Conventional MRI systems that operate at high field are expensive and require significant infrastructure to run, making them difficult to procure in LMIC settings. As a result, most neuroimaging studies are conducted in higher resourced hospitals and settings. Current commercially available ultra-low-field MRI systems, such as the Hyperfine Swoop, are portable and do not require substantial infrastructure or operator expertise to run. The systems also require significantly less power to run, making them more accessible in lower resourced settings. However, such systems were initially designed for clinical populations based in higher resource settings, and so their utility in global health is unknown.

The UNITY project launched in 2021 with the rollout of Hyperfine’s ultra-low field portable MRI scanner to locations such as Uganda, South Africa, and Pakistan, followed by more sites across sub-Saharan Africa, southern Asia. Sites include a mixture of research, primary, and tertiary care centres. Study cohorts encompass a spectrum of birth outcomes, maternal and child diet and nutritional status, socioeconomic characteristics, environmental adversities, and differing gender and social norms and equality. MRI and other complementary clinical assessments collected as part of these studies will be used to map neuro development from 0 – 5 years of age.

Our global network of clinical research partners is supported by a network of MR physicists and engineers, led by Professor Steve Williams at King’s College London. Working in close collaboration with Hyperfine, their goal is to develop the repertoire of low-field MR imaging methods and improve their sensitivity to targeted nutritional maternal & infant health interventions, based on the needs of our clinical partners. Members of the group are also responsible for developing a data quality control (QC) and quality assurance (QA) protocol to measure data consistency between sites and software versions.

In addition, the UNITY project includes analytical hubs for data analysis and neuromodelling. Their goals are to develop standardised image processing and analysis methods which can be run across all data via a cloud-based analysis platform and develop AI methods to improve image quality.

The network also has a training and capacity building portfolio, which aims to expand capacity at both the research and clinical level. This includes shared resources to support scanner delivery and implementation, running the scanner, QA procedures and MR image analysis. We have also partnered with collective minds radiology - a cloud-based platform where anonymised MR Images can be uploaded for community feedback, facilitating clinical interpretation by local radiologists. ISMRM (The International Society for Magnetic Resonance Imaging) is supporting knowledge exchange programs and educational training.