FIGURE SUMMARY
Title

Feeding State Modulates Behavioral Choice and Processing of Prey Stimuli in the Zebrafish Tectum

Authors
Filosa, A., Barker, A.J., Dal Maschio, M., Baier, H.
Source
Full text @ Neuron

Decision between Approach versus Avoidance Is Modulated by Feeding State

(A) Scheme of the behavioral setup.

(B and C) Examples of starved (B) and fed (C) 7 dpf zebrafish larvae demonstrating easy identification of individuals with food in their digestive tract (red arrows).

(D and E) Graphs depicting average valence (D) and activity (E) indeces of starved and fed 7 dpf larvae for various stimulus sizes (plotted on logarithmic scale). The valence index measures the tendency of animals to either pursue (positive valence) or avoid (negative valence) visual stimuli, without considering pursuit and avoidance efficiency. When the visual index is 0, larvae pursue or avoid circles with a 50% probability. The fish-visual stimulus interaction efficiency is measured with the activity index. *p < 0.05; t test with Benjamini-Hochberg correction. Data are presented as mean ± SEM.

See also Movies S1 and S2.

Serotonin Modulates Behavioral Choice Downstream of Feeding State

(A-C) Confocal images of a 7 dpf tph2:Gal4ff, UAS:GFP larva showing serotonergic neurons in the raphe nucleus (A), and immunofluorescence detection of phosphorylated ERK (pERK, B) and total ERK (tERK, C). A, anterior; P, posterior.

(D) Plot showing average pERK/tERK staining intensity ratios in raphe serotonergic neurons of 7 dpf fed or starved larvae. *p = 0.04, t test.

(E and F) Plots showing valence (E) and activity (F) indices of starved and fed larvae, treated with 1.5 µM fluoxetine, and untreated controls. *p < 0.05; **p < 0.01; p < ***0.001, t test with Benjamini-Hochberg correction. Asterisks indicate statistical significance only for the control fed and fluoxetine fed comparison. Statistical results are summarized in Table S2.

(G and H) Graphs depicting valence (G) and activity (H) indices for fish lacking serotonergic neurons, and control larvae. *p < 0.05; **p < 0.01, t test with Benjamini-Hochberg correction. Data are presented as mean ± SEM.

See also Table S2.

Feeding State Affects Visual Information Processing in the Tectum

(A) Scheme of the zebrafish retinotectal circuit and experimental setup.

(B) Two-photon image of the region of the tectum of a 7 dpf elavl3:Gal4, UAS:GCaMP5 larva used for imaging.

(C) ΔF/F traces of PVNs responsive to different visual stimulus sizes. Vertical red bars indicate the presence of visual stimuli.

(D) Normalized ΔF/F values (gray bars) obtained from the PVN responses shown in (C) and the corresponding WMR angles (vertical red lines).

(E and F) Graphs comparing cumulative percentages of WMR angles for PVNs in starved and fed (E) or starved and starved + odor-exposed (F) 7 dpf elavl3:Gal4, UAS:GCaMP5 larvae. ***p = 1.9E-16; n.s., not significant (two-sample Kolmogorov-Smirnov test). Data are presented as mean ± SEM.

The Influence of Feeding State Is Not Detectable in RGC Axons in the Tectal Neuropil

(A) Two-photon image of the tectal neuropil of a 7 dpf Isl2b:Gal4, UAS:GCaMP6s larva showing RGC axons. A, anterior; L, lateral; M, medial; P, posterior.

(B) Average distributions of the T scores obtained from the pixel-wise analysis of the activity of RGC axons, in response to visual stimuli of different sizes (in degrees of visual angle) in fed or starved 7 dpf Isl2b:Gal4, UAS:GCaMP6s larvae. The procedure for calculating the T scores is summarized in Figure S2.

(C) Average numbers of pixels active in response to visual stimuli of different sizes. Background noise was removed by subtraction of a component related to image time series obtained in the absence of visual stimulation (see Experimental Procedures for details). Error bars represent SEM. No statistically significant differences were detected between fed and starved animals (t test with Benjamini-Hochberg correction). n = number of trials.

See also Figure S2.

Feeding State Affects the Activity of Specific Populations of Neurons in the Tectum

(A) Two-photon image showing GABAergic neurons in the tectum of a 7 dpf gad1b:Gal4, UAS:GCaMP6s larva.

(B) Graph comparing cumulative percentages of WMR angles for tectal GABAergic neurons in starved and fed 7 dpf gad1b:Gal4, UAS:GCaMP6s larvae. ***p = 3.9E-5, two-sample Kolmogorov-Smirnov test. Data are presented as mean ± SEM.

(C) Two-photon image showing part of the tectum of a 6 dpf Gal4mpn354, UAS:GCaMP6s larva. The red and yellow dashed lines mark the brain midline and tectal neuropils, respectively.

(D) Graph showing numbers of Gal4mpn354 cells tuned to small (≤10°) or large (≥20°) visual stimuli. Dually tuned neurons responded to both small and large stimuli. Examples of the three types of responses are shown in Figure S3. *p = 0.02, Fisher’s exact test.

(E) Two-photon image showing part of the tectum of a 7 dpf Gal4s1038t, UAS:GCaMP6s larva.

(F) Plot showing numbers of Gal4s1038t neurons tuned to visual stimuli of different sizes in starved and fed animals. n.s., not significant, Fisher’s exact test.

A, anterior; L, lateral; M, medial; P, posterior; PVN, peri-ventricular neuron.

See also Figure S3.

The HPI Axis and the Serotonergic System Modulate Visual Information Processing in the Tectum

(A) Graph depicting cumulative percentages of WMR angles for PVNs in starved 7 dpf elavl3:Gal4, UAS:GCaMP5, grs357/s357 and starved control elavl3:Gal4, UAS:GCaMP5, gr+/+ larvae. p = 0.02, two-sample Kolmogorov-Smirnov test.

(B-D) Confocal images of a 7 dpf shha:gfp, tph2:Gal4ff, UAS:ntr-mCherry larva showing the presence of serotonergic innervation (red neurites in C and D) in the tectum neuropil (GFP-positive RGC axons in B and D). A, anterior; L, lateral; M: medial; NTR, nitroreductase; P, posterior; SAC, stratum album centrale; SFGS, stratum fibrosum et griseum superficiale; SGC, stratum griseum centrale; SO, stratum opticum.

(E) Graph comparing average cumulative percentages of WMR angles for PVNs of fed elavl3:Gal4, UAS:GCaMP5 larvae treated with 1.5 µM fluoxetine and untreated fed or starved controls. Fluoxetine abolished the satiety-induced change of PVN population response. ***p < 0.001, two-sample Kolmogorov-Smirnov tests. Data are presented as mean ± SEM.

Serotonin Activates a Population of Tectal Neurons Tuned to Small Visual Stimuli in Fed Larvae

(A and B) Two-photon images of a tectum of a fed 7 dpf elavl3:Gal4, UAS:GCaMP5 larva before (A) and after (B) application of 1.5 µM fluoxetine for 3.5-4 hr. PVN, peri-ventricular neuron.

(C) Bar graph showing the average percentages of neurons active only in the first (lost) or second (gained) recording session, or at both times (persistent). In the first session, fed 7 dpf elavl3:Gal4, UAS:GCaMP5 larvae were imaged. In the second session, the same animals were imaged again in the absence (control group) or presence of fluoxetine. *p < 0.04; n.s., not significant (t test with Benjamini-Hochberg correction). n = 13 (fluoxetine) and 9 (control) larvae.

(D) Average percentages of nonpersistent neurons (corresponding to the “lost” and “gained” groups in C) tuned to small-sized visual stimuli. *p = 0.05, paired t test. Error bars represent SEM.

(E) Scatterplot displaying WMR angles of persistent neurons in the first and second imaging sessions. Each circle represents a neuron. No statistically significant differences between the two groups were detected (p = 1, ANOVA). n = number of neurons. The green and purple lines represent linear regressions of the data relative to the fluoxetine-treated (r2 = 0.26) and control (r2 = 0.27) group, respectively. The dashed gray line marks the position occupied by data points relative to neurons whose response properties were the same in the two imaging sessions.

(F) Scheme summarizing the results. A satiety signal, potentially from the digestive system, activates the HPI axis, which, in turn, inhibits serotonergic neurons in the raphe nucleus sending axons to the tectum. Satiety may alter visuomotor transformations in the tectum in at least two ways (inset). First, it may increase avoidance of small objects by inhibiting a population of GABAergic neurons responsive to small- and medium-sized visual stimuli. Second, it decreases pursuit of small objects by inhibiting mpn354 PVNs. This modulation biases downstream motor circuits toward avoidance of small- and medium-sized objects. Additional fine-tuning by feeding state of other neural circuits could also contribute to the final behavioral choice (gray dashed arrows).

Acknowledgments
This image is the copyrighted work of the attributed author or publisher, and ZFIN has permission only to display this image to its users. Additional permissions should be obtained from the applicable author or publisher of the image. Full text @ Neuron