Multicenter MRI data becomes vulnerable to overfitting when the variability caused by differences between scanners is captured by (deep learning, biophysical, or signal representation) models.

Vendor-neutral pulse sequence development is an emerging open-source approach that offers an alternative to relying on proprietary vendor-native sequences and acquisition controllers. I am interested in applying this approach to standardize acquisitions for various MRI applications (primarily qMRI) with the goal of minimizing non-biological variability at the signal source across scanners from different vendors (e.g., Siemens, GE, Philips, and Canon).

I have experience developing vendor-neutral sequences using both RTHawk (JavaScript, C++) and Pulseq (MATLAB, Python) platforms.

🔗 Relevant article in MRM

⭐️ Significance First empirical evidence supporting the use of vendor-neutral acquisitions to reduce measurement variability across scanners from different vendors.

Whether based on MRI signal representations (e.g., Bloch equation that governs a multi-echo spin-echo experiment) or biophysical models (e.g., restricted intracellular diffusion), most qMRI parameter estimation and correction methods are developed and maintained in-house.

To address this challenge, I developed qMRLab, an open-source software package offering a comprehensive suite of qMRI methods for data fitting, simulation, and protocol optimization. qMRLab consolidates diverse qMRI implementations into a single platform, enhancing accessibility through extensive documentation, online executable notebooks, a user-friendly graphical interface, interactive tutorials, and informative blog posts.

🔗 Relevant article in JOSS

⭐️ Significance: The most popular qMRI toolbox on GitHub, with 157 stargazers, standardizing over 24 qMRI methods across 8 different categories.

Navigating a diverse range of open-source toolboxes for image reconstruction, as well as pre- and post-processing is needed to facilitate the practical use of vendor-neutral acquisitions.

These workflows, written in DSL2, are designed so that each step producing a derivative is defined as an independent process, mapped to a corresponding container provided by qMRFlow. By adhering to data standards for both k-space and image data, and leveraging Nextflow’s platform-agnostic executors, these workflows can be seamlessly deployed across cloud environments, high-performance computing (HPC) systems at scale, or workstations on any operating system at the scanner site.

To enhance data accessibility, promote large-scale collaborations, and accelerate progress in any application that could benefit from qMRI, I led the development of qMRI-BIDS, collaborating with over 30 researchers worldwide. The initial version of this BIDS extension standardized the units, metadata, and naming conventions of 18 parametric maps, and since then, other researchers have actively contributed data descriptions for new qMRI methods. This collaborative effort establishes qMRI-BIDS as a common ground for achieving interoperability in qMRI, enabling seamless integration and comparison of data across diverse studies and platforms.

qMRI-BIDS has unified access to thousands of datasets through platforms such as Zenodo, the Open Science Framework, and OpenNeuro.

Additionally, it has inspired enhancements to open-source data converters, such as ezBIDS, which now support converting DICOM images generated by various vendors according to the qMRI-BIDS specifications.

By offering free and standardized access to these datasets, qMRI-BIDS has transformed data sharing practices and promoted equity in the field, enabling researchers worldwide to benefit from valuable resources, regardless of their local infrastructure.

🔗 Relevant article in Scientific Data

• Container Orchestration: Designed and managed a Kubernetes cluster on bare metal to power our BinderHub deployment
• Cloud Infrastructure: Implemented multi-cloud architecture using OpenStack & OpenNebula with Terraform for infrastructure as code
• Web Development: Built full-stack applications using:
  • Ruby on Rails for our editorial management system
  • Flask/Celery/NGINX stack for high-performance API services
  • Modern CI/CD practices with extensive GitHub Actions workflows
• DevOps: Established monitoring, logging, and automated deployment pipelines
• System Design: Architected scalable solutions for compute-intensive scientific workloads

Epimuscular myofascial force transmission (EMFT) is a novel theory on how muscles transmit force through inter- and extramuscular medium, in addition to the myotendinous structures.

In-vivo studies on this theory is challenging due to the lack of non-invasive methods that can encode mechanical information at different joint angles or during muscle contraction.

To address this challenge, I developed a novel MRI method that combines diffusion tensor imaging (DTI) tractography with non-rigid registration of high-resolution anatomical images to quantify strain in the muscle fiber direction. One of the key findings from this work is heteregeneous strain distribution in the muscle fiber direction, which is a hallmark of EMFT effects, as well as an important determinant of muscle range of active force production.

Such comprehensive understanding of in-vivo mechanical interactions is crucial for the development of novel rehabilitation protocols and assistive devices. This knowledge has important implications both in sports science for optimizing athletic performance and in clinical conditions such as spastic cerebral palsy, where altered mechanical tissue properties significantly impact movement.

My recent work explores how qMRI can reveal key microstructural properties of muscle tissue, including extracellular volume fraction and collagen content. This multi-parametric approach provides complementary insights into the relationship between muscle structure and function.

• Submaximal plantarflexion
• Passive knee extension
• Robustness analysis