What your brain does effortlessly—like remembering a face or planning a route—is actually a costly computation, and a new framework from researchers at the University of Pennsylvania and Rutgers University can now measure that cost.
The Research
Suman S. Kulkarni, Jason Z. Kim, Panagiotis Fotiadis, Fabio Pasqualetti, and Dani S. Bassett posted a preprint on arXiv (q-bio.NC) in July 2026 that introduces a computational affordance landscape—a way to map out which computations a network readily supports based on its structure. They used control theory to define the cost of moving the network from one activity pattern to another, then computed that cost for all possible transitions.
Applying this to insect navigation circuits, they found that updating orientation is the least costly computation, matching known neural wiring. In human brain networks, they saw systematic differences: sensory networks have heterogeneous landscapes (fine-tuned for specific tasks), while association networks have homogeneous ones (geared for flexible, general processing).
When they trained recurrent neural networks on cognitive tasks, learning progressively increased landscape heterogeneity, reshaping the distribution of affordable computations. This suggests that learning physically rewires the cost structure of a network to favor the tasks it performs.
Why It Matters
This framework helps explain why some tasks feel easy and others hard—it's about the inherent cost of the computation on your brain's network. It also bridges structure and function, offering a quantitative way to see how your brain's wiring supports your thinking. For anyone interested in their own cognition, it underscores that your brain is not a blank slate; its structure predisposes you to certain abilities, which can be reshaped through learning and training.
What You Can Do
You can't change the overall architecture of your brain, but you can shift your own computational affordance landscape. Engage in varied cognitive tasks—like puzzles, memory games, or learning a new skill—to encourage your networks to become more heterogeneous and adaptable. This is the essence of brain training: not just getting better at one task, but rebalancing your brain's cost structure for new challenges.
Source: arXiv q-bio.NC
Curious about your own brain? Take our free adaptive IQ test or try 306 brain training levels.
