Tuesday 20 January 2026 Host: Peggy SerièsSpeakers: Lars Werne and Jeffrey ScholesThis workshop, entitled "Computational approaches to PTSD", will consist in three talks: a brief introduction given by Peggy Seriès, and two research talks.Title: PTSD Flashbacks as altered perceptual inference (Lars Werne)Abstract: In PTSD, trauma changes how individuals assess future situations.Flashbacks are a striking example of this; when faced with a reminder of their trauma -- such as a related but harmless visual stimulus -- individuals may vividly relive the event. Although subjects generally remain aware that they are recalling a memory, their emotional and physiological responses are often consistent with an extreme, ongoing threat. In this talk, I will present a computational framework that treats perceptual experience as an inference over contents, intensity, and sense of reality. I will compare different ways of capturing flashbacks as altered perceptual inference within this model, and discuss how these mechanisms may relate to empirical findings on PTSD, as well as to other symptoms such as nightmares and impaired verbal memory of the trauma. Title: Computational perspectives on behavioural differences in PTSD: A pilot multidimensional study on learning and decision-making (Jeffrey Scholes)Abstract: PTSD is a debilitating disorder linked to changes in behaviour due to an altered perception of the world following trauma. Computational modelling aims to improve our mechanistic understanding of what drives this disorder, for example, by shedding light on cognitive issues related to learning and decision-making.In this talk, I give a brief overview of computational approaches that have examined behavioural differences in PTSD. I highlight promising results, as well as issues in comparing approaches. I then outline a recent pilot study we have performed that utilises a multidimensional approach to disentangle variance in learning and decision-making that is specific to PTSD, from that which is specific to other disorders (anxiety, autism). By utilising a bifactor analysis on a battery of psychiatric self-report questionnaires, and fitting participant responses from a one-armed bandit task to various models, we nuance previous results and highlight the benefits of such a multidimensional approach Jan 20 2026 13.00 - 14.00 Tuesday 20 January 2026 Speakers: Lars Werne and Jeffrey Scholes IF, G.03
Tuesday 20 January 2026 Host: Peggy SerièsSpeakers: Lars Werne and Jeffrey ScholesThis workshop, entitled "Computational approaches to PTSD", will consist in three talks: a brief introduction given by Peggy Seriès, and two research talks.Title: PTSD Flashbacks as altered perceptual inference (Lars Werne)Abstract: In PTSD, trauma changes how individuals assess future situations.Flashbacks are a striking example of this; when faced with a reminder of their trauma -- such as a related but harmless visual stimulus -- individuals may vividly relive the event. Although subjects generally remain aware that they are recalling a memory, their emotional and physiological responses are often consistent with an extreme, ongoing threat. In this talk, I will present a computational framework that treats perceptual experience as an inference over contents, intensity, and sense of reality. I will compare different ways of capturing flashbacks as altered perceptual inference within this model, and discuss how these mechanisms may relate to empirical findings on PTSD, as well as to other symptoms such as nightmares and impaired verbal memory of the trauma. Title: Computational perspectives on behavioural differences in PTSD: A pilot multidimensional study on learning and decision-making (Jeffrey Scholes)Abstract: PTSD is a debilitating disorder linked to changes in behaviour due to an altered perception of the world following trauma. Computational modelling aims to improve our mechanistic understanding of what drives this disorder, for example, by shedding light on cognitive issues related to learning and decision-making.In this talk, I give a brief overview of computational approaches that have examined behavioural differences in PTSD. I highlight promising results, as well as issues in comparing approaches. I then outline a recent pilot study we have performed that utilises a multidimensional approach to disentangle variance in learning and decision-making that is specific to PTSD, from that which is specific to other disorders (anxiety, autism). By utilising a bifactor analysis on a battery of psychiatric self-report questionnaires, and fitting participant responses from a one-armed bandit task to various models, we nuance previous results and highlight the benefits of such a multidimensional approach Jan 20 2026 13.00 - 14.00 Tuesday 20 January 2026 Speakers: Lars Werne and Jeffrey Scholes IF, G.03
