Network Analysis of Neural Spike Times: Insights into Visual Processing in Mice Using the Allen Brain Dataset

The brain’s ability to process sensory information relies on complex networks of interconnected neurons. This thesis investigates the neural connectivity underlying visual processing using the Visual Coding – Neuropixels dataset from the Allen Brain Observatory. The experiment involved presenting various visual stimuli to mice while recording their neural activity using high-density Neuropixels probes. These probes simultaneously capture the spike times of hundreds of neurons across multiple brain regions, providing detailed insights into neural dynamics during visual perception. Focusing on three mice (Mouse28, Mouse1, and Mouse0), we analyze neural spike data across nine distinct visual stimuli, including spontaneous, natural scenes, and static gratings, to uncover patterns of functional connectivity at both neuronal and brain structure levels. Networks were constructed by identifying significant pairwise interactions between neurons based on their co-occurrence of spikes within a defined time bin. The resulting networks were analyzed to evaluate stimulus-specific connectivity patterns, overlap between stimuli, and the roles of neurons and brain structures. Additionally, brain structure-level graphs were created, where nodes represented anatomical regions and edges were weighted by inter-structural connectivity, providing a broader perspective on neural organization. Analysis of brain structures highlighted differences in inter-regional connectivity and emphasized the distributed yet organized topology of the visual system. Comparative analyses across mice demonstrated consistent network structures with individual variations that may reflect biological differences or experimental conditions. This work combines network science approaches with a comprehensive neural dataset to provide new insights into visual processing and the functional architecture of the mouse brain.