I'm reading a super interesting book by Chris Letheby, "Philosophy of Psychedelics" [1.], part of the International perspectives in philosophy and psychiatry series.
In my last post I was critical of the author for seemingly overlooking a major weakness of the Comforting Delusion Objection to the use of psychedelic therapy for a range of mental illnesses. The Objection is that if psychedelic therapy works by patients have experiences and forming beliefs about the world that aren't true, then this is in some way undesirable. I argued that since our everyday experience of the world isn't "true" either, it should not be considered philosophically privileged over psychedelic experiences and hence the main concerns should be the purely practical ones of safety and effectiveness.
Although Letheby doesn't engage on this philosophical point, it's actually much less interesting than the later discussion of how our experiences of the world arise, what psychedelic drugs likely do and why they are effective in treating illnesses like depression, OCD and PTSD. I'd like to give a sketch of this here, mainly as my own notes on this but if the reader also finds it useful, then fine :-)
The predictive processing brain
"[Predictive processing] depicts the brain as an interference engine that builds hierarchical models of the world in order to predict its future inputs." [1.]
As someone who's trained in using and thinking about models this makes a lot of sense, but probably not so accessible for many. Let's unpack that and see how it works.
Have you ever seen a snake, only to realise that it's actually just a piece of hose? Or read a word, then gone back, re-read it and realised that it doesn't say what you read the first time? It feels as if reality changed underneath you - a "glitch in the matrix".
In a very real sense, it did.
The predictive processing theory of brain function is that the world of our conscious everyday experience is entirely constructed by the brain - a "controlled hallucination", as Letheby puts it. It does this as a way of processing the surprisingly limited flow of sensory information from the environment into a coherent representation of the world, including ourselves at the centre of it.
Close one of your eyes. Can you see the blind spot in your visual field? There is a blank spot there where your optic nerve joins your retina, but you can't normally see it because your brain hallucinates the details to fill in the gap. Actually much of what you "see" isn't really there, it's your brain's model of what's there. That's how we can sometimes see a snake instead of a piece of hose, it's also how stage magic works.
Your brain continually updates this simulated world, comparing the sensory input with the expected input based on the model and updating the model as required to minimise the error between expectations and data. This is what's meant by "interference engine" - discrepancies between the model and the incoming sensory data create something like an interference pattern that the brain continually works to minimise by adapting the models or acting in the world.
Optical illusions are fun and slightly uncomfortable because your brain's continually trying to interpret the stimulus one way (which makes an error) and then another (which minimises that error but makes a different one). It's two faces, or it's a vase, then it's two faces again as your hallucinated world interpretation continually tries to minimise the error and match the sensory data. Is the dress black and blue? Or white and gold?
The world models are also hierarchical - basic elements like colours and objects at the bottom, up to more abstract situations and events comprised of the lower level features, right up to highly complex notions at the top of the hierarchy including a sense of self, narrative history, meaning, space and time. One of the main ideas in the predictive processing view of the brain is that perception starts at the high level models, ie our understanding of ourselves and our situation in the world. The high level models make predictions about what's likely to happen in the next layer down, and this propagates all the way down to the basic sensory data level. Any errors (discrepancies between predicted and actual sensory input) get propagated back up the hierarchy only as far as they need to go to update the model. In this way we save brain power and only become conscious of the unexpected.
This continuous modelling and error-checking is pretty much completely unconscious and "transparent" - meaning that we don't experience the modelling at all, but the results of the modelling. This is the brain-generated "controlled hallucination" that appears to us as if it's the real world. As Letheby puts it,
"All the furniture of our waking experience - the people, animals, plants, tables, chairs, and our own bodies and selves - is as thoroughly virtual, internally constructed, and simulatory as the fantastic creations of nocturnal dreams and psychotic hallucinations." [1.]
Wow. We really do live in The Matrix. The real world out there might as well be comprised of green computer code.
Here's where it gets really interesting. Your brain doesn't just compare the model with the sensory input, and update the model when there's a disagreement. It also rejects sensory input when it disagrees with the model. This actually makes sense - if there's an error, it could be a modelling error or it could be a measurement (sensing) error. Your brain has its own internal sense of the reliability of the different parts of the reality model based on having its own prior experience confirmed over and over again. For example, your experience has given a very high weight to the idea that two different objects can't occupy the same space at the same time, so if your sensory input says otherwise you'll probably assume it's something like a stray reflection or at worst a hallucination. These basic assumptions about what kind of things populate the world and how, are essential to be able to perceive and make sense of anything at all, to recognise the familiar, notice the unusual and to be alert to danger.
What psychedelics do (probably)
Different parts of the brain do different things, lots of different regions are wired together to perform different types of functions. There are two brain networks that seem to be important in this context, the Default Mode Network, and the Salience Network. I find the neurobiology a bit hard to follow, but these networks in the brain seem to be important at the very highest levels of the hierarchical modelling that results in our experience and perception of the world as noted above. These fundamental assumptions that the brain makes about "time, space, and causality, the laws of logic, and the existence of the self", rigorously tested and retested by a lifetime of experience, cascade from the Default Mode and Salience Networks, down through the lower levels of the model and quite literally determine how we can perceive the world. These most basic core beliefs are considered very trustworthy by the brain, and given a very high "weight" that suppresses perception of sensory or other inputs that are inconsistent with them.
Now, these brain networks happen to be particularly rich in serotonin-2A receptors. When you take a psychedelic drug, that binds on to the serotonin-2A receptors and has the effect of degrading the regular signalling that goes on those networks. The brain experiences this as a decrease in the reliability of the most fundamental beliefs and assumptions that operate at the very top of the modelling cascade. The modelling cascade that results in our experienced world. Our perception becomes less constrained and we experience all kinds of things that we cannot normally experience. The "hallucinated world" that we always inhabit, becomes a bit less error-corrected by our most basic unconscious beliefs.
A likely reason why psychedelics look so promising for conditions like anxiety, OCD, PTSD and depression is that all of these conditions involve some very foundational pathological beliefs about the self. These are very hard to shift because they are so strongly believed, that the predictive processing brain literally cannot perceive any evidence to the contrary. The psychedelic is probably therapeutic by decreasing the strength of those beliefs (by adding a whole lot of noise into the default mode network and the salience network), which allows some perception of the normally excluded reality - that the world isn't threatening, you're not a bad or inadequate person - in fact you, the world, and people in it are wonderful (and, actually, quite miraculous).
Once perceived, the experience is remembered, and it's that memory which then down-weights the pathological belief, even after the psychedelic has gone from the system. One this evidence has been perceived the beliefs soften and further perception can happen where previously it couldn't. And the condition improves.
1. Letheby, "Philosophy of Psychedelics", Oxford University Press, 2021
