Zigzag Persistence of Neural Responses to Time-Varying Stimuli

A team of researchers including Yuri Gardinazzi, Alessio Ansuini, and Matteo Biagetti has published a novel study applying advanced mathematical techniques to neuroscience. Their paper, 'Zigzag Persistence of Neural Responses to Time-Varying Stimuli,' introduces a method using topological data analysis (TDA) to decode how populations of neurons in the mouse visual cortex respond to complex, time-varying video stimuli from the Sensorium 2023 dataset. The core innovation is using zigzag persistent homology—a technique from algebraic topology—to build a frame-by-frame model of neural activity as a 'cubical complex' and track how its one-dimensional topological features, conceptualized as 'loops,' evolve over time. These loops represent coordinated, cyclical patterns of neural co-activation.

Technically, the team summarized these dynamic topological patterns into compact vector representations called 'persistence landscapes.' They then tested whether these mathematical signatures contained meaningful information by attempting to cluster neural responses from repeated trials of different videos. The results were clear: the topological descriptors reliably distinguished which video the mouse brain was processing. This work establishes a direct, interpretable connection between the evolving, high-dimensional activity of thousands of neurons and stable, computable topological signatures. It advances the use of TDA beyond static datasets, opening a new avenue for uncovering the neural code in complex, real-time dynamical systems like vision and potentially other sensory or cognitive processes.