At the resolution date, I will search among citations of Yang et al. 2018 (the most recent report on the largest RCT to date), for studies that meet the following criteria:
An RCT on a breastfeeding intervention with 300 or more participants which includes IQ (global score) reported at 10 years or older as an outcome variable in which the intervention increases exclusive or rates by 10% or more at any time point. If there are more than one RCTs that meet the criteria, I'll select the more robust one (e.g., the one with blinded assessment of IQ). The resolution is 'yes' if the difference between the control and intervention group IQ (global score) is 3 points or greater and the 95% CI do not include 0. If the IQ scores are compared at multiple ages, I'll take the oldest age measurement.
If RCT data are not available, I'll search for studies using an instrumental variable method. This could, for example, be a mendelian randomisation study or a study exploiting variation in breastfeeding support like this one. Other than study design, the criteria for selecting the study and the outcome variables are the same as for the RCT above.
