In brief, NO2 and PM were monitored at 40 and 20 sites, respectively, between October 2008 and November 2009. Each site was monitored for two consecutive weeks, three times (during warm, cold and intermediate temperature seasons). These discontinuous measurements were then adjusted for the actual long-term average of the observation period, according to data from one additional background site that measured air pollution concentrations using the same instruments continuously for a complete year. Using these data, land use regression models were developed and used to assign individual air pollution exposure estimates to each participant's home address at the age of 10 years. Details of this procedure have been previously published