Causal reasoning is domain-specific in the human brain
- 1MIT
Abstract
Causal reasoning enables humans to explain the past, predict the future, and intervene on the present. Despite extensive interest in causal inference by philosophers, cognitive scientists, and developmental psychologists, few studies have investigated causal reasoning in the human brain. Here we use fMRI to ask whether a unified brain system exists that implements causal inference across content domains, or whether instead distinct brain systems are engaged in causal reasoning about social agents and physical objects. Across three pre-registered experiments, we find that causal reasoning about social and physical phenomena engage distinct cortical regions. The fronto-parietal Physics Network (PN), previously implicated in physical reasoning, and the temporo-parietal junction (TPJ), previously implicated in social reasoning, both showed stronger responses to causal than non-causal conditions, but only within their respective domains. Selective responses emerged in both a reading comprehension task (Experiment 1) and a sentence matching task (Experiment 2), and for preventative as well as generative causal reasoning (Experiment 3). No region selectively engaged causal reasoning or showed common patterns of response across voxels for causality in both physical and social domains in whole-brain analyses. Thus, despite its central role in human cognition, causal reasoning is implemented in separate domain-specific regions rather than a single unified brain region.
Related
- Whole-brain modeling of dynamic causal circuits in human cognition using amortized variational inference
- Causal mapping of self-motion networks in the human brain
- MAST1 is neuron-specific regulator of mitochondrial trafficking and abundance in human brain tissue
- Division-Specific Organization of a Shared Functional Scaffold in the Early-Life Human Brain
- Network-specific metabolic cost of functional connectivity in the human brain
BibTeX
@misc{pramod2026causal,
title = {{Causal reasoning is domain-specific in the human brain}},
author = {R. Pramod and J. Chomik-Morales and L. Schulz and N. G. Kanwisher},
howpublished = {bioRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.27.762213},
url = {https://www.biorxiv.org/content/10.64898/2026.10.27.762213v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - Causal reasoning is domain-specific in the human brain
AU - R. Pramod
AU - J. Chomik-Morales
AU - L. Schulz
AU - N. G. Kanwisher
PY - 2026
DA - 2026/10/27
PB - bioRxiv
DO - 10.64898/2026.10.27.762213
UR - https://www.biorxiv.org/content/10.64898/2026.10.27.762213v1
AB - Causal reasoning enables humans to explain the past, predict the future, and intervene on the present. Despite extensive interest in causal inference by philosophers, cognitive scientists, and developmental psychologists, few studies have investigated causal reasoning in the human brain. Here we use fMRI to ask whether a unified brain system exists that implements causal inference across content domains, or whether instead distinct brain systems are engaged in causal reasoning about social agents and physical objects. Across three pre-registered experiments, we find that causal reasoning about social and physical phenomena engage distinct cortical regions. The fronto-parietal Physics Network (PN), previously implicated in physical reasoning, and the temporo-parietal junction (TPJ), previously implicated in social reasoning, both showed stronger responses to causal than non-causal conditions, but only within their respective domains. Selective responses emerged in both a reading comprehension task (Experiment 1) and a sentence matching task (Experiment 2), and for preventative as well as generative causal reasoning (Experiment 3). No region selectively engaged causal reasoning or showed common patterns of response across voxels for causality in both physical and social domains in whole-brain analyses. Thus, despite its central role in human cognition, causal reasoning is implemented in separate domain-specific regions rather than a single unified brain region.
ER -