Brain organoids reveal hidden loop structures in neural activity
Persistent homology finds topological rings in spontaneous organoid firing patterns
Deep Dive
Researchers applied persistent homology to microelectrode-array recordings of spontaneous activity from human and mouse cortical organoids (26–234 units). They found that loop structures (H1) in correlation-based networks rose significantly above null in 14 of 18 datasets. These loops are robust to random unit removal but break when core units are removed. Larger organoids also showed second homology (H2), indicating richer topological organization.
Key Points
- Loop structures (H1) found in 14 of 18 organoid datasets, significantly above null expectations
- Topology robust to random removal of units but collapses when core units are removed
- Larger networks also show second homology (H2), indicating richer multiscale organization
Why It Matters
Topological data analysis reveals hidden geometric order in brain organoids, offering a new tool for understanding neural computation and disease.