Integration of visual motion and locomotion in mouse visual cortex

3 years ago



This paper provides a compelling demonstration that multisensory integration occurs in primary visual cortex. The authors record extracellularly from V1 neurons in mice running on a ball in the presence or absence of a virtual scene. They find that many neurons are strongly modulated by running speed, even in the dark. By breaking the correlation between actual running speed and the translational speed of the virtual environment, they can measure the relative contributions of movement and vision to a neuron's firing rate. Most neurons' responses were predicted by a linear combination of running speed and virtual speed, but some were influenced by running or vision alone. Given how dramatically and systematically movement alters visual responses, these experiments should encourage us to reconsider the role of primary sensory areas in perception.



So much modeling and extrapolation about vision comes from anesthetized V1 recordings (or recordings in awake paralyzed cats). Recordings in awake, closer-to-freely-moving (treadmill) mice would have seemed like a sanity check, to make sure that the old rules hold in a brain that's active. The amount of V1 activity that has nothing to do with vision at all suggests that the old rules are an oversimplification. "Primary visual cortex" could actually be a misnomer - this study should encourage us to go back and record the whole visual pathway in moving animals.

Re: Novelty



Unit isolation can change on a behavioral timescale, especially in this prep where the mouse is putting so much pressure on his electrode implant. Not enough done to show how spikes change amplitude with head pressure in this setup. A lot of their effects may be contaminated by this issue.

Re: Rigor