Figure 1.

(A) Mirror antisaccade task. (B) Noorani and Carpenter (2014) model for antisaccades (reprinted with permission from Noorani and Carpenter, 2014). (C) Cutsuridis et al. (2014) neural network model of the superior colliculus (SC) for antisaccades in healthy and schizophrenia subjects (reprinted with permission from Cutsuridis et al., 2014). Neurons are represented as nodes. Short-range lateral excitation and long distance lateral inhibition was assumed between all nodes in the network. The left half of the network represented the left SC, whereas the right half represented the right SC. The left SC was activated by a reactive input Ir (error prosaccade decision signal), whereas the right SC was activated by a planned input Ip (antisaccade decision signal). The strengths of the inputs were not equal.