Hernán Anlló

Rationale: Anxiety affects 22–56% of patients with chronic obstructive pulmonary disease and predicts exacerbations, readmissions, and mortality. Mind-body adjuncts to pulmonary rehabilitation are recommended, but trials have lacked comparators isolating technique-specific effects and left determinants of individual benefit unexamined. Objectives: Testing whether adjunctive medical hypnosis reduces anxiety more than matched structured relaxation. Methods: Cluster-randomized, active-comparator trial with patients undergoing pulmonary rehabilitation for chronic obstructive pulmonary disease. Consecutive 4-week inpatient cohorts were randomized 1:1 to adjunctive hypnosis or structured relaxation matched on session length, practitioner pool, and framing. Primary outcome was the six-item State-Trait Anxiety Inventory at week 4 (intention to treat). Complementary multidimensional metrics of anxiety were obtained for robustness. Exploratory analyses tested outcome moderation by baseline sensory-affective dyspnea burden. Measurements and Main Results: Of 80 participants (24 clusters), 79 were analyzed (hypnosis 36, relaxation 43). Hypnosis did not differ from relaxation on the primary outcome (+0.95 points; 95% credible interval −5.85 to 7.42; posterior probability of clinically important hypnosis advantage 0.12) or its complementary metrics. Anxiety improved in both arms (pooled −5.72 points; −9.35 to −2.00). Exploratory analyses showed that the hypnosis advantage rose with baseline sensory and fell with affective burden, reaching −9.26 points at a high-sensory, low-affective profile (−23.37 to 4.71; posterior probability of clinically important hypnosis advantage 0.81). No deaths or serious adverse events occurred. Conclusions: Anxiety improved with no advantage of hypnosis over matched relaxation. Exploratory analysis suggests greater hypnosis benefit in patients with predominantly sensory dyspnea, proposing a candidate phenotype for stratified trials. Clinical trial registered with www.clinicaltrials.gov ( NCT04868357 ). Impact statement: Anxiety affects 22–56% of patients with chronic obstructive pulmonary disease and independently predicts exacerbations, readmissions, and death; current guidance recommends mind-body adjuncts to pulmonary rehabilitation for anxiety, yet their pooled benefit is heterogeneous, and no demographic or disease-severity variable explains who responds. By randomizing medical hypnosis against a structured-relaxation comparator matched on everything but hypnotic suggestion, this trial found no hypnosis benefit over relaxation for the pooled population, yet a post hoc multilevel moderator analysis revealed a clinically important hypnosis benefit that scaled with baseline sensory dyspnea burden. This hypothesis-generating responder profile is identifiable at screening with instruments already in routine respiratory practice, and points to the sensory–affective architecture of dyspnea as a plausible discriminant for matching adjunct to patient. At a Glance Commentary: Scientific Knowledge on the Subject: Pulmonary rehabilitation has a moderate anxiolytic effect in chronic obstructive pulmonary disease, with marked between-patient heterogeneity that demographic and disease-severity variables do not predict. Mind-body adjuncts have been tested almost exclusively against waitlist or usual care, designs that cannot separate technique-specific effects from the nonspecific benefits of structured psychological contact. Hypnosis attenuates both the sensory and affective dimensions of experimentally induced dyspnea, but whether this translates into a reliable anxiolytic benefit during rehabilitation, and for whom, is unknown. What This Study Adds to the Field: This cluster-randomized trial compared medical hypnosis with structured relaxation matched on session number, length, format, and self-management framing, delivered from verbatim scripts by the same certified practitioners, leaving hypnotic suggestion the only intended difference. In unselected patients, hypnosis conferred no advantage over relaxation on primary or confirmatory anxiety outcomes. An exploratory moderation analysis, however, suggested that hypnotic advantage grew with baseline sensory dyspnea burden, reaching three times the minimal clinically important difference at a high-sensory, low-affective profile. This potential candidate responder phenotype, pending stratified confirmation, is identifiable at screening with routine instruments in respiratory practice, and underscores the sensory–affective dyspnea profile’s potential as a discriminant for matching adjunct to patient.

L'Hypnose à travers l'attention: Dans le présent travail, nous proposons qu'un aperçu plus clair de l'interaction entre la suggestion hypnotique et l'attention aiderait à établir le point précis du chronogramme perceptif auquel les effets de l'hypnose interviennent, comment modulent-ils exactement le contrôle cognitif et dans quelle mesure la réponse hypnotique dépend-elle des ressources attentionnelles. Afin de répondre à ces questions expérimentales, nous avons développé trois projets de recherche: (1) les données normatives sur notre traduction en français pour l'Échelle de Susceptibilité Hypnotique de Groupe Harvard, (2) une évaluation des effets de la suggestion posthypnotique sur l'attention visuo-spatiale et (3) une évaluation sur la capacité de la suggestion hypnotique de moduler l'allocation automatique de l'attention accordée par le Anger Superiority Effect. Les résultats de notre première étude nous ont permis de noter avec fiabilité la susceptibilité hypnotique de plus de 500 participants pour les études qui ont suivi. Les résultats de notre deuxième étude indiquent que, pour les participants hautement susceptibles, la suggestion posthypnotique a perturbé avec succès les mécanismes d'attention précoce nécessaires à la stimulation de l'amorçage, ainsi que des jugements de visibilité subjectifs tardifs. Notre troisième étude a révélé que, grâce à une suggestion hypnotique, les participants hautement hypnotizables ont pu empêcher l'allocation automatique de l'attention vers des expressions de colère par un découplage stratégique du contrôle cognitif, mais seulement lorsque les ressources attentionnelles n'avaient pas été cooptées par des processus concurrents. Ensemble, nos résultats appuient les idées selon lesquelles l'hypnose émet ses effets grâce au contrôle cognitif, qui peut perturber les mécanismes attentionnels précoces et tardifs de manières distinctes et que la disponibilité des ressources attentionnelles détermine l'éventail d'action de l'induction et de la suggestion hypnotiques

Language is the cornerstone of teaching, but we still lack principled ways to identify what makesexplanations effective in terms of their composition. In a two-stage behavioural paradigm,“Teacher” participants learned two-armed bandit tasks through trial and error and wrote free-textlessons. “Pupil” participants received these lessons before completing the same tasks. Wefound that receiving a well-constructed lesson improved experiential learning. To understandwhat made those lessons “Good”, we uncovered their latent semantic dimensions through aninferential use of large language models. These dimensions (“Memorization”, “PatternRecognition”, “Option Ranking”, “Randomness”) predicted expert judges’ ranking of lessons andbehavioural outcomes. Further, artificially-generated Good and Bad lessons based on thesedimensions replicated the observed effect of good explanations on experiential performance.Overall, results show that the semantic structure of instructional language measurably shapeshow people learn from experience, and that those structures can be discovered, interpreted,and used to refine our understanding of the cognitive mechanisms underlying teaching andlearning.

Individuals often rely on the advice of more experienced peers to minimise uncertainty and increase success likelihood. In most domains where knowledge is acquired through experience, advisers are themselves continuously learning. Here we examine the way advising behaviour changes throughout the learning process, and the way individual traits and costs and benefits of giving advice shape this behaviour. We ran a series of experiments implementing a decision task within a reinforcement learning framework, where participants could decide to share their choices as advice to others. Participants were overall likely to share their choices as advice, even on the first trial before learning. Tendency to share advice and advice quality increased as advisers learned about the value of choices, and moved from exploratory to exploitative behaviour. The introduction of consequences to advising resulted in a shift of the overall tendency to give advice, lowering it when advising implicated monetary loss, and increasing it when advising held reputational value. Individual differences in social anxiety levels were associated with lower tendency to share exploratory decisions. Our results show that advisers tend to share choices that are backed by their own experience, but that this relationship can be altered by advice-consequences and individual traits.

Recent evidence indicates that reward value encoding in humans is highly context dependent, leading to suboptimal decisions in some cases, but whether this computational constraint on valuation is a shared feature of human cognition remains unknown. Here we studied the behaviour of n = 561 individuals from 11 countries of markedly different socioeconomic and cultural makeup. Our findings show that context sensitivity was present in all 11 countries. Suboptimal decisions generated by context manipulation were not explained by risk aversion, as estimated through a separate description-based choice task (that is, lotteries) consisting of matched decision offers. Conversely, risk aversion significantly differed across countries. Overall, our findings suggest that context-dependent reward value encoding is a feature of human cognition that remains consistently present across different countries, as opposed to description-based decision-making, which is more permeable to cultural factors.

In the present study, we investigate and compare reasoning in large language models (LLMs) and humans, using a selection of cognitive psychology tools traditionally dedicated to the study of (bounded) rationality. We presented to human participants and an array of pretrained LLMs new variants of classical cognitive experiments, and cross-compared their performances. Our results showed that most of the included models presented reasoning errors akin to those frequently ascribed to error-prone, heuristic-based human reasoning. Notwithstanding this superficial similarity, an in-depth comparison between humans and LLMs indicated important differences with human-like reasoning, with models’ limitations disappearing almost entirely in more recent LLMs’ releases. Moreover, we show that while it is possible to devise strategies to induce better performance, humans and machines are not equally responsive to the same prompting schemes. We conclude by discussing the epistemological implications and challenges of comparing human and machine behavior for both artificial intelligence and cognitive psychology.

In the present study, we investigate and compare reasoning in large language models (LLM) and humans using a selection of cognitive psychology tools traditionally dedicated to the study of (bounded) rationality. To do so, we presented to human participants and an array of pretrained LLMs new variants of classical cognitive experiments, and cross-compared their performances. Our results showed that most of the included models presented reasoning errors akin to those frequently ascribed to error-prone, heuristic-based human reasoning. Notwithstanding this superficial similarity, an in-depth comparison between humans and LLMs indicated important differences with human-like reasoning, with models’ limitations disappearing almost entirely in more recent LLMs’ releases. Moreover, we show that while it is possible to devise strategies to induce better performance, humans and machines are not equally-responsive to the same prompting schemes. We conclude by discussing the epistemological implications and challenges of comparing human and machine behavior for both artificial intelligence and cognitive psychology.

Individuals often rely on the advice of more experienced peers to minimize uncertainty and increase success likelihood. In most domains where knowledge is acquired through experience, advisers are themselves continuously learning. Here we examine the way advising behavior changes throughout the learning process, and the way that costs and benefits of giving advice shape this behavior. We ran a series of experiments implementing a decision task within a reinforcement learning framework, where participants could decide to share their choices as advice to others. Participants were overall likely to share their choices as advice, even on the first trial before learning. Tendency to share advice and advice quality increased as advisers learned about the value of choices, and moved from exploratory to exploitative behavior. The introduction of consequences to advising resulted in a shift of the overall tendency to give advice, lowering it when advising implicated monetary loss, and increasing it when advising held reputational value. Individual differences in social anxiety levels were associated with lower tendency to share exploratory decisions. Our results show that advisers tend to share choices that are backed by their own experience, but that this relationship can be altered by advice-consequences and individual traits.

In the present study, we investigate and compare reasoning in large language models (LLM) and humans using a selection of cognitive psychology tools traditionally dedicated to the study of (bounded) rationality. To do so, we presented to human participants and an array of pretrained LLMs new variants of classical cognitive experiments, and cross-compared their performances. Our results showed that most of the included models presented reasoning errors akin to those frequently ascribed to error-prone, heuristic-based human reasoning. Notwithstanding this superficial similarity, an in-depth comparison between humans and LLMs indicated important differences with human-like reasoning, with models’ limitations disappearing almost entirely in more recent LLMs’ releases. Moreover, we show that while it is possible to devise strategies to induce better performance, humans and machines are not equally-responsive to the same prompting schemes. We conclude by discussing the epistemological implications and challenges of comparing human and machine behavior for both artificial intelligence and cognitive psychology.