Event Details:
Alexander Attinger is a postdoc in the Giocomo lab. He will talk about manipulating neural ensembles in RSC and V1 in behaving mice using two-photon holographic stimulation. The seminar will be held in person only.
Retrosplenial cortex (RSC), a cortical association area, is a key node in the navigational network of the brain. Yet, how neural dynamics in RSC drive behavior is not well understood. Here, we use two-photon imaging combined with holographic optogenetic stimulation and virtual reality environments to investigate how functionally defined ensembles in RSC shape circuit dynamics and control behavior. We use a fully transgenic approach with new transgenic mouse lines for wide-spread, balanced and homogeneous co-expression of the red-shifted, soma-targeted opsin ChRmine and the indicator GCaMP6m. This approach leads to stable long-term expression of GCaMP and ChRmine in excitatory neurons throughout cortex, allowing us to track the activity of thousands of neurons across several weeks. We trained head-fixed mice to navigate a virtual hallway and memorize the identity of a briefly visible landmark, indicating the availability of a water reward at the of the hallway depending on the specific landmark presented. We then recorded the activity of neurons in L2/3 and L5 of RSC and found that neurons are active at preferred locations in the hallway and form stable representations with population dynamics that differentiate between rewarded and non-rewarded landmarks. By stimulating different functionally defined ensembles of neurons while mice perform the task, we are probing how neural representations in RSC influence visually-guided navigation and memory. Our data suggest that both ongoing dynamics in RSC and task performance are robust to perturbations of ensembles of excitatory neurons. In addition, location specific activity of targeted neurons remained stable even after repeated targeted activation. In contrast, we found that repeated activation of neurons in specific locations in a novel environment, biased the activity of targeted neurons towards the stimulated location. This bias was contingent on stimulation from first exposure, suggesting a rapid formation of stable and robust representations.