[Paper Review] Estimating measures of information processing during cognitive tasks using functional magnetic resonance imaging

Chetan Gohil, Oliver M. Cliff+3 more|arXiv (Cornell University)|Feb 3, 2026

Functional Brain Connectivity Studies0 citations

TL;DR

The paper introduces a framework to quantify information processing in task-based fMRI using active information storage (AIS), transfer entropy (TE), and net synergy, leveraging cross mutual information with resting-state data as reference. Applied to the HCP N-back task (470 subjects), it reveals module-level reorganization and links to performance.

ABSTRACT

Cognition is increasingly framed in terms of information processing, yet most fMRI analyses focus on activation or functional connectivity rather than quantifying how information is stored and transferred. To remedy this problem, we propose a framework for estimating measures of information processing: active information storage (AIS), transfer entropy (TE), and net synergy from task-based fMRI. AIS measures information maintained within a region, TE captures directed information flow, and net synergy contrasts higher-order synergistic to redundant interactions. Crucially, to enable this framework we utilised a recently developed approach for calculating information-theoretic measures: the cross mutual information. This approach combines resting-state and task data to address the challenges of limited sample size, non-stationarity and context in task-based fMRI. We applied this framework to the working memory (N-back) task from the Human Connectome Project (470 participants). Results show that AIS increases in fronto-parietal regions with working memory load, TE reveals enhanced directed information flows across control pathways, and net synergy indicates a global shift to redundancy. This work establishes a novel methodology for quantifying information processing in task-based fMRI.

Motivation & Objective

• Motivate quantitative characterization of information storage and transfer in cognition beyond activation and functional connectivity.
• Introduce AIS, TE, and net synergy as complementary information-processing measures for fMRI signals.
• Develop and apply a cross mutual information approach to handle non-stationarity and short task recordings by using resting-state data as a reference.
• Demonstrate the method on the Human Connectome Project N-back working memory task to reveal module-level information dynamics.

Proposed method

• Compute local (instantaneous) AIS, TE, and net synergy for each ROI and edge.
• Use cross mutual information with resting-state data as the reference distribution to estimate information-theoretic measures from task data.
• Deconvolve HRF to mitigate temporal blurring before information measures estimation.
• Parcellate brain into 333 ROIs and aggregate results at the Yeo functional module level.
• Perform first-level GLMs to obtain condition-specific measures and two task contrasts (2-back vs rest, 2-back vs 0-back).
• Assess statistical significance via non-parametric permutation testing with maximum t-statistic correction.

Figure 1: Calculation of condition-specific measures . A) The time series for some local (instantaneous) measure i. Examples include the BOLD signal, local AIS (for a region), and local MI (for an edge). B) Visualisation of the task design matrix where the blue line indicates the time points corresp

Experimental results

Research questions

• RQ1 How do AIS, TE, and net synergy change with increasing working memory load (2-back vs rest and 2-back vs 0-back)?
• RQ2 Do information processing measures reorganize at functional-module scales during the N-back task?
• RQ3 Are AIS, TE, or net synergy changes related to individual 2-back accuracy?
• RQ4 Does incorporating resting-state data as a reference via cross MI improve detection of task-related information dynamics?

Key findings

• AIS increases globally with working memory load, with largest increases in visual modules.
• TE shows widespread decreases across modules but increases in VIS A, DAN A, and CON B, indicating selective inter-module information flow changes.
• Net synergy shifts toward redundancy across most modules, with VIS B, CON A, and DMN A showing localized increases in synergy.
• During 2-back vs 0-back, AIS grows in fronto-parietal regions, and TE increases across modules, suggesting distributed information transfer under working memory demands.
• Stronger AIS in frontal regions and a shift toward redundancy predict higher 2-back accuracy.
• Cross MI (using resting-state as reference) provides contextualized insights into task-induced dependencies beyond conventional MI.
• Methodological contributions include HRF deconvolution and the cross MI framework for short, non-stationary task data.

Figure 2: Conventional measures for studying task fMRI data . For the 2-back vs rest condition (left) and 2-back vs 0-back condition (right): A) Change in mean activity, averaging over subjects. B) Change in conditional MI, averaged over subjects. This was calculated for using a conventional approac