• Thermodynamic Unification of Propositional Belief Evaluation: Subsuming Content-Independent fMRI Signatures Under Metabolic Active Inference

Neuroimaging studies of propositional evaluation reveal that the human brain relies on a remarkably consistent, content-independent network to assess truth, falsehood, and uncertainty. Anchored in the ventromedial prefrontal cortex (vmPFC), the anterior insula, and the anterior cingulate cortex (ACC), this circuitry operates identically whether we are evaluating mathematical equations, geographical facts, ethical claims, or religious doctrines (Harris et al., 2008, 2009). Yet classical neuroimaging taxonomies treat belief, disbelief, and uncertainty as static, cost-free labels. They leave an essential question unanswered: why do specific mental states impose distinct subjective effort, reaction-time delays, and metabolic costs? Here, we resolve this problem by unifying content-independent neuroimaging findings within the Thermodynamic State-Space Model (TSSM) of active inference. First, we derive explicit biophysical couplings that link variational free energy (F) and Kullback-Leibler divergence (DKL) directly to Landauer work bounds, adenosine triphosphate (ATP) hydrolysis flux (J̇ATP), and regional cerebral metabolic rate of oxygen consumption (CMRO2). Second, we explain vmPFC reward signaling as neurochemical reinforcement for the rapid collapse of free energy into familiar, prior-congruent basins. In contrast, we frame anterior insula disgust responses as interoceptive alarms triggered by costly prediction-error spikes (ε), which force the dorsal ACC (dACC) to inhibit default comprehension signals. Third, we expose a major methodological blind spot in forced-choice fMRI experiments. Across decision deadlines (τd), the single "undecided" category conflates two entirely different cognitive realities: low-energy apathetic disengagement (Uapathy) and high-energy active epistemic neutrality (Uactive), a frontoparietal control policy governed by a calibrated prefrontal metabolic depletion threshold (ζcrit ≈ 0.45). Finally, we propose a falsifiable 7-Tesla hybrid multi-echo fMRI and dual-voxel event-related proton functional magnetic resonance spectroscopy (1H-fMRS) protocol. By measuring trial-averaged regional lactate (ΔLac) and glutamate (ΔGlu) turnover over extended evaluation windows (τd = 30 s), this experiment can directly test the thermodynamic reality of human belief.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-21
DOI
https://doi.org/10.5281/zenodo.24425421
Primary Topic
Embodied and Extended Cognition
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

• Thermodynamic Unification of Propositional Belief Evaluation: Subsuming Content-Independent fMRI Signatures Under Metabolic Active Inference

Benjamin Ruby
Zenodo (CERN European Organization for Nuclear Research)
Embodied and Extended Cognition
preprint

• Thermodynamic Unification of Propositional Belief Evaluation: Subsuming Content-Independent fMRI Signatures Under Metabolic Active Inference

Benjamin Ruby
preprint en

Abstract

Neuroimaging studies of propositional evaluation reveal that the human brain relies on a remarkably consistent, content-independent network to assess truth, falsehood, and uncertainty. Anchored in the ventromedial prefrontal cortex (vmPFC), the anterior insula, and the anterior cingulate cortex (ACC), this circuitry operates identically whether we are evaluating mathematical equations, geographical facts, ethical claims, or religious doctrines (Harris et al., 2008, 2009). Yet classical neuroimaging taxonomies treat belief, disbelief, and uncertainty as static, cost-free labels. They leave an essential question unanswered: why do specific mental states impose distinct subjective effort, reaction-time delays, and metabolic costs? Here, we resolve this problem by unifying content-independent neuroimaging findings within the Thermodynamic State-Space Model (TSSM) of active inference. First, we derive explicit biophysical couplings that link variational free energy (F) and Kullback-Leibler divergence (DKL) directly to Landauer work bounds, adenosine triphosphate (ATP) hydrolysis flux (J̇ATP), and regional cerebral metabolic rate of oxygen consumption (CMRO2). Second, we explain vmPFC reward signaling as neurochemical reinforcement for the rapid collapse of free energy into familiar, prior-congruent basins. In contrast, we frame anterior insula disgust responses as interoceptive alarms triggered by costly prediction-error spikes (ε), which force the dorsal ACC (dACC) to inhibit default comprehension signals. Third, we expose a major methodological blind spot in forced-choice fMRI experiments. Across decision deadlines (τd), the single "undecided" category conflates two entirely different cognitive realities: low-energy apathetic disengagement (Uapathy) and high-energy active epistemic neutrality (Uactive), a frontoparietal control policy governed by a calibrated prefrontal metabolic depletion threshold (ζcrit ≈ 0.45). Finally, we propose a falsifiable 7-Tesla hybrid multi-echo fMRI and dual-voxel event-related proton functional magnetic resonance spectroscopy (1H-fMRS) protocol. By measuring trial-averaged regional lactate (ΔLac) and glutamate (ΔGlu) turnover over extended evaluation windows (τd = 30 s), this experiment can directly test the thermodynamic reality of human belief.

Zenodo (CERN European Organization for Nuclear Research)
Embodied and Extended Cognition
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon