Integrating the Behavioral and Neural Dynamics of Response Selection in a Dual-task Paradigm: A Dynamic Neural Field Model of Dux et al. ()

Buss, Aaron T.; Wifall, Tim; Hazeltine, Eliot; Spencer, John P. · 2013 · Journal of Cognitive Neuroscience

DOI: 10.1162/jocn_a_00496

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the neural mechanisms underlying the reduction of dual-task costs with practice, specifically investigating findings from Dux et al. (2009). Dux et al. demonstrated that while dual-task performance is initially slower than single-task performance, practice eliminates these costs. Crucially, fMRI data showed that the inferior frontal junction (IFJ) exhibited stronger hemodynamic responses during dual-task trials early in learning, which diminished to single-task levels as efficiency improved. Dux et al. concluded that this reflects increased processing efficiency in the IFJ. The current study aims to determine the specific neural mechanism driving this efficiency and whether it is isolated to cognitive control areas like the IFJ or distributed across sensory-motor systems. To answer these questions, the authors developed a Dynamic Neural Field (DNF) model of response selection. The model architecture consists of 2-D neural fields that bind stimulus features (visual or auditory) to response dimensions (manual or vocal). A cognitive control system, comprising dimensional attention units and "condition of satisfaction" (CoS) units, manages task selection and switching. Attention units boost activation in relevant response fields, while CoS units release attention once a response peak is formed. The model incorporates a Hebbian learning mechanism that accumulates memory traces during peak formation, thereby lowering activation thresholds for practiced stimulus-response mappings. The authors simulated the Dux et al. experimental paradigm, including eight sessions of mixed single- and dual-task trials, and used a linking hypothesis to generate simulated hemodynamic responses from the model's neural activity. The model successfully reproduced the behavioral data, capturing initial dual-task costs, their reduction over practice, and differential learning rates between task modalities. Dual-task costs emerged from mutual inhibition between competing attention units, which slowed peak formation. Hebbian learning compressed this competition in time, reducing reaction times. The model also quantitatively captured the neural data: simulated IFJ activation was higher during dual-task trials early in learning and decreased to single-task levels by the final session. To test the locus of efficiency, the authors restricted Hebbian learning to specific model components. They found that restricting learning to only the cognitive control system or only the sensory-motor fields failed to reproduce the observed reductions in dual-task costs. Only the unrestricted model, where learning occurred across both systems, matched the empirical data. The study concludes that the efficiency gains associated with dual-task practice are not isolated to the IFJ or other cognitive control regions. Instead, efficiency is distributed across both cognitive control and sensory-motor processing systems. The findings suggest that practice tunes task-related processes throughout the neural network, requiring less activation and reducing interference, rather than shifting processing to alternative pathways or segregating neural assemblies. This provides a computationally explicit account of how neural dynamics in the IFJ and broader sensory-motor systems interact to support efficient multitasking.

Key finding

A dynamic neural field model with distributed Hebbian learning successfully captures both the behavioral reduction in dual-task costs and the associated decrease in inferior frontal junction activation observed during practice.

Methodology

simulation_modeling

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via author_sweep_intake on 2026-05-28.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• dual task performance