Interesting Papers I've Read, May 11th 2023.

This is a recurring series, where I round up a handful of interesting scientific papers and comment on them.

Note for readers: This series of posts explicitly should not be sent out by the "newsletter" feature, I've it disabled, I think. Its also acting as a test of scheduled posts. I'm sorry if it ends up in your inbox and you didn't want it in your inbox. Still getting used to this CMS...

I also have such a recurring post on "interesting blog posts/articles", but I figured I'd keep the "peer reviewed journal articles" separate.

Subjects vary, but I read a fair amount of forensic papers. I also read a lot of papers about psychoactive substances. There is an overlap.

You can always find the papers themselves on Sci Hub or similar if the "open access" door slams shut.

Effect of lysergic acid diethylamide (LSD) on reinforcement learning in humans (2022).

In recent years, more and more studies have been done on potential applications of LSD, as we finally start to consider again that it may prove to be a useful medicine or theraputic agent instead of a vilified drug of abuse.

I occasionally find myself reading these studies, and regularly I find that I take issue with the experimental design of them.

This paper (DOI: 10.1017/S0033291722002963) outlines a study in which subjects were provided either a placebo (saline soln.), or 75 micrograms of LSD in saline via an IV, and given some learning tasks/tests designed to test if reinforcement learning (positive and negative stimuli) were improved, disimproved, or unaffected by the introduction of LSD.

They concluded that LSD improved learning in their tests, suggesting heightened neuroplasticity, and suggest this could have theraputic uses to help combat maladaptive behaviours.

All in all, a pretty big win, right?

Sort of. I have a couple of questions about the methodology - mostly the administration, "blinding", and such.

Firstly: the paper doesn't give any reason why IV was used over oral or sublingual administration. The paper simply says they got a cannula in the arm, and the LSD in saline, or saline only, was administered via it.

I can see absolutely no reason, beyond trying to make the whole experiment "seem more medical", for using IV administration over oral.

Maybe they thought it would lead to more consistent "effects" with less chance of LSD being degraded by metabolic processes along the way? I have not been able to find any study that claims any major difference in bioavailability.

Oral LSD has been used in previous studies - see "Acute LSD effects on response inhibition neural networks (2017)" (DOI: 10.1017/S0033291717002914), which does make me really question why IV was used here.

The "blinding" also interests me - the study notes that candidates were not told if they were getting LSD or placebo, but the test administrators did know. This makes it single blind as opposed to double blind.

However, given it is a dose of 75 micrograms of acid, and onset was noted within 15 minutes, I'd say that the patients knew damn well within a half an hour or so if they had gotten the saline or the LSD. Which makes it "kinda half blind" instead of single blind.

Furthermore - all of the subjects had previous experience with psychadelic substances. Which means they probably would know pretty quick if they were on the acid or not. If you know, you know.

Beyond just the small sample size (20 participants, of which 19 actually completed the whole thing), I don't yet have opinions on the rest of the methodology.

Can latent fingerprint disclose the sex of the donor? A preliminary test study using GC–MS analysis of latent fingerprints.

This paper (DOI: 10.1111/1556-4029.15260) was free when I read it, seems to be open access for now, but hey, its Wiley, so who knows when they will slam the door shut.

It explores an interesting question: if you have a fingerprint/fingermark left on an object at a crime scene, but it is smudged, and you can't get a DNA profile from epithelial tissue left behind on it or similar, can you get anything useful out of it?

Well, the investigators decided to see if they could determine the sex of the person who left the fingermarks behind by using GCMS (gas chromatography/mass spectroscopy), and noted that there were some differences in the ratios/amounts of some compounds left behind in the fingermarks that could be used to suggest the sex of the donor.

The caveats are pretty huge, of course. They were able to make the distinction with a statistically significant level of discrimination in a lab setting, under ideal conditions, with no contaminants. Kind of the spherical cow in a vacuum situation which probably won't happen in the wild very much.

Dose-response relationships of LSD-induced subjective experiences in humans (2023).

Yes, another LSD paper. This ones open access (DOI: 10.1038/s41386-023-01588-2), and explores the effects of dosage on the "subjective effect" (how fucking whack your experience is) when you take LSD.

This is a meta-study, the investigators didn't go and dose people up on a bunch of acid to find out what happens - instead they trawled data from other researchers work, did some analysis on it, and worked out some ideas from that.

They excluded data from recreational use (where dosage may be... unreliable), data where dose wasn't specified, data where multiple drugs were used, and only went with the "final data" from subjects in cases where multiple questionairres were filled out by subjects.

Dose ranges varied from 4 micrograms of LSD base (6.5 micrograms of the tartrate salt), to 200 micrograms - a pretty wide range of doses. They also helpfully reference all the studies they took data from, so I might write about some of those later.

Now, what I found interesting is that from reading their results, taking acid really has diminishing returns after about 100 microgram dose. Beyond that, you aren't really getting any more "fun effects", and the potential for a bad trip increases (fear, dread, etc to do with ego death/dissolution). This is, again, caveated with an observation that for a "full, mystical experience" to be delivered with certainty, you need to go hard, with a 200 microgram or so dose. This variance can really be put down to "set and setting" I suspect.

The other interesting point made is that at doses which are often regarded as microdoses - the sub-20 microgram range, there were still some, limited observable subjective effects, something a lot of skeptics with regards microdosing would have you believe doesn't happen. They do caveat this with the fact there is a gap in data between 20-50 micrograms, and mention that more data is really needed to draw any conclusion there.

Remember I complained about the IV use in the first study in this post? Turns out they found another study which used an IV - but most studies used oral/sublingual dosing. Again, it was a study involving Imperial College/Cardiff Uni that used IV. I strongly suspect that the British urge to inject LSD is something to do with the onerous requirements from the Home Office that any research involving drugs not even be remotely fun for the participants.

All in all, its exciting to see more and more research on LSD - I'll probably be posting about more of the studies as I read them. This one was a goldmine of references to other studies, a real jumping off point.

Screening unknown novel psychoactive substances using GC–MS based machine learning.

This paper (DOI: 10.1016/j.forc.2023.100499) is pretty neat.

So with novel psychoactive substances, when they come into the lab, you don't yet know the structure of them, or what the fuck they are. Just that it is some kinda new drug, and you need to start reverse engineering the structure/etc.

With a known substance, when you shove it into a GC/MS, you can often match the output peaks against a library of known data and get a match pretty quick. But what about with an unknown substance?

To grossly oversimplify how GC/MS works, you basically have two "parts". The gas chromatograph separates out the material into its individual components (to some extent, depends on gas used, column used, a bunch of stuff), emitting the individual parts of the mixture (hopefully) out the other end.

The mass spec is a bit more complicated, but to massively simplify it, it breaks the molecule up somewhat into smaller ionized components, and separates them based on size (mass), giving you a spectrum of "well, it had a bunch of shit at this size, a bunch of shit at this size..." and so on.

Now - when molecules break up in the mass spec, the charged sub-units are often quite predictable and take some usual sizes, so you can kinda sorta make some guesses as to what they might be.

What they did in this paper was feed a tonne of that data to a neural network, in the hopes that when presented with an unknown sample, it would be able to make somewhat of a guess as to what category/family of drug the sample belonged to, and achieved somewhat promising results.

This is pretty fucking cool, and I am excited to see this kind of work develop further.

The scientific controversy on the conversion of CBD into THC in the human stomach: Molecular modelling and experimental results compared.

This paper (DOI: 10.1016/j.forc.2023.100467) tries get answers for something that some of us have been wondering about for a while now. It doesn't get anything I'd call conclusive, but more data is always good.

Firstly, an introduction to the problem: under acidic conditions, CBD can undergo a variety of fun and exciting isomerization and ring closure reactions to produce other cannabinoids, including HHC or THC, depending on conditions. Both delta-8 and delta-9 THC seemingly can be produced this way.

This process occurs under an extremely wide variety of conditions, lets just say, there are a lot of Jessie's who have figured out how to conduct this process.

In the human body, the stomach is an acidic conditions. Given the reaction is quite forgiving in terms of reaction conditions, and the stomach is "eh, probably super good enough", the question becomes: if you eat CBD products, can they turn into just enough THC in your stomach to set off a drug test?

TL;DR after doing an assload of computational modelling and such, these scientists don't think so. Others disagree. Such is life in science land.

I'd like to see a simpler experiment done: test someone for THC/CBD at a point in time, then have them eat commercially available CBD infused products for a couple weeks at a "normal rate", then test them again. I've a sneaking suspicion real-world testing will go "actually, fuck your models" and produce some fun result.