The Thousand Brains Theory, proposed by Jeff Hawkins and colleagues, suggests that the neocortex is composed of thousands of repeating units called cortical columns, each acting as a sensorimotor system that learns models of objects. A new paper, "The Thousand Brains Theory 2.0," extends this framework to explain how long-range connections—including hierarchical feedforward/feedback and thalamic pathways—enable compositional learning and perspective transformation.
The Research
In this preprint, researchers from the Thousand Brains Project (Jeff Hawkins, Niels Leadholm, and Viviane Clay) review the anatomy of long-range neocortical connections, arguing they form a heterarchy rather than a hierarchy. They propose two key roles: First, the thalamus converts sensory and movement information from an egocentric (sensor-based) to an allocentric (model-based) perspective. Second, hierarchical connections enable columns to learn compositional models—representations built from subcomponents. The paper includes testable predictions for future experiments, linking theory to anatomy, neurophysiology, and behavior.
Why It Matters
This extension addresses previously unexplained aspects of brain function. Understanding how the cortex handles compositional objects—like recognizing a chair as a combination of seat, back, and legs—could inform AI architectures and educational strategies. For your own cognition, it highlights the importance of integrating multiple senses and movements when learning new concepts, supporting active, multi-modal learning.
What You Can Do
- Engage multiple senses: When learning something new, involve touch, movement, or sound alongside visual input.
- Break down complex ideas: Practice decomposing objects or concepts into their component parts, then assemble them mentally.
- Vary your perspective: Approach problems from different angles to build flexible, allocentric models.
Source: arXiv q-bio.NC
Curious about your own brain? Take our free adaptive IQ test or try 306 brain training levels.