Neurophysiological mechanisms of memory formation investigated with two-photon calcium imaging in the behaving mouse
Final Report Abstract
Early studies of the cellular mechanisms of memory formation and retrieval have shown that memories are stored in sparse cellular assemblies, called memory engrams, which are activated during memory formation and re-activated during memory recall. During memory formation, synaptic connections between these engram cells become strengthened, suggesting that Hebbian plasticity occurs at the synapse level. These studies looked at immediate early gene expression or synaptic activation profiles in fixed brain samples at specific time points after learning or memory retrieval. When adding more time points to the experimental design, this picture becomes more complex. For example, the analysis of memory cell ensembles of two distinct memories, revealed that the overlap of these cell ensembles depends on the time between the formation of the two memories, showing that the cells recruited for a specific memory depend on time and experience and are not predefined. Also, it was shown that remote memory ensembles in the cortex and the basolateral amygdala are not stable but dynamically evolve over time . When looking at the system even closer by recording single-cell activity from the amygdala of mice performing an auditory fear-conditioning task, the full dynamic nature of memory formation is revealed: The same sensory cues of the learning tasks activate different groups of cells from day-to-day, with only 55% overlap between activated cell populations on consecutive days and even less (35%) for five days, similar to what was found in the hippocampus. Regarding the flexibility of our brains to integrate distinct or update memories, the dynamic nature of single-cell processes involved is not surprising. But how can the brain store information long-term at all, while offering this great flexibility? Addressing this question is challenging because it involves the analysis of single-cell processes on a millisecond time scale, as well as long-term information storage across hours, days, or weeks. One method that is capable to look at both, is two-photon calcium imaging because it allows for chronic recordings of single-cell activity in dense cell populations of awake behaving animals. In this work here, experimental methods were designed, set up and tested for these experiments. This includes customdesigning and setting up the hardware for auditory fear-conditioning experiments of head-fixed, freely running mice, plus designing the custom-written software for hardware control, data acquisition and synchronization of animal behavior with two-photon calcium imaging. Moreover, surgical procedures for endoscopic GCaMP7f imaging in excitatory neurons of the lateral amygdala of the awake behaving mouse were established. The protocol of the behavioral task was developed and modified until the mice reliably learned associating a pure tone to the mild electric shock under head-fixation. To readout fear responses of the animals a new technique of analyzing the animals’ facial expression was established, by help of deep learning. Finally, a second scanner was added to the two-photon microscope, making use of the second output beam of the femto-second laser for single-cell stimulation. With these methodical advances, mice were successfully trained to associate a tone with a fearful event, the fear-response was read-out with video-recordings of the mouse faces, while singlecell activity was recorded from the lateral amygdala. The exact same cells were identified days or even weeks after the initial learning task and recorded during memory retrieval. In a next future step, the prepared hardware for single-cell stimulation needs to be tested in in vivo experiments and can potentially be used for all-optical interrogating of individual cells of the lateral amygdala during readout of animal behavior. Thus, this work lays the basis for a detailed analysis and manipulation of singlecell activity in the lateral amygdala during memory formation and retrieval and might shine insight into the complexity of storing information in this dynamic and complex system.
