中央研究院 資訊科學研究所

One of the ultimate goals of computational neuroscience is to quantitatively connect between complex neural circuits and behaviors. In the past decades, the touch response circuit in Caenorhabditis elegans has been extensively investigated in studies using genetically modified or laser-ablated worms. Synaptic connections, including chemical and electrical synapses, have been identified for most neurons in C. elegans. However, we still do not know whether the empirically observed touch responses can be derived from connectome reconstructed from databases. To address this issue, we defined the transmission abilities (or levels) of neurons in a firing-rate model in order to infer the behaviors of wild-type and ablated worms in response to posterior/nose/anterior touch stimuli. Our analysis showed that transmission abilities can be used to identify sensorimotor mapping from stimuli to movements and then to infer C. elegans behaviors under simulations based on the perspective of decision-making, and provide useful information about how chemical and electronic synapses should be combined in neural network movement analysis. The present study reveals an efficient tool that provided insights into the functions of complex neural circuits.