Abstract
Brain networks, localized or brain-wide, exist only at the cellular level, i.e., between specific pre- and post-synaptic neurons, which are connected through functionally diverse synapses located at specific points of their cell membranes. "Connectomics" is the emerging subfield of neuroanatomy explicitly aimed at elucidating the wiring of brain networks with cellular resolution and a quantified accuracy. Such data are indispensable for realistic modeling of brain circuitry and function. A connectomic analysis, therefore, needs to identify and measure the soma, dendrites, axonal path, and branching patterns together with the synapses and gap junctions of the neurons involved in any given brain circuit or network. However, because of the submicron caliber, 3D complexity, and high packing density of most such structures, as well as the fact that axons frequently extend over long distances to make synapses in remote brain regions, creating connectomic maps is technically challenging and requires multi-scale approaches, Such approaches involve the combination of the most sensitive cell labeling and analysis methods available, as well as the development of new ones able to resolve individual cells and synapses with increasing high-throughput. In this review, we provide an overview of recently introduced high-resolution methods, which researchers wanting to enter the field of connectomics may consider. It includes several molecular labeling tools, some of which specifically label synapses, and covers a number of novel imaging tools such as brain clearing protocols and microscopy approaches. Apart from describing the tools, we also provide an assessment of their qualities. The criteria we use assess the qualities that tools need in order to contribute to deciphering the key levels of circuit organization. We conclude with a brief future outlook for neuroanatomic research, computational methods, and network modeling, where we also point out several outstanding issues like structure-function relations and the complexity of neural models.