In the diagnosis of epilepsy long-term monitoring of electroencephalograms (EEG) data may be required to document and study interictal activities which appear as spikes in EEG signals. However visual inspection of EEG done by an expert neurologist is much time consuming. Here an automatic EEG spike detection method that uses morphological filter is described. The goal is to construct a database with data such as parameters of the detected spikes, the amount of different waves in EEG signal. The following analysis of the data is planned to be performed using methods of data mining to find the correlation of spiky signal areas with brain areas, the influence of amount of different waves and number of spikes in the signal on the nature of epilepsy. The grid environment is used for calculations.