Electroencephalography (EEG) is a valuable tool for the clinical localization of interictal spike sources. Typically, clinicians need to manually analyze and annotate all the data, a process that is time-consuming and prone to a high rate of false negatives. Recent advancements in mathematical algorithms and deep learning offer the possibility to automate this process. This work focuses on developing an algorithm using Fast Parametric Curve Matching (FPCM) filters to assist clinicians in classifying EEG data and detecting interictal spikes. The proposed method was trained on each FPCM coefficient, achieving an average ROC AUC of 0.967 ± 0.015, PR AUC of 0.9224 ± 0.021, and accuracy of 0.935 ± 0.012. These results suggest that while the method has potential, further development and optimization of the model architecture are necessary to fully realize its capabilities.