Although recent studies have elucidated cognitive and affective neural correlates of humor processing, they weren’t able to distinguish between the incongruity detection and incongruity resolution proposed by incongruity-resolution theory. In the present functional magnetic resonance imaging (fMRI) study, using a new caption-cartoon paradigm, three experimental conditions (humorous, nonhumorous, and unrelated cartoons) were designed to investigate the dissociation between detection and resolution of incongruity. The flow of stimulus presentation in each trial was as following. First, a fixation point (+) appeared on the center of the screen for 2 s, then the character context was presented for 2 s. After a variable time (2 or 4 s) of blank screen, the cartoon picture was presented. Subjects were required to make a “humorous/nonhumorous/unrelated” judgment about the relationship between the cartoon picture and the prior character context, and were instructed to press “1” key for pictures which they felt humorous with their right index finger, “2” key for pictures which they felt nonhumorous with their right middle finger and “3” key for pictures which they felt unrelated with previous caption context with their left index finger. Finally, after a variable time (2, 4, or 6 s) of blank screen, the next trial began. There were 3 runs and each had 45 trials, 15 trials for each condition. The different conditions in each block were displayed pseudo-randomly. The fMRI results showed that, comparing humorous to nonhumorous cartoons, activations were found in the left superior frontal gyrus (SFG), left inferior frontal gyrus (IFG), left middle frontal gyrus (MFG), left middle temporal gyrus (MTG), left amygdala, left and right precuneus, and right superior temporal gyrus (STG), etc. which were consistent with other fMRI studies about humor processing and corroborated the validity of the caption-cartoon paradigm. Most interestingly and importantly, comparing unrelated to nonhumorous cartoons, it showed activation in the right insula. This result suggested that the insula might be associated with incongruity detection. According to these results, the neural correlates of detection and resolution of incongruity might be separated in this study. We discussed the role of insula in context of its cognitive effect and proposed the insula and amygdala might covary contributing to incongruity detection within humorous cartoon processing. In addition, comparing nonhumorous to unrelated cartoons, the fMRI results showed the activation in left MFG and right medial frontal gyrus, right SFG, and right precuneus, etc. However the temporo-parietal junction (TPJ) area, which was activation in comparison of humorous to nonhumorous cartoons, was not activated. Because in comparison of nonhumorous to unrelated cartoons, there was no processes about incongruity detection and incongruity resolution but general resolution process of relationship between captions and cartoons, it is reasonable that the frontal areas and precuneus are associated with general resolution process and the TPJ area was involved in incongruity resolution process. Consistent with these analysis, the frontal areas, precuneus and TPJ area were not activated in comparison of unrelated to nonhumorous cartoons, because unrelated cartoons involved neither general resolution process nor incongruity resolution process.