Beyond Pixels: A Differentiable Pipeline for Probing Neuronal Selectivity in 3D
This work addresses the challenge for systems neuroscience researchers in isolating neuronal selectivity to physical scene properties, bridging inverse graphics with neuroscience, though it is incremental as it builds on existing differentiable rendering and mesh deformation techniques.
The paper tackles the problem of characterizing neuronal selectivity to 3D scene properties like shape and lighting, which is difficult with 2D pixel-based methods, by introducing a differentiable rendering pipeline that optimizes deformable meshes to maximize neuronal responses, enabling probing of interpretable 3D factors in monkey area V4.
Visual perception relies on inference of 3D scene properties such as shape, pose, and lighting. To understand how visual sensory neurons enable robust perception, it is crucial to characterize their selectivity to such physically interpretable factors. However, current approaches mainly operate on 2D pixels, making it difficult to isolate selectivity for physical scene properties. To address this limitation, we introduce a differentiable rendering pipeline that optimizes deformable meshes to obtain MEIs directly in 3D. The method parameterizes mesh deformations with radial basis functions and learns offsets and scales that maximize neuronal responses while enforcing geometric regularity. Applied to models of monkey area V4, our approach enables probing neuronal selectivity to interpretable 3D factors such as pose and lighting. This approach bridges inverse graphics with systems neuroscience, offering a way to probe neural selectivity with physically grounded, 3D stimuli beyond conventional pixel-based methods.