Modularity and neural coding from a brainstem synaptic wiring diagram

Neuronal wiring diagrams reconstructed from electron microscopic images are enabling new ways of attacking neuroscience questions. We address two central issues, modularity and neural coding, by reconstructing and analyzing a wiring diagram from a larval zebrafish brainstem. We identified a recurrently connected "center" within the 3000-node graph, and applied graph clustering algorithms to divide the center into two modules with stronger connectivity within than between modules. Outgoing connection patterns and registration to maps of neural activity suggested the modules were specialized for body and eye movements. The eye movement module further subdivided into two submodules corresponding to the control of the two eyes. We constructed a recurrent network model of the eye movement module with connection strengths estimated from synapse numbers. Neural activity in the model replicated the statistics of eye position encoding across multiple populations of neurons as observed by calcium imaging. Our findings show that synapse-level wiring diagrams can be used to extract structural modules with interpretable functions in the vertebrate brain, and can be related to the encoding of computational variables important for behavior. We also show through a potential synapse formalism that these modeling successes require true synaptic connectivity; connectivity inferred from arbor overlap is insufficient.