A Technique to Localize Activation in the Human Brain with Technetium-99m-HMPAO SPECT: A Validation Study Using Visual Stimulation

This study extends and validates a system for localizing brain activity changes based on fiducial markers, coregistration of SPECT and MRI structural images and atlas/MRI-assisted localization. Ten normal subjects underwent 99mTc-HMPAO SPECT during a resting eyes-closed baseline measurement and during visual stimulation (8-Hz reversing checkerboard). SPECT scans were registered with MRI scans obtained from each individual using a fiducial-based system that minimized z-axis and rotational errors, and registration was further refined along the x- and y-axes by superimposing corresponding axial SPECT and MRI slices. Regions of interest (ROIs) were located on MRI slices with the aid of an atlas. Corresponding loci on SPECT slices were chosen and incrementally adjusted such that the center of a ROI was located precisely at the maximum of activity in the visual cortex or the cortical gray matter ribbon. Activity in the calcarine cortex increased by 44.39% during visual stimulation (p < 0.001). Adjustment of ROI location in accordance with local activity maxima yielded superior results to a method relying strictly on atlas/MRI localization. Premotor cortex activity declined by 16.91% on the right (p < 0.01) and 13.85% on the left (p > 0.05), whereas no changes occurred in the somatosensory cortex. Changes in visual cortical activity were most comparable to previous functional MRI studies but also congruent with PET and SPECT findings. Using the locus of peak activity to aid in defining cortical ROIs improves the signal-to-noise ratio by reducing noise related to inevitable minor registration errors.