Hibicke et al. 2020 – Rats in tubes

Hibicke et al. 2020 engages with big questions about the nature of mind, healing, and psychopharmacology.

It does so by getting rats very high, and then, several days later, forcing them to swim.


I'm still digesting this paper, so for now I'll just make a couple points about method and philosophy of science.

When we study brains, drugs, and psychological models, we mostly care about the impact these things have on the subjective experience of humans.

(There's a tangent here about how instead we sometimes do science for its own sake... discovery as an aesthetic pursuit! Fair enough, but let's leave that to the side for now. "Science to better understand the nature of human subjective experience" captures a lot of what motivates current scientific inquiry.)

It is very difficult to directly study human subjective experience using objective methods, so a lot of scientific work is done on things that we think contribute to human subjective experience (e.g. blood oxygen levels in the brain), or on things that we think approximate things that contribute to human subjective experience (e.g. the behavior of rats).

Both of these approaches produce work that's fun to think about & contributes to our understanding of the world. But there are big questions around how far specific results can take us, i.e. how much any specific result can tell us about what we really care about at the end of the day: the nature of the subjective experience of humans.

We can make pretty pictures of how blood moves through the brain, but it's not trivial to connect those pictures to an understanding of how brain states relate to conscious experience. We can observe that spiders on LSD weave more regular webs than sober spiders, but how does that connect to the experiences that humans have on LSD? (This Slate Star Codex piece on adult neurogenesis does a great job drawing out some of the difficulties here.)

Okay, back to Hibicke et al. 2020. We're finding out that both ketamine and psilocybin can treat depression more quickly than our current standard-of-care antidepressants (depressed patients can respond to ketamine and psilocybin treatment within minutes or hours; patients usually take weeks to start responding to SSRIs).

This is interesting because ketamine & psilocybin have different mechanisms of action – psilocybin & other classical psychedelics are 5-HT2A receptor agonists, while ketamine is a NMDA receptor antagonist (though ketamine may also potentiate 5-HT2A response? I don't yet have enough neurochemistry to sort this all out).

Also interesting: ketamine's antidepressant effect tend to fade after a couple weeks, whereas psilocybin's antidepressant effect can last for months.

Why do these drugs act more quickly than traditional antidepressants? Why does the antidepressant effect of psilocybin last for so much longer (on average) than that of ketamine?

More profoundly: is the antidepressant effect of these drugs caused by the subjective experience of being on the drug? Or is it purely physiological?

These are big questions. Carefully designed human trials with longterm follow-ups & qualitative interviews could probably go a long towards answering them, or at least giving us better answers than we have currently.

Unfortunately, it's hard to create well-designed studies. And human trials with longterm follow-ups & qualitative interviews are extremely expensive and bureaucratically fraught, especially when psychoactive drugs are in the mix (just ask MAPS; see also My IRB Nightmare).

So a lot of research is done using "animal models" (in this case, giving drugs to rats). These are much cheaper & faster to run than human trials, so they're attractive to career researchers who have h-indices to think of.

But studying animals is obviously further away from what we care about (human subjective experience) than studying humans. So what can we learn from rats about the dynamics of psilocybin & ketamine's antidepressant effects in humans? How could we go about inferring anything?

Enter the forced swim test:


As you can gather from that image, in the forced swim test, rats are put in tubes full of water such that they can't touch bottom. If the rat just floats there ("immobility"), that's considered a "depression-like" response. If the rat swims around or tries to climb out, that's considered a less depressed response. If rats who were given an intervention swim and/or climb more than rats who were given a placebo, that's considered an "antidepressant-like effect" of that intervention.

The limitations here are probably obvious – a lab rat in tube of water is quite different from a human who scores highly on a QIDS questionnaire. What can we learn from the rat that applies to the human? What do we assume about the rat, and what do we assume about the human?

Here's some of what Hibicke et al. assumed about their rats:

Despite having a rudimentary default mode network,(34) and showing interest in mirrors, video feed, and still images of other rats,(35) rats are not commonly believed to be among the menagerie of self-aware animals. Thus, as far as modern science is able to determine, rats do not have a sense of self and are incapable of having existential anxiety, pondering the meaning behind their own existence or fearing a reality in which they, as individuals, do not exist. No one knows what the subjective experience of a rat is after being given a psychedelic, but it is highly unlikely that they are able to place that within the context of their life experiences and utilize that knowledge therapeutically to improve their affective state.

Notice the lack of citation after "not commonly believed to be among the menagerie of self-aware animals." The authors are simply asserting that rats are not self-aware, for some definition of self-awareness. (A good discussion of animal consciousness here, though note that consciousness & self-awareness are different.)

Another assumption from that passage: if the benefits of psychedelics are caused by the subjective experience of the psychedelic, those benefits arise from the subject utilizing that experience therapeutically. The content of the subjective experience is something to be used for therapeutic improvement by the subject, not something that happens to the subject or is done for them.

This is getting somewhat semantic, and a little far out. To bring it back to earth, let's just note that there's a temptation to conflate "the subjective experience of a psychedelic reduces depression" with "the subjective experience of a psychedelic reduces depression because the subject uses the experience to do so," but those are actually different claims.

Okay, that's enough for now. There's a lot here, and I should probably read more Kuhn before continuing on in this direction.