Mixed Bag #22: Nils Wendel on Learning Psychopharmacology
“Mixed Bag” is a series where I ask an expert to select 5 items to explore a particular topic: a book, a concept, a person, an article, and a surprise item (at the expert’s discretion). For each item, they have to explain why they selected it and what it signifies. — Awais Aftab
Nils Wendel, MD, is currently a 4th year psychiatry resident at Tufts University, soon to be an attending psychiatrist at Duke University. He doesn’t feel like much of an expert in anything but guesses that 4 years of undergraduate neurobiology, 4 years of medical school, and 4 years of psychiatry residency probably count for something at this point. He writes on Substack at Polypharmacy.
Book — Andy Clark, Surfing Uncertainty: Prediction, Action, and the Embodied Mind (2015)
How do you reason about psychopharmacology from first principles? Can you even do such a thing?
This is the fundamental problem with psychopharmacology, and it eats at me. All we can see are the results of our treatment. Sure, we can propose mechanisms all we want, but these attempts fail. No class of psychiatric drugs we use today was discovered based on first principles. Chlorpromazine, the first antipsychotic, was originally developed as an anesthetic agent because it lowered core body temperature. Its antipsychotic properties were only discovered when it was used as a replacement for blankets soaked in cold water as a method to treat agitation in the mentally ill. MAOIs, the first class of antidepressants, were originally tuberculosis drugs. When we do try to use first principles, we tend to fall flat. Trazodone was really supposed to be an alpha-1 antagonist, based off of a “psychic pain” model of the brain; it works, but only because it’s an SRI.
And yet, in the clinical world, discussions about drug mechanisms typically focus on particular receptors in particular regions of the brain. There is very little consideration given to the fact that (1) almost all of our psychiatric medications have affinities for many different receptors and their subtypes, and (2) they are entirely nonspecific with respect to the brain. Nobody has a good explanation for why dopamine is so heavily implicated in movement, psychosis, and addiction, but globally blocking D2 receptors has absolutely no effect on substance use disorders.
Surfing Uncertainty describes a (relatively) new paradigm for explaining the brain from the ground up: predictive processing. This theory contests that the brain is not a purely feed-forward system—that is, a system that builds a sensory recreation of the world and then interprets it—but a hierarchical system where top-down models compete to explain incoming sensory data. (As you might expect, it’s obviously a lot more complicated than that, but if you like this sort of thing, Andy Clark’s The Prediction Machine is a nice introduction.)
Predictive coding is a truly global theory of brain function, offering a framework that could explain how and why the brain works the way it does; from the level of simple interactions between neurons, all the way to complex communications between brain networks. It has totally revolutionized how I think about the brain, the nature of mental illness, what our drugs might be actually doing, and why. It also offers a framework by which we might start actually being able to reason about psychopharm from first principles. Maybe. Above all, it is epistemically humbling, a reminder of just how superficial our understanding of the brain is.
Concept—Functional Selectivity
In medical school, we’re taught that receptors have defined behaviors. Gi/o receptors inhibit adenylate cyclase. Gs receptors activate adenylate cyclase. Gq receptors activate phospholipase C. Sure, the receptor ligands change depending on the receptor family, you can have all sorts of agonists and antagonists, even some allosteric sites, but the basic function of the receptor is supposed to stay static… right?
“Haha, no, you idiot,” says Functional Selectivity, “When has anything in nature ever been simple? When has anything in the brain ever been simple?”
It turns out that, when bound by certain ligands, G protein-coupled receptors (GPCRs) can undergo a functional transformation and start behaving like totally different GPCR subtypes. A classic example of this is seen when LSD binds to 5-HT2A receptors, members of the Gq superfamily. LSD totally changes the nature of these receptors by:
Preferentially recruiting beta-arrestin instead of activating G-proteins (beta-arrestin, as the name suggests, inhibits the functionality of the receptors)
Activates arachidonic acid signaling
Does not induce any IP3 signaling—which is the canonical Gq signaling pathway.
Guess what else exhibits functional selectivity? The new antipsychotic xanomeline (Cobenfy).
Learning about functional selectivity instilled in me just how damn complicated psychopharmacology is and how little clinicians can really understand about what our drugs do from a first-principles point of view. It also got me wondering if our drugs are doing other complex things that we’re not thinking much about? Haha, yeah, of course there are: receptor heterodimerization, active, but unprofiled metabolites, receptor binding half-lives. It’s not simple, it will never be simple, and we should avoid pretending it ever could be.
Person — Stefan Leucht (head of the section of Evidence-Based Medicine in Psychiatry and Psychotherapy at the Technical University of Munich, Germany)
The psychopharmacology literature has a problem. It’s not very good at taking its statistical findings and translating them into a form that provides clinicians with useful information. What does a Cohen’s d of 0.5 look like in the real world? Is the difference between d=0.5 and d=0.47 clinically meaningful? How do changes in the scales used in research trials relate to effect size and real-world improvement?
I cannot tell you how many times I have gone to research questions exactly like those, planning to do a write-up for my next post, only to find that Stefan Leucht has beaten me to the punch. As annoying as it is, it is also a relief to know that there is someone like Stefan who seems to share my feeling that theorizing and mechanistic hypotheses are cute and all, but what really matters is making the clinical literature as clear as possible for the individuals who have to make prescribing decisions.
I didn’t make this connection until I wrote this, but I think Stefan and I write for very similar reasons, just in different contexts. I also want to make this information more legible to the average clinician, and I think writing outside of academia lets me take more legible—but still complex pieces of work like Stefan writes—and make them even more accessible.
Paper — Anything written by Scott Alexander, but if you want something specific: “The Control Group Is Out Of Control.”
I realize that Awais wanted this to be a journal article, but frankly I think that this piece by Scott is more useful for people to be aware of than almost anything else I could select. Scott does a fantastic job of storytelling, setting up a premise that we all nod along with. Yes, yes, obviously parapsychology is dumb and wrong. And yes, of course, if we just held researchers to higher standards and followed these Ten Easy Steps (You Won’t Believe How Effective They Are!!), parapsychology will be disproven once and for all. Except, of course, Scott does his brilliant writing jiu-jitsu thing about halfway through and you finish the essay feeling more enlightened, but also with a creeping sense of dread that you will never really be able to be particularly confident about anything ever again. Go read it, it’s really not that long.
Plus, it’s just fun to read. Yes, this essay about experimental design, posted on a blog in 2014, is downright enjoyable. Chances are, if you’re already reading Awais’s blog, you’re open to the idea that real, meaningful, and enjoyable works of scholarship and intellectual commentary can exist outside of the confines of peer-reviewed journals and published books. If you’re still new to this idea, or still a bit skeptical, Scott’s writing is perhaps the best pitch I can make that you’re missing out.
Surprise Item — Science Fiction
I know that certain people think that genre fiction isn’t “real” literature or whatever, but I read Lolita and Anna Karenina last year, so now I have the English Lit credentials to say, uh, fuck those people.
Why bring sci-fi up here, though? Well, listen, we humans are creatures of habit. We figure out how we like to do things and keep doing them that way. We find our little cultural niche that we are comfortable in and we tend to stay there. As psychiatrists, we find the drugs we like and stick with them, find the patients we like to treat and avoid the rest, find the diagnostic framework that we think makes the most sense and call the others foolishly naïve/reductionistic/<negative but erudite-sounding adjective here>, and so on. I’m not saying anything new here, I know. There are benefits to sticking to something, I know.
But! But. Psychiatry, perhaps more than any other discipline, demands that we develop the mental flexibility to step out of those cozy, comfortable mindsets and try and imagine what it is like to be somebody totally different. Someone who genuinely believes that their skin is black because their neighbors’ AC unit has been zapping them while they sleep, and how could a psychiatric hospitalization possibly fix that??? Someone who is so manically euphoric that the idea of taking medications to feel less than transcendently amazing is their definition of insanity. Someone who is so addicted to fentanyl that even though their life is obviously, objectively horrific—not to mention they don’t even like getting high anymore—continues to sacrifice what little they have left in order to use.
To do this, we need to engage in things on a regular basis that challenge us to imagine drastically different worlds, cultures, existences, technologies, types of intelligent life… you get the idea.
Issac Asimov’s classic Foundation trilogy asks what it would be like to live in a society where a “mathematics of sociology” can predict the rise and fall of empires… and what happens when it predicts that your civilization—one that spans the entire Milky Way—is about to collapse.
Ann Leckie’s Ancillary Justice pushes us to imagine what it would be like to have your consciousness spread amongst multiple bodies; could you maintain a single set of memories? Could you hide things from yourself?
In Children of Time, Adrian Tchaikovsky shows us what sort of a civilization of spiders with human-level intelligence might build, and how humans might try and communicate with them.
Science fiction’s other virtue is its ability to allow us to consider these ideas outside of our current cultural context and without the emotional valence typically attached to the sorts of customs and norms we have become attached to. Of course, we can’t quite leave everything at the door, but there is something about shifting away from the easily recognizable humans and experiences common in literary fiction—and into the typically distant and remote worlds of sci-fi—that allows us to feel much more comfortable as neutral observers.
Is any of that empirically true? No idea. Is that all a post-hoc justification to evangelize a genre that I am very emotionally attached to? At least partially.
See previous posts in the “Mixed Bag” series.
Psychiatry at the Margins is a reader-supported publication. Subscribe here.
 | A guest post by
|