Zaslat záznam emailem: Opening the black box of neural networks: methods for interpreting neural network models in clinical applications