DD-INR: Dynamics-Driven Implicit Neural Representation for Accelerated Whole-Brain Functional MRI Reconstruction
Researchers have developed DD-INR, a novel framework for reconstructing functional MRI (fMRI) data that has been acquired with accelerated sampling. This method specifically addresses the challenge of recovering subtle task-evoked brain activity signals, which are often missed by traditional reconstruction techniques that prioritize spatial accuracy over temporal fidelity. By separating static background information from dynamic changes and using an Implicit Neural Representation (INR) for the latter, DD-INR focuses computational resources on relevant activations, potentially enhancing the sensitivity and robustness of fMRI studies. AI
IMPACT This new framework could improve the sensitivity and robustness of fMRI studies by enabling more accurate reconstruction of brain activity from accelerated scans.