Fig. 4.
Within-subject variability in network amplitude. Figure A on the top shows amplitude changes across four runs, revealing that BOLD amplitudes in the primary sensory/motor networks were generally higher in the second scan compared with the first scan on both days. These finding cannot be explained by different phase encode directions for the runs (see supplementary results for more details). Figure B on the bottom shows amplitude changes within runs (across four 4-min blocks). The primary sensory/motor networks show a general increase in amplitude from the start of the run to the end, whereas a slight decrease in amplitude was observed in the DMN (IC 2; two-tailed t-tests showed that within-subject slopes were significant different from zero for each of the 4 runs: t1(818) = 10.0, p = 2.1*10−22; t2(818) = −7.1, p = 2.0*10−12; t3(818) = −8.8, p = 9.3*10−18; t4(818) = −3.6, p = 0.0003). The figure insets in the top right corner of A and B show changes in relative subject head motion between runs and within runs respectively. Error bars represent standard error of mean.