An Optimised Surface Projection Pipeline for Enhanced Analysis of Fetal Brain Connectivity
Poster No:
1270
Submission Type:
Abstract Submission
Authors:
Pablo Prieto Roca1, Logan Williams1,2, Sean Fitzgibbon3, Vanessa Kyriakopoulou2, Alice Davidson2, Alena Uus1,2, Antonis Makropoulos1, Andreas Schuh4, Lucilio Cordero-Grande5, Emer Hughes2, Anthony Price2, Eugene Duff6,3, Tomoki Arichi2,7,8, A. Edwards2,7,9, Daniel Rueckert10,4, Stephen Smith3, Joseph Hajnal2, Emma Robinson1,2, Vyacheslav Karolis2
Institutions:
1Biomedical Engineering Department, King's College London, London, United Kingdom, 2Centre for the Developing Brain, King's College London, London, United Kingdom, 3Wellcome Centre for Integrative Neuroimaging, FMRIB, University of Oxford, Oxford, Oxfordshire, United Kingdom, 4Department of Computing, Imperial College London, London, United Kingdom, 5Universidad Politécnica de Madrid & CIBER-BBN, Madrid, Spain, 6UK Dementia Research Institute, Department of Brain Sciences, Imperial College London, London, United Kingdom, 7MRC Centre for Neurodevelopmental Disorders, King’s College London, London, United Kingdom, 8Paediatric Neurosciences, Evelina London Children’s Hospital, Guy’s and St Thomas’ NHS Foundation Trust, London, United Kingdom, 9Neonatal Intensive Care Unit, Evelina London Children’s Hospital, London, United Kingdom, 10Klinikum rechts der Isar, Technical University of Munich, Munich, Bavaria, Germany
First Author:
Co-Author(s):
Logan Williams
Biomedical Engineering Department, King's College London|Centre for the Developing Brain, King's College London
London, United Kingdom|London, United Kingdom
Biomedical Engineering Department, King's College London|Centre for the Developing Brain, King's College London
London, United Kingdom|London, United Kingdom
Sean Fitzgibbon
Wellcome Centre for Integrative Neuroimaging, FMRIB, University of Oxford
Oxford, Oxfordshire, United Kingdom
Wellcome Centre for Integrative Neuroimaging, FMRIB, University of Oxford
Oxford, Oxfordshire, United Kingdom
Vanessa Kyriakopoulou, Dr
Centre for the Developing Brain, King's College London
London, United Kingdom
Centre for the Developing Brain, King's College London
London, United Kingdom
Alena Uus
Biomedical Engineering Department, King's College London|Centre for the Developing Brain, King's College London
London, United Kingdom|London, United Kingdom
Biomedical Engineering Department, King's College London|Centre for the Developing Brain, King's College London
London, United Kingdom|London, United Kingdom
Eugene Duff
UK Dementia Research Institute, Department of Brain Sciences, Imperial College London|Wellcome Centre for Integrative Neuroimaging, FMRIB, University of Oxford
London, United Kingdom|Oxford, Oxfordshire, United Kingdom
UK Dementia Research Institute, Department of Brain Sciences, Imperial College London|Wellcome Centre for Integrative Neuroimaging, FMRIB, University of Oxford
London, United Kingdom|Oxford, Oxfordshire, United Kingdom
Tomoki Arichi
Centre for the Developing Brain, King's College London|MRC Centre for Neurodevelopmental Disorders, King’s College London|Paediatric Neurosciences, Evelina London Children’s Hospital, Guy’s and St Thomas’ NHS Foundation Trust
London, United Kingdom|London, United Kingdom|London, United Kingdom
Centre for the Developing Brain, King's College London|MRC Centre for Neurodevelopmental Disorders, King’s College London|Paediatric Neurosciences, Evelina London Children’s Hospital, Guy’s and St Thomas’ NHS Foundation Trust
London, United Kingdom|London, United Kingdom|London, United Kingdom
A. Edwards
Centre for the Developing Brain, King's College London|MRC Centre for Neurodevelopmental Disorders, King’s College London|Neonatal Intensive Care Unit, Evelina London Children’s Hospital
London, United Kingdom|London, United Kingdom|London, United Kingdom
Centre for the Developing Brain, King's College London|MRC Centre for Neurodevelopmental Disorders, King’s College London|Neonatal Intensive Care Unit, Evelina London Children’s Hospital
London, United Kingdom|London, United Kingdom|London, United Kingdom
Daniel Rueckert
Klinikum rechts der Isar, Technical University of Munich|Department of Computing, Imperial College London
Munich, Bavaria, Germany|London, United Kingdom
Klinikum rechts der Isar, Technical University of Munich|Department of Computing, Imperial College London
Munich, Bavaria, Germany|London, United Kingdom
Stephen Smith
Wellcome Centre for Integrative Neuroimaging, FMRIB, University of Oxford
Oxford, Oxfordshire, United Kingdom
Wellcome Centre for Integrative Neuroimaging, FMRIB, University of Oxford
Oxford, Oxfordshire, United Kingdom
Emma Robinson, Dr
Biomedical Engineering Department, King's College London|Centre for the Developing Brain, King's College London
London, United Kingdom|London, United Kingdom
Biomedical Engineering Department, King's College London|Centre for the Developing Brain, King's College London
London, United Kingdom|London, United Kingdom
Introduction:
The human fetal period is a key time for the formation of the functional brain connectome [1,2]. Detailed exploration of this process is now possible with the availability of the open-access dataset from the Developing Human Connectome Project (dHCP) and advancements in fMRI in-utero methodology [3]. However, studying this is particularly challenging, partly due to the rapidly changing shape of the fetal brain, which complicates the alignment of cortical areas across subjects and different ages. In this context, surface-based methods markedly improve the accuracy of co-localizing cortical areas compared to volume-based methods [4]. Here we describe an optimized pipeline for projecting fetal volumetric functional data onto surfaces, enabling robust population-level inference.
Methods:
Resting-state in-utero dHCP fMRI data from 164 healthy developing fetuses (87 males and 77 females, age range of 24.5-38.5 gestational weeks) were acquired using a 3T Philips Achieva system, and underwent dynamic geometric and slice-to-volume motion corrections, and temporal denoising [1]. A recently developed spatiotemporal surface fetal brain atlas [5], which encompasses individual templates for each gestational week, ranging from 21 to 36 weeks, was used as a template space.
Our projection pipeline, illustrated in Figure 1, represents a modified version of the HCP pipeline for volume-to-surface mapping and incorporates tools from the Connectome Workbench software and FSL [6,7]. The process involves 5 steps: (1) the structural surface is mapped into native functional space; (2) individual functional data features are then mapped onto the surface through a ribbon-constrained volume-to-surface mapping process; (3) Surface matching between the individual surface and the 36th week of the spatio-temporal fetal surface template [5] was then calculated using Multimodal Surface Matching (MSM) [8] in two stages. First, (3.1) individual surfaces were mapped to the "nearest" template (from the closest gestational age); then (3.2) template-to-template registration was performed between consecutive weekly templates (e.g. from 25 to 26 weeks, and 26 to 27 weeks etc); finally (3), these mappings were concatenated together in order to map all individuals to the 36 week template in which the group-level analysis was done, the latter serves two primary purposes: 1) It allows a gradual transition between spaces, through expected cortical development patterns. This aspect is particularly emphasised by the kernel-weighted averaging across subjects utilised for the template construction [5]; 2) It enables the transition from any specific week's template to a desired group-level template in a single interpolation step, simplifying the workflow and enhancing efficiency.
To demonstrate the utility of our pipeline, we performed a group-level ICA with 25-dimensional factorization using FSL MELODIC [10].
Our projection pipeline, illustrated in Figure 1, represents a modified version of the HCP pipeline for volume-to-surface mapping and incorporates tools from the Connectome Workbench software and FSL [6,7]. The process involves 5 steps: (1) the structural surface is mapped into native functional space; (2) individual functional data features are then mapped onto the surface through a ribbon-constrained volume-to-surface mapping process; (3) Surface matching between the individual surface and the 36th week of the spatio-temporal fetal surface template [5] was then calculated using Multimodal Surface Matching (MSM) [8] in two stages. First, (3.1) individual surfaces were mapped to the "nearest" template (from the closest gestational age); then (3.2) template-to-template registration was performed between consecutive weekly templates (e.g. from 25 to 26 weeks, and 26 to 27 weeks etc); finally (3), these mappings were concatenated together in order to map all individuals to the 36 week template in which the group-level analysis was done, the latter serves two primary purposes: 1) It allows a gradual transition between spaces, through expected cortical development patterns. This aspect is particularly emphasised by the kernel-weighted averaging across subjects utilised for the template construction [5]; 2) It enables the transition from any specific week's template to a desired group-level template in a single interpolation step, simplifying the workflow and enhancing efficiency.
To demonstrate the utility of our pipeline, we performed a group-level ICA with 25-dimensional factorization using FSL MELODIC [10].
Results:
Figure 2 presents the outcomes of the Group-ICA with our pipeline, the first-ever performed on the fetal cortical surface. ICA components were excluded if upon visual inspection they were likely to be related to noise which was non-neural in origin. Although derived single components lack distributed network properties, several pairs of maps showed a distinctive interhemispheric symmetry, suggesting a coordinated emergence of activity across hemispheres.
Conclusions:
Our study's projection pipeline contributes to the field of fetal brain imaging by the mapping of volumetric functional data onto the cortical surface for subsequent application of a surface-based registration to template framework. This method has the potential of improving the analysis of in utero brain activity, offering insights into functional connectivity during the critically important fetal period. Such insights could be valuable for prenatal health and developmental neuroscience research, although further studies are required to fully understand their implications.
Lifespan Development:
Lifespan Development Other 1
Modeling and Analysis Methods:
fMRI Connectivity and Network Modeling
Image Registration and Computational Anatomy 2
Task-Independent and Resting-State Analysis
Keywords:
Data analysis
Data Registration
Development
FUNCTIONAL MRI
Spatial Warping
Statistical Methods
Structures
Workflows
1|2Indicates the priority used for review
Provide references using author date format
2. Moriah E. Thomason et al. “Cross-hemispheric functional connectivity in the human
fetal brain”. In: Science Translational Medicine 5.173 (Feb. 2013). issn: 19466234. doi:
10.1126/scitranslmed.300497
3. The Developing Human Connectome Project (dHCP). (2021). Approaches to the in-utero fMRI denoising in the Developing Human Connectome Project (dHCP). In: Organisation for Human Brain Mapping Annual Meeting, Vol. 2112.
4. Coalson, T. S., Van Essen, D. C., & Glasser, M. F. (2018). The impact of traditional neuroimaging methods on the spatial localization of cortical areas. Proceedings of the National Academy of Sciences of the United States of America, 115(27), E6356–E6365. doi: 10.1073/pnas.1801582115. ISSN: 10916490.
5. Karolis, V. et al. (2023). Developing Human Connectome Project spatio-temporal surface atlas of the fetal brain. doi: 10.12751/g-node.qj5hs7.
6. Human Connectome Project. (2021). Connectome Workbench. Available at: https://www.humanconnectome.org/software/workbench-command.
7. Jenkinson, M. et al. (2012). FSL. NeuroImage, 62(2), 782–790. doi: 10.1016/j.neuroimage.2011.09.015. ISSN: 10538119. Available at: https://linkinghub.elsevier.com/retrieve/pii/S1053811911010603.
8. Robinson, E. C. et al. (2018). Multimodal surface matching with higher-order smoothness constraints. NeuroImage, 167, 453–465. doi: 10.1016/j.neuroimage.2017.10.037. ISSN: 10959572.
9. Logan Z.J. Williams et al. “Structural and functional asymmetry of the neonatal cerebral
cortex”. In: Nature Human Behaviour 7.6 (June 2023), pp. 942–955. issn: 23973374. doi:
10.1038/s41562-023-01542-8.
10. Beckmann, C. F., & Smith, S. M. (2004). Probabilistic Independent Component Analysis for Functional Magnetic Resonance Imaging. IEEE Transactions on Medical Imaging, 23(2), 137–152. doi: 10.1109/TMI.2003.822821. ISSN: 02780062.