Does Nonverbal IQ Modulate the Relation Between Phonological Awareness and Reading Skill? Harmonization of Multi-Site fMRI-data

[hsuanwei-chen]

Title

Multi-site fMRI harmonization

Leaders

Isaac Chen @hsuanwei-chen

Collaborators

James R. Booth (james.booth@vanderbilt.edu)

Project Description

Research shows that reading skill (i.e. accuracy of reading words aloud) and phonological awareness (i.e. sensitivity to the sound structure of language) are bi-directionally related, but we do not know how non-verbal IQ (i.e. sensitivity to stimulus patterns) influences this relationship. The proposed project seeks to examine how nonverbal IQ modulates the effect of (1) behavioral differences in phonological awareness on the neural basis of reading skill and (2) behavioral differences in reading skill on the neural basis of phonological awareness.

Neuroimaging data for this project were collected across various sites using different scanners. Pooling fMRI data across multiple sites is critical for improving the generalizability of findings to a diverse population. However, multi-site fMRI data are often affected by non-biological variability, attributable to differences in scanner manufacturers, non-standardized imaging acquisition parameters, or other intrinsic factors. This source of variability may lead to limited statistical power or even result in spurious findings. The objective for Brainhack Vanderbilt 2025 is to develop a harmonization pipeline designed to reliably remove between-scanner variability and ensure reproducibility for downstream analyses.

Link to project repository/sources

https://github.com/hsuanwei-chen/task-VisRhyme_harmonization

Concerete Goals with Specific Tasks for Brainhack Vanderbilt 2025

1. Remove any identified site effects from task-based fMRI data using a combination of harmonization techniques, such as RAVEL and/or ComBAT.
2. Evaluate how quality of neuroimaging data (i.e. level of motion contamination) affects the effectiveness of harmonization.
3. Determine whether the preserved biologically relevant information can replicate previous findings on the neural basis of phonological awareness and reading skill. Extend this literature to determine how these relations are moderated by nonverbal IQ.
4. Create an automatic pipeline to preprocess multi-site fMRI data and derive reliable brain activity metrics (i.e. voxel activation).

Good first issues

1. issue one: Browse through the OpenNeuro datasets and learn about the differences in imaging acquisition parameters.

2. issue two: Read through the documentation for RAVEL and ComBAT.

Skills

1. A basic understanding of fMRI and how fMRI data is preprocessed.
2. Intermediate knowledge level of R and/or MATLAB preferred.
3. Good communication skills, such as active listening and the ability to reach out for help if needed.

Onboarding documentation

https://github.com/Jfortin1/RAVEL
https://github.com/Jfortin1/ComBatHarmonization

What will participants learn?

Participants will learn to identify factors that may bias results in aggregated fMRI data, to develop a pipeline to remove cross-scanner noise, and to reliably combine multi-site fMRI data for potential future retrospective and prospective studies.

Public data to use

https://openneuro.org/datasets/ds001486/versions/1.3.1
https://openneuro.org/datasets/ds001894/versions/1.4.2
https://openneuro.org/datasets/ds002236/versions/1.1.1

Number of collaborators

3

Credit to collaborators

Collaborators contributing to methods have the potential to be included as co-authors in future publications.

Image

Project Summary

Reading skill and phonological awareness are bi-directionally related, but the role of non-verbal IQ in this relationship is unclear. Multi-site task-based fMRI was collected in children, and this project aims to develop a harmonization pipeline to reliably remove between-scanner variability.

Type

method_development, pipeline_development

Development status

0_concept_no_content

Topic

reproducible_scientific_methods, other

Tools

BIDS, SPM

Programming language

Matlab, R

Modalities

fMRI

Git skills

1_commit_push

Anything else?

We are open to try other available harmonization techniques and are not limited to the programming languages above. Please reach out to us if you have any innovative ideas or to share your expertise!

See below for data descriptor articles for data used in this project:

LytleBitan.DB.2020.pdf
LytleMcNorgan.SD.2019.pdf
PellicioniLytle.SD.2019.pdf

See references for how RAVEL and ComBAT have been used in fMRI data:

Eshaghzadeh Torbati, M., Minhas, D. S., Ahmad, G., O’Connor, E. E., Muschelli, J., Laymon, C. M., Yang, Z., Cohen, A. D., Aizenstein, H. J., Klunk, W. E., Christian, B. T., Hwang, S. J., Crainiceanu, C. M., & Tudorascu, D. L. (2021). A multi-scanner neuroimaging data harmonization using RAVEL and ComBat. NeuroImage, 245, 118703. https://doi.org/10.1016/j.neuroimage.2021.118703

Yu, M., Linn, K. A., Cook, P. A., Phillips, M. L., McInnis, M., Fava, M., Trivedi, M. H., Weissman, M. M., Shinohara, R. T., & Sheline, Y. I. (2018). Statistical harmonization corrects site effects in functional connectivity measurements from multi‐site fMRI data. Human Brain Mapping, 39(11), 4213–4227. https://doi.org/10.1002/hbm.24241

Things to do after the project is submitted and ready to review.

• Add a comment below the main post of your issue saying: Hi @brainhack-vandy/project-monitors my project is ready!

[hsuanwei-chen]

Hi @brainhack-vandy/project-monitors my project is ready!