What do you think about short-term recreational drugs that impact serotonin (e.g., MDMA, psychedelics)? These and particularly MDMA do not blunt emotion. Would you have to just say that serotonin has many roles and these modulate some other role?
MDMA has a methamphetamine-like quality and also increases dopamine and norepinephrine, so it's not clear how much to ascribe the positive affect of MDMA to that.
Yes it is fair to say that serotonin has many roles related to its many receptor subtypes.
It's more complicated than "serotonin inhibits emotions". SSRI do, yes. But serotonin acutely, itself? No, analogy from psychedelics and MDMA that suddenly increase serotonergic function - emotions are amplified massively! Only with consistent stimulation of serotonergic receptors by SSRI, after 2 weeks or so their sensitivity decreases and this coincides with therapeutic effectiveness. So does serotonin inhibit emotion? No, not really, it amplifies them, but if you overkill the system, then it shuts down, and emotion is reduced.
Don't classical psychedelics (like LSD and psilocybin) amplify emotions? And don't they act by being serotonin agonists? "Amplifies emotions" seems like a decent descriptor of what they do to mood. Someone feels a little anxious, suddenly they're having a full-blown panic attack. Someone is feeling alright, suddenly they're feeling the happiest they've ever been.
But obviously, this is opposite to your theory of how serotonin is working. What do you make of all this? Is there some reason why serotonin analogues would contradictorily increase emotions?
I didn’t read any of your citations, so I apologize if they would all have make the answer to this obvious: How do you either reconcile this with or refute the “computational psychiatry” model of serotonin mediating the confidence of predictions? It’s easy to say that exogenous serotonin receptor agonists causing hallucinations is incomparable to endogenous serotonin’s activity in the brain, but why would less precise and/or confident predictions - to any degree - have the consequence of less emotion? Or, if it doesn’t, why is this model superior?
Thanks for your thoughtful comment and encouragement.
I think these views are consistent with each other; it's three blind men and an elephant situation. Though I don't know much about the psychedelic angle.
My understanding of the computational view is that serotonin reduces precision/confidence (very over-generalized) of predictions, which again puts it opposed to other neuromodulators like dopamine and acetylcholine.
We could think of emotions as highly weighted, rapidly activated predictions (heuristics). If you see an angry bear, you REALLY want to run away right now. It is difficult to overcome this prior, and acting on a high confidence prediction tends to pay off compared to sitting and musing about it.
So "serotonin makes predictions less precise/confident" is another way of saying it turns down the valence or weighting of those emotional impulses and promotes long-term consideration and planning.
Here are a few quotes from Colombo's "Serotonin, Predictive Processing, and Psychedelics (2022)":
"Specifically, according to one of these hypotheses, serotonin represents expectations about negative outcomes (i.e., punishments or negative rewards) and, on the basis of these representations, it would compute appropriate inhibitory motor responses (Dayan & Huys, 2009). If this hypothesis is true, increased serotonin activity will promote the downregulation of impulsive emotional reactions to expected punishment or provocation, boosting patience and coping with anxiety."
And:
"According to another hypothesis developed within the RL framework, serotonin computes the time and resources available for action, learning and development”...Higher serotonin activity in a decision-making task would signal that there is enough time available for deeper deliberation, more explorative choices and longer-term predictions of the likely outcomes of different choices."
More broadly, I find the predictive processing theory and computational approaches unintuitive. Not saying they're wrong, just hard to understand. I'm not that mathy, but I believe that predictive processing is conceptually similar (and mathematically equivalent?) to control theory. Feedback loops are easier to grasp, and my model of the entire nervous system is a stack of nested control loops. I think this is a restatement of predictive processing theory, not a competitor, so for instance the "prediction error" is just error (SP-PV). This is how we already talk about physiology in the body (cardiac and endocrine systems) so it's more consistent with the rest of biology and medicine.
I’m not sure it’s right to equate emotions with heavily weighted predictions or that this squares with computational models of mental illness, but that’s not what your post was about and computational theory leaves a lot unclear, so I don’t want to drag the conversation down that path.
Does “serotonin inhibits emotions” make accurate predictions about what SSRIs can and can’t treat? Some are also used for anxiety, but, to my understanding, the closest there is to a disorder in which all emotions, with little or no respect to content or direction (unlike anxiety), are insufficiently inhibited is borderline personality disorder: do SSRIs treat that? Conversely, do SSRIs treat any disorders which themselves inhibit emotions?
Feedback to this piece has made me want to understand predictive processing better, particularly how it conceptualizes emotions. So I am going to look into that topic more.
I agree that BPD might be aptly described as "way too much emotions." Interestingly, there is not good evidence for any medications treating BPD, though individuals with BPD often end up on multiple meds, including SSRI, mood stabilizer, antipsychotics, benzos, stimulants, etc. There are bunch of systematic reviews on pharmacological treatment for BPD, which find weak evidence for a handful of med classes, thought details vary. I don't have a good explanation for this, except that BPD has many manifestations and potential endpoints to look at, which makes it hard to study.
Disorders of too little emotion is an interesting idea. Depression can certainly lead to flatness and apathy. My gestalt which I reference in the piece is that this type of depression does not respond as well to SSRIs as to bupropion, which is more activating. I also think about OCPD (no meds used for that) and antisocial personality (lack of guilt?) where meds are also not used except for impulsive aggression.
I'm currently on an NDRI inhibitor. Do we have a good idea neurologically of how those treat depression? I feel like it puts a floor on my mood, mainly.
Thanks for commenting. I assume you're referring to bupropion. Norepinephrine is a non-specific neurotransmitter that generally acts to increase alertness, arousal, awareness re the external environment, things like that.
There are several theories about dopamine; I like Jaak Panksepp's work; he argues that dopamine is the basis of the SEEKING system; which is a basic affective system that prompts animals to "engage in enthusiastic investigations of their environments" and correlates to feelings like excitement, interest, and anticipation. I think of this as "mood" in the purest sense.
Thanks for commenting. I don't see a conflict here. For one, I think you can just as easily say that thoughts come from emotions. But everything runs in both directions. A very, very simple model might have 4 layers of control, each with the potential to both activate and inhibit the layers above and below. The layers could be cognition ("cognitive control" is a common term), emotional feelings, the body, and the environment. Each is affecting and being affected by the others. You end up with a hugely complex dynamical system that can fall into many different equilibria.
What do you think about short-term recreational drugs that impact serotonin (e.g., MDMA, psychedelics)? These and particularly MDMA do not blunt emotion. Would you have to just say that serotonin has many roles and these modulate some other role?
MDMA has a methamphetamine-like quality and also increases dopamine and norepinephrine, so it's not clear how much to ascribe the positive affect of MDMA to that.
Yes it is fair to say that serotonin has many roles related to its many receptor subtypes.
It's more complicated than "serotonin inhibits emotions". SSRI do, yes. But serotonin acutely, itself? No, analogy from psychedelics and MDMA that suddenly increase serotonergic function - emotions are amplified massively! Only with consistent stimulation of serotonergic receptors by SSRI, after 2 weeks or so their sensitivity decreases and this coincides with therapeutic effectiveness. So does serotonin inhibit emotion? No, not really, it amplifies them, but if you overkill the system, then it shuts down, and emotion is reduced.
Don't classical psychedelics (like LSD and psilocybin) amplify emotions? And don't they act by being serotonin agonists? "Amplifies emotions" seems like a decent descriptor of what they do to mood. Someone feels a little anxious, suddenly they're having a full-blown panic attack. Someone is feeling alright, suddenly they're feeling the happiest they've ever been.
But obviously, this is opposite to your theory of how serotonin is working. What do you make of all this? Is there some reason why serotonin analogues would contradictorily increase emotions?
Thanks for explaining why people can't become endlessly happier by taking higher doses of SSRIs.
I didn’t read any of your citations, so I apologize if they would all have make the answer to this obvious: How do you either reconcile this with or refute the “computational psychiatry” model of serotonin mediating the confidence of predictions? It’s easy to say that exogenous serotonin receptor agonists causing hallucinations is incomparable to endogenous serotonin’s activity in the brain, but why would less precise and/or confident predictions - to any degree - have the consequence of less emotion? Or, if it doesn’t, why is this model superior?
Thanks - glad to see you writing!
Thanks for your thoughtful comment and encouragement.
I think these views are consistent with each other; it's three blind men and an elephant situation. Though I don't know much about the psychedelic angle.
My understanding of the computational view is that serotonin reduces precision/confidence (very over-generalized) of predictions, which again puts it opposed to other neuromodulators like dopamine and acetylcholine.
We could think of emotions as highly weighted, rapidly activated predictions (heuristics). If you see an angry bear, you REALLY want to run away right now. It is difficult to overcome this prior, and acting on a high confidence prediction tends to pay off compared to sitting and musing about it.
So "serotonin makes predictions less precise/confident" is another way of saying it turns down the valence or weighting of those emotional impulses and promotes long-term consideration and planning.
Here are a few quotes from Colombo's "Serotonin, Predictive Processing, and Psychedelics (2022)":
"Specifically, according to one of these hypotheses, serotonin represents expectations about negative outcomes (i.e., punishments or negative rewards) and, on the basis of these representations, it would compute appropriate inhibitory motor responses (Dayan & Huys, 2009). If this hypothesis is true, increased serotonin activity will promote the downregulation of impulsive emotional reactions to expected punishment or provocation, boosting patience and coping with anxiety."
And:
"According to another hypothesis developed within the RL framework, serotonin computes the time and resources available for action, learning and development”...Higher serotonin activity in a decision-making task would signal that there is enough time available for deeper deliberation, more explorative choices and longer-term predictions of the likely outcomes of different choices."
More broadly, I find the predictive processing theory and computational approaches unintuitive. Not saying they're wrong, just hard to understand. I'm not that mathy, but I believe that predictive processing is conceptually similar (and mathematically equivalent?) to control theory. Feedback loops are easier to grasp, and my model of the entire nervous system is a stack of nested control loops. I think this is a restatement of predictive processing theory, not a competitor, so for instance the "prediction error" is just error (SP-PV). This is how we already talk about physiology in the body (cardiac and endocrine systems) so it's more consistent with the rest of biology and medicine.
I’m not sure it’s right to equate emotions with heavily weighted predictions or that this squares with computational models of mental illness, but that’s not what your post was about and computational theory leaves a lot unclear, so I don’t want to drag the conversation down that path.
Does “serotonin inhibits emotions” make accurate predictions about what SSRIs can and can’t treat? Some are also used for anxiety, but, to my understanding, the closest there is to a disorder in which all emotions, with little or no respect to content or direction (unlike anxiety), are insufficiently inhibited is borderline personality disorder: do SSRIs treat that? Conversely, do SSRIs treat any disorders which themselves inhibit emotions?
Thanks!
Feedback to this piece has made me want to understand predictive processing better, particularly how it conceptualizes emotions. So I am going to look into that topic more.
I agree that BPD might be aptly described as "way too much emotions." Interestingly, there is not good evidence for any medications treating BPD, though individuals with BPD often end up on multiple meds, including SSRI, mood stabilizer, antipsychotics, benzos, stimulants, etc. There are bunch of systematic reviews on pharmacological treatment for BPD, which find weak evidence for a handful of med classes, thought details vary. I don't have a good explanation for this, except that BPD has many manifestations and potential endpoints to look at, which makes it hard to study.
Disorders of too little emotion is an interesting idea. Depression can certainly lead to flatness and apathy. My gestalt which I reference in the piece is that this type of depression does not respond as well to SSRIs as to bupropion, which is more activating. I also think about OCPD (no meds used for that) and antisocial personality (lack of guilt?) where meds are also not used except for impulsive aggression.
Your questions are very insightful, thanks!
I'm currently on an NDRI inhibitor. Do we have a good idea neurologically of how those treat depression? I feel like it puts a floor on my mood, mainly.
Thanks for commenting. I assume you're referring to bupropion. Norepinephrine is a non-specific neurotransmitter that generally acts to increase alertness, arousal, awareness re the external environment, things like that.
There are several theories about dopamine; I like Jaak Panksepp's work; he argues that dopamine is the basis of the SEEKING system; which is a basic affective system that prompts animals to "engage in enthusiastic investigations of their environments" and correlates to feelings like excitement, interest, and anticipation. I think of this as "mood" in the purest sense.
Thanks for commenting. I don't see a conflict here. For one, I think you can just as easily say that thoughts come from emotions. But everything runs in both directions. A very, very simple model might have 4 layers of control, each with the potential to both activate and inhibit the layers above and below. The layers could be cognition ("cognitive control" is a common term), emotional feelings, the body, and the environment. Each is affecting and being affected by the others. You end up with a hugely complex dynamical system that can fall into many different equilibria.