C. elegans foraging as a model for understanding the neuronal basis of decision-making

Abstract

Animals have evolved to seek, select, and exploit food sources in their environment. Collectively termed foraging, these ubiquitous behaviors are necessary for animal survival. As a foundation for understanding foraging, behavioral ecologists established early theoretical and mathematical frameworks which have been subsequently refined and supported by field and laboratory studies of foraging animals. These simple models sought to explain how animals decide which strategies to employ when locating food, what food items to consume, and when to explore the environment for new food sources. These foraging decisions involve integration of prior experience with multimodal sensory information about the animal's current environment and internal state. We suggest that the nematode Caenorhabditis elegans is well-suited for a high-resolution analysis of complex goal-oriented behaviors such as foraging. We focus our discussion on behavioral studies highlighting C. elegans foraging on bacteria and summarize what is known about the underlying neuronal and molecular pathways. Broadly, we suggest that this simple model system can provide a mechanistic understanding of decision-making and present additional avenues for advancing our understanding of complex behavioral processes.

Keywords: Dietary choice; Exploitation; Exploration; Patch-leaving.

Fig. 1

Strategies and mechanisms for food search. (a) A representative track of C. elegans initiating a reorientation (pirouette) while moving down a concentration gradient, a hallmark of klinokinesis. The animal shown locomotes forward at 0 s, reverses by 6 s, and, by 12 s, reorients its body to face a new direction with continued forward movement. (b) A representative track of a curved path (weathervane) initiated during movement perpendicular to a concentration gradient, a process known as klinotaxis. The animal shown locomotes forward at 0 s, initiates a turning bias towards higher concentration by 6 s, and, by 12 s, locomotes up the concentration gradient. (c) A simplified anatomical diagram of C. elegans showing the head and tail, dorsal and ventral sides, and pharynx. (d) Diagram of an example circuit for the encoding of several chemical stimuli within the C. elegans nervous system. A subset of sensory neurons (triangles), interneurons (hexagons), motor neurons (circles), and muscle groups (rectangles) known to be involved in the klinotaxis and klinokinesis responses to NaCl, 2-nonanone, and isoamyl alcohol (IAA) are shown. (e) Example traces of neuronal activity from calcium imaging experiments of the bilaterally symmetric sensory neuron pair, ASEL and ASER. In response to the addition or removal of NaCl, ASEL and ASER respond as on- and off-cells, respectively. The neuronal activity plotted was taken and adapted from [78]

Fig. 2

Overview of C. elegans foraging decisions. (a) C. elegans navigate towards an odor source in an environment where a spatial concentration gradient of a food-related odor is present. C. elegans odor-guided navigation is characterized by a decrease in the probability of reorientation with increasing rate of change in concentration (klinokinesis) and an increase in the turning bias with an increase in the concentration gradient (klinotaxis). (b) C. elegans conduct an initial local search followed by a more exploratory global search in an environment where no food-related sensory cues are present. The transition from local to global search is characterized by a decrease in the probability of reorientation and an increase in the search area as a function of time since the animal’s last encounter with food. (c) Detection of food is characterized by an abrupt slow-down upon encounter with the patch edge. (d) In dietary choice assays, preference for different bacterial food types is often not initially observed as animals must sample bacteria before ascribing it a subjective value. Therefore, the probability of residing on a higher quality patch develops over time. This dietary choice behavior is driven by modulation of exploratory and exploitative behavioral states with animals more likely to be quiescent or dwelling on high quality patches and roaming on low quality patches. (e) When animals forage in a group, assessment of within-patch spatial heterogeneity leads to resource partitioning with gregarious strains preferring the dense bacterial patch border and solitary strains distributing themselves proportionately with the bacterial density. (f) C. elegans may decide to leave or stay upon encounter with the bacterial patch edge. The probability of leaving is higher for less dense and lower quality bacterial patches and further increases as a function of resource depletion as the bacteria is consumed over time

Fig. 3

Foraging decisions are often cyclical. A decision tree of the C. elegans foraging decisions described in this review is shown. Corresponding elements to Fig. 2 are indicated. Actions regarding alternative behaviors, area-restricted search, navigation, resource partitioning, and exploitation are described in blue-hued boxes. Decisions regarding an animal’s current motivation, food search, food detection, dietary choice, spatial distribution, and patch-leaving are described in purple-hued boxes. Decisions result in binarized consequences (yes – green check; no – red X) leading to the next decision or action