Recent improvements on computer processing capacity and imaging technologies have opened a new horizon for making extensive connectivity maps of nervous systems. The electron microscopy (EM)-based connectomics made it possible to identify individual synaptic connections of neurons comprehensively in large areas, and has gradually gained a unique and important position in the field of neuroanatomy. Through connectomics projects, various circuit models for information processing have been established regardless of species or systems. Such connectivity information has provided a robust platform for functional analyses of the circuits.
Connectome researches can be categorized largely into two types; one of these is called the dense reconstruction, which aims to identify all neurons and their synaptic connections within a region of interest comprehensively. The volume of a region of interest can be as large as 100~1,000 μm³. This approach can reveal the neuronal connectivity of a whole system, however, a long time and significant amount of human power will be required to process the huge amount of data. The alternate way is the sparse reconstruction. This is an approach to identify neuronal circuits by selectively focusing on limited number of neurons of interest. Compared to the dense reconstruction, time and effort required for analysis can be drastically reduced with this method, as the neurons to analyse are specified in advance.
Applying the sparse reconstruction method, we analysed neuronal connections in the lobula, a neuropil in the optic lobe of the fruit fly brain, to identify motion information processing pathways. We expressed Gal4-driven membrane-targeted horseradish peroxidase (HRP) in a type of the cells in the circuit, to trace and reconstruct their highly dendritic terminals efficiently. By combining the connectivity information and transcript profiling data of the neurons, a candidate substrate for a motion detection circuit in the lobula was identified as a result. Using these cases as examples, we compare the two strategies, and discuss their advantages, technical difficulties and the solutions.