Using Diffusion Models to Demonstrate the Difference Between Saccades and Antisaccades

The executive control network of attention [Posner, Peterson, 1990] is responsible for different tasks including inhibiting reflexive responses, resolving conflicts, and detecting errors. One way to research this function in tandem with the others in through the attentional networks test [the ANT; Fan и др., 2002]. This experimental paradigm allows looking into the interaction between executive control network and other networks, such as alerting and orienting. To investigate how the executive control network of covert attention is connected with eye movements, we further developed one version of ANT [Callejas, Lupiáñez, Tudela, 2004] by switching the response modality from manual to saccadic and using antisaccades as the executive control component. Further, a drift-diffusion model [Ratcliff, 1978; Ratcliff, McKoon, 2008] was fit to the baseline condition data (unalerted neutrally cued trials) to determine the parameters that differentiate reaction times in saccades versus antisaccades. Antisaccades were found to require more time needed for preprocessing. The nondecision component for antisaccades was bigger than for prosaccades, while the bias component was smaller which means that participants tended to avoid trying to guess a correct response in antisaccadic trials more often than in saccadic trials.