Childhood Poverty and Stress Reactivity Are Associated With Aberrant Functional Connectivity In Large-Scale Brain Networks
Supplemental Information
Supplemental Methods
Participants
In this sample, 26 of the participants came from families that were at or below an income-to-needs ratio of 1, which is the U.S. federal poverty line. The income-to-needs ratio is an annually adjusted per-capita index. The other 26 participants came from families with income-to-needs ratios two to four times the poverty line, which represents the level of most American families. See Evans and Kim (2012) for further details on participant recruitment and protocols. Additional data were collected when participants reached mean ages of 13.36 years (Wave 2), and 17.47 years (Wave 3). Of 52 participants in the current study, 49 participants completed the fMRI and TSST tasks and had usable data. Average age was 23.69 (SD=1.31, range 20-27) and 55.1% (27 out of 49) were male. The average income-to-needs ratio at Wave 1 was 1.8 (SD=1.1) and 3.2 (SD=3.0) at age 24.
Trier Social Stress Test (TSST)
The TSST was developed at the University of Trier for induction of moderate psychosocial stress under laboratory conditions with special emphasis on changes in HPA activity. In numerous studies, the TSST has proven to elicit significant changes in cardiovascular parameters, endocrine parameters, and subjective stress ratings. In our adaptation, participants were told that they would assume the role of a job applicant (for a job tailored to their interests) and give a brief job talk in front of a panel, followed by a thinking task (consisting of mental arithmetic). They were escorted to a bare room in front of two panel members and a video camera. The members of the panel were presented as being particularly astute at monitoring non-verbal behavior, and participants were told that their presentation would be recorded for later analysis. Panel members were instructed to maintain neutral expressions throughout the task, and to refrain from providing positive feedback. Participants were given three minutes to prepare a speech, while standing at a table in front of the panel. They were permitted use of paper and pencil during preparation but were informed that notes would not be used during their presentation. They were then invited to use the next five minutes to talk. They were prompted to continue or were asked prepared questions if they were unable to fill the time. When the five minutes were completed, they were asked to subtract 13 from 1,022 serially, as accurately and as rapidly as possible. If an error occurred, they were asked to start over. This continued for five minutes. The TSST was performed at approximately 2:00 PM, and always at least one hour after the participant had finished eating lunch.
fMRI Resting-State Paradigm
Resting-state scans always occurred prior to tasks. Participants were positioned in the MR scanner and their heads comfortably restrained to reduce head movement. Heart rate and respiration measurements were acquired for group comparisons (via Independent-Samples Kruskal-Wallis tests). A black fixation cross on a white background was displayed in the center of the screen for 8 minutes. Participants were instructed to relax and keep their eyes open and fixed on the cross. Participants viewed the fixation cross through MR-compatible liquid crystal display goggles (NordicNeuroLabs
fMRI Image Acquisition and Preprocessing
After a T1 image (T1-overlay) was obtained, a T2*-weighted, echoplanar acquisition sequence [GRE; repetition time, 2000 ms; echo time, 25 ms; flip angle, 90°; field of view (FOV), 22 cm; 42 slice; thickness/skip, 3.0/0 mm matrix size equivalent to 64 x 64] was collected. After discarding five initial volumes to permit thermal equilibration of the MRI signal, 240 volumes were acquired over 8 minutes. After acquiring the functional volumes, a high-resolution anatomical T1-weighted image with a three dimensional gradient recalled echo was acquired with TR = 9.8 ms, TE = 459 ms, FA = 80, FOV = 256 mm, 180 slices with 288 x 288 matrix per slice, 1 mm slice. Scans were reconstructed, motion corrected, slice-time corrected using a middle slice (slice 21) as a reference, realigned to the first scan in the experiment to correct for head motion, and co-registered with the high-resolution sagittal images. After motion correction, realigned functional images were spatially normalized to a functional template and resampled to 2 x 2 x 2 mm voxels. Images were then spatially smoothed using a Gaussian filter with a full-width half-maximum value of 8mm. Motion parameters (mean displacement, mean angle, maximum displacement, maximum angle) were compared across groups via Independent-Samples Mann-Whitney U tests, and runs with any single movement greater than 3 mm or 3 degrees were excluded.
Motion Scrubbing
During motion scrubbing, individual frames with excessive head motion (0.5 mm for framewise displacement and 0.5% ΔBOLD for DVARS) were censored from the time series, following Power et al. (2012).One subject required the removal of 167 frames, and thus was excluded from the analysis. Of the 51 subjects that were retained, an average of 1.06 frames per subject was removed (SD = 2.46, range = 0-14). There were no group differences in framewise displacement (FD; p = .2) or number of frames removed per subject (p = .3). Average subject-specific motion was added as a covariate in the second-level model. This practice has been shown to provide substantial additional cleansing of motion-related effects (see Fair et al, 2012). Motion was calculated from the 6 motion parameters as follows: Translational displacements (x, y, z) were measured in millimeters, and rotational displacements (roll, pitch, and yaw) were converted from degrees to millimeters by calculating displacement on the surface of a 50 mm radius sphere, which approximates the distance from the cerebral cortex to the center of the head (see Power et al, 2012, pg. 3). Framewise FD was calculated as the sum of the absolute value of the derivatives of these 6 parameters. Mean FD was calculated from the mean of the framewise FDs.
Region of Interest (ROI) Analysis
Regions of interest were defined using the automated anatomical labeling atlas (AAL). Anterior insula was partitioned from a whole insula mask using linear interpolation based on the location of the middle insular gyrus, following Aupperle and colleagues (2011). Activation threshold and cluster size were determined using AlphaSim (Ward, 2000) to correspond to a false positive rate of p0.05, corrected for multiple comparisons within ROIs. A priori ROIs of anatomical PCC (k=797), hippocampus (k=932), bilateral frontal medial orbital gyrus (vmPFC) (k=1575), anterior insula (k=1357), amygdala (k=248), and supplementary motor area (SMA; k=4518) were used as masks. Cluster size thresholds were separately calculated for each region using AlphaSim. Images were thresholded using a voxelwise threshold of p0.01 uncorrected with a minimum cluster size (produced by AlphaSim) of 39 connected voxels for PCC, 49 voxels for hippocampus, 69 voxels for vmPFC, 63 connected voxels for anterior insula, 9 voxels for amygdala, and 148 voxels for supplementary motor area. The corrected voxel-wise probabilities were: PCC p0.004, hippocampus p0.003, vmPFC p0.006, anterior insula p0.004, amygdala p0.004, and SMA p0.002. Only the activations within the ROIs that survived the volume and voxel correction criteria were extracted and used for further analysis. Reported voxel coordinates correspond to standardized Montreal Neurologic Institute (MNI) space.
Supplemental Results
Between-Network Connectivity
Compared to the childhood middle-income group, the childhood low-income group showed reduced connectivity between dACC/SMA seed and PCC [(-2, -48, 18), k=119, Z = 3.34, p.05, SVC]. There were no other group differences in between-network connectivity.
References
Aupperle RL, Ravindran L, Tankersley D, Flagan T, Stein NR, Simmons AN, et al (2011). Pregabalin influences insula and amygdala activation during anticipation of emotional images. Neuropsychopharmacology36(7): 1466-1477.
Evans GW, Kim P (2012). Childhood poverty and young adults' allostatic load: the mediating role of childhood cumulative risk exposure. Psychol Sci23(9): 979-983.
Fair DA, Nigg JT, Iyer S, Bathula D, Mills KL, Dosenbach NU, et al (2012). Distinct neural signatures detected for ADHD subtypes after controlling for micro-movements in resting state functional connectivity MRI data. Front Syst Neurosci6: 80.
Power JD, Barnes KA, Snyder AZ, Schlaggar BL, Petersen SE (2012). Spurious but systematic correlations in functional connectivity MRI networks arise from subject motion. Neuroimage59(3): 2142-2154.
Ward B (2000). Simultaneous inference for fMRI data.