2.6 |. Lesion network mapping analysis

We first evaluated the functional connectivity network derived from the lesion location, which was represented as a three-dimensional lesion volume, or mask. This analysis relied on a recently developed method that infers the lesion-associated networks using normative resting-state functional MRI (rs-fMRI) data (Lesion network mapping analysis; Boes et al., 2015; Fischer et al., 2016). A publicly available rs-fMRI data set was used that included 98 healthy right-handed subjects (age 22 ± 3.2 years, 50 females) (Buckner, Roffman, & Smoller, 2014), as used previously for similar analyses (Boes et al., 2015; Fischer et al., 2016). Rs-fMRI data were processed in accordance with the strategy of Fox et al., (2005) as implemented in Van Dijk et al., (2009) and described in detail elsewhere (Fox, Buckner, White, Greicius, & Pascual-Leone, 2012). Briefly, subjects completed two 6.2-min rs-fMRI scans during which they were asked to rest in the scanner (3T, Siemens) with their eyes open (TR = 3,000 ms,TE = 30 ms, FA = 85°, 3 × 3 × 3 mm voxels, FOV = 216,47 axial slices with interleaved acquisition and no gap). Functional data were acquired at 3 × 3 × 3 mm voxel size and spatially smoothed using a Gaussian kernel of 4 mm full-width at half-maximum. The data were spectrally filtered (0.009 Hz < f < 0.08 Hz) and several nuisance variables were removed by regression, including: (a) six movement parameters computed by rigid body translation and rotation during preprocessing, (b) mean whole brain signal, (c) mean brain signal within the lateral ventricles, and (d) the mean signal within a deep white matter ROI. Inclusion of the first temporal derivatives of these regressors within the linear model accounted for the time-shifted versions of spurious variance.

After preprocessing, we performed lesion network mapping to identify the lesion-associated networks for all participants with brain injury (n = 29). For each participant, the lesion mask was used as a seed region to calculate rsFC in the healthy participant sample with rs-fMRI data (n = 98). To compute rsFC, the mean BOLD time series for each lesion mask seed region was correlated with all other voxels in the brain for all healthy participants, yielding 98 rsFC maps of voxelwise correlation values. The 98 rsFC maps (transformed to Z-scores) were then thresholded to examine only positive correlations (Z-scores) and entered into a voxelwise one-sample t test (FLAME in FSL). Similar to a previous study (Albazron et al., 2019), the Z-score maps associated with each lesion were thresholded at a value of 8 and binarized. This lesion network mapping analysis resulted in binarized lesion-associated network maps for all brain-injured participants (n = 29).

In order to identify networks associated with reduced mind-wandering in our brain-injured participants, we performed a voxel-based lesion network mapping analysis using npm (Rorden, Karnath, & Bonilha, 2007; https://people.cas.sc.edu/rorden/mricron/stats.html). We used the voxelwise Brunner and Munzel test with binary lesion-associated network masks and continuous self-reported trait daydreaming scores from each of our brain-injured participants. In contrast to the lesion symptom mapping analysis, this test shows the lesion-associated networks that are related to reduced trait daydreaming. We also performed a power analysis to evaluate our ability to detect significant associations based on the lesion-derived network distribution of the sample. Power maps were family-wise error (FWE) corrected at a threshold of p_FWE < 0.05 (Figure S1).