Fig. 7: ROI-based correlation network analysis.

A Colored-coded Pearson correlation network through line projections in 3D mouse brain contours, which is based on the ROIs defined by Pupillary-light-responses (PLR)-based fMRI analysis (pontine reticular nuclei (PRN), median raphe nucleus (MnR), superior colliculus (SC), visual cortex (VC), retrosplenial cortex (RSP), hippocampus (CA1), anterior cingulate cortex (ACA), and lateral septal area (LS)). The color-coded line projections present the paired ROIs with statistically stronger based on one-sample t-test of wild-type (WT) and Alzheimer’s disease (AD) mice (WT: n = 9 mice with 21 sessions, AD: n = 9 mice with 21 sessions, two-sided t-test p < 0.01), and with significantly different correlation coefficients between WT and AD mice (WT-AD, two-sample two-sided t-test, p < 0.01), B The correlation matrices of WT and AD mice, as well as the differential matrix (WT-AD), presenting an alternative way to show the correlation networks based on the PLR-fMRI specific ROIs. C The permutation test histogram shows the distribution of number of significantly different connections between WT and AD mice across 5000 permutation tests with randomly selected ROIs, which was fitted with a Gaussian distribution profile (red line, μ = 2.25, σ = 5). Red star indicates the number of significantly different correlation based on the PLF-fMRI specific ROIS, which is ~μ + 7σ of the Gaussian distribution. D A pie chart highlights the ROIs presenting the most different correlation patterns (i.e., the top 20 paired ROIs with the largest coefficient difference at two-sided t-test p < 0.01) between WT and AD mice, of which the PRN accounts for 45%. Source data are provided as a Source Data file.