PING

PING is a cross-sectional study of typically developing children, adolescents and young adults that pairs multimodal brain MRI with cognitive testing, developmental history and genome-wide genotyping. It was run by the Center for Human Development at UC San Diego with nine partner sites and funded by NIDA and NICHD, with the aim of linking genetic variation to brain structure, connectivity and cognition across development. Data collection is finished and the resource is now held in the NIMH Data Archive.

Composition

More than 1,700 participants were enrolled, and the shared repository keeps the 1,493 who have usable data for at least two data types (780 male). Of these, 1,239 have acceptable MRI (645 male), 1,453 have cognitive scores and 1,391 have genotypes. Ages range from 3 to 21 years. The NDA collection record lists 1,494 shared subjects. Children with neurological disorders, head trauma, preterm birth, autism spectrum disorder, bipolar disorder, schizophrenia or intellectual disability were excluded; learning disability and ADHD were not exclusion criteria.

Acquisition

Scanning took place at 10 US sites on 3T GE, Siemens and Philips systems with harmonized protocols. A standard session held a 3D T1-weighted scan and a 3D T2-weighted fast spin echo scan, both with prospective motion correction, a high angular resolution diffusion scan and a resting-state BOLD fMRI scan, the last two with built-in distortion correction. Cognition was measured with the NIH Toolbox Cognition Battery, and saliva DNA was genotyped on the Illumina Human660W-Quad BeadChip.

Annotations

There are no manual lesion or structure labels. The repository includes processed measures from a FreeSurfer-based pipeline (cortical thickness, surface area, subcortical volumes) and diffusion tensor measures for atlas-defined fiber tracts, plus genetic ancestry estimates. Raw scans were rated for quality by trained technicians.

Known limitations

The study is cross-sectional. Some site review boards did not allow full sharing, so part of the raw imaging and genetic data is restricted. Imaging is pooled from several scanner models and vendors, so scanner effects need modelling.