This work presents a new data-driven method for the identification of functionally connected regions in the rat brain, using agglomerative clustering based on the discrete wavelet transform (DWT). The proposed approach is evaluated on resting state fMRI data and no a priori assumptions about the distribution of the signals or anatomical ROIs are made. The coefficients of the DWT are used as features in the clustering algorithm, and the performance of the classifier is evaluated as the capability to produce clusters that best correlate with known anatomical regions in the sensorimotor cortex of the brain. Wavelet features that best represent salient characteristics in the spectrum of the voxel signals are found to produce best clustering results.