Blueprint of visual information processing in the mouse brain

15 June 2022

A research team led by Vincent Bonin at NERF (empowered by imec, KU Leuven and VIB) has assessed more than 30,000 nerve cells in the mouse brain to learn more about how they are specialized in processing visual information. The results, published in Nature Communications, provide fine-grained insights into how the brain is able to see what our eyes perceive.

When navigating the world around us, it is vital we can correctly process visual information. Yet, researchers still have only a limited view of how the visual processing streams in our brain help us accomplish this amazing feat.

Prof. Vincent Bonin and his team in Leuven study these neural circuits of vision: “Our main goal is to elucidate the biological mechanisms underlying visual processing and visually-guided behavior. Our research is centered on the visual cortex, which is part of the outer layer of the brain, as well as its connections, and how these circuits contribute to sensory perception and behavior.”

The response of 30,000 brain cells

To get a clear picture of what happens in our brain with visual information, scientists primarily need more and better data, says Bonin. “There is a lack of data sets spanning multiple visual areas to uncover information processing streams, and the sparse, singular recording data sets obtained from individual areas often fall short in revealing their functional diversity.”

That is why Xu Han, who recently completed his PhD at the Bonin lab, set out to measure the activity of more than 30,000 neurons from 8 different visual areas in the mouse brain in response to a broad array of visual stimuli. With this massive effort, Han and his colleagues wanted to map the response from tens of thousands of neurons in terms of their ability to encode orientation, spatiotemporal contrast, and visual motion speed.

Highly structured, highly distributed

The researchers found that while all mouse visual cortical areas conveyed diverse types of visual information, they had a distinct bias in terms of the number of neurons that are tuned to particular features.

Han explains: “Higher visual areas in the mouse brain form complementary neural representations when it comes to visual cues at distinct spatial and temporal scale, including motion speed and spatial patterns. The resulting parallel processing streams specialized in analyzing fine spatial patterns, fast motion or slow motion, are critical for visual behavior such as object recognition, foraging and navigation.”

Zooming in on individual cells, some features formed a continuum while other visual features are clearly encoded by distinct tuning types.

Bonin: “Our data underscore the highly structured and highly distributed nature of cortical representations of visual cues, driving specialization of different areas and information streams.”

As such, the study opens a new window on the organizing rules of neural circuits underpinning sensory processing, perception, and behavior. “By combining functional imaging, tracing and manipulation of neural circuits, we provide a cellular-resolution blueprint for the organizing principles underlying visual processing in the brain.”

Publication

Diversity of spatiotemporal coding reveals functional specialization of visual streams in the mouse cortex
​Han, Vermaercke & Bonin, Nat Comms 2022