Since gray and white matter have a blood uptake ratio of about 4:1, more pronounced activity may be found in the cortical rim and deeper gray structures such as the thalamus, hippocampus, and basal ganglia. The voxel value ratios in regions corresponding to gray and white matter in acquired images are about 2:1 in practice because of partial volume effects due to finite imaging resolution. With the assessment of the nature of the information the resultant SPECT data gives about cerebral perfusion, anatomical knowledge of the distribution of the cerebral perfusion allows for some mental assessment of the location of abnormal metabolism. Again, the gross distribution of cerebral perfusion may be investigated, but detailed information about the orientation and location of structures with respect to anatomy is still absent. The location of a point on a graph may be used in an analogy. Unless the axes are properly labelled and tick marks are visible, all that may be said is that the point is in the mid-left or right, or at about the center of the graph, etc.. The labelled axes and tick marks represent the anatomical detail provided by MR images. With the knowledge of the information provided by SPECT images (e.g. preferential uptake in gray matter) and the approximate dimensions of different anatomical structures, the location of a point anywhere in the graph is at best approximated by using prior knowledge of the expected function to be plotted. Similarly, anatomical placement of abnormal perfusion in smaller structures, say the hippocampus, is more difficult with the limited anatomical information provided by SPECT. To extend the analogy further to the context of the evaluation of error, improper placement of the tick marks and incorrect labelling of the axes, registration error, will result in an improper determination of the point's location on the graph. If the error in the placement of the tick marks is known, then the error in the point's location may be more correctly assessed.