Establishment of an effort discounting paradigm in C. elegans.

A. Pharyngeal pumping frequency in suspensions of normal (CON) and elongated (CEPH) bacteria. Symbols (box and whisker plot): filled circles, single-worm data points; diamonds, means; numbers, sample size.

B1. Quantification of Oil-Red-O staining of lipid in worms cultured on normal (CON) or elongated (CEPH) bacteria. Median worms from each condition are shown.

B2. Mean pixel intensity of Oil-Red-O staining of lipid in worms cultured on normal (CON) or elongated (CEPH) bacteria. Diamonds, means; numbers, sample size.

C-D. Protein and carbohydrate in worms cultured on normal (CON) and elongated bacteria (CEPH). Columns of data points of the same color are biological replicates run in parallel. Symbols: filled circles, technical replicates (n = 6); diamonds, means of technical replicates.

E. Lipid concentration in worms cultured on normal (CON) and elongated bacteria (CEPH). Symbols: filled circles, biological replicates (n = 3); diamonds, means of biological replicates. F1. T-maze assay for relative attractiveness of normal and elongated bacteria. Food patches contained sodium azide as a paralytic agent.

F2. Mean preference index in the T-maze assay. Error bars, ± 95% CI.

Statistical information. Bold p values denote significant effects.

Food preference in T-maze assays.

A. B. T-maze assay for preference for normal (Control) versus elongated bacteria (cephalexin).

B. Effect on preference of raising the optical density of elongated bacteria (CEPH). Mean preference index is shown for assays in which three different optical densities (OD) of elongated bacteria (CEPH) were paired with normal bacteria (CON) at OD 1.0. Worms are indifferent when elongated bacteria are at OD 1.5. Numbers, sample size. Error bars ± 95% CI

C. Tests of predicted indifference points. Mean preference index is shown for three different pairs of optical densities of normal (CON) and elongated (CEPH) bacteria. The grey bar is the indifference point found in panel B. The black bars are novel indifference points found by scaling the optical density of elongated bacteria by the discount factor of 2/3 obtained from the grey bar. Numbers, sample size. Error bars ± 95% CI.

Kinetic analysis of food-patch leaving assays.

A. Food-patch leaving assays for normal (control) and elongated (cephalexin-treated) bacteria. Open circles, arenas; colored circles, food. A representative track of a single worm is superimposed on each diagram.

B. Three-state kinetic model of behavior in the assay. Each state represents the indicated zone in the arena.

C. Probability of states F, B, and O for normal (CON) and elongated (CEPH) bacteria. Significance was assessed using compositional statistics as described in Fig. 3––figure supplement 1.

D. Rate constants for normal (CON) and elongated (CEPH) bacteria.

E. Locomotion speed for normal (CON) and elongated (CEPH) bacteria on or off food.

Effect of dopamine signaling mutations on preference for normal versus elongated bacteria.

Data are mean preferences measured in T-maze assays. Dopamine signaling was reduced by a nonsense mutation in cat-2 or by deletions of three dopamine receptor genes (dop). Dopamine signaling was increased by a deletion in dat-1. Numbers, sample size. Error bars ± 95% CI.

Preference for normal versus elongated bacteria in N2 and five wild-isolate strains.

Data are mean preferences measured in T-maze assays. N2 and DL238 are significantly different from zero. Numbers, sample size. Error bars ± 95% CI.

Standard models of effort discounting.

There are three main models of effort discounting, each with a unique equation relating effort and reward. 𝑉0 is the value of the reward if it were given effort-free. 𝐸 is the level of effort required to obtain the effortful reward. 𝑉(𝐸) is the discounted value of 𝑉0under conditions of indifference, given level of effort 𝐸. The quantity 𝛼 is a positive constant representing the chooser’s sensitivity to effort. The quantity 𝑠 in the generalized hyperbolic model is the shape parameter. Discount factors are constant at fixed 𝐸. In the graph, 𝑉0 = 100, 𝛼 = 0.05, and 𝑠 = 1.5.

Statistical test of the effect of cephalexin on state probabilities in Figure 3C.

State probabilities are an instance of compositional data, meaning data that sum to a constant (1 in this case). Such data require special treatment because a decrease in one variable means an increase in at least one other variable. Accordingly, data were first subjected to an isometric log ratio transform (see Materials and Methods). The transformation quantifies the change in balance between subsets of the data. Balances are indicated by the numbers in the figure. Here, the control balance between 𝑃Fand the reciprocal probabilities 𝑃Band 𝑃O(1.47) is decreased by cephalexin treatment (0.85). A t-test on individual worm balances revealed that 𝑃Fsignificantly decreased while the joint 𝑃Band 𝑃Odata significantly increased. In contrast, a similar approach showed that the balance between 𝑃Band 𝑃Owas not significantly affected.

This indicates that 𝑃B and 𝑃O increased by similar amounts.