Table 1 Studies on air pollution and MRI-detected alterations in brain structure and function.
Exposure period (Year) | Exposure duration (Year) | Period at MRI assessment (Year) | Pollutant/ Method of assessment | Brain alterations | Reference |
|---|---|---|---|---|---|
Lifetime | Lifetime | Childhood Mean age ± SD: ~10.7 ±2.3 years 2 MRI assessments 1 year apart. Mean age ±SD at baseline: ~7.0 ±0.6 years | Lived in a highly polluted city vs low polluted city | Prefrontal WM hypersensitivity ↓ WM volumes in temporal and parietal (R) lobe at 1 year follow-up among children residing in a highly polluted city compared to children from a low polluted city | Mexico City and Polotitlán area study n ≈ 30 |
Childhood 9–10 years (2016–2018) | 1 year (2016) | Childhood 9–10 years (2016–2018) | PM2.5 Spatiotemporal model | ↓ SA in frontal pole (R), cuneus (L) ↑ SA in lateral orbitofrontal (R) ↓ CT in lateral orbitofrontal (L), superior frontal (L), inferior temporal (R), parahippocampus (R), rostral anterior cingulate (L), caudal anterior cingulate (L), posterior cingulate (L), isthmus (L), insula (R) ↑ CT lateral orbitofrontal (R), paracentral (R), middle temporal (L), rostral anterior cingulate (R), caudal anterior cingulate (R), posterior cingulate (R) ↑ volumes in accumbens (L), pallidum (R), thalamus (R) ↓ volumes in pallidum (L), putamen (L) | [33] ABCD study n = 10,341 |
Childhood 9–10 years (2016–2018) | 1 year (2016) | Childhood 9–10 years (2016–2018) | PM2.5 Spatiotemporal model | ↑ rN0 (nonlinear) in the cingulum hippocampal portion (L), uncinate fasciculus (L), and fornix (L) ↑ rN0 (linear) in the uncinate fasciculus (R), the fornix (R), superior longitudinal fasciculus (L) ↓ MD (nonlinear) in the anterior thalamic radiations (L), cingulum hippocampal portion (L), fornix (L), superior longitudinal fasciculus (L), uncinate (L), inferior longitudinal fasciculus (R), and uncinate (R) ↓ MD (linear) in the inferior fronto-occipital (L), inferior longitudinal fasciculus (L), cingulum hippocampal portion (R), fornix (R) | [32] ABCD study n = 7602 |
Childhood/Preadolescence 9–13 years | 2 years prior to first MRI assessment | Preadolescence/ adolescence 11–15 years 2 MRI assessments 2 years apart | PM2.5 Spatiotemporal models | At 2 years follow-up higher PM2.5 exposure was associated with the following changes: ↑ WM volume in caudate/corpus callosum (L), cingulum (L), inferior fronto-occipital fasciculus, inferior frontal gyrus (R), inferior temporal gyrus (R) ↑ GM volume in precentral gyrus (L), cerebellum (L), medial orbitofrontal cortex ↓ WM volume in inferior temporal gyrus (L), angular gyrus (L), posterior thalamic radiation (L), middle frontal gyrus (L), hippocampal cingulum (L), postcentral gyrus (R) ↓ GM volume in insula (L), cingulate gyrus (R), caudate (R), cerebellum (L), fusiform gyrus, precentral gyrus, middle frontal gyrus | [34] San Francisco and San Jose Bay Area study n = 115 |
Prenatal | Whole pregnancy (PM2.5) 48-h during last trimester (PAH) | Childhood/Adolescence 6–14 years | PM2.5, PAH Spatiotemporal models (PM2.5) Personal air monitors (PAH) | Exposure to PM2.5 ↓ WM surface in lateral pre/procentral gyrus, superior frontal gyrus, middle frontal gyrus (L), middle temporal gyrus (L), inferior parietal lobule (L), anterior cingulate cortex, posterior cingulate cortex (R) ↑ WM surface in medial and dorsal pre/procentral gyrus, medial superior frontal gyrus, lateral superior temporal gyrus (R), dorsal superior parietal gyrus ↓ CT in superior parietal gyrus, pre/procentral gyrus ↑ CT in superior frontal gyrus, inferior frontal gyrus (L), superior temporal gyrus (L), inferior temporal gyrus, middle temporal gyrus, inferior parietal lobule (L), anterior cingulate cortex, posterior cingulate cortex (R), fusiform and lingual gyrus ↑ FA in caudate, lenticular nucleus, insula, brainstem, thalamus, cingulate gyrus, superior corona radiate ↑ ADC in inferior fronto-occipital fasciculus, anterior corona radiata, vertical occipital fasciculus Exposure to PAH ↓ WM surface in inferior temporal gyrus, middle temporal gyrus, inferior parietal lobule (L), anterior cingulate cortex (R), posterior cingulate cortex (L) ↑ WM surface in pre/procentral gyrus, superior frontal gyrus, dorsal middle frontal gyrus, ventral fusiform gyrus, ventral lingual gyrus ↓ CT in superior frontal gyrus, middle frontal gyrus, inferior frontal gyrus (L), pre/procentral gyrus, superior temporal gyrus (R), middle temporal gyrus (R) ↑ CT in middle temporal gyrus (L), anterior cingulate cortex (R), fusiform and lingual gyrus ↑ FA in middle orbitofrontal gyrus, cerebellum, hippocampus, globus pallidus, putamen, thalamus, corpus callosum, internal capsule ↓ ADC in internal capsule, corpus callosum | [27] CCCEH study n = 332 |
Prenatal | 1st, 2nd, 3rd trimester and whole pregnancy | Childhood/ Preadolescence 8–12 years (2012–2014) | PM2.5 LUR models Measurements from monitors at site were collected between Oct 2008 and Apr 2011 | ↓ volume in total, anterior and body corpus callosum with higher PM2.5 exposure during the 3rd trimester. Associations did not survive false discovery rate correction. No association for WM, GM and lateral ventricles | [29] BREATH n = 186 |
Childhood/Preadolescence (2012) 7–11 years 8–12 years | 1 year (2012) | Childhood/ Preadolescence 7–11 years 8–12 years (2012–2014) | NO2, PAH, BPA, EC, copper Monitors at site Two 1-week periods separated by two semesters | Exposure to PAH, BPA and NO2 ↓ Caudate volume No association for putamen and globus pallidus volumes Exposure to EC No association for caudate, putamen, and globus pallidus volumes Exposure to copper ↑ GM concentration in the caudate nucleus No association for putamen and globus pallidus ↑ FA predominantly in caudate nucleus ↓ rsFC between the frontal lobe opercula and the caudate nuclei, and vice versa | BREATH study n ≈ 200 |
Childhood/ Preadolescence (2012) 8–12 years | 1 year (2012) | Childhood/ Preadolescence 8–12 years | Pollution index: weighted average of pooled indoor and outdoor NO2 and EC Monitors at site Two 1-week periods separated by two semesters | ↓ rsFC between regions belonging to the DMN ↑rsFC between the medial frontal cortex and the frontal operculum at the lateral boundary of the DMN ↓ deactivation in the supplementary motor area and somatosensory cortex in the study deactivation map | [31] BREATH study n = 263 |
Prenatal (2001–2006) | Whole pregnancy | Childhood 6–10 years | NO2, PMcoarse, PM2.5, PM2.5abs LUR models 2-week measurements in three different seasons (warm, cold, and intermediate) from monitors at site between Feb 2009 and Feb 2010 | ↓ CT in praecuneus (R), pars opercularis (R), pars orbitalis (R), rostral middle frontal (R), superior frontal (R), cuneus (L) with higher exposure to PM2.5 ↓ CT in lateral orbitofrontal (R) with higher exposure to PM2.5coarse ↓ CT in fusiform (L) with higher exposure to PM2.5abs | [22] Generation R study n = 783 |
Prenatal (2001–2006) | Whole pregnancy | Preadolescence 9–12 years | NOx, NO2, PM10, PMcoarse, PM2.5, PM2.5abs, PAH, OC, copper, iron, silicon, zinc, OP LUR models 2-week measurements in three different seasons (warm, cold, and intermediate) from monitors at site between Feb 2009 and Feb 2010 | ↑ volumes of putamen and pallidum with higher exposure to PMcoarse ↑ volume of cerebellum with higher exposure to PM10, PMcoarse, PM2.5, PM2.5ab ↓ volume of hippocampus with higher exposure to PAH, copper ↓ volume of amygdala with higher exposure to OC, silicon ↓ volume of corpus callosum with higher exposure to OP ↓ CT in postcentral gyrus (R) with higher exposure to OC (marginally nonsignificant) ↓ CT in rostral middle frontal gyrus (R) with higher exposure to copper and PM2.5abs (marginally nonsignificant) ↓ FA in forceps minor, corticospinal tract, superior longitudinal fasciculus (R) with higher exposure to PM2.5 ↑ MD in cingulum bundle, forceps minor, superior longitudinal fasciculus (L), inferior longitudinal fasciculus (L) with higher exposure to silicon ↑ rsFC between brain regions of the same brain hemisphere, predominantly in the auditory association, dorsolateral prefrontal, somatosensory and motor, anterior cingulate and medial prefrontal, dorsal stream visual, and insular and frontal opercular cortices with higher exposure to NO2 | [22] Generation R study n = 3133 |
Childhood 6–10 years | 1 year | Preadolescence 9–12 years | NOx, NO2, PM10, PMcoarse, PM2.5, PM2.5abs, PAH, OC, copper, iron, silicon, zinc, OP LUR models 2-week measurements in three different seasons (warm, cold, and intermediate) from monitors at site between Feb 2009 and Feb 2010 | ↓ volume of hippocampus with higher exposure to PMcoarse, OP ↑ volume of nucleus accumbens with higher exposure to zinc ↓ volume of corpus callosum with higher exposure to OC ↓ CT in lingual gyrus (L) with higher exposure to copper and OP (marginally nonsignificant) ↑ SA in precentral gyrus (R) with higher exposure to zinc and OP ↑SA in pericalcarine cortex (L) and precuneus (L) with higher exposure to zinc ↓ SA in pars triangularis (R) with higher exposure to PMcoarse (marginally nonsignificant) ↓ FA in corticospinal tract (L), uncinated fasciculus, superior longitudinal fasciculus (R), inferior longitudinal fasciculus (R) with higher exposure to NOx ↑ MD in cingulum bundle (L) with higher exposure to OP ↑ MD in cingulum bundle, forceps minor, superior longitudinal fasciculus, inferior longitudinal fasciculus, uncinated fasciculus with higher exposure to zinc | [22] Generation R study n ≈ 3000 |
Prenatal and childhood | Whole pregnancy and childhood at periods: 0–2 years 2–5 years 5–9 years | Preadolescence 9–12 years | NOx, NO2, PM10, PMcoarse, PM2.5, PM2.5abs LUR models 2-week measurements in three different seasons (warm, cold, and intermediate) from monitors at site between Feb 2009 and Feb 2010 | Exposure to NOx from 0–2 and 2–5 years ↑ rsFC in areas in auditory association, premotor, orbital and polar frontal, inferior parietal, and posterior cingulate cortices, in the ventral diencephalon, and in the MT+ complex and neighbouring visual areas Exposure to NO2 during pregnancy and from 0–2 years ↑ rsFC in areas in in auditory association, ventral diencephalon, and insular and frontal cortices opercular, somatosensory and motor and early auditory, dorsal stream visual and superior parietal Exposure to PMcoarse from 2–5 and 5–9 years ↑ rsFC in areas in anterior cingulate and medial prefrontal cortices and in the MT+ complex and neighbouring visual areas Exposure to PM2.5abs from 0–2 and 2–5 years ↑ rsFC in areas in insular and frontal opercular, auditory association, lateral temporal, somatosensory and motor, anterior cingulate and medial prefrontal, and posterior cingulate cortices, and in the MT+ complex and neighbouring visual areas | [22] Generation R study n = 2197 |
Prenatal and childhood | Whole pregnancy and childhood at periods: 0–3 years 3–6 years 6 – age of MRI | Preadolescence 9–12 years | NOx, NO2, PM10, PM2.5, PM2.5abs LUR models 2-week measurements in three different seasons (warm, cold, and intermediate) from monitors at site between Feb 2009 and Feb 2010 | Exposure to NOx from 3–6 years ↑ rsFC in regions of the visual and task positive networks: MT+ complex and neighbouring visual areas—inferior frontal cortex and MT+ complex and neighbouring visual areas—insular and frontal opercular cortex Exposure to NO2 from 0–3 years ↑ rsFC in regions of the visual, auditory and task positive networks: dorsal stream visual cortex – superior parietal cortex and auditory association cortex – insular and frontal opercular cortex Exposure to PM2.5abs from 0–3 years ↑ rsFC between brain regions of several networks (19 of 22): visual - visual, visual – auditory, visual – task positive, visual – task negative, auditory – task positive, auditory – task negative, and task negative – task negative ↓ rsFC between brain regions of visual – task positive networks and task positive – task negative networks (3 of 22): MT + complex and neighbouring visual areas – superior parietal cortex, posterior cingulate cortex – superior parietal cortex, and frontal opercular cortex – lateral temporal cortex | [79] Generation R study n = 2197 |
Prenatal and childhood | 48-h during last trimester 5 years of age | Childhood Mean age ± SD: 8.0 ± 1.3 years | PAH Personal air monitors during last trimester Urine samples in childhood | Prenatal exposure to PAH ↓ local volume in the middle frontal gyrus, medial orbitofrontal gyrus, inferior frontal gyrus, superior frontal gyrus, pre-central gyrus, post-central gyrus, supramarginal gyrus, middle temporal gyrus, superior temporal gyrus, mesial superior parietal gyrus, praecuneus, cuneus, cingulate gyrus, gyrus rectus in the left hemisphere. ↓ WM surface extending throughout the left hemisphere. No association with cortical thickness Postnatal exposure to PAH ↓ WM surface in dorsolateral prefrontal regions, especially over the superior frontal gyri | [26] CCCEH study n = 40 |
Lifetime | From birth (2001–2003) to 12 years of age | Childhood 12 years | EC (high vs low exposure group) LUR models Measurements from 27 monitors at site between 2001 and 2006, and simultaneous 24-hour sampling at 4–5 sites over different seasons | ↓ CT in the medial frontal gyrus, ventromedial prefrontal cortex (R), paracentral lobule, postcentral gyrus, superior frontal gyrus, precentral gyrus (L), inferior parietal lobule (L), superior parietal lobule (L), anterior cingulate (L), cingulate (R), precuneus (L), fusiform gyrus (L) ↓ GM volume in the cerebellum, precentral gyrus, inferior parietal lobule | [28] CCAAPS study n = 135 |
Adulthood (1998–2001) | 1 year (2000) | Adulthood (1999–2005) ≥ 60 years Median age [IQR]: 68.0 [9.0] years | PM2.5 Spatiotemporal model | ↓ total cerebral brain volume No association with hippocampal volume | [40] Framingham Offspring Study n = 943 |
Adulthood (2006–2010) | 1 year (2010) | Adulthood (2014) 44–80 years | NOx, NO2, PM10, PMcoarse, PM2.5 LUR models | ↓ total GM volume with higher exposure to any of the investigated pollutants ↓ GM volume in the frontal pole and operculum cortex (L) with higher exposure to PM10 ↓ GM volume in the frontal pole, operculum cortex (L), and orbital cortex (R) with higher exposure to NOx ↓ GM volume in the frontal pole (R) and operculum cortex (L) with higher exposure to NO2 Exposure to PM2.5 ↓ total WM volume ↓ GM volume in the frontal pole, orbital cortex (R), operculum cortex (L) Exposure to PMcoarse ↓ GM volume in the frontal pole (R), superior gyrus (L), operculum cortex (L) ↓ volume in the thalamus (L) | UK Biobank n ≈ 18,290 |
Adulthood (1999–2005-6) | 3, 8, and 10 years (1999–2005-6) | Adulthood (2005–2006) 71–89 years | NO2, PM2.5, diesel PM Spatiotemporal model | 3-year cumulative exposure to NO2 ↓ GM volumes in the prefrontal cortex ↓ volumes in the anterior cingulate gyrus, insula, amygdala, limbic medial temporal lobe, basal ganglia 3-year cumulative exposure to PM2.5 ↓ WM volumes in the anterior and posterior extreme/ external capsule, calcarine gyri ↓ GM volumes in the superior, middle, medial frontal gyri, inferior frontal gyrus (L), superior parietal lobule, occipital poles ↓ volume in the anterior cingulate gyrus ↑ volumes in the thalamus, putamen, globus pallidus, posterior insula No association with volumes of corpus callosum, hippocampus, temporal lobe 8-year cumulative exposure to PM2.5 ↓ total WM volume and in the frontal, parietal, temporal lobes, corpus callosum No association with hippocampal, basal ganglia volumes and GM volumes across the cerebral cortex 10-year cumulative exposure to diesel PM ↑ ventricular volume U-shaped associations were observed for total WM volumes and in frontal, parietal and temporal lobes ↓ total GM volumes and in frontal, parietal and temporal lobes | WHIMS n = 764, 1403, 1365 n = 1403 (8- and 10-year assessment) |
Adulthood (1990–1998, 1999–2007, and 1990–2007) | 5–20 years assessed at three 8-year periods (1990–1998, 1999–2007, and 1990–2007) | Adulthood (2011–2013) Mean age ±SD: 76.0 ±5.0 years | PM10, PM2.5 Spatiotemporal model | ↓ deep-GM volumes ↓ volumes in total brain, frontal and parietal lobe in one of the study centres with higher exposure to PM2.5 No association with hippocampal volume | [45] ARIC study n = 1753 |
Adulthood (2006–2008) 50–80 years | 2 years (2006–2008) | Adulthood (2011–2015) 55–85 years | NOx, NO2, PM10, PM2.5, PM2.5abs LUR models Three separate 2-week periods (to cover different seasons) between Oct 2008 and Oct 2009 | Local atrophy in inferior parietal lobule (R) with higher exposure to NOx, PM10, and PM2.5 Local atrophy in posterior cingulate cortex and praecuneus (R) with higher exposure to NOx, NO2, and PM10 No association with local atrophy in the dorsolateral prefrontal cortex | [46] 1000BRAINS n ≈ 615 |
Adulthood (2000–2003) Mean age: 56.1 years | 3 years (2000–2003) | Adulthood (2011–2015) 56–85 years Mean age: 67.4 years | NO2, PM10, PM2.5, PM2.5abs LUR models Three separate 2-week periods (to cover different seasons) between Oct 2008 and Oct 2009 | ↓ Intra-network rsFC and segregation index in the dorsal attention network with higher exposure to NO2 ↓ Intra-network rsFC in the ventral attention network with higher exposure to PM10 and PM2.5 ↑ Inter-network rsFC in the visual network with higher exposure to PM2.5abs ↓ segregation index in the ventral attention network with higher exposure to PM2.5abs | [54] 1000BRAINS n = 574 |
Adulthood (2015–2019) | 1 year | Adulthood (2015–2019) Mean age ±SD: 49.5 ± 13.3 years | NO2, PM2.5, ozone Hybrid kriging-LUR models | ↑ volume in rostral middle frontal (L), supramarginal (L), transverse temporal (L), pars opecularis (R) with higher exposure to NO2 ↑ volume in pars triangularis (L) and CT in fusiform (R) with higher exposure to PM2.5 ↑ pars orbitalis volume (L) with higher exposure to ozone No association with WM and GM volumes | [47] Taiwanese sleep study n = 4866 |
Adulthood (2014–2017) | 5 years prior to the recruitment intervals (NO2, PM10) (2010–2014, 2011–2015, and 2012–2016) 1 year (PM2.5) (2015) | Adulthood (2014–2017) Mean age ±SD: 67.3 ± 6.4 years | NO2, PM10, PM2.5, PAH metabolites Kriging model Annual concentrations of PM10 and NO2 at monitoring sites between 2001 and 2016. Same method used for PM2.5 Urine samples (PAH) | Exposure to NO2 ↓ volume in caudate, pallidum, amygdala, nucleus accumbens ↓ CT in the frontal cortex, lateral temporal cortex, inferior parietal cortex, posterior cingulate, insula, parahippocampal gyri, fusiform gyri ↑ CT occipital cortex, postcentral gyri (L) Exposure to PM10 ↓ volume in pallidum, putamen, amygdala, nucleus accumbens ↓ CT in the lateral temporal cortex, inferior parietal cortex, prefrontal cortex, posterior cingulate, insula, parahippocampal gyri, fusiform gyri ↑ CT occipital cortex, postcentral gyri Exposure to PM2.5 ↓ volume in nucleus accumbens ↓ CT in the lateral temporal cortex, inferior parietal cortex, prefrontal cortex, insula, parahippocampal gyri, fusiform gyri ↑ CT occipital cortex, postcentral gyri Exposure to PAH (highest vs lowest quartile) ↓ CT in parietal, temporal and insular lobes in men ↓ CT in frontal and parietal lobes in women ↓ volumes in the caudate in men, and pallidum in women | EPINEF study n = 957 n = 528 (PAH) |
Adulthood (2016) | 1 year (2015) | Adulthood (2016) 20–48 years | NO2, NOx, PM10, and PM2.2 Berlin’s Senate Department for Urban Development and Housing | Exposure to NO2 and NOx No association was observed Exposure to PM2.5 and PM10 ↓ stress-related activation in frontoinsular cortex, hippocampus, amygdala, ventral striatum, inferior parietal cortex, thalamus, precuneus, posterior cingulate cortex, anterior cingulate cortex, dorsolateral, ventrolateral, and ventromedial prefrontal cortex. The associations were more pronounced for PM2.5 than PM10 | [56] Berlin neuroimaging study n = 42 men |
Adulthood | 120 min | Adulthood Mean age ±SD: 27.4 ±5.5 years | Diesel exhaust (intervention arm) Filtered air (control arm) | No differences in the default mode network rsFC for post- compared to pre-diesel exhaust exposure ↑ rsFC in the middle temporal gyrus (R), occipital fusiform gyrus (R) for post-filtered air compared to pre-filtered air ↑ rsFC in the angular gyrus (R), frontal pole, middle frontal gyrus, middle temporal gyrus, praecuneus cortex, temporal pole (L) for post-filtered air compared to post-diese; exhaust | [253] Randomised cross-over study n = 25 |
