Abstract: Objective To find biological markers to predict the mathematical cognitive ability in order to set patients free from the pain and time-consuming behavioral tests. Methods 86 patients with stroke or brain traumatic injuries were recruited and acquired T1 and resting-state functional MRI imaging data. And a mathematical task (7 calculation items, 2 counting items) and a word-reading task (140 items) was also finished. The partial correlative analysis was made between the score of mathematical task and the amplitude of low frequency fluctuation of each voxel of the whole brain with the word-reading performance as controlling task, and AlphaSim correction method was used with corrected P<0.05 (single voxel level: P<0.05; cluster size: >110 voxels). Results There were 5 cerebral regions whose amplitude of low frequency fluctuation significantly correlated with mathematical performance: left inferior parietal lobule (161 voxels, rpeak=0.34), left precuneus/superior parietal lobule (141 voxels, rpeak=0.31), left middle temporal gyrus (359 voxels, rpeak=0.34), left middle frontal gyrus (491 voxels, rpeak=0.36), and right middle frontal gyrus (156 voxels, rpeak=0.32). Conclusion The amplitude of low frequency fluctuation of left inferior parietal lobule, precuneus/superior parietal lobule, middle temporal gyrus, middle frontal gyrus, and right middle frontal gyrus could be used as predictors of mathematical cognitive ability for brain-damaged patients.

Key words: brain damage, resting-state, amplitude of low frequency fluctuation, mathematical cognition