Table 3 Exposure Details

Harris et al. (2015) [7]

Prenatal (third trimester)

12.3 ± 2.6

Residential addresses of the birthing parent reported at study visits and on annual questionnaires were geocoded and spatially joined to pollution estimate models.

Satellite aerosol optical depth measurements at the 10x10 km grid scale for years 2000-2010 from the Moderate Resolution Imaging Spectroradiometer aboard the Earth Observing System satellites. Additional inputs to PM_2.5 concentration measurements from the US Environmental Protection Agency and Interagency Monitoring of Protected Visual Environment networks, along with data on area point sources of PM_2.5, land use, locations of major roads, and meteorology

Porta et al. (2016) [9]

Postnatal

19.5 ± 2.2

Participants' residential addresses at birth and all of their following residential addresses through the age of cognitive assessment were geocoded and spatially joined to pollution estimate models.

Land use regression models developed within the European Study of Cohorts for Air Pollution Effects (ESCAPE). Rome particulate matter levels were measured at 20 sites from 2010-2011 over three separate 2-week periods: cold (January-March), warm (June-September), and intermediate (April-June) seasons. Results from the tree measurements at each site were averaged, adjusting for temporal variations using centrally located background reference sites, which took measurements for an entire year (cross-validation R² = 0.79).

Wang et al. (2017) [10]

Postnatal

13.7 ± 6.7

Residential addresses for families were prospectively collected through self-reports every 2-3 years. Addresses were geocoded to match residencies by exact parcel locations or specific street segments of participating families

Daily PM_2.5 concentrations were obtained from the US Environmental Protection Agency Technology Transfer Network for the years 2000-2014. A spatial-temporal model based on the measured PM_2.5 concentration was constructed which had high consistency (R² = 0.74-0.79) to estimate the monthly average for each subject’s geocoded residential location.

Seifi et al. (2021) [12]

Childhood

Low 39.0± 16.9 Intermediate 58.0 ± 23.9 High 84.2 ± 32.2

Three low-privileged geographic locations (A, B, C) were selected to conduct monitoring. An equal number of participants from the three respective locations were randomly selected to undergo IQ testing

Real-time measurements of PM_2.5 mass concentrations were provided from environmental dust monitors based on an optical scattering method. Indoor and outdoor exposure was simultaneously measured using direct reading equipment

Ni et al. (2022) [11]

Prenatal (pregnancy average)

8.75 ± 2.0

Residential addresses were collected from participants at enrollment and updated at each subsequent point of contact

Point-based PM_2.5 levels were estimated from a spatial-temporal model on a 2-week scale. This model used monitoring data from regulatory networks, further enhanced with air pollution measurements from intensive research cohort-specific monitors. A geographic information system was used to identify covariates representing land use characters that could reflect spatial variability in air pollution distributions and the dimension-reduced regression covariates were obtained using partial least squares from more than 400 geographic variables.

Sun et al. (2023) [8]

Prenatal (first trimester)

38.8 ± 6.2

Geographical coordinates of participants based on birthing parent’s residential addresses. During follow-up visits, migration was taken into consideration by averaging exposure levels if multiple residences were reported.

Satellite-based modeling and aerosol optical depth retrieval and GEOS-Chem simulations. Ground measurements from approximately 1000 monitors were used for cross-validation. Predictions were highly consistent with the real-time measurements (R² = 0.78)