Cluster analysis is based on the observation that it is more likely that activations should form clusters rather than being spatially isolated events. Therefore, a SPM, thresholded at a particular statistic level, is further analysed to identify all the unique clusters of voxels. There is, however, the possibility that even these clusters might be a chance occurence. The theory of Gaussian random fields [1] allows us to model the probability that a cluster occurs as a chance process and so, to form our final result, we evaluate each cluster accordingly rejecting those clusters whose probability of being due to chance is unacceptably high.