This article describes an experiment using an adaptive neuro-fuzzy inference system (ANFIS) to classify epileptiform events in electroencephalographic (EEG) signals. The experiment uses Wavelet Transform (WT) to extract features from the signal. The extracted features are statistically calculated from the resulting wavelet coefficients and these data are used as inputs in a grid and clustered ANFIS. The grid ANFIS is composed by twenty inputs with three memberships each. The clustered ANFIS is composed by twenty inputs with 27 memberships grouped in 27 rules. The results are presented. This work also describes concepts regarding epilepsy and their behavior in EEG signals and does review previous works.