The developing Human Connectome Project (dHCP) automated resting-state functional processing framework for newborn infants

Abstract

Highlights:1. An automated and robust pipeline to minimally pre-process highly confounded neonatal fMRI data 2. Includes integrated dynamic distortion and slice-to-volume motion correction 3. A robust multimodal registration approach which includes custom neonatal templates 4. Incorporates an automated and self-reporting QC framework to quantify data quality and identify issues for further inspection 5. Data analysis of 538 infants imaged at 26-45 weeks post-menstrual age 2 Abstract:The developing Human Connectome Project (dHCP) aims to create a detailed 4-dimensional connectome of early life spanning 20 to 45 weeks post-menstrual age. This is being achieved through the acquisition of multi-modal MRI data from over 1000 in-and ex-utero subjects combined with the development of optimised pre-processing pipelines. In this paper we present an automated and robust pipeline to minimally pre-process highly confounded neonatal resting-state fMRI data, robustly, with low failure rates and high quality-assurance. The pipeline has been designed to specifically address the challenges that neonatal data presents including low and variable contrast and high levels of head motion. We provide a detailed description and evaluation of the pipeline which includes integrated slice-to-volume motion correction and dynamic susceptibility distortion correction, a robust multimodal registration approach, bespoke ICA-based denoising, and an automated QC framework. We assess these components on a large cohort of dHCP subjects and demonstrate that processing refinements integrated into the pipeline provide substantial reduction in movement related distortions, resulting in significant improvements in SNR, and detection of high quality RSNs from neonates.

Functional and diffusion MRI reveal the functional and structural basis of infants’ noxious-evoked brain activity

“…Using near-infrared spectroscopy to measure haemodynamic responses to pain in the perinatal period, the amplitude of these responses was also observed to progressively increase with age 15 . And studies using fMRI to observe the developmental progression of resting-state activity have also found increased sensory and motor (and cognitive) network functional connectivity strength and activity amplitude with increasing age 24,25 . While the structural correlates tested in our nociception-paradigm infant dataset demonstrated statistically significant negative correlation between noxious response amplitude and white matter MD, clear trends existed for both MD and FA.…”

“…While analogous studies have not been conducted in the newborn infant population, large-scale resting-state networks are detectable using fMRI from birth and appear to correspond to adult canonical resting-state and task-response networks 12,13 , suggesting a similar correspondence could exist at this early stage of development. Using a cohort of 18 healthy newborn infants, we replicate large-scale resting-state networks that have previously been characterised in an independent age-matched subset of the dHCP dataset 14 . The amplitudes of spontaneous activity of these networks were then used to predict the infants' cerebral haemodynamic response amplitudes in response to an experimental nociceptive stimulus, in a cross-validated manner.…”

“…Both the noxious-response and resting-state fMRI data were preprocessed using an extended version of the dHCP fMRI preprocessing pipeline 24,41 . The functional data were corrected for motion and distortion using FSL's EDDY 54,55 , which included slice-to-volume motion correction 56 and susceptibility-by-movement distortion correction 57 .…”

“…To define a robust set of core resting-state networks in the infants' resting-state fMRI data, resting-state networks identified in the 18-infant nociception-paradigm dataset were compared to those identified in a subset of the dHCP dataset, which had previously been produced as part of the dHCP 14 . A robust set of core resting-state networks was defined as those replicated across datasets.…”

The intrinsic neonatal hippocampal network: rsfMRI findings

“…Minimal preprocessing for the resting-state fMRI included (Fitzgibbon et al, 2016) distortion correction, motion correction, 2-stage registration of the MB-EPI functional image to the T2 structural image, temporal high-pass filtering (150s high-pass cutoff), and ICA denoising using FSL's FIX (Salimi-Khorshidi, et al, 2014). In addition to this minimal preprocessing, we smoothed the data (Gaussian filter, FWHM = 3mm) across the gray matter, and applied a bandpass filter at 0.009-0.08 Hz.…”