Ooh, I love the new predictions at the bottom in the nifty boxes! Also, I just want to say how much I appreciate you being back. I get genuinely happy every time I see there's something new to read, because I know it's something worth reading and is absolutely thought-provoking. Thanks so much, Scott!

Expand full comment

I wonder if duration of discontinuation syndrome sheds any light on the long buildup claim. Looks like the standard answer is that brain zaps last a few weeks tops, but I seem to recall them lasting months, and I see other reports of that when I dig beneath the surface. Seems like something that could easily be underestimated.

Expand full comment

> Also, all these mice studies are great, but in human studies, SSRIs barely work. The best meta-analyses of thousands of depressed patients manage to pick up a small signal, assuming the pharma companies haven’t figured out some way to fake us all out.

This is the part that bugs me the most. How can you claim to have discovered the mechanism behind an effect that, as far as I can tell, may not even exist?

Expand full comment

It's well know that compounds with basic amines (e.g. amantadine, monamines, etc) accumulate in lysosomes. The basic mechanism is simple: uncharged molecules generally pass through cell membranes more easily than charged molecules. At physiological pH a small fraction of basic amines (~0.1%) will be uncharged. In lysosomes however, which are 2-3 pH units lower, only 0.001-0.0001% will be uncharged.

In a slightly simplified scenario where only uncharged compounds can pass through membranes, there will be roughly 100-1000x more compounds capable of entering a lysosome than of exiting one at any given moment. This essentially means lysosomes are sinks for basic compounds.

I'm not sure how this would affect overall brain accumulation, especially over week-long periods, but it does make the AD accumulation bit seem a bit more plausible.

Expand full comment

Scott, how open are you to giving scientific talks about these issues in a university setting? (Either in person or more likely Zoom). I organize the lab meetings for my neuroscience lab at Hebrew University (Jerusalem) and we'd be happy to get your perspective. (I've recently been working on models/ideas that involve BDNF and plasticity so this is especially interesting to me).

Expand full comment

There are quite a few sentences that end with question marks without much good reason in this piece? It's the first time I've noticed that in Scott's writing?

Expand full comment

How *do* you figure out whether a mouse is depressed? (if our animal models were cetacean rather than rodentine, would our inventories find porpoises more depressed than dolphins?)

Expand full comment

I'm casually familiar with how UpToDate, NICE review and synthesize new findings, but have no idea how Wikipedia does it. How do they review literature? Can't anyone edit those pages? I usually trust Wikipedia for basic facts, but don't have a good gauge for its quality in evolving fields.

Also, I think you should post your predictions in Metaculus (or another prediction market when better ones come along). You might even help get researchers more interested in prediction markets.

Expand full comment

What about some bio/psycho interaction? For example - initial serotonin elevation dampens amygdala reactivity, but only when the calmed down amygdala meets the world and sees people aren't as bad as it perceived before does the depression start to remit. This will explain the time needed for anti-depressant effect to take place, and maybe some biological findings (such as elevated BDNF) are secondary to effective re-learning and not due to a direct anti-depressant effect. In other words - set and setting matter in anti-depressant treatment too. Browning et al. have an interesting paper where they can predict response in an early stage, depending on altered emotional processing https://www.sciencedirect.com/science/article/pii/S0924977X18319631

supporting the claim that the medication only alters some basic emotional processing property, but that the anti-depressant effect is dependent on meeting the world with different glasses

Expand full comment

Lovely, mnemonic in the caption!

[epistemic status: cheerfully ignorant conjecture, like most of what I believe, but don't really understand]

BDNF binds to cholesterol? Cool.

More evidence for the "cholesterol is good, you haters!"-side.

That side appears less crazy to me than "cholesterol is bad for you", but I have a strong contrarian bias.

[or rather I just think, that majority belief is usually insane group-think, since we live in the insane timeline]

Can explain those findings better:


I think the carnivore diet has been claimed to have antidepressant effects, too.

Not a great signal though, too radically different in general, too few people doing it.


Is there any other "high cholesterol"-diet to look at, for figuring out whether it's suitable as an intervention?

Eat all the bacon and eggs they have and die happy of a heart attack.

Good deal, right?

Expand full comment

"Pictured: BDNF binds to TrkB. The IRS confiscates 1/2 of it as taxes, which radicalizes the receptor and makes it join Gab (see footnote 1), where it tweets out an SOS message to the Ras of Ethiopia. But the left wing of the receptor joins the Palestine Liberation Council and moves to California (see footnotes 2+). California has sunshine and good beaches, so you stop feeling depressed."

Where are the footnotes for this? I want to learn more about GAB and the adventures of the Palestine Liberation Council :)

Expand full comment

> Also, all these mice studies are great, but in human studies, SSRIs barely work. The best meta-analyses of thousands of depressed patients manage to pick up a small signal, assuming the pharma companies haven’t figured out some way to fake us all out.

I’m extremely confused by both this and the “month-to-take-effect” part. I know well a couple people with depression. For both of them, they used to have pretty bad depression. They’ve been on ADs for years and they’re now basically normal. If they miss even a single pill, however, they consistently go back to seriously depressed for a couple days before returning to normal. Is this something-something adjusted to the medication + placebo effect? That seems not enough to explain the extreme effects.

Generally, I’ve seen that ADs seem to either work well for people or not really work at all (or trying a few might find a successful one). So it’s very strange that metanalyses don’t seem to find an effect on anyone.

Expand full comment

Substances based on adhesmine (pyridoxine) and sulbutiamine (Arcalion) at dosages above 300mg per day seem to produce good effects in the medium term, after 30 to 60 days, or in a longer period for more severe depressions.

Expand full comment

> Neurons are less likely to respond to stimuli by connecting to other neurons.

That sounds like something for feeling dumb, but not as a reason for the bad mood.

There are a lot of papers, that show that depression is a systemic disfunction in cells' methabolism. Every neuron has pretty complex ion machinery, including moving ions through membrane and storing them in places like endoplasmic reticulum and mitochondria.

Both these organelles are involved in depression:

Role of endoplasmic reticulum stress in depression


Mitochondria and Mood: Mitochondrial Dysfunction as a Key Player in the Manifestation of Depression


The reason why it takes months to activate the depression response to SSRI could be that SSRI do not directly fix problems like Endoplasmic Reticulum stress. They could work more profound. For example:

fluoxetine-induced decrease in mitochondrial ATP production results in the emptying of the ER, leading to capacitative calcium entry. Furthermore, Ca2+ quickly accumulated in the mitochondria, leading to mitochondrial Ca2+ overload


And that kind of mechanism of emptying ER (by blocking ATP or by activating ER IP3 receptors via 5-HT2) may lead to normalization of the cell function:

On the other hand, loss of luminal Ca2+ causes ER stress and activates an unfolded protein response, which, depending on the duration and severity of the stress, can reestablish normal ER function or lead to cell death.


ER stress and its XBP1 stress signaling is also linked to the BDNF activity:

The Role of Brain-derived Neurotrophic Factor (BDNF)-induced XBP1 Splicing during Brain Development

Xbp1 was activated in neurites in response to brain-derived neurotrophic factor (BDNF), followed by subsequent translocation of the active Xbp1 into the nucleus. BDNF-dependent neurite outgrowth was significantly attenuated in Xbp1−/- neurons.


Expand full comment

Is it intentional that the word "something" appears repeatedly in the first paragraphs of this blog post? It look like either placeholders the author used in a draft and never corrected before posting, or a way to bypass explaining some complicated neurobiological process, or some sort of random error when the article posted. In any case, it undermines credibility and you may want to fix.

Expand full comment

Wait! Doesn't it take a very skilled rodent therapist to determine whether the mouse was depressed in the first place and to evaluate extent of recovery?

Expand full comment

> But why are so many monoaminergic substances antidepressants? Where are all the papers saying "we tested this new SSRI, it did a great job inhibiting serotonin reuptake, but there was no antidepressant effect whatsoever"? Surely some substances that are SSRIs don't also interact with this new TrkB receptor domain?

Perhaps the binding site for whatever protein is being inhibitted by these SSRIs is very similar to the newly discovered binding site in the TrkB receptor. If so, then any substances which fits one binding site will also be very likely to fit the other, which would make it very hard to find an SSRI which isn't also doing this TrkB thing.

Expand full comment

There are some TRK inhibitors approved for treatment of cancers driven by TRK fusion (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6859818/) . Some of them even are used to treat brain cancer. For example, entrectinib was originally developed as an inhibitor of ALK for the treatment of ALK-driven lung cancer with a twist that it can get to the brain to pursue metastasis their. As the field became overcrowded (https://en.wikipedia.org/wiki/ALK_inhibitor), the developers re-purpused the drug to hit TRK (apparently many ALK inhibitors inhibit TRKA/B/C, and entrectinib was one of such drugs). Finally they got an approval. It's fun to observe how the view on BBB penetration changed from glee (wow, we hit brain mets) to horror (omg, TRK might be important for depression something-something) to optimism (we can treat TRK driven cancers in the brain). Long story short, depression is one of the AEs for entrectinib (https://www.accessdata.fda.gov/drugsatfda_docs/label/2019/212725s000lbl.pdf) , but it's not reported for the other BBB-penetrating TRK inhbitor larotrectinib (https://ascopubs.org/doi/abs/10.1200/jco.2019.37.15_suppl.2006).

Expand full comment

I just wanted to say I love this use of question marks in non-questions. It's probably a known thing but I saw it here for the first time and, as everything with Scott's writing, it was very apt and lucid, easy to move past the initial annoyance jolt. By using it on all three sentences in a paragraph, Scott manages to suppress the inevitable "typo?" response.

Expand full comment

It doesn't effect the point that if it takes studies of a zillion people to find an effect for SSRIs in people how come tiny studies can detect an effect in mice, but I'm a bit surprised you've summed up the former as 'SSRIs barely work', given that I thought you were arguing against exactly that reading of the SSRIs-in-humans data here: https://slatestarcodex.com/2018/11/07/ssris-an-update/

Have you changed your mind, or am I wrong in thinking that post is incompatible with 'SSRIs barely work' being a reasonable, helpful summary of the data?

Expand full comment

Has the number of synapses per neuron in depressed people been measures, and if so how does it compare with the number of synapses per neuron of non-depressed people? Does this vary according to area of the brain?

I wonder whether there is any correlation between synapses and anything else, such as age, job, or IQ? For example, maybe artists have more synapses in the visual cortex, and footballers have more in the areas controlling leg movements.

Expand full comment

Speaking as a guy who started graduate school studying TrkB signaling and who now studies accumulation of junk in the lysosome, I think you're safe as houses on both predictions. I don't know what it was (and still is?) about studying TrkB, but I found the literature on it to be unusually prone to really basic mistakes and implausible findings.

As an example of the former, there were LOTS of studies talking about using K252a as a "specific inhibitor of Trk kinases." Which it is--if you consider only receptor tyrosine kinases. There are a number of very important serine/threonine kinases, including phosphorylase kinase and CamKII, which it inhibits at the same concentration. As an example of the latter, the new hotness when I was starting out was the finding, published in Nature, that TrkB directly modulates voltage-gated sodium channels through some non-phosphorylation-based mode of action (the kinetics weren't right for the effects seen to be due to ordinary kinase activity).


Even at the time, this seemed squirrely to me--too fast, too big an effect, and too hard to square with what else was known about TrkB and voltage-gated sodium channels. And, indeed, almost twenty years further on, it is not central to the understanding of the function of that sodium channel and is considered more as an anomaly and probably an artifact than as a central finding the field (see discussion here https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4996523/ for an example).

I'll join you at 95%+ that this turns out to be the same.


Expand full comment

Are there any direct interactions between serotonin and TrkB? I'm assuming that would have been picked up long ago, but Scott doesn't list it as one of their controls for the new binding site.

Expand full comment

Enjoyable read. However, there are other, non serotonin related, methods of achieving antidepressant effects (ketamine, exercise, socialization, psilocybin, MDMA, etc.). Therefore, I don’t think you can hang your hat on serotonin being a necessary player. I may have missed something! Really appreciate your insight though and I hope I get a response back!!!

Expand full comment

Just reporting from a 2022 paper :

The paper is titled "Ketanserin reverses the acute response to LSD in a randomized, double-blind, placebo-controlled, crossover study in healthy subjects" with DOI:10.1093/ijnp/pyac075

The main conclusions that are relevant here are:

* ketanserin (a 5HT2A antagonist) can be given 1H *after* LSD and will completely remove subjective effets of the intoxication.

* mixing ketanserin and LSD shows no psychological impairement but does not reduce LSD's spike in BDNF levels.

IMO this has important implications on how we view 5HT2A and paves the way to new research.

Expand full comment