Fig. 3.
The relationship between self-reported physical exercise in leisure (PEL) and gray matter volume (GMV), gyrification, and cortical folding. All clusters were significant at (TFCE) pFWE < 0.05. (A) PEL scores were significantly associated with GMV in the left frontal pole (red cluster). (B) GMV parameter estimates were extracted from voxels within the left frontal pole (IGE < Healthy) and visualized using box-and-whisker plots. Group means are represented by vertical boxes, with standard deviation indicated by whiskers. The scatterplot shows data from the left frontal pole cluster where the interaction of GMV and PEL was statistically significant. The line of best fit is described by the following regression equation: GMV parameter estimate = 0.008*PEL − 0.02, p = 0.005. (C) PEL scores were significantly associated with cortical gyrification in 4 clusters: 1) superior parietal lobule, 2) superior frontal gyrus, 3) precuneus, and 4) a large cluster consisting of the precentral (40 %), paracentral (33 %), and superior frontal (14 %) gyri and part of the precuneus (12 %). (D) Vertex-wise gyrification indices were extracted and plotted for the superior parietal lobule (IGE < Healthy). The magnitude and direction of group differences are visualized using box-and-whisker plots. The scatterplot shows gyrification indices within the superior parietal lobule where the interaction of gyrification and PEL was statistically significant. The line of best fit for this interaction in the superior parietal lobule is best described by the following regression equation: Gyrification index = 0.2664*PEL + 22.82, p = 0.045. (E) PEL scores were significantly associated with fractional dimensionality (FDf) in a cluster consisting of the left caudal anterior cingulate (42 %), left superior frontal gyrus (40 %) and left posterior cingulate (18 %). (F) Vertex-wise FDf values were extracted from the left caudal anterior cingulate (IGE < Healthy) and visualized using box-and-whisker plots. The scatterplot shows data from this region, where the interaction of FDf and PEL was significant. The line of best fit is described by the following regression equation: FDf in the left caudal anterior cingulate = 0.0647*PEL + 2.30, p = 0.03.