Abstract
The diffusion model (DM) encapsulates the dynamics of perceptual decision within a ‘diffusion field’ that is defined by a basis with sensory-evidence (SE) and time vectors. At the core of the DM, it assumes that a decision is not made until an evidence particle drifts in the diffusion field and eventually hits one of the two pre-fixed bounds defined in the SE axis. This assumption dictates when and which choice is made by referring to when and which bound will be hit by the evidence particle. What if urgency pressures the decision system to make a choice even when the evidence particle has yet hit the SE bound? Previous modeling attempts at coping with time pressure, despite differences in detail, all manipulated the coordinate of SE bounds. Here, we offer a novel solution by adopting another bound on the time axis. This ‘dual-bound’ model (DBM) posits that decisions can also be made when the evidence particle hits a time bound, which is determined on a trial-by-trial basis by a ‘perceived time interval’ – how long the system can stay in the ‘diffusion’ field. The classic single-bound model (SBM) exhibited systematic errors in predicting both the reaction time distributions and the time-varying bias in choice. Those errors were not corrected by previously proposed variants of the SBM until the time bound was introduced. The validity of the DBM was further supported by the strong across-individual correlation between observed precision of interval timing and the predicted trial-by-trial variability of the time bound.