Episode Details

https://astralcodexten.substack.com/p/on-cerebralab-on-nuttcarhart-harris

[epistemic status: extremely speculative]

George at CerebraLab has a new review of Nutt and Carhart-Harris's paper on serotonin receptors (I previously reviewed it here). Two points stood out that I had previously missed:

First of all - predictive coding identifies suffering with prediction error. This conflicts with common sense. Suppose I tell you I'm going to stab you in the arm, you agree that I'm going to stab you in the arm, and then I stab you in the arm, and it hurts a lot. You predicted what would happen correctly, but you still suffered. The theory resolves this with a distinction between common-sense-level and neurological predictions: your brain is "set" to expect normal neurological feedback from your arm, and when it gets pain signals instead, that's a violated prediction, and this is the level on which prediction error = suffering. But there are other cases where the common-sense and neurological sense of predictions are more congruent. When you first step into a cold shower, you feel suffering, but after you've been in it a while you adjust your "predictions" and it's no longer as unpleasant. If you unexpectedly lost $25,000 it would come as an extremely unpleasant shock, but when you predictably have to pay the taxman $25,000 each year you grumblingly put up with it.

The theory of "active inference" adds another layer of complexity here; it posits that sometimes your brain automatically resolves prediction error through action. If you were expecting to be well-balanced, but actually you're off-balance, you'll reflexively right yourself until you're where you expected to be. At its limit, this theory says that all action takes place through the creation and resolution of prediction errors - I stand up by "predicting" on a neurological level that I will stand up, and then my motor cortex tries to resolve the "error" by making me actually stand.

(one remaining problem here is why and how some prediction errors get interpreted as rewards. If you get $1 million one day because you're a CEO and it's payday and that's how much you make every payday, you will not be especially happy. If you get $1 million because you're an ordinary middle-class person and a crypto billionaire semi-randomly decides to give you $1 million one day, you will be very happy. This has been traced to reward being dopamine-based prediction error in the nucleus accumbens, and the CEO was predicting his windfall while the gift recipient wasn't. This suggests there's still something we don't understand about prediction error and suffering).

So one question is: for some given prediction error, how much do I suffer vs. adjust my predictions and stop feeling it vs. take action to resolve it?

George's take on Carhart-Harris & Nutt is that this is influenced by the balance of 5-HT1A vs. 5-HT2A receptors - two different kinds of serotonin receptor. 5-HT1A is (to vastly oversimplify) the main target of antidepressants. The more strongly it's stimulated, the more likely you are to resolve prediction error by adjusting your predictions - the equivalent of stepping into a freezing shower, but then acclimating so that it feels okay. Suppose you're depressed/anxious/upset because your boss keeps yelling at you. With enough 5-HT1A activation, you're better able to - on a neurological level - adjust your world-model to include a prediction that your boss will yell at you. Then when your boss does yell at you, there's less prediction error and less suffering. This is good insofar as you're suffering less, but