A General Theory of Sensory Attractors and Internal Stiffness Geometry**(AGI)

Despite decades of progress in neuroscience, the mechanism by which the brain transforms noisy sensory input into coherent percepts remains poorly understood. Traditional frameworks—Bayesian models, deep learning analogies, predictive coding—describe percepts as interpretations of data, but do not explain the physical mechanism by which the brain collapses ambiguous signals into stable percepts. This work proposes a mechanistic, physics-based explanation: > Perception is the relaxation of sensory input into the nearest attractor of a learned stiffness geometry. This geometry is encoded in neural connections via Hebbian plasticity and forms a low-dimensional manifold in the high-dimensional sensory space. Percepts correspond to stable geometric basins (attractors). Ambiguous input triggers relaxation toward one of them; mis-tuned geometry produces distortions; competing attractors yield bistable perception. This view naturally explains illusions, hysteresis, adaptation, and perceptual stability. The theory is validated in two ways: Real human ECoG data demonstrating attractor manifolds during auditory bistability. A behavioral mis-tuning experiment demonstrating perceptual drift predicted by the model.