Related Experiment Videos
Testing the validity of conflict drift-diffusion models for use in estimating cognitive processes: A
Corey N White1, Mathieu Servant2, Gordon D Logan2
1Department of Psychology, Syracuse University, 409 Huntington Hall, Syracuse, NY, 13244, USA. cnwhite@syr.edu.
Psychonomic Bulletin & Review
|March 31, 2017
Summary
Cognitive models help understand individual differences in conflict processing. This study validates drift-diffusion models for conflict tasks, finding constraints improve parameter recovery for reliable cognitive measures.
Area of Science:
- Cognitive psychology
- Computational neuroscience
Background:
- Individual differences in conflict processing are crucial but difficult to measure using traditional behavioral data like reaction times (RTs) or error rates.
- These behavioral measures are confounded by factors such as response caution and bias, limiting their interpretability.
- Cognitive models, particularly drift-diffusion models, offer a way to decompose decision processes and provide more specific measures of cognitive function.
Purpose of the Study:
- To assess the validity of recently developed drift-diffusion models for conflict tasks in capturing individual differences.
- To evaluate the performance of these models under conditions of limited data collection through a parameter-recovery study.
- To provide practical recommendations for using these models in analyzing conflict processing data.
Main Methods:
- A parameter-recovery study was conducted to test the validity of three drift-diffusion models for conflict tasks: the dual-stage two-phase model, the shrinking spotlight model, and the diffusion model for conflict tasks.
- Model parameter validity was assessed using various data-fitting methods and differing numbers of experimental trials.
- The study systematically evaluated how well the models could recover their own parameters from simulated behavioral data.
Main Results:
- Each tested model demonstrated limitations in accurately recovering valid parameters, especially with limited data.
- The study found that imposing constraints on the models could significantly mitigate these parameter recovery issues.
- The effectiveness of parameter recovery varied across the different models and fitting procedures.
Conclusions:
- Drift-diffusion models show promise for analyzing individual differences in conflict processing but require careful application.
- Constraints are essential for improving the reliability and validity of parameter estimates from these models, particularly in data-limited scenarios.
- The findings offer guidance on the appropriate use and limitations of specific cognitive models for conflict tasks.