Neither VBM or TBM provides a means to distinguish “true” gray matter loss (i.e. dendritic pruning, or volume reduction in tissue that was present before) versus white matter encroachment. In both methods, the quantity of each tissue type is inferred based on the intensity of the MRI signal. In VBM, this is done explicitly, because the MRI intensity histogram is compiled, and the mean intensities of the respective tissues are estimated from the data. Based on that, and using statistical information on the likely location of different tissues in a standard space, the most likely classification of each voxel is made (i.e. to gray or white matter or some other class, e.g. CSF). The labeling of gray matter depends ultimately on the MRI signal intensity, and in general a white matter encroachment in the cortex will increase the signal intensity of the cortex on T1-weighted MRI, making it more likely to be classified as white matter. If so, the cortical changes may still reflect dendritic pruning or increased myelination. In TBM, although the gray matter is not explicitly segmented, the
