Extending Evidence Accumulation Models to Bounded Continuous Self-report Data

2026-04-28

AI Summary

This study addresses the challenge that traditional evidence accumulation models face in handling bounded, non-circular continuous self-report data commonly encountered in psychology. To bridge this gap, the authors propose two novel models—the Half-Circle Diffusion Model (HCDM) and the Beta Drift Diffusion Model (BDDM)—which extend the evidence accumulation framework to such data for the first time. Both models jointly account for the distribution of choices and response times while offering parameters with clear cognitive interpretations. Efficient parameter estimation and model comparison are achieved through likelihood-free amortized Bayesian inference (ABI) and amortized Bayesian model comparison (ABMC). Empirical evaluations demonstrate that both models accurately fit the data, with BDDM exhibiting superior performance under conditions of extreme response accuracy. The accompanying code and example datasets are publicly available.

Abstract

Evidence accumulation models (EAMs) provide a powerful framework for inferring latent cognitive processes from choice and reaction time data. While EAMs are traditionally limited to binary choices, recent developments have extended them to rotationally symmetric continuous responses via the circular diffusion model \citep{smith2016diffusion} and the spatially continuous diffusion model \citep{ratcliff2018decision}. Yet, such extensions are limited in scope, as many psychological constructs are measured on bounded non-rotational scales. In this paper, we bridge this gap by presenting and comparing two adaptations designed for bounded continuous data: the Half-Circular Diffusion Model (HCDM) and the Beta Drift Diffusion Model (BDDM). Using an empirical dataset, we demonstrate a complete methodological workflow for parameter estimation and model selection using Amortized Bayesian Inference (ABI) and Amortized Bayesian Model Comparison (ABMC). These likelihood-free methods bypass the need for analytical likelihood functions, making these complex models accessible for practical data analysis. Our results indicate that both models accurately capture the joint distribution of choices and their reaction times and yield interpretable parameters that can be reliably recovere. The BDDM provides a superior fit in scenarios characterized by very high or very low response precision. To facilitate the adoption of these methods in experimental psychology, we provide fully documented code and example datasets. This work extends the EAM framework to a new application context, the bounded continuous self-report data, offering researchers a user-friendly toolkit for modeling the cognitive dynamics of continuous responses.

Problem

• Evidence Accumulation Models
• Bounded Continuous Data
• Self-report
• Diffusion Models
• Cognitive Modeling

Innovation

• Evidence Accumulation Models
• Bounded Continuous Responses
• Beta Drift Diffusion Model
• Amortized Bayesian Inference
• Likelihood-free Inference