A Modular Computational Framework for Integrating Motion and Chromatic Representations in Mice Retina

This thesis presents a modular computational framework for integrating motion-sensitive and chromatic response representations in the mouse retina. Although motion and color processing have been extensively studied, they are typically characterized using different experimental paradigms and incompatible representational forms. Motion sensitivity is described through spatiotemporal receptive fields, while chromatic sensitivity is summarized using temporal response kernels. This work addresses the challenge of combining these heterogeneous representations into a unified and interpretable computational system. The proposed framework operates on fixed experimental datasets and incorporates chromatic information into motion receptive fields through structured, region-wise modulation of center and surround components. Anatomical alignment across independent datasets is achieved using normalized inner plexiform layer (IPL) depth distributions and electron microscopy–defined bipolar cell types as a bridging reference, enabling anatomically constrained pairing of motion and chromatic clusters. Stimulus-driven simulations are performed using linear convolution under moving-bar, looming, receding, and full-field flicker stimulation. Model predictions are evaluated against experimentally recorded flicker responses using correlation-based metrics. Results show that chromatic weighting systematically modulates the timing and magnitude of motion-evoked responses while preserving receptive-field structure. These findings demonstrate that independently characterized motion and color representations can be integrated within a principled and interpretable computational framework for studying early visual processing.

Author dc:creator

• Resham Patil (24399695)

• Computer Science

Statement dc:rights

• In Copyright

DOI dc:identifier

https://doi.org/10.25417/uic.32994170.v1

OAI identifier oai:identifier

oai:figshare.com:article/32994170