Using a dynamic headspace system with Tenax trap, GC-MS, GC-olfactometry (GC-O), and multivariate analysis, the aroma chemistry of six distinctly different rice flavor types (basmati, jasmine, two Korean japonica cultivars, black rice, and a nonaromatic rice) was analyzed. A total of 36 odorants from cooked samples were characterized by trained assessors. Twenty-five odorants had an intermediate or greater intensity (odor intensity >or= 3) and were considered to be major odor-active compounds. Their odor thresholds in air were determined using GC-O. 2-Acetyl-1-pyrroline (2-AP) had the lowest odor threshold (0.02 ng/L) followed by 11 aldehydes (ranging from 0.09 to 3.1 ng/L), guaiacol (1.5 ng/L), and 1-octen-3-ol (2.7 ng/L). On the basis of odor thresholds and odor activity values (OAVs), the importance of each major odor-active compound was assessed. OAVs for 2-AP, hexanal, ( E)-2-nonenal, octanal, heptanal, and nonanal comprised >97% of the relative proportion of OAVs from each rice flavor type, even though the relative proportion varied among samples. Thirteen odor-active compounds [2-AP, hexanal, ( E)-2-nonenal, octanal, heptanal, nonanal, 1-octen-3-ol, ( E)-2-octenal, ( E, E)-2,4-nonadienal, 2-heptanone, ( E, E)-2,4-decadienal, decanal, and guaiacol] among the six flavor types were the primary compounds explaining the differences in aroma. Multivariate analysis demonstrated that the individual rice flavor types could be separated and characterized using these compounds, which may be of potential use in rice-breeding programs focusing on flavor.