نهمین کنگره نقشه برداری مغز

نمایش مقاله

Title:

Information Cascades and Trade-Flow Herding in Prediction Markets: A Free-Energy Interpretation

Subject:

Artificial Intelligence and Brain Mapping

Presenter Author:
Name:	Mohammad ZadehEbrahim

Author:
Name 1:	Mohammad ZadehEbrahim - Corresponding - Presenter
Affiliation 1:	University of Tehran, Faculty of Psychology and Educational Science, Department of Cognitive Science

Name 2:
Affiliation 2:

Name 3:
Affiliation 3:

Name 4:
Affiliation 4:

Name 5:
Affiliation 5:

Name 6:
Affiliation 6:

Name 7:
Affiliation 7:

Name 8:
Affiliation 8:

Name 9:
Affiliation 9:

Name 10:
Affiliation 10:

Abstract

Original Research

introduction :

The Free-Energy Principle (FEP) and its corollary, Active Inference, were formalized by Karl Friston — the same researcher behind Statistical Parametric Mapping (SPM) and Dynamic Causal Modelling (DCM), now standard tools across brain-mapping research. FEP offers a unifying, biologically grounded account of how adaptive systems, from single neurons to whole organisms, resolve uncertainty by minimizing prediction error. Recent simulations extend this framework beyond the brain: when artificial agents minimize expected free energy while exchanging beliefs over a social network, their collective behavior reproduces "herding" — individuals abandoning private evidence to follow others' observed choices. This raises a direct question for brain-mapping research: Do the same free-energy and network-topology principles known to organize neural processing also shape population-scale human behavior? We test this using Polymarket, a platform where large populations continuously trade on the outcomes of future events, offering a naturalistic, large-scale behavioral dataset.

Methods:

We extracted timestamped decision events from Polymarket via its public data interfaces. From this stream we derived two signatures analogous to features extracted from neural time series: a Trade-Flow Imbalance index (quantifying directional bias in collective choice, decomposed into local and network-wide components) and a Burstiness index (quantifying irregularity in inter-decision timing, conceptually parallel to inter-event-interval measures used in spike-train analyses). We test whether network-wide directional bias predicts the direction of the next decision (Hypothesis H1) using logistic/probit regression and permutation testing. These patterns are evaluated against explicit predictions from active-inference models, in which agents minimize expected free energy while exchanging beliefs across a network, reproducing herding-like statistical signatures from topology alone. To close the loop back to the brain, we are conducting a parallel graph-theoretic analysis of resting-state functional connectivity from the Human Connectome Project, testing whether the same topological properties (clustering, path length, hub structure) shape both simulated belief propagation and real brain network organization.

Results:

Preliminary exploration is consistent with a statistically significant, network-wide directional bias predicting subsequent decision direction, supporting H1. Burstiness is expected to scale with this bias, mirroring clustering signatures from active-inference simulations. Full statistical results will be finalized alongside preliminary Human Connectome Project graph metrics, enabling a side-by-side comparison of network signatures in large-scale human behavior and brain structural organization.

Conclusion:

This work treats a large-scale, naturally occurring behavioral dataset as a population-level complement to neuroimaging studies of social conformity — testing, at a scale no scanner could achieve, whether free-energy and network-topology principles believed to organize neural processing also organize collective human decision-making. It aims to motivate future brain-mapping designs, including fMRI/EEG paradigms built around naturalistic decision tasks, for directly probing the neural basis of herding.

Keywords:

Brain Mapping; Connectome; Bayes Theorem; Decision Making; Economics, Behavioral