Structural Connectome Matrix & Graph Theory Studio

Diffusion MRI · Whole-Brain Streamline Adjacency, Small-World Topology, Rich-Club Hubs & Targeted Lesion Resilience

Core Principle: Whole-brain tractography bridges voxelwise diffusion measurements to macroscopic systems neuroscience. Grouping streamline counts between parcellated cortical regions yields an \(N \times N\) adjacency matrix \(A_{ij}\). The human brain balances local segregation (high clustering \(C\)) with global integration (short path length \(L\)) to form a small-world network (\(\sigma > 1\)). Highly connected "rich-club" hubs provide rapid inter-modular communication but represent critical points of vulnerability.

Default Mode Network (DMN)

Somatomotor Network (SMN)

Visual Network (VIS)

Executive / Salience (FPN)

Subcortical Hubs

Char. Path Length (L)

Global Integration

Clustering Coeff (C)

Local Segregation

Small-Worldness (σ)

σ = (C/C_rand) / (L/L_rand)

Global Efficiency (E)

Harmonic Mean Path

Connectivity Matrix A_ij

Hover or click an anatomical node to inspect degree, community module, and connections.

Streamline Density Threshold 12%

Dense (Keep Weak) Sparse (Strict Core)

Network Resilience & Lesions:

Test the Hypothesis: Random lesions simulate diffuse axonal injury; network efficiency degrades gradually. Targeted attacks strike rich-club hubs (Precuneus, Thalamus); path length surges and communication collapses.

Topological Principles of Brain Networks

1. From Streamlines to Adjacency Matrix:

A structural connectome models the macroscale wiring of the brain. Whole-brain tractography generates millions of streamlines. By mapping cortical gray matter into parcellated Regions of Interest (ROIs), we count the number of streamlines connecting region \(i\) to region \(j\), forming the symmetric adjacency matrix \(A_{ij}\). Thresholding removes false-positive streamlines resulting from tracking errors in crossing-fiber zones.

2. Small-World Architecture:

Random networks have short path lengths but negligible clustering. Lattice networks have high clustering but long path lengths. Brain networks exhibit small-world organization: high local clustering \(C \gg C_{\text{rand}}\) (neighboring regions are densely interconnected for specialized local processing) combined with short characteristic path lengths \(L \approx L_{\text{rand}}\) (long-range association fasciculi enable rapid whole-brain communication), producing \(\sigma = (C/C_{\text{rand}}) / (L/L_{\text{rand}}) > 1\).

3. Rich-Club Phenomenon:

A subset of high-degree brain nodes (including the precuneus, superior frontal cortex, posterior cingulate, and thalamus) are more densely connected to one another than expected by degree distribution alone. This "rich club" forms a central high-capacity backbone for integrating information across disparate functional domains.

4. Attack Vulnerability & Clinical Pathology:

Brain networks demonstrate a classic "scale-free-like" resilience profile: highly robust against random injury, but extraordinarily vulnerable to targeted attacks. In Alzheimer's disease and traumatic brain injury, amyloid deposition and axonal shearing disproportionately affect metabolic hubs, rapidly degrading global network efficiency:

E_{\text{glob}} = \frac{1}{N(N-1)}\sum_{i \neq j} \frac{1}{d_{ij}}