Reanalysis of our positron emission tomography study on the reading of Japanese Kanji (morphograms) and Kana (phonograms) with Statistical Parametric Mapping (SPM) Version 96 (Cogn. Brain Res. 9 (2000) 111) were described in comparison with our previous region of interest (ROI)-based analysis. Some areas were activated in one analysis but not in the other, which was due to the methodological differences. Furthermore, we conducted cognitive subtraction analysis between Kanji and Kana and between Kana words and nonwords and confirmed that Kana words activated the lateral occipital gyri more extensively than Kanji words.