Separating the roles of orthographic and lexical information in the neural correlates of evidence accumulation for lexical decision-making

According to computational theories of evidence accumulation, decision making is driven by the accumulation of evidence over time. To date, most studies only explored the impact of single source of evidence on the evidence accumulation process. In contrast, our study makes use of the fact that processing of Chinese characters takes place at different levels; radical information, orthographic information and lexical information are separated. This allows us to examine how these different sources of evidence enter the accumulation process, and in what temporal order this occurs. Drift diffusion model fitting showed that the drift rate is correlated with the strength of evidence for deciding whether Chinese characters reflect words. There was a negative correlation between the drift rate and individual differences in judgments of word-likeness and familiarity. Importantly, we could observe neural correlates of the evidence accumulation process of lexical decisions. Specifically, we found that the N1 (which represents orthographic processing) and N400 (which represents semantic processing) contribute to the evidence accumulation process. These results show that first the decision is driven by orthographic information—before 250 ms post-stimulus, followed by lexical information around 400 ms. Together, this study clarifies how different sources of evidence are combined in decision making.