Northwestern University Physicists Discover That Stronger Social Ties Can Unexpectedly Trap New Ideas in Loops

Northwestern University physicists find that strong network connections can trap ideas in loops. Discover why weakening ties is essential for innovation.

By: AXL Media

Published: Apr 28, 2026, 8:59 AM EDT

Source: Information for this report was sourced from EurekAlert!

The Paradoxical Nature of Connectivity and Innovation

While sticking to familiar social or professional circles may provide a sense of security, new theoretical research suggests it effectively stifles the reach of novel concepts. Physicists at Northwestern University have developed a mathematical framework demonstrating that tight-knit networks often act as traps, keeping information circulating within the same small loops. This study, published in Communications Physics, indicates that for an idea or behavior to proliferate across a broad system, the entities within that network must actively move away from established, high-strength connections.

Integrating Hebbian Principles Into Network Theory

The research team, led by Assistant Professor István Kovács, utilized the concept of Hebbian learning to explain these network dynamics. Originally a psychological principle used to describe how neurons that "fire together wire together," Hebbian learning explains how connections strengthen through repeated use. Unlike traditional static network models, this new framework allows links between nodes to adapt based on experience. This adaptation fundamentally alters the architecture of the network over time, revealing how local interactions dictate the global spread of activity.

Escaping the Algorithmic Death Spiral

One of the most striking findings of the simulations was the emergence of "death spirals," similar to the circular paths formed by blind fire ants following pheromones. In a network driven by positive reinforcement, activity tends to circle back along the same routes as those paths become "stronger" and more attractive. According to Kovács, this positive incentive creates a feedback loop that prevents activity from reaching the periphery of the network. To avoid this stagnation, the system requires a mechanism where connections can weaken, forcing the activity to seek out and explore entirely new paths.