We have developed flexible, active, multiplexed recording devices for high resolution recording over large, clinically relevant areas in the brain. While this technology has enabled a much higher-resolution view of the electrical activity of the brain, the analytical methods to process, categorize and respond to the huge volumes of seizure data produced by these devices have not yet been developed. This paper examines a series of segmentation, feature extraction, and unsupervised clustering methods for interictal and itcal spike segmentation and spike pattern clustering. We first applied advanced video analysis techniques (particularly region growing and motion analysis) for spike segmentation and feature extraction. Then we examined the effectiveness of several different clustering methods for identifying natural clusters of the spike patterns using different features. These methdos have been applied to in-vivo feline seizure recordings. Based on both the similarity with a human clustering result and on the ratio of the intracluster vs. inter-cluster correlations, we found the best results by clustering using a Dirichlet Process Mixture Model on the correlation matrix of the spikes extracted using video segmentation. Effective clustering of spike patterns and subsequent analysis of the temporal variation of the spike pattern is an important step towards understanding how seizures initiate, progress and terminate.