An Exploration of Human Brain Connectomics - based on work by Zhou, et al.

• Download the individual connectomes MATLAB Object

• Use ConnectomeToCSV.m to recreate the .csvs in /connectomes

• /plots/sc_vs_random_degree_cc_gre.PNG

This 3D scatter plot compares the network properties of brain data from 70 human subjects against 70 "random" networks. In graph theory and neuroscience, these metrics help us understand how "efficiently" a brain is wired.

Here is a breakdown of what you are looking at:

1. The Variables (The Axes)

The plot uses three metrics to describe each network:

• Avg. Degree (X-axis): This measures the average number of connections each "node" (in this case, brain voxels) has.
• Avg. Clustering Coefficient (Y-axis): This measures how much the neighbors of a node are connected to each other. High clustering means the network has many local "cliques" or clusters.
• Global Resource Efficiency (Z-axis): This measures how effectively information travels across the entire network. A higher value usually indicates shorter paths between any two random points.

---

2. The Data Groups

While the legend in the top left is blank, we can infer the groups based on the title "70 Subjects (SC, 68 voxels) vs 70 random":

• The Blue Cluster (Human Brain Data): Notice how the blue dots are tightly packed together. This tells us that across these 70 subjects, the brain's structural connectivity (SC) is remarkably consistent. They have relatively low clustering and a specific, narrow range of efficiency and degree.
• The Orange Curve (Random Networks): These points follow a distinct, sweeping path. This represents a "null model." By generating random networks with varying degrees of connectivity, researchers can see where the human brain sits in comparison.

---

3. Key Takeaways

• Biological Consistency: The tight grouping of the blue dots suggests that "normal" brain architecture isn't random; it follows a very specific structural "recipe" that is conserved across individuals.
• Efficiency vs. Cost: The human brain (blue) sits at a different coordinate than many of the random networks. Often, these plots are used to show that the brain achieves a "Small-World" configuration—balancing high local processing (clustering) with high global integration (efficiency) while minimizing the "cost" of long-distance wiring.