SPM/Correlation

See this thread on the SPM mailing list: