One common concern in laminar fMRI studies is the large signal amplitude on the pial surface, which could be potentially affected and contaminated by large veins on the cortical surface, or partial volume effects from large voxel sizes. Different methods have been employed to resolve the large vein problem. Simply avoiding the first layer compartment is the easiest way [21]. Indeed, if we did consider our hypotheses by excluding the first layer, they would be confirmed by our results. Another alternative approach is restricting the laminar analysis to strongly activated clusters in each subject [18]. A novel pial vein pattern analysis by optical imaging was suggested by Chen et al. to remove voxels associated with large veins, and the vein-free fMRI exhibited clear laminar specificity [13]. However, around 40% of activated voxels in the primary visual cortex was excluded for this study. In addition, optical imaging technique used by Chen et al. was invasive, and hence is not suitable for human studies. In this study, we approached this issue in terms of HRF differences across layers. We reasoned that any
