Multi-site Diffusion MRI Data Harmonization of Human Connectome Project Lifespan and Disease Studies
Poster No:
2227
Submission Type:
Abstract Submission
Authors:
Suheyla Cetin-Karayumak1, Ryan Zurrin1, Fan Zhang2, Kang Ik Cho3, Lauren O'Donnell4, Yogesh Rathi4
Institutions:
1Harvard Medical School and Brigham and Women's Hospital, Boston, MA, 2University of Electronic Science and Technology of China, Chengdu, Sichuan, 3Department of Psychiatry, Psychiatry Neuroimaging Laboratory, Brigham and Women's Hospital, Harvard, Boston, MA, 4Harvard Medical School, Boston, MA
First Author:
Co-Author(s):
Kang Ik Cho
Department of Psychiatry, Psychiatry Neuroimaging Laboratory, Brigham and Women's Hospital, Harvard
Boston, MA
Department of Psychiatry, Psychiatry Neuroimaging Laboratory, Brigham and Women's Hospital, Harvard
Boston, MA
Introduction:
The Human Connectome Project (HCP) is a multi-site neuroimaging initiative that aims to study the connections of the human brain [1–3]. The HCP lifespan project explores how brain connectivity changes during typical development and aging, while the HCP disease projects explore how brain connectivity changes in various neurological and psychiatric disorders. However, combining neuroimaging data from multiple sites requires careful handling of scanner-related measurement bias before further analysis.
Despite consistent imaging protocols across sites in the HCP, intrinsic hardware variabilities and software versions can introduce scanner-related bias [4,5]. This bias is particularly significant in diffusion MRI (dMRI), reducing statistical power and reliability of multi-site dMRI data analysis. Harmonization is an image processing technique that standardizes dMRI datasets from different sources, enabling pooling of data for joint analysis [4,6]. This is especially important for neuroimaging studies of psychiatric disorders, where effect sizes associated with psychiatric disorders are often small [7]. This study summarizes our harmonization efforts for the HCP lifespan and disease projects.
Despite consistent imaging protocols across sites in the HCP, intrinsic hardware variabilities and software versions can introduce scanner-related bias [4,5]. This bias is particularly significant in diffusion MRI (dMRI), reducing statistical power and reliability of multi-site dMRI data analysis. Harmonization is an image processing technique that standardizes dMRI datasets from different sources, enabling pooling of data for joint analysis [4,6]. This is especially important for neuroimaging studies of psychiatric disorders, where effect sizes associated with psychiatric disorders are often small [7]. This study summarizes our harmonization efforts for the HCP lifespan and disease projects.
Methods:
a) Dataset and dMRI data preprocessing: We sourced unprocessed dMRI data from the NIMH Data Archive (NDA), which included HCP lifespan and disease datasets. Figure 1 provides a participant overview by study. Diffusion MRI data was acquired on 11 scanners: 3T Siemens Prisma or Prisma fit scanners with similar acquisition parameters. We applied the same dMRI data preprocessing steps in all datasets using FSL's eddy and topup tools, along with a deep learning-based brain masking tool [8], to prepare the data for harmonization.
b) DMRI data harmonization: We applied our retrospective harmonization algorithm, which aligns dMRI data from different scanners directly on the dMRI data by leveraging unique tissue properties using Rotation-Invariant-Spherical-Harmonics (RISH) features [4,9]. We selected 30 healthy subjects from each scanner, matched by age, sex, and IQ, to one reference dataset to which all other scanner data were harmonized. We chose the reference device based on its wide range of characteristics (e.g., age, sex) and sample size, which allowed us to obtain representative samples across all the different scanners. We used the RISH features of these subjects to create templates representing scanner differences. After determining these mappings, we applied these templates to the full dMRI dataset for harmonization. More details of these steps can be found in [4,6].
c) Harmonization Performance: Various dMRI measures, such as Return To Origin Probability (RTOP) and Fractional-Anisotropy (FA), were calculated pre- and post-harmonization and compared with the reference dataset as a baseline. We assessed the harmonization performance in 30 matched subjects. Averages of the measures were calculated over the whole brain white matter skeleton and 42 white matter regions of interest [10]. We first compared the original and harmonized datasets to the reference dataset using unpaired t-tests. To further verify the harmonization, we conducted unpaired t-tests using 30 newly matched subjects, which were not part of creating templates.
b) DMRI data harmonization: We applied our retrospective harmonization algorithm, which aligns dMRI data from different scanners directly on the dMRI data by leveraging unique tissue properties using Rotation-Invariant-Spherical-Harmonics (RISH) features [4,9]. We selected 30 healthy subjects from each scanner, matched by age, sex, and IQ, to one reference dataset to which all other scanner data were harmonized. We chose the reference device based on its wide range of characteristics (e.g., age, sex) and sample size, which allowed us to obtain representative samples across all the different scanners. We used the RISH features of these subjects to create templates representing scanner differences. After determining these mappings, we applied these templates to the full dMRI dataset for harmonization. More details of these steps can be found in [4,6].
c) Harmonization Performance: Various dMRI measures, such as Return To Origin Probability (RTOP) and Fractional-Anisotropy (FA), were calculated pre- and post-harmonization and compared with the reference dataset as a baseline. We assessed the harmonization performance in 30 matched subjects. Averages of the measures were calculated over the whole brain white matter skeleton and 42 white matter regions of interest [10]. We first compared the original and harmonized datasets to the reference dataset using unpaired t-tests. To further verify the harmonization, we conducted unpaired t-tests using 30 newly matched subjects, which were not part of creating templates.
Results:
We selected the scanner with deviceid=166007 as the reference due to its extensive collection of healthy controls and diverse age distribution (N=30) in the template creation process. The harmonization's effectiveness for all datasets was evidenced by the elimination of any existing statistical differences across datasets during harmonization (before harmonization: p<0.01; after harmonization: p>0.3).
Conclusions:
The harmonized HCP dMRI data of 2545 subjects will be available in the NDA for large-scale analysis. The enhancement in statistical power will aid in better characterizing connectomes in individuals with specific disorders compared to healthy controls and in identifying neuroanatomical changes related to each disease.
Modeling and Analysis Methods:
Diffusion MRI Modeling and Analysis 2
Neuroinformatics and Data Sharing:
Databasing and Data Sharing 1
Keywords:
MRI
Open Data
White Matter
WHITE MATTER IMAGING - DTI, HARDI, DSI, ETC
Other - Harmonization, Human Connectome Project
1|2Indicates the priority used for review
Provide references using author date format
2. Harms, M. P. (2018), Extending the Human Connectome Project across ages: Imaging protocols for the Lifespan Development and Aging projects. Neuroimage 183, 972–984.
3. Somerville, L. H. (2018), The Lifespan Human Connectome Project in Development: A large-scale study of brain connectivity development in 5-21 year olds. Neuroimage 183, 456–468.
4. Cetin Karayumak, S. (2019), Retrospective harmonization of multi-site diffusion MRI data acquired with different acquisition parameters. Neuroimage 184, 180–200.
5. Ning, L. (2020), Cross-scanner and cross-protocol multi-shell diffusion MRI data harmonization: Algorithms and results. Neuroimage 221, 117128.
6. Cetin-Karayumak, S. (2023), Harmonized diffusion MRI data and white matter measures from the Adolescent Brain Cognitive Development Study. bioRxiv doi:10.1101/2023.04.04.535587.
7. Cetin-Karayumak, S. (2020), White matter abnormalities across the lifespan of schizophrenia: a harmonized multi-site diffusion MRI study. Mol. Psychiatry 25, 3208–3219.
8. Palanivelu S*, Cetin Karayumak S* (2020), CNN based diffusion MRI brain segmentation tool. doi:10.5281/zenodo.3665739.
9. Cetin Karayumak S*, Billah T* (2019), Multi-site Diffusion MRI Harmonization. (2019) doi:10.5281/zenodo.2584275.
10. Varentsova, A. (2014), Development of a high angular resolution diffusion imaging human brain template. Neuroimage 91, 177–186.