Thuy Thanh Dao

We developed a multi-shell diffusion-weighted imaging approach for separate ventricular and subarachnoid cerebrospinal fluid thermometry. The approach uses measured image noise to assess agreement among temperature estimates. In 12 participants, temperature was estimated from four non-zero diffusion weightings. For each voxel, disagreement between these estimates was compared with that expected from image noise, accounting for their shared reference image. Voxels meeting this consistency criterion were retained for regional analysis. Mean diffusion-derived temperature (± SD) was 38.4 ± 0.4 °C in the lateral ventricles and 36.0 ± 1.1 °C in the subarachnoid space. Ventricular estimates were consistently higher than subarachnoid estimates across all participants, with a mean difference of 2.4 ± 1.2 °C (95% confidence interval 1.6–3.1 °C). This difference persisted under alternative noise models and temperature estimators. Subarachnoid estimates increased with retained voxel count; in a post hoc analysis of the six participants with the largest retained counts, ventricular estimates remained higher by 1.6 ± 0.7 °C. Absolute temperature accuracy and the physiological origin of the regional difference require independent validation. This approach provides an internal consistency check for CSF thermometry and supports separate assessment of ventricular and subarachnoid temperatures.

Interactive visualization is essential for collaborative analysis of complex neuroimaging data. Modern workflows span multiple computing environments including notebooks, integrated development environments, and standalone applications. All these workflows rely on visualization for quality control, exploration, and interpretation. However, visualization tools are typically siloed, forcing users to switch between environments and disrupting interactive analysis. Here, we address this challenge by integrating a shared visualization module directly into these diverse environments, providing a consistent visualization experience across formats, representations, and usage contexts. We developed a generalizable ecosystem of open-source tools built around the NiiVue rendering engine, supporting four primary use cases. First, we provide integrated file viewer capability for VS Code and JupyterLab. Second, we showcase three different architectures for standalone desktop applications. Third, we integrate NiiVue for Python notebooks, Julia, and R. Finally, we integrate NiiVue into traditional web applications. This unified ecosystem enables immediate, in-place inspection of neuroimaging data across the majority of modern computing platforms, eliminating the need to export data to external viewers. The tools support interactive inspection, dashboard creation, and the development of custom analysis applications within users’ existing workflows. The components are validated by automated continuous-integration testing, including unit tests and browser visualization tests against stored reference targets. By embedding visualization directly into commonly used computing environments, this ecosystem lowers the barrier to interactive data exploration and streamlines neuroimaging workflows. Visualization becomes an integral component of automated, reproducible, and shareable analysis pipelines, supporting more efficient collaboration and interpretation of complex neuroimaging data. By sharing a common rendering core, file format support, and interaction model, users can transfer visualization skills seamlessly across environments.

Neuroimaging education presents both conceptual and technical challenges: one must master complex foundational principles while navigating technical requirements including software installation, dependency management, and cross-platform compatibility. These tasks are valuable learning opportunities, but they can be time-consuming, which may shift focus away from learning about core neuroimaging concepts. While neuroimaging education benefits from many established resources, interactive notebooks can further integrate instruction with execution, providing a hands-on complement to existing educational materials. Here, we showcase NeurodeskEDU, an emerging learning resource developed using Neurodesk, an open-source, containerised neuroimaging platform, which enables neuroimaging workflows to be executed in a fully reproducible fashion through interactive Jupyter Notebooks. The notebooks integrate code, visualisations, narrative documentation, and direct access to open datasets into analysis workflows. Containerised software delivery reduces installation overhead while ensuring consistent execution across computing platforms. Community contributions undergo automated testing via GitHub Actions to ensure notebook quality and functionality, while persistent DOIs establish notebooks as formally citable scholarly outputs. This new community learning resource provides multiple deployment modes: zero-install cloud execution, downloadable notebooks for local use, static HTML documentation, and interactive slideshows for teaching scenarios. Through three use cases, we demonstrate how such interactive notebooks consolidate fragmented educational workflows, enable seamless multi-tool integration, and support scalable deployment from workshops to self-directed learning. By establishing educational materials as scholarly outputs with systematic quality assurance and version control, this work addresses ongoing challenges in computational publishing: maintaining reproducibility as software evolves, facilitating community contributions, and providing formal academic recognition for educational scholarship.

Neuroimaging education presents both conceptual and technical challenges: one must master complex foundational principles while navigating technical requirements including software installation, dependency management, and cross-platform compatibility. These tasks are valuable learning opportunities, but they can be time-consuming, which may shift focus away from learning about core neuroimaging concepts. While neuroimaging education benefits from many established resources, interactive notebooks can further integrate instruction with execution, providing a hands-on complement to existing educational materials. Here, we showcase NeurodeskEDU, an emerging learning resource developed using Neurodesk, an open-source, containerised neuroimaging platform, which enables neuroimaging workflows to be executed in a fully reproducible fashion through interactive Jupyter notebooks. The notebooks integrate code, visualisations, narrative documentation, and direct access to open datasets into analysis workflows. Containerised software delivery reduces installation overhead while ensuring consistent execution across computing platforms. Community contributions undergo automated testing via GitHub Actions to ensure notebook quality and functionality, while persistent DOIs establish notebooks as formally citable scholarly outputs. This new community learning resource provides multiple deployment modes: zero-install cloud execution, downloadable notebooks for local use, static HTML documentation, and interactive slideshows for teaching scenarios. Through three use cases, we demonstrate how such interactive notebooks consolidate fragmented educational workflows, enable seamless multi-tool integration, and support scalable deployment from workshops to self-directed learning. By establishing educational materials as scholarly outputs with systematic quality assurance and version control, this work addresses ongoing challenges in computational publishing: maintaining reproducibility as software evolves, facilitating community contributions, and providing formal academic recognition for educational scholarship.

Accurate magnetic resonance imaging (MRI) segmentation is crucial for clinical decision-making, but remains labor-intensive when performed manually. Convolutional neural network (CNN)-based methods can be accurate and efficient but often generalize poorly to MRI’s variable contrast, intensity inhomogeneity, and sequences. Although the transformer-based Segment Anything Model (SAM) has demonstrated remarkable generalizability in natural images, existing adaptations often treat MRI as another imaging modality, overlooking these modality-specific challenges. We present SAMRI, an MRI-specialized SAM trained and validated on 1.1 million labeled MR slices spanning whole-body organs and pathologies. We demonstrate that SAM can be effectively adapted to MRI by fine-tuning its mask decoder using a two-stage strategy, reducing training time by 94% and trainable parameters by 96% compared to full-model retraining. Across diverse MRI segmentation tasks, SAMRI achieves a mean Dice of 0.87, delivering state-of-the-art accuracy across anatomical regions and robust generalization on unseen structures, particularly small clinically important structures. In addition, we provide a complete training-to-inference pipeline and a user-friendly local graphical interface that enables interactive application of pretrained SAMRI models on standard machines, facilitating practical deployment for real-world MRI segmentation.

Neuroimaging research has increasingly embraced open science practices, including the use and sharing of publicly available datasets. However, researchers still face significant challenges in accessing and preparing such datasets, from navigating through different storage systems and tools to complying with data privacy regulations. Moreover, the varied dependencies used by these tools often hinder reproducibility and accessibility, which in turn can negatively impact the reliability of analyses. In this paper, we show how Neurodesk streamlines both the use of open data and the preparation of datasets for sharing publicly. By integrating data access and standardized processing, Neurodesk maximizes the utility of open data initiatives while promoting accessibility and equity in the scientific community. We demonstrate Neurodesk’s capabilities through real-world use cases and outline ongoing developments toward secure, scanner-integrated data processing, further supporting privacy-preserving collaboration in neuroscience.

Computational tools are essential to neuroimaging research, yet their development, use, and citation are inconsistently addressed in current scientific publishing frameworks. Despite increasing adoption of the FAIR and FAIR4RS principles, challenges remain in ensuring the reproducibility, accessibility, and attribution of research software. This article presents how Neurodesk, an open-source, community-driven platform improves the transparency, portability, and sustainability of neuroimaging workflows. Neurodesk uses containerisation to encapsulate complete software environments that are portable across operating systems and computing infrastructures. Each container and workflow can be versioned and assigned a persistent DOI to support formal citation and long-term access. Through concrete use cases, we illustrate how Neurodesk facilitates code review, supports methodological extension, and fosters sustainable tool development. We argue that platforms like Neurodesk provide a practical infrastructure for enacting cultural and technical shifts in publishing, making research software a primary research output. By integrating attribution, standardisation, and reproducibility into everyday research workflows, Neurodesk exemplifies how future scientific publishing can more effectively recognise and preserve computational contributions.

Neuroimaging research requires purpose-built analysis software, which is challenging to install and may produce different results across computing environments. The community-oriented, open-source Neurodesk platform (https://www.neurodesk.org/) harnesses a comprehensive and growing suite of neuroimaging software containers. Neurodesk includes a browser-accessible virtual desktop, command-line interface and computational notebook compatibility, allowing for accessible, flexible, portable and fully reproducible neuroimaging analysis on personal workstations, high-performance computers and the cloud.