Event-related potentials (ERP) and electroencephalogram (EEG) analysis should consider the spatial and temporal dynamics of the electrical activity. Extraction of the relevant information is crucial for improving the signal to noise ratio in order to get efficient tools for clinical purposes. The Spatio-Temporal Parcellation (STEP) algorithm characterizes a single subject and a group by set of events. Event is defined as an extreamum point in the spatio-temporal amplitude space and its associated surroundings. Clustering is applied on all events of all subjects in order to get the group characteristics. Two groups of normal subjects underwent the auditory oddball task. By implementing the STEP algorithm, it was possible in both groups to successfully differentiate the evoked responses to the Novel Vs. Target stimuli with statistical significance. The STEP algorithm holds promise as a tool for diagnosis, follow-up and drug development of different neurological conditions.