Perfusion brain single photon emission computed tomography (SPECT) is currently used to evaluate patients with cognitive impairments. Although widely available, it has been reported to be significantly less sensitive than F-18 FDG positron emission tomography. Optimization of SPECT parameters using nonuniform attenuation correction (NUAC) and scatter correction (SC) might improve the accuracy of the method. This study assessed the effect of x-ray CT-based NUAC and triple-energy window (TEW-SC) on brain SPECT compared with Chang-uniform (UAC). A total of 31 patients with memory complaints underwent Tc-99m ECD SPECT/CT. CT-NUAC+TEW-SC and Chang-UAC were applied and compared. The images were spatially normalized to a default SPECT template supplied with Statistical Parametric Mapping software (SPM2). A paired t test image was then reconstructed. Regional cerebral blood flow measurements were apparently reduced in the frontal white-matter and in the frontotemporal cortex when CT-NUAC+TEW-SC were used. These findings need to be considered when interpreting Tc-99m ECD SPECT after applying CT-AC+TEW-SC. Further prospective studies with clinical correlations are needed.