The BOLD-contrast fMRI is widely used to study brain function in the past decade. It has been suggested that new techniques such as FAIR (flow-alternating inversion recovery) and VASO (vascular space occupancy) have better localization on cortical gray matter. Previous evidence compared the activation areas based on T1-weighted image. However, such anatomical images are not suffered from susceptibility artifacts as those EPI-based fMRI techniques. Coregistration between these images may lead to bias during comparison. In the present study, we used EPI-based T1 mapping to identify whether the voxels in the activation map are gray or white matter. In order to determine the typical ranges of T1 values within different brain tissue, a linear regression has been calculated between the brightness of T1-weighted image and T1 values obtained from less distorted areas in EPI images. The human brain activations in response to 8Hz checkerboard visual stimulation were measured by three fMRI techniques. These experiments were conducted on a Bruker MedSpec 3T system with a quadrature birdcage head coil. Three fMRI experiments using BOLD sequence (TR/TE= 3000/30 ms), FAIR sequence (TR/TI/TE= 3000/1200/10.03 ms), or VASO sequence (TR/TI/TE= 3000/1200/6.72 ms) were performed on human subject. In addition to T1-weighted image, an EPI-based sequence with 25 different TI was carried out for T1 mapping. The typical range of T1 values are about 1200~1500 ms within gray matter and about 750~1050 ms within white matter. Voxels with intermediate T1 values between gray and white matter were estimated their proportion by linear interpolation. Functional images at each pixel were processed using correlation coefficient (CC) analysis. Our preliminary results indicated that VASO and FAIR has better localization on gray matter (51% and 54% of activated voxels) than BOLD (35%) (CC>0.5 in FAIR and BOLD and CC<-0.3 in VASO). This finding further confirmed the cortical specificity of VASO by more precise comparison using EPI-based T1 mapping.