A new review article co-authored by Prof. Héctor M. Manrique, from the Comparative Minds Research Group (UdG), offers a neuroscientific framework to explain how discriminatory and dehumanising cognition develops and persists. Published in BioSystems, the work brings together researchers from the Universitat de Girona, the Universidad de Zaragoza, University College London (UCL), and the University of Murcia.
Why Do Prejudices Resist Change?
Despite decades of anti-discrimination education and legislation, prejudiced attitudes remain stubbornly present across societies. Why is it so difficult to change what people believe about others? A new study by the Comparative Minds Research Group offers a surprising answer: it may have less to do with ignorance than with the way the brain is wired to avoid uncertainty.
The research draws on Active Inference, a computational framework rooted in the Free Energy Principle, which proposes that the brain is essentially a prediction machine — one that constantly tries to minimise "surprisal", that is, the gap between what it expects and what it actually experiences. When that drive goes awry in social contexts, the result can be a mind that filters out any evidence that challenges its view of the world — including evidence that contradicts its prejudices.
Central to the paper is the concept of Zones of Bounded Surprisal (ZBS), previously developed by Manrique and Walker (2023) to explain why chimpanzees, unlike humans, rarely learn by imitating others. Here, the concept is applied to the social domain. "A discriminatory mind is characterised by a narrow ZBS bandwidth— a rigid cognitive filter that prevents belief updating even when contradictory evidence is directly encountered," explains Prof. Manrique, lead author of the study.
"This paper is the result of applying the theoretical tools we have been developing in comparative cognition to one of the most urgent problems in contemporary societies. Collaborating with Karl Friston allowed us to ground the analysis in first principles of neuroscience and thermodynamics, connecting the biology of the brain to the sociology of discrimination." — Héctor M. Manrique.
Two Types of Biased Mind
Not all discriminatory thinking is the same, and one of the paper's key contributions is drawing a clear line between two qualitatively different profiles. A discriminatory mind is characterised by rigid but potentially modifiable beliefs: the person tends to avoid evidence that challenges their prejudices, remains passive when witnessing wrongdoing within their own group, and gravitates towards social environments — including media bubbles — that confirm what they already think. Uncomfortable as this profile is, it leaves some room for change.
A dehumanising mind is a different matter altogether. It emerges when discriminatory beliefs become so deeply ingrained —often absorbed from early childhood— that they stop being opinions and become part of the person's very identity. At that point, the brain no longer treats prejudice as a belief to be tested against reality; it treats it as a condition of its own existence. In active inference terms, confirming those beliefs becomes an autopoietic imperative — something the individual must do to feel they are still the kind of person they believe themselves to be.
The paper also identifies several mechanisms that keep both profiles locked in place. Discriminatory behaviour can easily become a self-fulfilling prophecy: acting with hostility towards an out-group provokes a defensive reaction, which is then read as proof that the prejudice was justified all along. At the same time, people actively construct social niches —from the neighbourhoods they live in to the content they consume online— that shield them from disconfirming experiences. Most strikingly, the research suggests that adults with narrow ZBS bandwidths in positions of authority can effectively transmit their rigid worldview to children and subordinates, passing discriminatory cognition from one generation to the next.
A Brain Problem, Not Just a Social One
The study draws a thought-provoking parallel between extreme discriminatory thinking and clinical conditions such as schizophrenia, OCD, and depression —all of which involve an excessive resistance to updating beliefs in the face of new evidence. This does not mean that racists are mentally ill, but it does suggest that the cognitive machinery underlying extreme prejudice shares important features with known psychopathologies, and may therefore respond to similar therapeutic approaches.
In this spirit, the authors propose that techniques from Acceptance and Commitment Therapy (ACT) and Mindfulness-Based Cognitive Therapy (MBCT) could help achieve what they call cognitive defusion: the process of disentangling deeply held beliefs from personal identity, so that a person can begin to observe their own prejudices as thoughts rather than truths. The goal is not to replace one belief with another, but to loosen the grip that belief has on the self.
"Tackling dehumanising minds is not simply a matter of providing correct information. It requires creating social environments that make surprise manageable —spaces where encountering difference no longer feels like an existential threat." — Héctor M. Manrique.
At the societal level, the paper revisits the landmark case of U.S. school desegregation in 1954, arguing that enforced exposure to disconfirming social environments can, over time, widen ZBS bandwidths and gradually erode entrenched prejudices —even when attitude change was not the starting point.
Reference:
Manrique, H. M., Friston, K. J., & Walker, M. J. (2026). An active inference explanation of discriminatory cognition with regard to social attitudes and harmful behaviour. BioSystems, 264, 105793. https://doi.org/10.1016/j.biosystems.2026.105793