Evaluation and calibration of functional network modeling methods based on known anatomical connections

► We compare network models in detecting direct connections in resting-state fMRI data. ► Correlation methods, Patel's Kappa, and Bayesian Network methods performed best. ► Using 8.5 and 34 min of data, the top methods achieved ~ 60% and ~ 80% accuracy. ► Increasing the size of the network decreased the performance of multivariate methods. ► We propose calibration based on known anatomical connections in future studies.