Can Systematic Drift Rate Variability Replace Random Variability in the Diffusion Decision Model?

The full diffusion decision model (DDM) assumes that the rate of evidence accumulation varies across trials, which allows the model to account for slow errors and asymptotic accuracy. This across-trial drift rate variability, however, has been criticised for being difficult to estimate and ad hoc. To examine whether random drift rate variability corresponds to meaningful variations in the quality of decision evidence, we assessed whether the model-estimated drift rate variability parameter can be partitioned by trial-level systematic drift rate information. Using a large recognition memory dataset with electroencephalography (EEG) recordings (n = 132), we systematically linked drift rate to individual trials using exogenous experimental factors—such as word frequency and study-test lag—along with endogenous factors using EEG data. Using simulations, we first demonstrated that, when slow errors arise solely due to across-trial drift rate variability, the random variability can be well partitioned and replaced by systematic variability on the trial-level. However, for the experimental data, the inclusion of systematic variability resulted in little decrease in the random across-trial drift rate variability parameter. These findings indicate that, while the quality of decision-relevant evidence (and hence drift rate) is expected to vary across trials, other mechanisms that produce slow errors are likely present but not implemented in the full DDM.