The Linear Ballistic Accumulator (LBA) model, introduced by Brown and Heathcote (2008), is a prominent framework in cognitive psychology for modelling decision-making and response times. Unlike the Drift Diffusion Model (DDM), which relies on a single accumulator tracking relative evidence, the LBA assumes independent accumulators for each response option that race toward a common decision threshold.
Crucially, the LBA is "ballistic", meaning that once evidence accumulation begins, it proceeds at a constant linear rate without within-trial noise. The variability in decision times and accuracy arises entirely from between-trial variability in the starting point of accumulation and the drift rate.
The LBA is defined by the following key parameters:
The LBA offers several advantages over other accumulation models. Its primary strength lies in its mathematical tractability. Because there is no within-trial noise, the LBA has closed-form analytic solutions for both response times and accuracy, making it extremely fast to fit to empirical data.
Additionally, its multi-accumulator architecture makes it naturally suited for tasks with more than two response options, whereas the standard DDM is strictly limited to binary choices.