Manifolds for the neural representation of 3D spatial experience

Neural representations of 3D pose are modeled as manifolds derived from SE(3), with topology shaped by movement constraints, object geometry, and visual features.

• Why it matters: A unified framework is needed to explain how brains across insects to primates represent three-dimensional position and orientation within a common theory.
• What they did: The framework derives low-dimensional pose manifolds for macaque parietal cortex object orientation and for head direction in fruit bats, mice, and monkeys, then uses them to build topology-respecting tuning functions.
• The result: The resulting taxonomy and manifold-guided tuning functions enable experimental tests and provide a common approach linking 3D pose geometry to neural coding across species and brain systems.