as suspending reality-testing, etc). For more on this, read the The Entropic Brain. I’m serious, it’s really good.

Perhaps the mechanism for improved learning is improved plasticity by a reduction in the strength of priors, which Carhart-Harris has argued more recently (2019):

The hypothesized flattening of the brain’s (variational free) energy landscape under psychedelics can be seen as analogous to the phenomenon of simulated annealing in computer science—which itself is analogous to annealing in metallurgy, whereby a system is heated (i.e., instantiated by increased neural excitability), such that it attains a state of heightened plasticity, in which the discovery of new energy minima (relatively stable places/ trajectories for the system to visit/reside in for a period of time) is accelerated (Wang and Smith, 1998). Subsequently, as the drug is metabolized and the system cools, its dynamics begin to stabilize—and attractor basins begin to steepen again (Carhart-Harris et al., 2017). This process may result in the emergence of a new energy landscape with revised properties (Fig. 1).

In other words, serotonin, via 5-HT2A, may relax priors about the world, making it easier to find other ways of viewing it.

I also think serotonin has a connection to fasting. This post was getting a bit long so I concluded the argument about dopamine and serotonin on that page.


  1. David Epstein’s excellent book Range is largely about this trade-off, and he very much comes down on the side of the former, i.e., going as wide as possible and specialising late.


I'm Bryan Kam. I'm thinking about complexity and selfhood. Please sign up to my newsletter or see more here.