Feasibility, image quality, and analytic usability of mobile 64 mT ultra-low-field MRI for infant brain imaging at 3 and 12 months in southern Malawi.

Mobile ultra-low-field MRI (ULF-MRI) could reduce inequities in infant neuroimaging in sub-Saharan Africa, but evidence of its feasibility, image quality, and analytic usability remains limited. We evaluated the feasibility, image quality, and analytic usability of mobile 64 mT ULF-MRI for non-sedated infant brain imaging at three and twelve months of age within a longitudinal trial platform in southern Malawi. Feasibility was assessed through visit-level scan uptake and sequence completion; image quality through artifact-based quality control (QC) and radiologist interpretability review; and analytic usability through multi-structure volumetry and correspondence across independent processing workflows. Across 810 eligible visits (3 months: 410; 12 months: 400), 654 MRI sessions were completed (80.7%; 86.6% at 3 months; 74.8% at 12 months), with high sequence acquisition success among completed scans. Of 654 completed sessions, 495 entered artifact-based QC and 442/495 (89.3%) met predefined full-brain quality criteria for volumetric processing; motion-related degradation was the principal determinant of exclusion. Radiologist review (n = 426) rated 396/426 scans (93.0%) as analysable. Volumetry was derived for 215 infants at 3 months and 227 at 12 months, with 99 paired observations demonstrating measurable age-related differences in tissue volumes (grey matter +39.0 ± 7.2%; white matter +48.8± 9.5%). An independent cloud-based pipeline showed high correspondence for supratentorial tissue and total intracranial volume at 12 months (r ≈ 0.97). These findings demonstrate that non-sedated infant ULF-MRI can acheive high uptake, interpretable imaging, and scalable volumetric processing in a low-resource setting. The results support its analytic readiness for integration into developmental neuroscience research where conventional MRI is unavailable.