Neural Geometry

Neural geometry is the source’s way of treating collective neural activity as shapes, trajectories, and manifolds in high-dimensional space. In Claire Isabel Webb & Nina Miolane: The Geometry of Consciousness, Nina Miolane explains that each neuron’s firing rate can be treated as a dimension, so the population’s state at one moment becomes a point and its activity over time becomes a trajectory.

Key Claims

• Population Coding can reveal low-dimensional structure hidden inside apparently disorganized high-dimensional firing patterns.
• The Spatial Navigation Torus is the central example: grid-cell-like periodic activity produces a torus in both biological and artificial systems.
• Geometry is not only visualization in the talk; it is meant to support explanation, prediction, and testable mathematical theory.
• Reward and salience can deform neural geometry by changing representational resolution around important locations.
• Sleep-state geometry gives a possible empirical handle on Consciousness Measurement, because head-direction activity changes structure between wake/REM and non-REM sleep.

Connections

• Nina Miolane - speaker who presents the geometric approach.
• Mathematical Theory Of Intelligence - broader explanatory ambition.
• Population Coding, Spatial Navigation Torus, and Fourier Spatial Encoding - core technical cluster.
• Consciousness Measurement - boundary where geometry becomes relevant but not decisive.
• Representation Learning and AI Interpretability By AI - AI-facing implications of geometric representations.