Whole-brain dynamics underlying spontaneous behavior in C. elegans

Abstract

Understanding how the brain drives behavior requires knowing the activity of every neuron in the circuit and the behavior that is produced from it. While obtaining such data in higher level organisms is currently unfeasible, the nematode Caenorhabditis elegans provides an attractive platform for studying the neural basis of behavior. With only 302 neurons and an optically transparent body, C. elegans is particularly amenable for observing an entire neural circuit. We have developed an imaging system that allows us to record calcium activity from every neuron in the worm’s brain while it is freely behaving. Additionally, we have developed a time-independent algorithm for automatically segmenting and tracking each neuron across a recording.

Using the methods we have developed for whole-brain imaging in freely behaving C. elegans, we sought to determine the size, separation, and dynamics of the circuits driving locomotion and turning. To do so, we used machine learning techniques to predict these behaviors from the neural activity, identifying the neurons involved in each of these behaviors. We find that neither behavior requires integration over previous neural activity, suggesting that they are both driven by the current magnitude of activity of their circuit components. Additionally, we find that the circuits for locomotion and turning are largely distinct, although the locomotion circuit does contain some information about turning. These results suggest that these two behaviors are driven by the magnitude of neural activity of separate circuits.