alcohol aversion on sema gets framed two ways: either it’s nausea (boring, mechanical) or it’s dopamine reward blunting (exciting, implies AUD treatment potential). both get cited constantly. neither one is the whole story, and there’s a third mechanism that explains the timeline better than either. gastric emptying is significantly slowed on sema. that’s not controversial. what doesn’t get discussed: delayed gastric emptying changes ethanol absorption kinetics. alcohol clears the stomach slower, which means absorption is slower but also potentially more erratic. net effect for a lot of people is that the same 2-3 drinks hit harder and faster once they do absorb, which at a gut level creates something close to a conditioned aversion response. your brain logs ‘this thing that used to feel pleasant now feels off’ and updates accordingly. this is meaningfully different from reward pathway blunting, which is the mechanism getting cited in the AUD research. the dopamine circuit angle is plausible but too tangled to isolate – you’d need a clean window where gastric motility is controlled for, and that’s not what any of the observational reports are doing. the reason this matters: if it’s pharmacokinetic rather than reward-pathway, the aversion may not be durable off-drug. if it’s actually blunting the wanting signal at a dopaminergic level, that’s a different story. personally ran sema for about five months and noticed the same thing. two drinks felt like four. whether that’s a GI kinetics problem or something upstream – genuinely don’t know, and I’d be skeptical of anyone who says they do without controlled data.
the pharmacokinetic framing is underutilized and the post earns its point there. but “if it’s pharmacokinetic, the aversion may not be durable off-drug” doesn’t follow as cleanly as it looks, bc conditioned aversions can persist well past the unconditioned stimulus that generated them, that’s pretty basic pavlovian territory. if your brain reliably logs two drinks as functionally four for several months on sema, that association can outlast the gastric motility effect even if absorption kinetics normalize after discontinuation. the mechanism question and the durability question are separate problems, and conflating them undersells how sticky conditioned responses tend to be once established.
Where I’d give Priya the point completely is that the mechanism question and the durability question really are separate, and it’s a tidy correction to my original framing. Conditioned aversions outlasting the thing that caused them is well documented, and the interoceptive version (the Garcia effect, taste aversion from a single nauseating pairing) is famously sticky in a way that makes your case even stronger than basic Pavlov does. The caveat I’d add is that stickiness isn’t a fixed property, it depends a lot on what happens after you come off. Conditioned responses extinguish with repeated unreinforced exposure, so someone who stops sema and has a few normal drinks where two no longer feels like four is actively unlearning the association. The person who just avoids alcohol entirely off-drug never gives the response a chance to extinguish, so theirs looks durable for a completely different reason. Same starting condition, opposite trajectories, and neither tells you anything clean about the underlying mechanism. There’s also a wrinkle in my own post that cuts against tidy conditioning either way. I called the absorption erratic, and erratic pairing is a noisier teaching signal than a reliable one. If two drinks landed like four on Tuesday but felt nearly normal on Saturday, the brain has a harder time forming a crisp association in the first place. So I’d soften “may not be durable” to “durability is going to vary enormously person to person depending on post-drug exposure,” which is honestly a less satisfying claim than the one I made, but closer to what the observational reports can actually support.
variable reinforcement schedules tend to produce more extinction-resistant conditioning, not less, so the “erratic pairing is a noisier teaching signal” framing runs into the partial reinforcement extinction effect pretty directly. intermittent Tuesday/Saturday inconsistency might actually produce a stickier aversion than a reliable signal would. softening the durability claim is still the right call, but the noise-in-the-signal argument doesn’t quite get you there on its own