New AI Makes Brain Scans Four Times Faster Without Losing Detail
Shorter scans mean less waiting, less squirming, and fewer kids needing sedation.
Researchers built an AI called ART-Net to refine fast brain MRI scans. A shortcut called SENSE4 — four-fold accelerated sensitivity encoding — can shorten brain MRI acquisition time, but it may amplify noise and leave residual aliasing artifacts after conventional reconstruction. ART-Net, an Anatomy-aware Residual Attention Network, refines those accelerated reconstructions. In a prospective paired study, 80 participants underwent fully sampled and four-fold accelerated SENSE T1-weighted MRI, split into training, validation, and independent test cohorts of 45/5/30 participants. On the independent test cohort, ART-Net achieved the highest peak signal-to-noise ratio (31.03 +/- 2.88 dB) and structural similarity index (0.963 +/- 0.022) among the evaluated methods, and showed improved anatomical fidelity, gradient fidelity, regional contrast preservation, and radiologist-rated structural quality. The authors conclude that ART-Net improved agreement between SENSE4 and fully sampled T1-weighted images while maintaining segmentation-derived anatomical measurements.
- Brain MRI scans can be sped up four times over, then cleaned up by software — no new machine required.
- In a study of 80 people, the AI's fast scans came closest to full-quality scans of any method tested.
- Measurements of brain regions used to spot shrinkage, like in Alzheimer's, stayed accurate.
Why It Matters
Faster scans mean shorter waits, fewer sedated children, and cheaper imaging — if it holds up beyond one hospital.