In analysis apparatuses such as ECG apparatuses (4), signal-recording electrodes (6A-6J) could be erroneously attached to a patient. The present invention describes a new method for discovering erroneously attached electrodes (6A-6J). At least one artificial neural network (18A-18C) is taught by being fed measurement signals from both correctly recorded measurements and from erroneously recorded measurements. The artificial neural network (18A-18C) is then able to identify erroneous attachments with great accuracy from recorded measurement signals. <IMAGE>