ARTICLE Communicated by Francisco Pereira Comparing Classification Methods for Longitudinal fMRI Studies Tanya Schmah schmah@cs.toronto.edu Department of Computer Science, University of Toronto, Toronto, Ontario, M5S 3G4, Canada Grigori Yourganov gyourganov@rotman-baycrest.on.ca Rotman Research Institute of Baycrest Centre and Institute of Medical Science, University of Toronto, Toronto, Ontario, M6A 2E1, Canada Richard S. Zemel zemel@cs.toronto.edu Geoffrey E. Hinton hinton@cs.toronto.edu Department of Computer Science, University of Toronto, Toronto, Ontario, M5S 3G4, Canada Steven L. Small small@uchicago.edu Department of Neurology, University of Chicago, Chicago, IL 60637, U.S.A. Stephen C. Strother sstrother@rotman-baycrest.on.ca Rotman Research Institute of Baycrest Centre, Department of Medical Biophysics, and Institute of Medical Science, University of Toronto, Toronto, Ontario, M6A 2E1, Canada We compare 10 methods of classifying fMRI volumes by applying them to data from a longitudinal study of stroke recovery: adaptive Fisher’s linear and quadratic discriminant; gaussian naive Bayes; support vector machines with linear, quadratic, and radial basis function (RBF) ker- nels; logistic regression; two novel methods based on pairs of restricted Boltzmann machines (RBM); and K-nearest neighbors. All methods were tested on three binary classification tasks, and their out-of-sample classification accuracies are compared. The relative performance of the methods varies considerably across subjects and classification tasks. The best overall performers were adaptive quadratic discriminant, support Neural Computation 22, 2729–2762 (2010) C  2010 Massachusetts Institute of Technology