Fig. 5: Module-specific time predictors and their performance.

Dataset was the same as in Fig. 4 with K = 3. Three module-specific time predictors were trained, each using data from the pertinent module, and its performance pertaining to hourly time prediction was tested in all 3 modules, in order to assess their same-module and cross-module time predictability. The lower dashed line in each panel represents the chance level. Data are shown as mean ± SEM (n = 5000 trials).