Table 1.
Summary studies.
| References | Study design and MRI sample | Exposure | Outcome | Main findings | Confounders adjusted/controlled | Conclusions |
|---|---|---|---|---|---|---|
| OUTDOOR AIR POLLUTION AND BRAIN IMAGING | ||||||
| Peterson et al. (34) | 40 right-handed children born to nonsmoking Dominican and African-American women 18–35 years old residing in Washington Heights, Harlem, or the South Bronx in NYC were recruited 1998–2006 through local prenatal care clinics and followed in utero to 7–9 years of age | Personal air monitors to measure 8 airborne PAHs (benz[a]anthracene, chrysene, benzo[b]fluroanthene, benzo[k]fluroanthene, benzo[a]pyrene, indeno[1,2,3-cd]pyrene, disbenz[a,h]anthracene, and benzo[g,h,i]perylene) and determine maternal exposures over 48 h period during 3rd trimester, with mean ± standard deviation (SD) of 5.13 ± 6.20 ng/m3. Spot urine samples collected from child at age 5 used to measure postnatal urinary PAH metabolites, with N = 38 and mean ± SD of 14999.4 ± 19795.8 ng/L. Source of pollutants: traffic emissions |
3T GE, with 8 channel head-coil; T1 weighted: TR = 4.7 ms, TE = 1.3 ms, acceleration factor = 2, 256 × 256 matrix, 160 slices, 1 mm thickness; ANALYZE 8.0 and in-house software; structural MRI's, surface mapping. Correction using False Discovery Rate (p < 0.05). | PAH correlated inversely with morphological measures in frontal, superior, temporal, parietal, and rostrocaudal extent of the mesial surface mostly in the Left hemisphere; underlying white matter mainly driving effect. Pearson correlation coefficients and 95% confidence intervals: Prenatal PAH levels: Left lateral superior prefrontal cortex: −0.57 (−0.75, −0.304); Left superior temporal gyrus: −0.51 (−0.71, −0.23); Left precuneus: −0.50 (−0.71, −0.22); Left medial prefrontal cortex: −0.56 (−0.75, −0.30) Postnatal (age 5) PAH levels: Left dorsal-lateral prefrontal cortex: −0.47 (−0.69, −0.18); Left medial dorsal prefrontal cortex: −0.52 (−0.72, −0.25) |
Child's age and sex; prenatal cotinine levels, measures of postnatal PAH exposure at age 5 years, handedness | Prenatal PAH relates to less left hemisphere white matter; while postnatal PAH contributes to less white matter in dorsal prefrontal regions |
| Pujol et al. (35) | 263 children (mean age 9.7 ± 0.9 years old; range: 8–12.1 years) from the BREATHE project who were selected from 39 schools in Barcelona (schools selected based on modeled NO2 values) and had been in the school for at least 18 months | Pollutant levels in school courtyards were sampled twice during 1-week periods (8 h/day) separated by 6 months in the warm (year 2012) and cold (year 2012/2013) periods. Elemental carbon (measured in PM2.5) was measured with High-Volume samplers and NO2 with passive dosimeters (96 h period), with mean ± SD of 0.92 ± 0.30 μg/m3 and range of 0.42–1.92 μg/m3. A single traffic-related pollutant indicator was computed using the weighted average of these two, using the following formula: | 1.5T GE with 8 channel head-coil; T1 weighted: TR = 11.9 ms, TE = 4.2 ms, Flip = 15°, 256 × 256 matrix, 134 slices, 1.2 mm thickness; SPM (version 8) VBM and FREESURFER (version not reported). DTI: 25 directions with b-value = 1,000 s/mm2, TR = 8,300 ms, TE = 94 ms, Flip = 90°, 128 × 128 matrix, 26 slices, 5 mm thickness with no gap, spin-echo single-shot EPI; FMRIB FDT, accounted for head motion. |
No significant association was identified between air pollution and any anatomical, DTI or metabolic brain measurement. Traffic-related air pollution was significantly associated with: (1) weaker functional connectivity between regions belonging to the DMN (i.e., between the medial frontal cortex and the angular gyrus bilaterally); (2) stronger functional connectivity between the medial frontal cortex seed region and the frontal operculum at the lateral boundary of the DMN; (3) lower deactivations (rest > task map) during passive viewing and listening in the supplementary motor area and somatosensory cortex. | Age, sex, academic achievement, difficulties scores, obesity, parental education, home and school vulnerability index, distance from home to school and public/non-public school category | Traffic-related air pollution was associated with functional brain changes, but there was no relationship with brain anatomy, white matter microstructure, or membrane metabolism |
| [(EC/group median) + (NO2/group median)/2] Source of pollutants: traffic emissions | RS-FMRI (eyes closed): TR = 2,000 ms, TE = 50 ms, Flip = 90°, 64 × 64 matrix, 22 slices, 4 mm with 1.5 mm gap, 180 EPI volume; SPM8 preprocessing with 8 mm Gaussian filter; high-pass filter (~0.008 Hz), white matter, CSF, and global signal removal; scrubbed for motion; Functional connectivity maps using 3.5 mm spheres of frontal lobe (n = 4) and caudate (n = 2) seeds; Correction with Monte Carlo and a family-wise error rate (p < 0.05) and Bonferroni per connectivity map (n = 4). Sensory FMRI task: TR = 2,000 ms, TE = 50 ms, Flip = 90°, 64 × 64 matrix, 22 slices, 4 mm with 1.5 mm gap, 120 EPI volume; block design (ABABABAB) of 30 s fixation and 30 s of visual-auditory input; Correction with Monte Carlo and a family-wise error rate (p < 0.05). Proton (1H) Spectroscopy: PROBE-SV and STEAM sequence, TR = 2000 ms, TE = 30 ms, 128 signal averages, voxel = 23 × 14 × 14 mm in the left frontal white matter; NAA FWHM 0.09 ppm and creatine RMS >8. |
T-values (non-adjusted vs. adjusted by age and sex): RS-FMRI Medial frontal seed map: Left Lateral frontal cortex (+ correlation): 3.6 vs. 4.1; Left Parietal cortex (– correlation): 3.5 vs. 3.2; Right Parietal cortex (– correlation): 3.5 vs. 3.2 RS-FMRI Dorsal frontal seed map: Left Parietal cortex (+ correlation): 3.6 vs. 3.5; Right Lateral frontal cortex (– correlation): 3.7 vs. 3.5; Right Insula (– correlation): 4.3 vs. 4.3 RS-FMRI Posterior cingulate seed map: Left Lateral frontal cortex (+ correlation): 3.3 vs. 3.2 RS-FMRI Supplementary motor area seed map: Left Prefrontal cortex (+ correlation): 4.4 vs. 4.6; Right Prefrontal cortex (+ correlation): 3.5 vs. 3.6; Left Parietal cortex (+ correlation): 3.3 vs. 3.4; Right Parietal cortex (+ correlation): 4.1 vs. 3.9; Left Anterior cingulate cortex (– correlation): 3.6 vs. 3.7 Sensory FMRI task: Right Somatosensory cortex (+ correlation): 3.7 vs. 3.8; Left Premotor cortex (+ correlation): 3.7 vs. 3.6 |
||||
| Pujol et al. (36) Brain and Behavior | 263 children (mean age 9.7 ± 0.9 years old; range: 8–12.1 years) from the BREATHE project who were selected from 39 schools in Barcelona (schools selected based on modeled NO2 values) and had been in the school for at least 18 months | Copper (measured in PM2.5) levels in school courtyards were sampled with High-Volume samplers twice during 1-week periods (8 h/day) separated by 6 months in the warm (year 2012) and cold (year 2012/2013) periods and analyzed using ICP-MS, and then averaged to obtain the yearly exposure levels, with mean ± SD of 8.7 ± 3.0 ng/m3 and range of 3.7–13.8 ng/m3. Source of pollutants: traffic emissions, industrial sources, and railway |
1.5T GE with 8 channel head-coil; T1 weighted: TR = 11.9 ms, TE = 4.2 ms, Flip = 15°, 256 × 256 matrix, 134 slices, 1.2 mm thickness; SPM (version 8) VBM (5 FWHM) and FREESURFER (version not reported). DTI: 25 directions with b-value = 1,000 s/mm2, TR = 8,300 ms, TE = 94 ms, Flip = 90°, 128 × 128 matrix, 26 slices, 5 mm thickness with no gap, spin-echo single-shot EPI; FMRIB FDT, accounted for head motion. RS-FMRI (eyes closed): TR = 2,000 ms, TE = 50 ms, Flip = 90°, 64 × 64 matrix, 22 slices, 4 mm with 1.5 mm gap, 180 EPI volume; SPM8 preprocessing with 8 mm Gaussian filter; high-pass filter (~0.008 Hz), white matter, CSF, and global signal removal; scrubbed for motion; Functional connectivity maps using 3.5 mm spheres of frontal lobe (n = 4) and caudate (n = 2) seeds; Correction with Monte Carlo and a family-wise error rate (p < 0.05). |
Higher copper levels were associated with: (1) higher gray matter concentration in the striatum, specifically in the caudate nucleus, with no effect on tissue volume; (2) higher FA in white matter close to the caudate nucleus and in the caudate nucleus; (3) changes in the complex architecture of neural tissue diffusion; (4) reciprocal reduction of functional connectivity between the caudate nucleus and the frontal lobe operculum (bilaterally). Beta coefficients and 95% confidence intervals: Gray matter density: Left caudate nucleus: β = 0.3 (0.1, 0.5) DTI Fractional Anisotrophy (FA): Left caudate nucleus: β = 0.1 (0.05, 0.2) RS-FMRI: Left frontal cortex to Left caudate: β = −0.1 (−0.2, −0.1) |
Age, sex, academic achievement, academic difficulties score, obesity, parental education, home and school vulnerability index, public/nonpublic school category, socioeconomic status, elemental carbon, other toxic agents (Pb, Mn, Sb, and Fe) | Association between environmental copper exposure in children and alterations of caudate structure and function |
| Mortamais et al. (37) | 242 children (median age 9.7 years; range: 8–12 years) from the BREATHE project who were selected from 35 of 39 schools in Barcelona (schools selected based on modeled NO2 values) and on average at school for 6.5 years before study; removed 19 participants b/c parents said smoked at home | Pollutant levels in school courtyards were sampled twice during 1-week periods (8 h/day) separated by 6 months: Jan to June 2012 and Sept 2012 to Feb 2013. PAHs (measured in PM2.5) was measured with High-Volume samplers. Outdoor and indoor PAHs: (benz[a]anthracene (BAAN), chrysene (CHRYS), benzo[b+j+k]fluroanthene (BFL), benzo[e]pyrene (BEP), benzo[a]pyrene (BAP), indeno[1,2,3-c,d]pyrene (IP), and benzo[g,h,i]perylene (BGP)) levels were obtained by averaging the two one-week measures, which were seasonalized, with BAP mean ± SD of 99 ± 62 pg/m3 and range of 20–304 pg/m3, and a PAHs mean ± SD of 1,458 ± 704 pg/m3 and range of 597–3,235 pg/m3. Weekly-averaged NO2 concentrations with passive dosimeters were also obtained. Source of pollutants: traffic emissions |
1.5T GE with 8 channel head-coil; T1 weighted: TR = 11.9 ms, TE = 4.2 ms, Flip = 15°, 256 × 256 matrix, 134 slices, 1.2 mm thickness; FREESURFER volumes (version not reported). No correction for multiple comparisons. | No associations with brain parenchymal fraction, putamen, or globus pallidus volumes. Higher outdoor school PAHs was associated with caudate volumes. Beta coefficients and 95% confidence intervals per IQR increase: Caudate nucleus, mm3: Total PAHs: β = −132.9 (−245.0, −20.8); p = 0.020; BAP: β = −150.6 (−259.1, −42.1); p = 0.007 Boys (n = 123): Total PAHs: β = −19.2 (−211.0, 172.6); p = 0.844; BAP: β = −48.6 (−219.8, 122.5); p = 0.577 Girls (n = 119): Total PAHs: β = −192.3 (−364.1, −20.5); p = 0.028; BAP: β = −212.0 (−382.8, −41.2); p = 0.015 No significant interactions between sex and PAHs observed for brain volumes Null findings: Brain parenchymal fraction, %: Total PAHs: β = 0.3 (−0.2, 0.7); p = 0.240; BAP: β = 0.0 (−0.4, 0.5); p = 0.929. Putamen, mm3: Total PAHs: β = 26.4 (−128.3, 181.2); p = 0.738; BAP: β = 13.0 (−139.2, 165.2); p = 0.867; Globus pallidus, mm3: Total PAHs: β = 4.5 (−46.1, 55.2); p = 0.861; BAP: β = −16.9 (−64.9, 31.0); p = 0.488 |
Age, sex, ICV, maternal education, home socioeconomic vulnerability index, residential exposure to NO2 and PM2.5, classroom noise | Exposure to PAHs associated with subclinical changes on the caudate |
| Guxens et al. (38) | 783 children (6–10 years old) from the Generation R Study, a population-based birth cohort in Rotterdam, The Netherlands, who were born between April 2002 and Jan 2006. Oversampled based on certain maternal exposures during pregnancy (i.e., cannabis, nicotine, selective serotonin reuptake inhibitors, depressive symptoms, and plasma folate levels) and child behavior problems (i.e., ADHD, pervasive developmental problems, dysregulation problems, and aggressive problems) | Air pollution monitoring campaigns took place over three 2-week periods for NO2 in 80 sites and PM10, PM2.5, and PM absorbance (a proxy for elemental carbon) in 40 sites in 2009 to 2010 across The Netherlands and Belgium. Coarse PM calculated as the difference between PM10 and PM2.5. The 3 measurements were averaged, adjusting for temporal variation. LUR models were then used to estimate air pollution exposure levels at mothers' geocoded residential address during the entire fetal period. Background monitoring network sites (n = 7) were used to back-extrapolate to the fetal period, accounting for changes of home address during pregnancy. a range of 16.8–28.1 μg/m3, the coarse (PM10) median value was 11.8 μg/m3 with a range of 9.2–17.8 μg/m3, the proxy for elemental carbon (PM absorbance) median value was 1.9 × 10−5 m−1 with a range of 1.2–3.6 10−5 m−1. Source of pollutants: intra-ubran road traffic, traffic on highways, and industrial and other point sources |
3T GE with 8 channel head coil; T1 weighted: TR = 10.3 ms, TE = 4.2 ms, Flip = 16°, 0.9 × 0.9 mm in-plane resolution, 186 slices, 0.9 mm thickness; FREESURFER (version 5.1) volumes and cortical thickness. Corrected using Monte-Carlo null-Z w/10,000 iterations (p < 0.1). | No associations were seen between air pollution exposure during fetal life and global brain volume measures at 6–10 years of age. Significant associations between air pollution exposure during fetal life and regional cortical thickness at 6–10 years of age. Fully adjusted beta coefficients and 95% confidence intervals representing the differences in thickness (mm) per each increase of 5 mg/m3 of PM2.5, 5 mg/m3 of PM10, and 10−5 m−1 of PMabsorbance: PM2.5: Right precuneus region: β = −0.045 (−0.062, −0.028); p < 0.001; Right Pars opercularis region: β = −0.024 (−0.033, −0.014); p < 0.001; Pars orbitalis region, right: β = −0.028(−0.043, −0.012);p = 0.001; Right rostral middle frontal region: β = −0.029 (−0.041, −0.018); p < 0.001; Right superior frontal region: β = −0.029 (−0.043, −0.016); p < 0.001; Left cuneus region: β = 0.022 (−0.035, −0.009); p = 0.002 PM10: Right Lateral orbitofrontal region: β = −0.037 (−0.059, −0.016); p = 0.001 PMabsorbance: Left fusiform region: β = −0.105 (−0.160, −0.049); p < 0.001 Null findings: NO2: Total brain volume: β = 124 (−1118, 1,375); p = 0.84; Cortical gray matter volume: β = −60 (−853, 733); p = 0.88; Cortical white matter volume: β = 199 (−287, 685); p = 0.42; Subcortical gray matter volume: β = 36 (−17, 89); p = 0.18; Ventricular volume: β = 4 (−57, 64); p = 0.90 PM2.5: Total brain volume: β = −3079 (−7790, 1,632); p = 0.20; Cortical gray matter volume: β = −2598 (−5583, 387); p = 0.09; Cortical white matter volume: β = −268 (−2096, 1,559); p = 0.77; Subcortical gray matter volume: β = −60 (−258, 138); p = 0.55; Ventricular volume: β = −96 (−323, 131); p = 0.40; PM10: Total brain volume: β = −4868 (−10337, 822); p = 0.09; Cortical gray matter volume: β = −3542 (−7059, 8); p = 0.05; Cortical white matter volume: β = −1129 (−3215, 1127); p = 0.34; Subcortical gray matter volume: β = −92 (−325,148); p = 0.46; Ventricular volume: β = −100 (−372, 168); p = 0.45; PMabsorbance: Total brain volume: β = −2861 (−18745, 24,467); p = 0.79; Cortical gray matter volume: β = −2683 (−16377, 11,012); p = 0.70; Cortical white matter volume: β = 5,807 (−2566, 14,180); p = 0.17; Subcortical gray matter volume: β = 418 (−497, 1,334); p = 0.36; Ventricular volume: β = −64 (−1108, 979); p = 0.90 |
Parental educational levels, monthly household income, parental countries of birth, parental ages, maternal prenatal smoking, maternal prenatal alcohol use, maternal parity, family status, and maternal psychological distress, calculated pre-pregnancy BMI, maternal IQ when children were 6; child's gender, age at scanning, genetic ancestry | Children exposed to higher PM levels during fetal life had thinner cortex in several brain regions of both hemispheres |
| Alemany et al. (39) | 163 children (mean age 9.3 ± 0.82 years old; range: 8–12.1 years) from the BREATHE project who had genetic data available. Children were selected from 38 schools in Barcelona and 1 adjacent city, Sant Cugat del Vallés (schools selected based on modeled NO2 values) and had been in the school for at least 18 months. 22.7% (n = 37) were APOE-e4 carriers | Pollutant levels in school courtyards were sampled twice during 1-week periods (8 h/day) separated by 6 months: Jan to June 2012 and Sept 2012 to Feb 2013. PAHs (measured in PM2.5) was measured with High-Volume samplers. Outdoor and indoor PAHs: (benz[a]anthracene (BAAN), chrysene (CHRYS), benzo[b+j+k]fluroanthene (BFL), benzo[e]pyrene (BEP), benzo[a]pyrene (BAP), indeno[1,2,3-c,d]pyrene (IP), and benzo[g,h,i]perylene (BGP)) levels were obtained by averaging the two one-week measures, which were seasonalized, with PAHs mean ± SD of 1546.29 ± 775.08 pg/m3. Elemental carbon (measured in PM2.5) was measured with High-Volume samplers and NO2 with passive dosimeters (96 h period), with NO2 mean ± SD of 47.74 ± 12.95 μg/m3. Source of pollutants: traffic emissions |
1.5T GE with 8 channel head-coil; T1 weighted: TR = 11.9 ms, TE = 4.2 ms, Flip = 15°, 256 × 256 matrix, 134 slices, 1.2 mm thickness; FREESURFER (version 5.3). No correction for multiple comparisons. | Beta coefficients and 95% confidence intervals of basal ganglia volumes by annual average of air pollution exposure to PAHs, EC, and NO2 (per IQR increase): Caudate nucleus, mm3: Total PAHs: β = −120.1 (−211.2, −29); EC: β = −110 (−250.5, 30.6); NO2: β = −265.1 (−474.3, −56) Putamen, mm3: Total PAHs: β = 9.1 (−115.2, 133.5); EC: β = −15.2 (204.7, 235.1); NO2: β = 93.4 (−364.3, 177.5) Globus pallidus, mm3: Total PAHs: β = 1.1 (−37.2, 39.5); EC: β = −22.2 (−101.6, 57.1); NO2: β = −35.6 (−133.9, 62.7) In e4 carriers and non-carriers: Caudate nucleus, mm3: Total PAHs: e4 carrier β = −590.2 (−1032.3, −148.15) vs. non-carrier β = −137.1 (−322.0, 47.84); p = 0.04; EC: e4 carrier β = −375.0 (−776.81, 26.88) vs. non-carrier β = −49.8 (−236.6, 137.1); p = 0.11; NO2: e4 carrier β = −737.9 (−1201.3, −274.5) vs. non-carrier β = −157.6 (−388.8, 73.6); p = 0.03 Putamen, mm3: Total PAHs: e4 carrier β = −13.6 (−272.5, −245.3) vs. non-carrier β = 19.45 (−456.4, 495.3); p = 0.90; EC: e4 carrier β = −31.0 (−444.3, 382.25) vs. non-carrier β = 52.9 (−206.3, 312.1); p = 0.55; NO2: e4 carrier β = −63.75 (−586.3, 458.8); non-carrier β = −82.4 (−405.4, 240.5); p = 0.74 Globus pallidus, mm3: Total PAHs: e4 carrier β = −43.1 (−212.9, 126.75)vs. non-carrier β = −5.1 (−81.0, 70.88); p = 0.40; EC: e4 carrier β = −13.1 (−166.3, 140.2) vs. non-carrier β = −26.10 (−102.0, 49.8); p = 0.97; NO2: e4 carrier β = −49.6 (−237.9, 138.6) vs. non-carrier β = −44.15 (−138.7, 50.35); p = 0.79 |
Age, sex, ICV, maternal education, maternal smoking, exposure to tobacco at home, home socioeconomic vulnerability index, residential exposure to NO2 and PM2.5. | Association between annual average air pollution exposure to PAHs and NO2 and smaller caudate volumes was larger in children carrying the APOE-e4 allele compared to non-carriers |