Potential Synaptic Connectivity of Different Neurons onto Pyramidal Cells in a 3D Reconstruction of the Rat...

Most existing connectomic data and ongoing efforts focus either on individual synapses (e.g., with electron microscopy) or on regional connectivity (tract tracing). An individual pyramidal cell (PC) extends thousands of synapses over macroscopic distances (∼cm). The contrasting requirements of high-resolution and large field of view make it too challenging to acquire the entire synaptic connectivity for even a single typical cortical neuron. Light microscopy can image whole neuronal arbors and resolve dendritic branches. Analyzing connectivity in terms of close spatial appositions between axons and dendrites could thus bridge the opposite scales, from synaptic level to whole systems. In the mammalian cortex, structural plasticity of spines and boutons makes these "potential synapses" functionally relevant to learning capability and memory capacity. To date, however, potential synapses have only been mapped in the surrounding of a neuron and relative to its local orientation rather than in a system-level anatomical reference. Here we overcome this limitation by estimating the potential connectivity of different neurons embedded into a detailed 3D reconstruction of the rat hippocampus. Axonal and dendritic trees were oriented with respect to hippocampal cytoarchitecture according to longitudinal and transversal curvatures. We report the potential connectivity onto PC dendrites from the axons of a dentate granule cell, three CA3 PCs, one CA2 PC, and 13 CA3b interneurons. The numbers, densities, and distributions of potential synapses were analyzed in each sub-region (e.g., CA3 vs. CA1), layer (e.g., oriens vs. radiatum), and septo-temporal location (e.g., dorsal vs. ventral). The overall ratio between the numbers of actual and potential synapses was ∼0.20 for the granule and CA3 PCs. All potential connectivity patterns are strikingly dependent on the anatomical location of both pre-synaptic and post-synaptic neurons.