Gradient Biotech

Connectome analysis

Analyze a precomputed connectivity matrix — compute Brain Connectivity Toolbox graph metrics, explore interactive visualizations, and generate analytical interpretation.

Research question

What is the topology of this brain network? Which nodes are hubs, how segregated and integrated is the network, and does it exhibit small-world properties?

Who this is for

  • Neuroscience researchers with precomputed ROI-to-ROI or voxel connectivity matrices
  • Computational neuroscience groups migrating from MATLAB BCT workflows
  • Labs needing quick graph-theoretic summaries without maintaining local toolchains

Data requirements

DataRequiredPurpose
Square connectivity matrixYesGraph metrics and visualization input
Atlas region labelsNoHeatmap and 3D brain view annotations
Subject metadataNoContext for report and interpretation

Supported formats: CSV, NPZ, NumPy .npy, MATLAB .mat.

Workflow

Create study → Upload connectivity matrix → Graph metrics → Visualize → Interpret (optional)

Step 1 — Upload matrix

Create a neurology study and upload a symmetric connectivity matrix from the study home page. No matrix handy? See Sample data for a minimal example to test with. You are redirected to the dataset workspace after upload.

Step 2 — Graph metrics

On the dataset workspace, click Compute graph metrics. Key parameters:

ParameterDefaultPurpose
thresholdautoEdge weight cutoff for binary graph construction
n_rand_smallworld100Null model count for small-worldness

Review global metrics — clustering coefficient, characteristic path length, global efficiency, modularity, small-worldness — and per-node BCT measures in the results table.

Step 3 — Visualize

When the run completes:

  • Connectivity heatmap — matrix values with optional threshold
  • Network graph — force-directed layout colored by module assignment
  • 3D brain view — when atlas coordinates are available for region labels

Step 4 — Interpret (optional)

Use Interpret on the dataset workspace after graph metrics complete for a plain-language summary citing your computed values. If Interpret isn't available, ask your administrator to enable AI interpretation.

Expected outputs

  • Global BCT metrics (clustering, modularity, efficiency, small-worldness, path length)
  • Per-node centrality and module assignments
  • Interactive heatmap, graph layout, and 3D brain views
  • AI narrative explaining network topology with metric citations

Example insight

The network shows elevated modularity with reduced global efficiency compared to typical resting-state norms, suggesting increased segregation with weaker long-range integration.

Related guides