BrainRing: Interactive FC Visualizer

• BrainRing is a free, open-source tool that visualizes brain functional connectivity using interactive chord diagrams with built-in atlas support.
• It streamlines workflows by integrating data import, real-time SVG rendering, and manual edge editing into a single, installation-free HTML file.
• The tool delivers publication-ready figures with customizable styling and multilingual support, enhancing both usability and accessibility.

BrainRing is a free, open-source, browser-based interactive tool for visualizing brain functional connectivity (FC) as interactive chord diagrams. It was introduced to remove a practical bottleneck in neuroimaging figure production: existing solutions such as Circos and BrainNet Viewer either require complex configuration files, scripting, or proprietary software environments, whereas BrainRing places the full workflow in a single HTML file that runs directly in a modern web browser, with no installation, backend server, or programming knowledge required (Fan et al., 28 Mar 2026). The tool is designed for fast, editable, publication-ready FC visualization, rendering regions as arcs on a circle and connectivity as curved links inside the ring so that anatomical organization and network structure can be inspected in one compact representation.

1. Purpose and design rationale

BrainRing is positioned as a solution for brain functional connectivity visualization. In the underlying formulation, FC analysis is central to understanding how distributed brain regions coordinate activity, and chord diagrams are treated as an intuitive representation of connectivity among brain regions on a circular ring (Fan et al., 28 Mar 2026). The motivation for the software is explicitly practical rather than theoretical: the paper identifies friction in existing workflows for producing publication-quality connectivity figures.

The contrast with prior tools is central to its rationale. Circos is described as powerful but originally built for comparative genomics, so applying it to brain data requires domain-specific configuration files, custom karyotypes, and iterative script-based rendering. BrainNet Viewer is described as focusing on 3D brain surfaces in MATLAB, requiring a licensed MATLAB environment, not supporting chord diagrams, and offering limited interactive control. Custom Python or R scripts are described as flexible but as requiring substantial programming effort and as typically producing static outputs that must be regenerated after each tweak (Fan et al., 28 Mar 2026).

From that problem definition, the software adopts four stated design principles: Zero installation, Real-time interaction, Atlas-aware design, and Publication-ready output. This combination suggests that BrainRing is best understood not merely as a visualization script or figure template, but as a specialized FC diagram environment whose main contribution is to collapse preparation, editing, rendering, and export into a single interactive workflow.

2. Software architecture and interaction model

BrainRing is implemented as a self-contained single-file HTML app of roughly 185 KB, using D3.js v7 for SVG rendering. Atlas data, color schemes, internationalization dictionaries, and program logic are embedded directly in the file, which allows the tool to run fully offline (Fan et al., 28 Mar 2026). The absence of a backend server is therefore not only a deployment property but also a computational one: the complete rendering and interaction loop is contained within the browser.

The rendering model is reactive. When any parameter changes, BrainRing re-renders the SVG chord diagram, typically in under 50 ms on modern hardware. The paper emphasizes that there is no “run” button; every change triggers immediate re-rendering. In practical terms, this shifts the workflow away from batch-style figure generation toward continuous interactive adjustment.

The interface is divided into three panels. The left panel contains atlas selection, label configuration, arc color mode, separator settings, manual edge input, click-to-connect mode, data import, region reorder, and hierarchical browser. The center panel contains the live chord diagram canvas. The right panel contains theme toggle, color presets, background and color-depth controls, edge filtering, edge color management, and sliders for layout and typography parameters (Fan et al., 28 Mar 2026). This panel structure reflects the paper’s emphasis on direct manipulation: data definition, structural organization, rendering, and stylistic control are exposed simultaneously rather than serialized into separate preprocessing and plotting stages.

3. Built-in atlas support and atlas-aware organization

A defining feature of BrainRing is its built-in support for 8 widely-used brain atlases, a capability presented as one of its main distinctions from Circos, which has no built-in brain atlas support (Fan et al., 28 Mar 2026). The supported atlases include both anatomical and functional parcellations.

Atlas	Category	Hierarchy
Brainnetome 246	anatomical	Lobe $\rightarrow$ Gyrus $\rightarrow$ Subregion
AAL-90	anatomical	Lobe $\rightarrow$ Region
AAL-116	anatomical	Lobe $\rightarrow$ Region
Schaefer-100	functional	Yeo7 $\rightarrow$ Parcel
Schaefer-200	functional	Yeo7 $\rightarrow$ Parcel
Schaefer-400	functional	Yeo7 $\rightarrow$ Parcel
Power-264	functional	14 Networks
Dosenbach-160	functional	7 Networks

This atlas-aware structure is not limited to static labeling. The hierarchical browser in the interface exposes the parcellation structure directly, and BrainRing supports multiple label levels, namely Lobe, Gyrus, Subregion, and Index (Fan et al., 28 Mar 2026). In the Brainnetome 246 setting, for example, the hierarchy extends to subregion level, while in the Schaefer atlases the hierarchy is organized by Yeo7 network and parcel.

The inclusion of atlas semantics within the visualization environment has consequences for both usability and interpretability. It means that region identity, grouping, and display order are not external bookkeeping tasks delegated to user-written scripts. A plausible implication is that BrainRing reduces the mismatch between statistical FC outputs and figure-level anatomical organization, since atlas metadata are already encoded within the rendering system.

4. Data input pathways and edge construction

BrainRing supports several input modes intended to accommodate both automated pipelines and manual figure editing (Fan et al., 28 Mar 2026). The first is FC matrix upload. Users can upload an $N \times N$ symmetric functional connectivity matrix in CSV, TXT, or TSV format, where $N$ matches the selected atlas. The delimiter is auto-detected among comma, tab, or space. After loading, the threshold is automatically adjusted to show a meaningful number of edges, and regions involved in supra-threshold connections are selected for display.

The second pathway is sparse-network input via edge lists in the format

$i,\, j,\, weight$

with $\rightarrow$0 and $\rightarrow$1 as 1-indexed region identifiers. The paper notes that this is useful for graph-analysis pipelines such as those built with the Brain Connectivity Toolbox (Fan et al., 28 Mar 2026).

A third pathway is direct interoperability with Circos. BrainRing accepts the standard Circos link format

$\rightarrow$2

and applies per-edge colors specified by color=r,g,b directly. This allows reuse of existing Circos workflows without preprocessing.

The fourth pathway is manual edge specification. Users can enter region pairs such as

$\rightarrow$3

or batches such as

$\rightarrow$4

BrainRing also provides a click-to-connect mode in which a source arc and then a target arc are clicked to create an edge; the mode resets after each pair. All manually added edges appear in a management panel with full region labels, exemplified in the paper by

$\rightarrow$5

and each edge can be deleted individually (Fan et al., 28 Mar 2026).

Taken together, these input modes show that BrainRing is not restricted to a single data provenance. It supports dense FC matrices, sparse graph outputs, Circos-derived files, and hand-edited diagram construction within the same interface.

5. Visualization controls, edge management, and region reordering

BrainRing is built for interactive styling of chord diagrams and exposes control over color, layout, labeling, and individual edges (Fan et al., 28 Mar 2026). It includes six built-in color schemes, including HAH, which reproduces 246 unique per-subregion colors from standard Circos neuroimaging templates. A color intensity slider ranges from $\rightarrow$6 to $\rightarrow$7, and arc colors can be configured as flat or gradient within each lobe.

Layout parameters are also explicitly controllable. The tool allows adjustment of ring width (4–100 px), arc padding, separator mode (none, lines, or gaps), gap size (20%–300%), rotation (0°–360°), chart size (30%–100%), and arc stroke width and stroke color (Fan et al., 28 Mar 2026). These options are directly tied to the publication-oriented goal of the software, since they expose the graphical degrees of freedom that are often hard-coded or only indirectly accessible in script-based workflows.

Labeling is hierarchical and editable. BrainRing supports the four label levels noted above, and gyrus labels can be shown directly on the ring with adjustable font sizes. Individual gyrus labels can be clicked to rename, recolor, resize, or hide them. A dark/light theme toggle changes only the canvas background, allowing editing in dark mode while exporting a white-background figure for publication (Fan et al., 28 Mar 2026).

Edge management is one of the tool’s most distinctive capabilities. It supports filtering by FC threshold and by Top-$\rightarrow$8 selection, including strongest, weakest, and absolute modes. It also supports manual edge addition, click-to-connect creation, clicking an existing edge to open a color picker, per-edge color customization, and a list view of all manual edges with delete buttons (Fan et al., 28 Mar 2026). This gives fine-grained control over which connections are visible and how they are encoded.

BrainRing additionally supports gyrus-level reordering. Users can move a gyrus group to any position on the ring, and when this occurs the FC matrix, selection state, manual edges, and edge colors are automatically remapped to preserve consistency. This is a substantive feature rather than a cosmetic one: region order in chord diagrams materially affects readability, and automatic remapping prevents structural inconsistencies during manual rearrangement.

6. Export, multilingual support, and relation to existing tools

BrainRing exports figures as SVG and PNG, with PNG rasterized at 4× resolution. Exports preserve the current background color and all visual settings, so the output is already publication-ready and requires no post-processing (Fan et al., 28 Mar 2026). This directly serves the paper’s stated objective of producing publication-ready figures within seconds rather than the minutes-to-hours workflow typical of script-based approaches.

The interface is fully bilingual, supporting Chinese and English. The paper specifies that 79 static UI elements and 36 dynamic text strings are translated, and that the language can be switched at any time via a header toggle (Fan et al., 28 Mar 2026). This multilingual support is unusual among domain-specific neuroimaging visualization utilities and broadens direct usability without altering the underlying rendering model.

In comparative positioning, the paper emphasizes that BrainRing differs from Circos by requiring no installation, being browser-based, providing instant preview, including built-in brain atlases, supporting click-to-edit interaction, offering a bilingual UI, simplifying edge management, and exporting directly to SVG and PNG. At the same time, it retains one of Circos’s important strengths, namely per-edge coloring, and can import Circos link files directly. Relative to BrainNet Viewer, BrainRing addresses chord diagrams rather than 3D brain surface visualization, avoids MATLAB dependence, supports interactive parameter changes, and provides publication-ready export (Fan et al., 28 Mar 2026).

These comparisons clarify the tool’s scope. BrainRing is not presented as a general replacement for all neuroimaging visualization environments; it is presented as a practical, open-source alternative for Circos-style and MATLAB-based workflows when the target representation is a brain connectivity chord diagram. The software is freely available at https://github.com/XiuFan719/brain-connectivity-viz, with a live demo at https://XiuFan719.github.io/brain-connectivity-viz/ (Fan et al., 28 Mar 2026).