A cognitive process modeling framework for the ABCD study stop-signal task

Weigard, Alexander; Matzke, Dóra; Tanis, C.C.; Heathcote, Andrew · 2022 · Developmental Cognitive Neuroscience

DOI: 10.1016/j.dcn.2022.101191

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 a critical methodological issue in the Adolescent Brain Cognitive Development (ABCD) Study, a large-scale longitudinal neuroimaging project. The study’s stop-signal task, designed to measure response inhibition, violates the "context independence" assumption required by standard non-parametric methods for estimating stop-signal reaction time (SSRT). Specifically, the visual stop signal masks the choice stimulus, altering the processing of the "go" process on stop trials compared to go trials. This violation threatens the validity of SSRT estimates, prompting calls for caution or task redesign. The authors propose a novel cognitive process modeling framework, the RDEX-ABCD model, to accurately estimate inhibitory ability and other mechanistic parameters despite this design flaw. The RDEX-ABCD model integrates the racing-diffusion ex-Gaussian (RDEX) framework with a specific account of visual masking effects. It models the go process as a race between evidence accumulators for matching and mismatching choices, while the stop process finishing times follow an ex-Gaussian distribution. To address the context independence violation, the model introduces two new parameters: a baseline processing speed ($v_0$) and a growth rate ($g$) for discriminative information. This structure posits that at short stop-signal delays (SSDs), where the stimulus is heavily masked, evidence accumulation relies primarily on processing speed, resulting in guess-like responses. As SSD increases, discriminative information grows linearly until it reaches asymptotic levels equivalent to unmasked go trials. The model also includes parameters for trigger failures (inattention to the stop signal) and go failures (omissions). The authors validated the model using simulation studies to assess parameter recovery and applied it to empirical ABCD data, utilizing Bayesian estimation to handle the limited number of stop trials per participant. The results demonstrate that the RDEX-ABCD model successfully accounts for key behavioral trends in the ABCD data, including the decrease in choice accuracy at short SSDs and the inhibition function. Simulation studies revealed that failing to account for context independence violations leads to erroneous inferences in realistic scenarios. In contrast, the RDEX-ABCD model provided accurate and precise estimates of SSRT and additional mechanistic parameters, such as attention to the stop signal and cognitive efficiency. The model comparison favored the linear growth specification over more complex non-linear alternatives, which showed signs of overfitting. Furthermore, the analysis confirmed that trigger failures are non-trivial in the ABCD sample, highlighting the importance of including this mechanism in the model. The significance of this work lies in providing a robust methodological solution for analyzing existing ABCD data without requiring task redesign. By offering a parsimonious explanation for the impact of visual masking, the RDEX-ABCD model allows researchers to draw valid and nuanced inferences about response inhibition and related neurocognitive mechanisms. This framework advances the field by enabling the use of large-scale developmental data for clinical and neuroscientific research, ensuring that estimates of inhibitory control are not biased by the task's design features. The availability of open code and data facilitates further application and validation of this approach in future studies.

Key finding

The RDEX-ABCD cognitive process modeling framework effectively accounts for context independence violations in the ABCD stop-signal task, enabling accurate estimation of inhibitory control parameters where standard methods fail.

Methodology

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.

• executive function