publications

The inverse base-rate effect is a robust bias in how people resolve ambiguity between competing categories, and the most prominent theories explain it through prediction error. Across two experiments we progressively removed the elements of the predictive-learning design that supply such error signals: first by moving to observational learning, then to an unsupervised procedure in which category labels were not presented. The effect persisted–the irrational bias is independent of supervised learning procedures. We propose a new theory, OSCAR, that integrates core computational principles of the best-validated models and operates on self-generated feedback akin to pattern completion. OSCAR extends the learning dynamics underlying the response bias to observational and unsupervised procedures. Evaluated on a large preexisting supervised dataset in addition to the two new experiments reported here, OSCAR performs competitively against alternatives, and is the first model that reproduces the pattern of individual differences seen in humans across all three procedures. The model provides an explanation of hitherto unexplained eye-tracking data, something none of the alternative accounts provide.

@article{dome2026errorless,
  title = {Errorless {{Irrationality}}: {{A}} Unified Computational Account of the Inverse Base-Rate Effect across Predictive, Observational, and Unsupervised Procedures},
  shorttitle = {Errorless {{Irrationality}}},
  author = {Dome, Lenard and Wills, Andy J.},
  year = {2026},
  month = aug,
  number = {arXiv:2608.06149},
  eprint = {2608.06149},
  primaryclass = {q-bio.NC},
  journal = {arXiv.},
  publisher = {arXiv.},
  doi = {10.48550/arXiv.2608.06149},
  archiveprefix = {arXiv},
  keywords = {auto-associator network, neural network, irrationality, computational modelling, inverse base-rate effect},
  pdf = {https://arxiv.org/pdf/2608.06149}
}

Introspective accuracy shapes self-determined decisions, yet how it changes across the lifespan is unresolved. The evidence at the extremes of adolescence and old age points in conflicting directions. In a unique lifespan dataset N=6,000 spanning 10-90 years, we traced age-related effects of perceptual metacognition. Using piecewise regression, we found that metacognitive sensitivity, tracking how well confidence distinguishes between correct and erroneous decisions, followed a biphasic trajectory with a peak in adolescence and a steady decline towards older age. Critically, we found two distinct drivers for this biphasic trajectory: whilst confidence in correct decisions increased early in life, plateauing after adolescence, confidence in erroneous responses continued to rise throughout life, driving the metacognitive decline later in life. This means that metacognition is stratified rather than unitary, where correct and error confidence operates with distinct lifespan profiles, a critical insight when considering how to support impoverished metacognition at either end of the lifespan.

@article{dome2026age,
  title = {Age-related shifts in metacognition are driven by converging confidence for correct and error judgments across the lifespan},
  url = {osf.io/preprints/psyarxiv/ugb9s_v1},
  year = {2026},
  month = jul,
  pdf = {osf.io/preprints/psyarxiv/ugb9s_v1},
  doi = {10.31234/osf.io/ugb9s_v1},
  journal = {PsyArXiv},
  author = {Dome, Lenard and Moses-Payne, Madeleine and Skvortsova, Vasilisa and Hauser, Tobias U.},
  keywords = {metacognition, lifespan, perceptual decision making, aging}
}

Mental health symptoms, like those in obsessive-compulsive disorder (OCD), show pronounced fluctuations over time. However, little is known about the underlying factors driving these fluctuations. Whilst theoretical accounts propose that cognitive processes drive symptom variability, cross-sectional work is unable to effectively probe these mechanisms. Based on prevailing notions of doubt and underconfidence being cognitive hallmarks of OCD, we investigated whether metacognitive processes correspond to OCD symptoms. We employed a novel smartphone-based momentary assessment design wherein participants with OCD (N = 137) logged instances of experiencing OCD symptoms across 2-weeks. Participants also completed daily scheduled assessments, and played a perceptual metacognitive task every two days. Our results showcase the validity of our novel approach and, importantly, reveal that metacognition drives OCD symptoms. Concretely, we found directional effects whereby reductions in both self-confidence and task-based metacognition preceded occurrences of OCD symptom logs, but not vice versa. Thus, our novel micro-longitudinal study design provides compelling evidence for cognitive distortions (both self-reported and task-based) preceding symptom escalations, signifying the potential for designing novel, just-in-time interventions.

@article{marzuki2026drops,
  title = {Drops in Confidence Precede Symptoms of Obsessive-Compulsive Disorder},
  author = {Marzuki, A. A. and Kosina, L. A. and Dome, L. and Hewitt, S. R. C. and Hauser, T. U.},
  year = {2026},
  month = jul,
  journal = {Nature Communications.},
  volume = {17},
  number = {1},
  pages = {8333},
  publisher = {Nature Publishing Group},
  issn = {2041-1723},
  doi = {10.1038/s41467-026-76671-5},
  urldate = {2026-08-13},
  copyright = {2026 The Author(s)},
  keywords = {metacognition,obsessive-compulsive disorder,computational psychiatry,ecological momentary assessment}
}

The Computational Psychiatry Modelling (cpm) toolbox is a Python library for theory-driven modelling in computational psychiatry and cognitive (neuro-)science. cpm integrates a wide range of established computational approaches in a single framework. It is designed to be accessible to both expert and non-expert modellers in order to conduct cutting-edge computational modelling following the best practices of computational modelling and reporting. The toolbox provides a flexible and modular architecture that adjusts to different needs, and covers a wide range of tasks (such as risky decision-making, reward/punishment learning, perceptual metacognition), models (including those based on associative, reinforcement learning, and signal detection theories) and methods (such as hierarchical hyperparameter estimation using empirical and variational Bayesian techniques). The cpm toolbox thus provides the means for novices and experts in computational modelling alike to rapidly model their data and to adhere to best practices. We believe that building such an overarching, customisable tool can facilitate access to computational modelling and thus jumpstart computational approaches in psychiatry and beyond.

@article{dome2026cpm,
  title = {cpm: {{A}} Python Library for Theory-Driven Modelling in Computational Psychiatry},
  author = {Dome, Lenard and Hezemans, Frank H. and Kadri, Kenza and Wagner, Ben J. and Webb, Andrew and Hauser, Tobias U.},
  year = {2026},
  month = may,
  journal = {PLOS Computational Biology.},
  volume = {22},
  number = {7},
  pages = {1--31},
  publisher = {Public Library of Science},
  doi = {10.1371/journal.pcbi.1014481},
  keywords = {computational psychiatry, computational modelling, python, open-source software, cognitive science}
}

Large Language Models (LLMs) are increasingly used to predict human behavior from plain-text description of experimental tasks that range from judging disease severity to consequential medical decisions. While these methods promise quick insights without complex psychological theories, we reveal a critical flaw: they often latch onto accidental patterns in the data that seem predictive but collapse when faced with novel experimental conditions. Testing across multiple behavioral studies, we show these models can generate wildly inaccurate predictions, sometimes even reversing true relationships, when applied b‘eyond their training context. Standard validation techniques miss this flaw, creating false confidence in their reliability. We introduce a simple diagnostic tool to spot these failures and urge researchers to prioritize theoretical grounding over statistical convenience. Without this, LLM-driven behavioral predictions risk being scientifically meaningless, despite impressive initial results.

@article{schlegelmich2026snake,
  author = {Schlegelmilch, René and Dome, Lenard},
  journal = {Behavioral Sciences.},
  title = {Snake oil or Panacea? How to Misuse AI in Scientific Inquiries of the Human Mind.},
  year = {2026},
  month = jan,
  volume = {16},
  issue = {2},
  doi = {10.3390/bs16020219},
  keywords = {large-languge models, out-of-distribution generalization, model adequacy, extrapolation failure},
  pdf = {https://www.mdpi.com/2076-328X/16/2/219/pdf}
}

Models are often evaluated when their behavior is at its closest to a single, sometimes averaged, set of empirical results, but this evaluation neglects the fact that both model and human behavior can be heterogeneous. Here, we develop a measure, g-distance, which considers model adequacy as the extent to which models exhibit a similar range of behaviors to the humans they model. We define g as the combination of two easily interpretable dimensions of model adequacy: accommodation and excess flexibility. We apply this measure to five models of an irrational learning effect, the inverse base-rate effect (IBRE). g-distance identifies two models, a neural network with rapid attentional shifts (NNRAS) and a dissimilarity-similarity generalized context model (DGCM18), that outperform the previously most supported model (EXIT). We show that this conclusion holds for a wide range of beliefs about the relative importance of excess flexibility and accommodation. We further show that a pre-existing metric, the Bayesian Information Criterion (BIC), misidentifies a known-poor model of the IBRE as the most adequate model. Along the way, we discover that some of the models accommodate human behavior in ways that seem unintuitive from an informal understanding of their operation, thus underlining the importance of formal expression of theories. We discuss the implications of our findings for model evaluation generally, and for models of the inverse base-rate effect in particular, and end by suggesting future avenues of research in computational modeling.

@article{dome2025gdistance,
  title = {g-{{Distance}}: {{On}} the Comparison of Model and Human Heterogeneity.},
  shorttitle = {g-{{Distance}}},
  author = {Dome, Lenard and Wills, Andy J.},
  year = {2025},
  journal = {Psychological Review},
  volume = {132},
  number = {3},
  month = jun,
  pages = {632--655},
  publisher = {American Psychological Association},
  issn = {0033-295X},
  doi = {10.1037/rev0000550},
  urldate = {2025-04-16},
  langid = {english},
  pdf = {https://psyarxiv.com/ygmcj},
  selected = {true},
  keywords = {computational modelling, irrationality, model adequacy, heterogeneity, inverse base-rate effect}
}

The inverse base-rate effect (IBRE) is an irrational phenomenon in predictive learning. It occurs when people try to generalize what they have experienced to novel and ambiguous events. This irrational generalization manifests as a preference for rare, unlikely outcomes in the face of ambiguity. At least two formal mathematical models of this irrational preference (EXIT, NNRAS) lead to a counter-intuitive prediction: the effect reduces under concurrent load. We tested this prediction across two experiments (N1 = 72, M1age = 20.12; N2 = 160, M2age = 20.88). We confirm the prediction, but only when participants were under an obvious time constraint. This empirical confirmation is as surprising as the prediction itself—irrationality reduces under increased task demands. Further, our data are more consistent with the NNRAS model than with EXIT, the most prominent model of the IBRE to date.

@article{dome2025better,
  title = {Better generalization through distraction? Concurrent load reduces the size of the inverse base-rate effect},
  journal = {Psychonomic Bulletin & Review.},
  author = {Dome, Lenard and Wills, Andy J.},
  year = {2025},
  month = feb,
  doi = {10.3758/s13423-025-02661-1},
  pdf = {https://doi.org/10.31234/osf.io/eskr9},
  supp = {{https://osf.io/u8kcb/}, {https://osf.io/ng8kj/}},
  bibtex_show = {true},
  keywords = {inverse base-rate effect, predictive learning, cognitive load, irrationality, computational modelling}
}

The inverse base-rate effect is a robust irrational bias that arises when people face ambiguity. The most prominent theories of this irrational bias depend on prediction error. In this study, we gradually removed elements of a predictive learning design to test the extent to which error-driven processes underlie this bias. In our first experiment, we removed explicit feedback by implementing the inverse base-rate effect in an observational learning procedure. In our second study, we further removed any causal relationship between stimulus features and category labels by moving towards an unsupervised learning procedure. This removed any information participants could use to identify category labels. In both experiments, the inverse base-rate effect persisted and remained robust. This outcome suggests that this irrational bias is independent of supervised learning procedures. We propose that any theories and models of the inverse base-rate effect must manage information encoding and connection updates without explicit prediction error. We end by proposing two clear paths for future investigations.

@inproceedings{dome2023errorless,
  title = {Errorless irrationality: removing error-driven components from the inverse base-rate effect paradigm},
  author = {Dome, Lenard and Wills, Andy J.},
  booktitle = {Proceedings of the Annual Meeting of the Cognitive Science Society.},
  volume = {45},
  number = {45},
  year = {2023},
  pdf = {https://psyarxiv.com/936bj/},
  doi = {10.31234/osf.io/936bj},
  bibtex_show = {true},
  pages = {237-243},
  supp = {{https://osf.io/auwvt/}, {https://osf.io/2tmc4/}},
  keywords = {inverse base-rate effect, predictive learning, observational learning, unsupervised learning, irrationality}
}

There is a growing interest in alternative explanations to thedual-system account of how people learn category structuresvarying in their optimal decision bounds (unidimensional andinformation-integration structures). Recognition memory per-formance and hippocampal activation patterns in these tasksare two interesting findings, which have not been formally ex-plained. Here, we carry out a formal simulation with SUS-TAIN (Love, Medin, & Gureckis, 2004), an adaptive modelof category learning, which had great success in accountingfor recognition memory performance and fMRI activity pat-terns. We show, for the first time, that a formal single-systemmodel of category learning can accommodate recognition per-formance after learning and is consistent with fMRI data ob-tained while participants learned these structures

@inproceedings{dome21sustain,
  first_author = {true},
  title = {SUSTAIN captures category learning, recognition, and hippocampal activation in a unidimensional vs information-integration task},
  author = {Dome, Lenard and Edmunds, Charlotte E. R. and Wills, Andy J.},
  keywords = {categorization, recognition memory, formal model, computational modelling, adaptive clustering},
  booktitle = {Proceedings of the Annual Meeting of the Cognitive Science Society.},
  volume = {43},
  year = {2021},
  pages = {3013-3019},
  address = {Austin, TX: Cognitive Science Society},
  pdf = {https://escholarship.org/uc/item/5r98q3dr},
  supp = {https://osf.io/jc9xs/},
  url = {https://escholarship.org/uc/item/5r98q3dr},
  osf = {https://osf.io/jc9xs/},
  bibtex_show = {true}
}

It is generally assumed that the Rescorla and Wagner (1972) model adequately accommodates the full results of simple cue competition experiments in humans (e.g. Dickinson et al., 1984), while the Bush and Mosteller (1951) model cannot. We present simulations that demonstrate this assumption is wrong in at least some circumstances. The Rescorla-Wagner model, as usually applied, fits the full results of a simple forward cue-competition experiment no better than the Bush-Mosteller model. Additionally, we present a novel finding, where letting the associative strength of all cues start at an intermediate value (rather than zero), allows this modified model to provide a better account of the experimental data than the (equivalently modified) Bush-Mosteller model. This modification also allows the Rescorla-Wagner model to account for a redundancy effect experiment (Uengoer et al., 2013); something that the unmodified model is not able to do. Furthermore, the modified Rescorla-Wagner model can accommodate the effect of varying the proportion of trials on which the outcome occurs (i.e. the base rate) on the redundancy effect (Jones et al., 2019). Interestingly, the initial associative strength of cues varies in line with the outcome base rate. We propose that this modification provides a simple way of mathematically representing uncertainty about the causal status of novel cues within the confines of the Rescorla-Wagner model. The theoretical implications of this modification are discussed. We also briefly introduce free and open resources to support formal modelling in associative learning.

@article{spicer2021representing,
  title = {Representing uncertainty in the Rescorla-Wagner model: blocking, the redundancy effect, and outcome base rate},
  author = {Spicer, Stuart G. and Wills, Andy J. and Jones, Peter M. and Mitchell, Chris J. and Dome, Lenard},
  keywords = {associative learning, prediction error, predictive learning, uncertainty, computational modelling},
  year = {2021},
  journal = {Open Journal of Experimental Psychology and Neuroscience},
  url = {https://doi.org/10.46221/ojepn.2021.6623},
  doi = {10.46221/ojepn.2021.66233},
  pdf = {https://www.ojepn.com/images/2021/2021-6623.pdf},
  osf = {https://osf.io/7u6re/},
  supp = {https://osf.io/7u6re/},
  bibtex_show = {true}
}

The aim was to quantify ego depletion and measure its effect on inhibitory control. Adults (N = 523) received the letter “e” cancellation ego depletion task and were subsequently tested on Stroop task performance. Difficulty of the cancellation task was systematically manipulated by modifying the text from semantically meaningful to non-meaningful sentences and words (Experiment 1) and by increasing ego depletion rule complexity (Experiment 2). Participants’ performance was affected by both text and rule manipulations. There was no relation between ego depletion task performance and subsequent Stroop performance. Thus, irrespective of the difficulty of the ego depletion task, Stroop performance was unaffected. The widely used cancellation task may not be a suitable inducer of ego depletion if ego depletion is considered as a lack of inhibitory control.

@article{wimmer2019letter,
  title = {Is the Letter Cancellation Task a Suitable Index of Ego Depletion?},
  author = {Wimmer, Marina C. and Dome, Lenard and Hancock, Peter J. B. and Wennekers, Thomas},
  ketwords = {ego-depletion, inhibitory control, letter cancellation, replication, conceptual questions},
  journal = {Social Psychology},
  year = {2019},
  publisher = {Hogrefe Publishing},
  doi = {10.1027/1864-9335/a000393},
  url = {https://doi.org/10.1027/1864-9335/a000393},
  pdf = {https://github.com/lenarddome/publications/blob/main/pdf/wimmer2019ego-depletion.pdf},
  supp = {https://doi.org/10.6084/m9.figshare.7575581},
  figshare = {https://doi.org/10.6084/m9.figshare.7575581},
  pages = {345-354},
  bibtex_show = {true},
  keywords = {ego-depletion, inhibitory control, letter cancellation, replication, conceptual questions}
}