Discovering the circuit and molecular basis of inter-strain and inter-species differences in learning

Objective

Across the animal kingdom, closely-related species can differ in their learning ability, as can different individuals of the same species, but the underlying causes of these differences are poorly understood. Differences in learning could be due to different structural and functional properties of learning circuits which in turn result from different patterns of gene expression. However, identifying homologous learning circuit neurons in different species for comparing their synaptic connectivity, cellular-resolution functional properties and transcriptomes is challenging in larger brains. We have identified a strain of Drosophila melanogaster larvae that learns faster (after fewer training trials) than others and a closely related Drosophilid species that learns faster than the D. melanogaster strains. We will use this tractable genetic model system to compare learning circuits in faster/better and slower/worselearner strains and species, in particular, their synaptic resolution connectomes; cellular-resolution activity maps; and transcriptomes. We will then screen the differentially expressed genes to identify those that improve learning and determine the way in which they affect connectivity and functional properties of learning circuits. This project will reveal the architectural features of learning circuits that enhance learning and memory, as well as the molecules that can transform slow/worse- into fast/better-learners and their mechanisms of action. Uncovering the underlying structural, functional and genetic causes of variability in learning will not only have a major impact on neuroscience but could also potentially inspire the development of better architectures and algorithms for artificial intelligence and provide new avenues for treating learning and memory deficits.