NBS-Predict is a prediction-based extension of the network-based statistic. In NBS-Predict, we combine Network-based Statistics (Zalesky et al., 2010) and feature engineering for connectome-based prediction. Briefly, NBS-Predict allows identifying graph components with associated prediction accuracies by combining a graph theory with machine learning algorithms (e.g., support vector machine, decision tree) in a cross-validation structure. Additionally, NBS-Predict comes with a graphical user interface (GUI) that does not require any programming expertise. As such, it allows for a broad audience of researchers to potentially benefit from advanced but easy to interpret machine-learning applications to facilitate the search for biomarkers.