Interactive Dashboards Project

This repository contains my current prototype work for the Brainhack project on interactive MRI surface dashboards.

Project goal

The goal of this prototype is to explore how to build an interactive dashboard for MRI surface-based stats maps. The intended dashboard should allow users to:

• select a hemisphere
• select an ROI from an atlas
• load and visualize the corresponding surface-based stats map on a template brain

This project also explores tool choices for this workflow, especially Marimo, Nilearn, and ipyniivue / Niivue.

What this prototype currently does

The current notebook explores how to:

• load fsaverage cortical surfaces with Nilearn
• read FreeSurfer .annot atlas files with NiBabel
• visualize atlas ROIs on surfaces such as pial_left and infl_left
• create an interactive brain view with nilearn.plotting.view_surf
• display a separate text legend for ROI names
• browse ROI names with a dropdown in Marimo

What currently works

• fsaverage surface loading with Nilearn
• .annot atlas loading with NiBabel
• ROI name extraction from the atlas
• interactive brain visualization with nilearn.plotting.view_surf
• separate text legend for ROI names
• ROI dropdown selection in Marimo

What still needs work

• linking hemisphere and ROI selection to a full stats-map workflow
• improving the overall dashboard layout
• testing or debugging ipyniivue further for mesh-based interactivity
• checking whether hover and click interactions can be used in the intended workflow

Main files

• notebooks/prototype_dashboard.py
  Main Marimo notebook containing the prototype

• data/atlas/lh.HCP-MMP1.annot

• data/atlas/rh.HCP-MMP1.annot
  Atlas files used in the prototype

• assets/hcp-mmp1-atlas_left-hemisphere.png

• assets/hcp-mmp1-atlas_left-pial.png
  Example screenshots from the current prototype

Project structure

interactive-dashboards-project/
├── README.md
├── pyproject.toml
├── uv.lock
├── notebooks/
│   └── prototype_dashboard.py
├── assets/
│   ├── hcp-mmp1-atlas_left-hemisphere.png
│   └── hcp-mmp1-atlas_left-pial.png
└── data/
    └── atlas/
        ├── lh.HCP-MMP1.annot
        └── rh.HCP-MMP1.annot
# interactive-dashboards-project
Create interactive dashboards using ipyniivue and Marimo

## Getting started
1. Install uv
2. Fork/clone repo
3. Install project dependencies `uv sync`
4. Edit file with `uv run marimo edit init_app.py`
5. Explore functionality of Python package ipyniivue
6. Get Glasser atlas and plot on fsaverage high resolution brain