Visual information processing in the brain - from retina to cortex

My work focuses on neuronal information processing in vision. We study what the retina—the image processor of the eye—tells the brain and how the brain makes sense of this information. By conducting extensive neuronal recordings and leveraging advanced machine learning algorithms, we explore how the visual input is represented by neuronal populations in the retina and the visual cortex, using both mice and primates as model species. My research covers, among others, the functional classification of neuron types, the neuronal correlates of color vision, and species-specific differences in visual processing. Through integrating neuroscience with AI tools, my work deepens our understanding of the neuronal underpinnings of vision, opening new avenues for the development of artificial vision systems and treatments for visual disorders.

Using AI TOOLs to study biological vision

My interdisciplinary research stands at the intersection of artificial intelligence (AI) and neuroscience, and uses AI tools to unravel the neuronal correlates of biological vision. By combining large-scale neuronal recordings with convolutional neural networks (CNN), we aim to understand how visual neurons process the high-dimensional visual input. Specifically, we use state-of-the-art CNN models as “digital twin” of specific visual areas, allowing us to in-silico synthesize most exciting inputs for single visual neurons and to systematically characterize the neurons’ stimulus selectivity in the context of different latent variables, such as distinct object features (shape or texture) as well as internal brain state. This approach not only provides a deeper insight into the brain's visual processing capabilities but also bridges the gap between biological systems and artificial vision technologies, opening new pathways for advancements in both fields.