Ryan Smith

Ryan Smith

@rssmith.bsky.social

Associate Professor at the Laureate Institute for Brain Research. My lab focuses on computational neuroscience and psychiatry, emotion-cognition interactions, prospective planning, exploration, and interoception.

438 followers218 following70 posts

Ryan Smith
@rssmith.bsky.social

Congratulations to Ko-Ping Chou and Navid Hakimi on their new paper doing a formal, extensive empirical comparison of reinforcement learning and active inference in fitting/explaining behavior on a 3-armed bandit task in transdiagnostic and healthy samples (authors.elsevier.com/c/1nboy53nb3...).

authors.elsevier.comPlease wait whilst we redirect youAll content on this site: Copyright © 2026 Elsevier B.V., its licensors, and contributors. All rights are reserved, including those for text and data mining, AI training, and similar technologies. For all open access content, the relevant licensing terms apply.
Ryan Smith
@rssmith.bsky.social

One interesting result was that, at the trial level, RL was better at predicting participants' win-stay choices while active inference was better at predicting their lose-switch choices (and all switch choices). Active inference was also a bit better at predicting when they chose the best option.

Ryan Smith
@rssmith.bsky.social

This is a new paper testing Bayesian computational mechanisms of improvements in cardiac interoception due to a training program developed by @sarahgarf.bsky.social and colleagues. Led by a very talented grad student @chatrins.bsky.social (www.sciencedirect.com/science/arti...)

sciencedirect.comA hierarchical Bayesian model reveals increased precision weighting for afferent cardiac signals, and reduced anxiety, as a function of interoceptive trainingInteroceptive interventions offer a promising avenue for improving mental health conditions, which commonly feature bodily or interoceptive symptoms. …
Ryan Smith
@rssmith.bsky.social

Congrats to Navid on his new paper in Molecular Psychiatry! Computational modeling found that learning rates and uncertainty sensitivity could distinguish between affective disorders and substance use disorders. Forgetting rate was instead affected transdiagnostically. www.nature.com/articles/s41...

nature.comComputational mechanisms of learning and forgetting differentiate affective and substance use disorders - Molecular PsychiatryMolecular Psychiatry - Computational mechanisms of learning and forgetting differentiate affective and substance use disorders

Ryan Smith reposted

Andreea Diaconescu
@cognemo.bsky.social

We’re bringing computational psychiatry to #ACNP2026 Our panel shows how RL, DDMs, active inference, and neural models can unpack symptom heterogeneity across anxiety, psychosis, ADHD, depression, and suicide risk-moving beyond diagnoses to mechanisms. @acnporg.bsky.social #ComputationalPsychiatry

Ryan Smith reposted

LuvLifts
@weluvmike.bsky.social

This is ‘Really’ FUTURISTIC!! #MentalHealth I would be Soo ‘Honored’ to Help in anyway EvenIF ‘That way’ is By SHARING these Advancements!!! @rssmith.bsky.social @mpwpaulus.bsky.social

Springer@springer.springernature.com

🧠📈 How can computational models help decode interoception in #MentalHealth? This chapter explores how sensing the body’s internal signals links to psychiatric disorders— how computational psychiatry may unlock new treatment targets. bit.ly/49cjj2P @rssmith.bsky.social @mpwpaulus.bsky.social

Ryan Smith
@rssmith.bsky.social

Big congrats to Marishka on this. I know how much time and effort went into it. Really happy to see it out! For those interested, she replicated prior longitudinal results using an approach-avoidance conflict task and showed how model parameters afforded out-of-sample clinical prediction.

Computational Psychiatry@cpsyjournal.bsky.social

New paper in CPsy - 'Computational Mechanisms of Approach-Avoidance Conflict Predictively Differentiate Between Affective and Substance Use Disorders' from Marishka Mehta and the group of @rssmith.bsky.social doi.org/10.5334/cpsy...

doi.orgComputational Mechanisms of Approach-Avoidance Conflict Predictively Differentiate Between Affective and Substance Use Disorders | Computational Psychiatry
Ryan Smith
@rssmith.bsky.social

Agreed. The video examples are super interesting. I wish the questions were asked in a more controlled way, but I can’t fault for some limitations of this rarely possible type of work. It seemed quite somatic and abrupt. No spontaneous descriptions of emotion proper either. Only when given a word.

Nicole Rust@nicolecrust.bsky.social

(buried in a reply, but) This is mind-blowing & deserves a post. Scroll to "video" in this article to watch what happens when the human hypothalamus is stimulated: it's a combo of embarrassment/shame and intense body sensations emanating from the heart. WOW. www.brainstimjrnl.com/article/S193...

brainstimjrnl.comComplex negative emotions induced by electrical stimulation of the human hypothalamusStimulation of the ventromedial hypothalamic region in animals has been reported to cause attack behavior labeled as sham-rage without offering information about the internal affective state of the an...

Ryan Smith reposted

Communications Psychology
@commspsychol.nature.com

Computational modeling revealed that individuals with methamphetamine use disorder show reduced information-seeking and slower belief updating when trying to maximize long-term reward. @torutakahashi.bsky.social @rssmith.bsky.social www.nature.com/articles/s44...

nature.comIndividuals with methamphetamine use disorder show reduced directed exploration and learning rates independent of an aversive interoceptive state induction - Communications PsychologyComputational modeling revealed that individuals with methamphetamine use disorder show reduced information-seeking and slower belief-updating when trying to maximize long-term reward. This might crea...

Ryan Smith reposted

Nadine Dijkstra
@nadinedijkstra.bsky.social

I am so excited to share that our paper 'A neural basis for distinguishing imagination from reality' is now published in @cp-neuron.bsky.social! 🧠✨ See thread below! doi.org/10.1016/j.ne...

doi.orgRedirecting
Nadine Dijkstra@nadinedijkstra.bsky.social

I am beyond excited to share our new preprint ‘A neural basis for distinguishing between imagination and reality’ with Thomas von Rein, @peterkok.bsky.social and @smfleming.bsky.social! osf.io/preprints/ps... a thread 🧵

osf.ioOSF

Ryan Smith reposted

Jon Roiser
@jonroiser.bsky.social

Replying to @rssmith.bsky.social

this is excellent - congratulations! In particular nice to see the very robust basic pruning behavioural effect replicate in this version of the task (originally designed for the scanner). It seems that pruning may end up being more relevant to SUD than MDD, @docqhuys.bsky.social!

Ryan Smith reposted

Quentin Huys
@docqhuys.bsky.social

Thank you @clairelavalley.bsky.social and @rssmith.bsky.social. Striking to see aversive pruning altered in people with methamphetamine use, particularly given the results in depression (dx.doi.org/10.1017/S003...).

dx.doi.orgA comparison of ‘pruning’ during multi-step planning in depressed and healthy individuals | Psychological Medicine | Cambridge CoreA comparison of ‘pruning’ during multi-step planning in depressed and healthy individuals - Volume 52 Issue 16
Ryan Smith@rssmith.bsky.social

Huge congrats to @clairelavalley.bsky.social on this new paper. www.nature.com/articles/s41.... Results show that methamphetamine users show a specific deficit in model-based planning, in which they fail to simulate plans with large short-term losses (elevated aversive pruning). Brief thread below!

nature.comComputational mechanisms underlying multi-step planning deficits in methamphetamine use disorder - Translational PsychiatryTranslational Psychiatry - Computational mechanisms underlying multi-step planning deficits in methamphetamine use disorder
Ryan Smith
@rssmith.bsky.social

The groups also differed in scores on a cognitive reflectiveness test, and a mediation model suggested that differences in reflectiveness partially accounted for greater pruning in the methamphetamine users. The effect of anxiety induction on craving was also positively associated by pruning levels.

Ryan Smith
@rssmith.bsky.social

Huge congrats to @clairelavalley.bsky.social on this new paper. www.nature.com/articles/s41.... Results show that methamphetamine users show a specific deficit in model-based planning, in which they fail to simulate plans with large short-term losses (elevated aversive pruning). Brief thread below!

nature.comComputational mechanisms underlying multi-step planning deficits in methamphetamine use disorder - Translational PsychiatryTranslational Psychiatry - Computational mechanisms underlying multi-step planning deficits in methamphetamine use disorder
Ryan Smith
@rssmith.bsky.social

HCs and methamphetamine users completed an established multistep planning task. From certain starting positions, the optimal path required transitioning through a large loss. Unlike other model parameters (planning horizon, reward sensitivity), group differences were selective to aversive pruning.

Ryan Smith
@rssmith.bsky.social

Congratulations to co-first author @clairelavalley.bsky.social on validating this new performance-based measure of social emotion regulation ability: emotionandpsychopathology.org/journal/arti... In psychopathology, this may be crucial for maintaining social support. We hope others find it useful!

emotionandpsychopathology.org The Inter-Personal Affect Regulation Test (IPART): Development and Validation of a Performance-Based Assessment of the Ability to Improve the Emotions of Others | Journal of Emotion and Psyc...
Ryan Smith
@rssmith.bsky.social

One final important note - an earlier preprint from this study instead reported that anxious depression showed elevated directed exploration than HCs that was reduced by the anxiety induction. This was due to a coding error we caught during revision. We apologize for this unfortunate mistake.

Ryan Smith
@rssmith.bsky.social

In general, results suggest state anxiety may interfere with reflection on relative uncertainty and reduce DE in healthy individuals, while consistently elevated anxiety in anxious depression may lead to compensatory strategies that reduce sensitivity to this effect.

Ryan Smith
@rssmith.bsky.social

We combined the Horizon Task with a somatic (interoceptive) anxiety induction using breathing resistance. Both healthy individuals and those diagnosed with anxious depression participated (N=58 and 61, respectively).

Ryan Smith
@rssmith.bsky.social

Anxiety induction reduced directed (but not random) exploration in healthy individuals, but not in those diagnosed with anxious depression. Directed exploration was also associated with scores on a cognitive reflection test.

Ryan Smith
@rssmith.bsky.social

Reinforcement Learning and Active Inference are two frameworks used in computational psychiatry, but these are rarely directly compared empirically. In this new article, we aimed to compare these in a more systematic manner by fitting each to multiple datasets: papers.ssrn.com/sol3/papers....

papers.ssrn.comA Systematic Empirical Comparison of Active Inference and Reinforcement Learning Models in Accounting for Decision-Making Under UncertaintyReinforcement Learning (RL) and Active Inference (AInf) are related computational frameworks for modeling learning and choice under uncertainty. However, differ
Ryan Smith
@rssmith.bsky.social

We first compare their relative ability to account for behavior on a three-armed bandit task in a sample from the US and a sample from Taiwan. We then look at how the parameters correlate with each other and how strongly their mechanistic explanations overlap.

Ryan Smith
@rssmith.bsky.social

We also test their sensitivity to task manipulations, and how strongly parameters in each model correlate with affective measures, trait measures, and psychiatric symptoms. We hope these results offer useful insights into the relative strengths and weaknesses of each framework.

Ryan Smith reposted

Charles Verdonk
@ch-verdonk.bsky.social

I'm excited to attend the #ECNP Workshop on Applied Neuroscience and present our findings on markers of abnormal gastrointestinal interoception predicting relapse in anorexia nervosa. Catch Poster P107 on Thursday, March 20, & my oral presentation on Friday, March 21, at 9:10 AM

Ryan Smith
@rssmith.bsky.social

Congrats to Marishka Mehta on her new chapter! She did a great job summarizing the current state of the field and important future directions. link.springer.com/chapter/10.1... "Computational Approaches for Uncovering Interoceptive Mechanisms in Psychiatric Disorders and Their Biological Basis"

link.springer.comComputational Approaches for Uncovering Interoceptive Mechanisms in Psychiatric Disorders and Their Biological BasisInteroception, the process of detecting, perceiving, and interpreting signals from within the body, is essential for physiological regulation and adaptive behavior. A growing body of research undersco...
Ryan Smith
@rssmith.bsky.social

Excited to congratulate Marishka Mehta on her new preprint (www.medrxiv.org/content/10.1...) combining modeling, machine learning, and longitudinal clinical data. "Computational Mechanisms of Approach-Avoidance Conflict Predictively Differentiate Between Affective and Substance Use Disorders"

medrxiv.orgComputational Mechanisms of Approach-Avoidance Conflict Predictively Differentiate Between Affective and Substance Use DisordersPsychiatric disorders are highly heterogeneous and often co-morbid, posing specific challenges for effective treatment. Recently, computational modeling has emerged as a promising approach for charact...

Ryan Smith reposted

Micah G. Allen
@micahgallen.com

@rssmith.bsky.social and I are working on a symposium proposal on interoceptive and embodied computational psychiatry for this - do let us know if you would like to join as we have open slots. Focus is on quantification and intervention.

Quentin Huys@docqhuys.bsky.social

Poster submissions for the Computational Psychiatry Conference 2025 in Tübingen are now open. Deadline is 7th February. Symposium submissions are open until 15th January. www.cpconf.org. Please RB.

cpconf.orgComputational Psychiatry ConferenceTübingen, Germany (July 14-16, 2025)

Ryan Smith reposted

Ido Aizenbud
@idoai.bsky.social

Human cortical pyramidal neurons are larger, with more elaborate branching, and distinct nonlinear biophysical properties compared to rat cortical pyramidal neurons. Are they more functionally complex? Could that boost the human brain’s computational power? and is that what makes us human? (1/11)

You are up to date.