Whole integration of neural connectomics, dynamics and bio-mechanics for identification of behavioral sensorimotor pathways in Caenorhabditis elegans

Abstract

The ability to fully discern how the brain orchestrates behavior requires the development of successful computational approaches to integrate and inform in-vivo investigations of the nervous system. To effectively assist with such investigations, computational approaches must be generic, scalable and unbiased. We propose such a comprehensive framework to investigate the interaction between the nervous system and the body for the nematode Caenorhabditis elegans (C. elegans). Specifically, we introduce a model that computationally emulates the activity of the complete somatic nervous system and its response to stimuli. The model builds upon the full anatomical wiring diagram, the connectome, and integrates it with additional layers including intra-cellular and extra-cellular bio-physically relevant neural dynamics, layers translating neural activity to muscle forces and muscle impulses to body postures. In addition, it implements inverse integration which modulates neural dynamics according to external forces on the body. We validate the model by in-silico injection of currents into sensory-and interneurons known to play a role in locomotion behaviors (e.g. posterior/anterior touch) and by applying external forces on the body. We are able to generate characteristic baseline locomotion behaviors (forward and backward movements). Inclusion of proprioceptive feedback, implemented through inverse integration, shows that feedback can entrain and sustain movements initiated by neural or mechanical triggers. We further apply neural stimuli, experimentally known to modulate locomotion, and show that our model supports natural behavioral responses such as turns, reversals and avoidance. The proposed model can be utilized to infer neural circuits involved in sensorimotor behavior. For this purpose, we develop large-scale computational ablation approaches such as (i) ablation survey and (ii) conditional ablation. Our results show how an ablation survey can identify neurons required for a ventral turning behavior. We also show how conditional ablation can identify alternative novel neural pathways, e.g. propose neurons which facilitate steering behavior towards olfactory attractants.The outcomes of our study show that the framework can be utilized to identify neural circuits, which control, mediate and generate natural behavior.

Deep Reinforcement Learning for Neural Control

“…The reason is that neural activity supported by the connectome and its relation to the target function are nonlinear dynamic processes [29]. The complexity is evident from C. elegans nervous system which connectome was fully mapped, however, ablations of the connectome cause outcomes that are difficult to predict without examination of neural activity and behavior [30][31][32].…”

“…As an example, we use the computational neuro-mechanical model of C. elegans to set up our experiments. The model consists of a simulation of a full nervous system activity with the body [31,32]. In order to mimic the typical chemotaxis environment, we extend the model to support spatial gradient stimuli which introduces stimulus into olfactory neurons in realistic way [36].…”

Statistical Perspective on Functional and Causal Neural Connectomics: A Comparative Study

“…While the anatomical connectome includes the backbone information on possible ways of how neurons could interact, it does not fully reflect the “wiring diagram of the brain”, which is expected to incorporate the dynamic nature of neurons' activity and their interactions (Lee and Reid, 2011 ; Kopell et al, 2014 ; Kim et al, 2019a , b ). In particular, the anatomical connectome does not correspond to how the anatomical structure relates to brain function since each anatomical connectivity map can encode several functional outcomes of the brain (Bargmann and Marder, 2013 ).…”

Statistical Perspective on Functional and Causal Neural Connectomics: A Comparative Study

“…While the anatomical connectome includes the backbone information on possible ways of how neurons could interact, it does not fully reflect the "wiring diagram of the brain", which is expected to incorporate the dynamic nature of neurons' activity and their interactions [5][6][7][8]. In particular, the anatomical connectome does not correspond to how the anatomical structure relates to brain function since each anatomical connectivity map can encode several functional outcomes of the brain [9].…”