Multi-echo Acquisition and Thermal Denoising Advances Precision Functional Imaging
Julia Moser,Steven M. Nelson,Sanju Koirala,Thomas Madison,Alyssa K. Labonte,Cristian Morales Carrasco,Eric Feczko,Lucille A. Moore,Jacob T Lundquist,Kimberly B Weldon,Gracie Grimsrud,Kristina Hufnagle,Weli Ahmed,Michael J. Myers,Babatunde Adeyemo,Abraham Z Snyder,Evan M. Gordon,Nico U F Dosenbach,Brenden Trevo-Clemmens,Bart Larsen,Steen Moeller,Essa Yacoub,Luca Vizioli,Kamil Ugurbil,Timothy O. Laumann,Chad M. Sylvester,Damien A. Fair
DOI: https://doi.org/10.1101/2023.10.27.564416
2024-10-09
Abstract:The characterization of individual functional brain organization with Precision Functional Mapping has provided important insights in recent years in adults. However, little is known about the ontogeny of inter-individual differences in brain functional organization during human development. Precise characterization of systems organization during periods of high plasticity is likely to be essential for discoveries promoting lifelong health. Obtaining precision fMRI data during development has unique challenges that highlight the importance of establishing new methods to improve data acquisition, processing, and analysis. Here, we investigate two methods that can facilitate attaining this goal: multi-echo (ME) data acquisition and thermal noise removal with Noise Reduction with Distribution Corrected (NORDIC) principal component analysis. We applied these methods to precision fMRI data from adults, children, and newborn infants. In adults, both ME acquisitions and NORDIC increased temporal signal to noise ratio (tSNR) as well as the split-half reliability of functional connectivity matrices, with the combination helping more than either technique alone. The benefits of NORDIC denoising replicated in both our developmental samples. ME acquisitions revealed longer and more variable T2* relaxation times across the brain in infants relative to older children and adults, leading to major differences in the echo weighting for optimally combining ME data. This result suggests ME acquisitions may be a promising tool for optimizing developmental fMRI, albeit application in infants needs further investigation. The present work showcases methodological advances that improve Precision Functional Mapping in adults and developmental populations and, at the same time, highlights the need for further improvements in infant specific fMRI.
Neuroscience