In my view the set of mechanisms that cause most mental illness are already well-established, but for some reason aren’t commonly all put together or popular. I’m very curious for people’s thoughts, especially on what fraction of mental illness this model describes.
Most cases of mental illness are individually unique complex adaptive system attractor states whose reinforcing elements are something like:
1. Learned (predictive processing) bayesian priors (e.g. a belief like “I am bad” associated with a fear response that drives thought and behavior) that emerged as an adaptive response to a different context than the individual is currently in. The different context is often an emotionally dangerous childhood, which often result in sets of priors that we call things like “attachment disorders.”
2. Attention (predictive processing) and its lack (avoidance). Incoming sensory evidence is typically sufficient to update (called reconsolidation in memory research) priors that are no longer adaptive, but that only works when the strength (how much attention is payed to it) of the sensory evidence is within some range of the strength of the prior. So it fails when there isn’t sufficient attention to the sensory evidence or when the strength of the prior is too high.
3. Priors predicting certain levels of threat and powerlessness activate the set of autonomic nervous system functions called the defense cascade, i.e. sympathetic/parasympathetic hypo- and hyper-activation. This is fight or flight, freeze, arousal, tonic immobility, and collapsed immobility.
4. A bunch of other stuff whose role is harder to understand, like medical history, medication, environment, genes, sleep quality, mental illness symptoms, and a variety of low-level neurobiological dynamics from which priors and attention emerge.
Together, certain sets of priors and avoidance create stable and maladaptive attractor states. The priors involved often erroneously predict threat and powerlessness and therefore activate defense cascade states.
You can permanently destabilize these maladaptive attractor states (and associated defense cascade activation) by updating/reconsolidating the reinforcing priors by activating them and juxtaposing them with strong contradictory evidence that the person may otherwise be avoiding, e.g. the best examples of psychotherapy. This juxtaposition creates prediction error, which updates the priors reinforcing the maladaptive attractor state. Of course this is easier said than done; it’s hard to pay attention to contradictory evidence when you’re in flight or flight or tonic immobility, you feel that feeling emotions is a threat itself, you don't even know what the relevant prior is, etc. I personally think MDMA therapy is particularly helpful here because it seems to facilitate prediction error for most or all maladaptive priors that are activated/triggered during a session and often works during fight or flight or tonic immobility.
Maybe there’s some use in categorizing most mental illness into boxes, but in this model you would still cure it contradictory evidence specific to each case (or possibly through practices that facilitate universally applicable prediction error), so why bother other than for legal reasons.
To be clear, this model doesn’t fully describe disorders involving certain clear biological issues like psychosis (aberrant salience), a swath of neurological disorders, brain damage, etc. This model also doesn't explain anything about how, when, and why a lot of psychiatric drugs work and don't work.
Thank you for this comment! I agree that this is a promising way of thinking about mental heath problems and I find it appealing too, as do a lot of other people who are drawn to complex dynamic systems thinking. I recently posted (book review of Nicole Rust’s Elusive Cures) about how clinical neuroscience is moving in this direction too. However, I would not say that mechanisms understood in this manner are “already well-established.” We have not yet moved beyond theoretical appeal and plausibility and methodological fervor to actual empirical demonstrations of well-defined hypotheses. And there are significant challenges involved in getting to that point.
Thanks for the reply! I wonder if some of our different interpretations of establishment come from perspective. I started out in the psychotherapy prediction error/memory reconsolidation world surrounding "Unlocking the Emotional Brain" by Bruce Ecker. I think people there more readily think that the framing I laid out is more established, or common knowledge, at least the "previously adaptive priors that are no longer adaptive," and "facilitating prediction error/memory reconsolidation is the solution" parts. That group might have less strong views about the parts of the framework involving the defense cascade, attention, and complex system attractor states. As far as I can tell the defense cascade stuff is also somewhat established, but I'm less confident about the specifics of the attractor state dynamics and how attention plays into that.
I imagine actually verifying these attractor state dynamics might be experimentally difficult! Especially since each case seems at least a bit unique.
Thank you, Awais, for a thorough and thoughtful survey of this fraught topic--one I have been grappling with for four decades! I would like to suggest another way of looking at the concept of "validity" in psychiatry--one that shifts the discussion from etiological considerations to pragmatic, instrumental and ethical issues. I discuss this under the rubric of "instrumental validity" in this article:
"Following the pragmatic tradition of William James and John Dewey, I define “instrumental
validity” as that property of a diagnostic criteria set which bears on how fully it achieves a par-
ticular aim or goal. Now - to hyper-condense along argument - I believe that the fundamental
goal of general medicine and psychiatry is to reduce certain kinds of human suffering and inca-
pacity..."
With respect to general medicine, I think that the issue of "validity" does arise fairly frequently, albeit (usually) without the polemical flourishes that typically accompany discussions in psychiatry. For example, there is intense interest and discussion regarding the validity of the entity commonly called "Long Covid" See:
There are other debates that border on the "validity" question in general medicine and neurology, such as whether Persistent Idiopathic Facial Pain is a valid diagnosis (this was formerly known as "atypical facial pain")
What do you think of other medical conditions that don't based on subjective symptoms and don't really have biological anchors. Migraines, IBS, CFS, chronic pain etc. the dispute of Thier validity I never see come into question
There are definitely disputes around etiology or causes of such conditions but rare to see them framed as being about their “validity,” which again suggests something peculiar is happening in psychiatric debates. Some of it is a bias against mental illness but it also speaks to the complexity and ambiguity of mental health problems in terms of their conceptualization (illnesses, life problems, personality problems, spiritual problems, etc) as well as their relationship to biology.
In my view the set of mechanisms that cause most mental illness are already well-established, but for some reason aren’t commonly all put together or popular. I’m very curious for people’s thoughts, especially on what fraction of mental illness this model describes.
Most cases of mental illness are individually unique complex adaptive system attractor states whose reinforcing elements are something like:
1. Learned (predictive processing) bayesian priors (e.g. a belief like “I am bad” associated with a fear response that drives thought and behavior) that emerged as an adaptive response to a different context than the individual is currently in. The different context is often an emotionally dangerous childhood, which often result in sets of priors that we call things like “attachment disorders.”
2. Attention (predictive processing) and its lack (avoidance). Incoming sensory evidence is typically sufficient to update (called reconsolidation in memory research) priors that are no longer adaptive, but that only works when the strength (how much attention is payed to it) of the sensory evidence is within some range of the strength of the prior. So it fails when there isn’t sufficient attention to the sensory evidence or when the strength of the prior is too high.
3. Priors predicting certain levels of threat and powerlessness activate the set of autonomic nervous system functions called the defense cascade, i.e. sympathetic/parasympathetic hypo- and hyper-activation. This is fight or flight, freeze, arousal, tonic immobility, and collapsed immobility.
4. A bunch of other stuff whose role is harder to understand, like medical history, medication, environment, genes, sleep quality, mental illness symptoms, and a variety of low-level neurobiological dynamics from which priors and attention emerge.
Together, certain sets of priors and avoidance create stable and maladaptive attractor states. The priors involved often erroneously predict threat and powerlessness and therefore activate defense cascade states.
You can permanently destabilize these maladaptive attractor states (and associated defense cascade activation) by updating/reconsolidating the reinforcing priors by activating them and juxtaposing them with strong contradictory evidence that the person may otherwise be avoiding, e.g. the best examples of psychotherapy. This juxtaposition creates prediction error, which updates the priors reinforcing the maladaptive attractor state. Of course this is easier said than done; it’s hard to pay attention to contradictory evidence when you’re in flight or flight or tonic immobility, you feel that feeling emotions is a threat itself, you don't even know what the relevant prior is, etc. I personally think MDMA therapy is particularly helpful here because it seems to facilitate prediction error for most or all maladaptive priors that are activated/triggered during a session and often works during fight or flight or tonic immobility.
Maybe there’s some use in categorizing most mental illness into boxes, but in this model you would still cure it contradictory evidence specific to each case (or possibly through practices that facilitate universally applicable prediction error), so why bother other than for legal reasons.
To be clear, this model doesn’t fully describe disorders involving certain clear biological issues like psychosis (aberrant salience), a swath of neurological disorders, brain damage, etc. This model also doesn't explain anything about how, when, and why a lot of psychiatric drugs work and don't work.
I flesh this out in much more, fully-cited, detail in Chapter 2 of https://www.researchgate.net/publication/394097304_Open_MDMA_An_Evidence-Based_Mixed-Methods_Review_Theoretical_Framework_and_Manual_for_MDMA_Therapy.
Thank you for this comment! I agree that this is a promising way of thinking about mental heath problems and I find it appealing too, as do a lot of other people who are drawn to complex dynamic systems thinking. I recently posted (book review of Nicole Rust’s Elusive Cures) about how clinical neuroscience is moving in this direction too. However, I would not say that mechanisms understood in this manner are “already well-established.” We have not yet moved beyond theoretical appeal and plausibility and methodological fervor to actual empirical demonstrations of well-defined hypotheses. And there are significant challenges involved in getting to that point.
Thanks for the reply! I wonder if some of our different interpretations of establishment come from perspective. I started out in the psychotherapy prediction error/memory reconsolidation world surrounding "Unlocking the Emotional Brain" by Bruce Ecker. I think people there more readily think that the framing I laid out is more established, or common knowledge, at least the "previously adaptive priors that are no longer adaptive," and "facilitating prediction error/memory reconsolidation is the solution" parts. That group might have less strong views about the parts of the framework involving the defense cascade, attention, and complex system attractor states. As far as I can tell the defense cascade stuff is also somewhat established, but I'm less confident about the specifics of the attractor state dynamics and how attention plays into that.
I imagine actually verifying these attractor state dynamics might be experimentally difficult! Especially since each case seems at least a bit unique.
Thank you, Awais, for a thorough and thoughtful survey of this fraught topic--one I have been grappling with for four decades! I would like to suggest another way of looking at the concept of "validity" in psychiatry--one that shifts the discussion from etiological considerations to pragmatic, instrumental and ethical issues. I discuss this under the rubric of "instrumental validity" in this article:
https://scispace.com/pdf/toward-a-concept-of-instrumental-validity-implications-for-4j3yc4xt2d.pdf
The core of my thesis is as follows:
"Following the pragmatic tradition of William James and John Dewey, I define “instrumental
validity” as that property of a diagnostic criteria set which bears on how fully it achieves a par-
ticular aim or goal. Now - to hyper-condense along argument - I believe that the fundamental
goal of general medicine and psychiatry is to reduce certain kinds of human suffering and inca-
pacity..."
With respect to general medicine, I think that the issue of "validity" does arise fairly frequently, albeit (usually) without the polemical flourishes that typically accompany discussions in psychiatry. For example, there is intense interest and discussion regarding the validity of the entity commonly called "Long Covid" See:
https://www.psychiatrictimes.com/view/what-long-covid-can-teach-psychiatry-and-its-critics
There are other debates that border on the "validity" question in general medicine and neurology, such as whether Persistent Idiopathic Facial Pain is a valid diagnosis (this was formerly known as "atypical facial pain")
Best regards,
Ron
Ronald W. Pies MD
What do you think of other medical conditions that don't based on subjective symptoms and don't really have biological anchors. Migraines, IBS, CFS, chronic pain etc. the dispute of Thier validity I never see come into question
There are definitely disputes around etiology or causes of such conditions but rare to see them framed as being about their “validity,” which again suggests something peculiar is happening in psychiatric debates. Some of it is a bias against mental illness but it also speaks to the complexity and ambiguity of mental health problems in terms of their conceptualization (illnesses, life problems, personality problems, spiritual problems, etc) as well as their relationship to biology.
IDK if this is specifically about validity, but Van den Bergh wrote an influential paper interpreting these symptoms through the predictive processing framework: https://orbilu.uni.lu/bitstream/10993/30368/1/Van_den_Bergh_et_al_Symptom_Perception_NBBR_2017.pdf