An Evidence Accumulation Model Analysis of Prior Probability and Spatial Attention Biases

Abstract Both the locus of our attention and our prior knowledge bias the decisions we make. These biases can be quantitatively investigated using formal models of decision-making such as evidence accumulation models (EAMs). EAMs describe how evidence towards the possible choices is accumulated until a decision is made. Previous studies investigating the biasing effect of spatial attention and prior probability manipulations on EAMs’ parameters supported the idea that the rate of evidence accumulation (the so-called drift rate) is mainly influenced by spatial attention cues. Conversely, the starting point (i.e., the prior bias) was found to be primarily driven by prior probability manipulations. Despite the numerous studies which investigated spatial attention and prior probability manipulations in isolation, few studies investigated their joint influence on EAMs. To this end, we perform a model comparison analysis of hierarchical diffusion decision models (DDM) and racing diffusion models (RDM) on data from three experiments. The experiments featured spatial attention and prior probability manipulations, both in isolation, to assess their selective influence on drift rate and starting point, respectively, and in conjunction to test for possible interactions between the two. We conclude that neither spatial attention nor prior probability manipulations selectively influence the drift rate and starting point. Moreover, when used in conjunction, the two manipulations show a weak interaction on the starting point.