The preoptic area and dorsal habenula jointly support homeostatic navigation in larval zebrafish

Abstract

Animals must maintain physiological processes within an optimal temperature range despite changes in their environment. Through behavioral assays, whole-brain functional imaging, and neural ablations, we show that larval zebrafish, an ectothermic vertebrate, achieves thermoregulation through homeostatic navigation-non-directional and directional movements toward the temperature closest to its physiological setpoint. A brain-wide circuit encompassing several brain regions enables this behavior. We identified the preoptic area of the hypothalamus (PoA) as a key brain structure in triggering non-directional reorientation when thermal conditions are worsening. This result shows an evolutionary conserved role of the PoA as principal thermoregulator of the brain also in ectotherms. We further show that the habenula (Hb)-interpeduncular nucleus (IPN) circuit retains a short-term memory of the sensory history to support the generation of coherent directed movements even in the absence of continuous sensory cues. We finally provide evidence that this circuit may not be exclusive for temperature but may convey a more abstract representation of relative valence of physiologically meaningful stimuli regardless of their specific identity to enable homeostatic navigation.

Keywords: behavioral strategy; dorsal habenula; ectothermic vs. endothermic; homeostasis; innate behavior; preoptic area; sensory context; thermoregulation; valence; zebrafish.

Graphical abstract

Figure 1

Larval zebrafish maintain homeostasis in a linear shallow thermal gradient (A) Top: example trajectories of three fish for the entire duration of the experiment (15 min). Actual temperature measured in the water is pictured with different shades of red. The highest temperature (33°C) is represented on the left by the dark red. Bottom: histogram of animals’ x-position in the arena (n = 26, gray traces), median of the population (black trace) and temperature measurement (red trace). (B) Trajectories of all the fish split in 3-min bins. (C) Top: sketch of an example trajectory. Bottom: change of temperature perceived over time for each swim event in the top sketch. Time is coded by the shades of green (from dark to light green). Arrows on top point to a movement in a ICxt (improving context), NCxt (no context), and a WCxt (worsening context), respectively. (D) Sketch of information available to the fish for behavioral choice using the sensory context provided by the previous swim. (E) Summary of all the stimulus and behavioral variables used in this study for a representative individual fish. On the left: the entire duration of the experiment. On the right: close up of the trajectory and the relevant variables highlighted in the gray box. The four graphs show: (i) temperature experienced by an example fish, (ii) ΔTemperature (difference in temperature between current and last movement) experienced (color coding represents: worsening of conditions [in brown, away from physiological setpoint], improvement [in turquoise, toward physiological setpoint], or no change [in gray]), (iii) the absolute value of reorientation for each swim (color coding represents: turn [in red, when reorientation is higher than 30°], or forward [in blue, when reorientation is lower than 30°]), and (iv) the direction of turns, either left or right. See also Figure S1.

Figure 2

Homeostatic navigation is performed by combining an adirectional with a directional strategy (A) Relative distribution of the absolute angle turned when fish experienced a WCxt or an ICxt (mean ± standard error of the mean, Mann-Whitney nonparametric test with Bonferroni correction for multiple comparisons). (B) Normalized increase in turn fraction depending on sensory context (median ± standard error of the median, Mann-Whitney nonparametric test). (C) Direction of swim during WCxt according to the direction of the previous swim (mean ± standard error of the mean, Mann-Whitney nonparametric test). Positive values imply that the swim tends to be in the same direction as the previous one. Diamond on the right is a shuffle obtained by assigning the sign for each value randomly. (D) Left: behavioral setup. Right top: stimulus protocol for temporal gradient experiment. Right bottom: average normalized turning rate (red) and forward-swim rate (blue) for n = 40 fish. (E) Increase in turning fraction depending on sensory context for experiment 2, similar to (B) (median ± standard error of the median, Mann-Whitney nonparametric test). (F) Turn correlation upon WCxt for experiment 2, similar to (C) (mean ± standard error of the mean, Mann-Whitney nonparametric test). (G) Mean transition matrix for sensory context across fish population. (H) Top: example U-maneuver performed during experiment 1. Bottom: sketch of a U-maneuver. (I) Difference in temperature experienced by fish during the execution of U-maneuvers, starting from in a WCxt (mean ± standard error of the mean). See also Figures S2 and S3.

Figure 3

Identification of neurons encoding sensory context in a whole-brain screen (A) Left: schematic representation of the head-restrained preparation for the lightsheet setup. Top right: protocol as in Figure 2D. Bottom right: average normalized turning rate (red) and forward-swim rate (blue) for n = 11 fish (mean ± standard error of the mean). (B) Left: schematic representation of the multimodal experiment. Right: short protocol used for the lightsheet imaging experiments. (C) From left to right: anatomy of an example fish during the temperature and salt sessions. Yellow, green, cyan, and purple arrows indicate four different patterns of neurons visible in both sessions. Raw activity of 10% of neurons (randomly selected) in a dataset for both temperature and salt session. Arrangement of the traces has been done according to the rostro-caudal position of selected ROIs. (D) Temporal trajectories in PC space of the neuronal activity of reliable neurons, color-coded based on stimulus context (brown WCxt and turquoise ICxt). (E) Projection onto PC space of all reliable neurons color-coded according to their cluster identity. (F) Mean activity of each cluster (mean ± standard error of mean) and fraction of cells per group (tot n = 3,270). (G) Example raw traces from the three clusters and one motor ROI (all in black), the approximate stimulus profile (green) and the tail trace showing behavior (pink). (H) Relative frequency distribution of the difference in mean fluorescence change during the single and double pulse stimulation for each cluster. Negative values imply that the fluorescence increased with increasing absolute temperature. (I) Scatter plot of the average deviation from baseline fluorescence during single (on the left) and double pulse (on the right) split by sensory context (WCxt and ICxt) for the neurons in the three clusters (mean ± standard error of mean). (J) Left: anatomical distribution of reliable ROIs according to cluster identity. Different panels depict different projections. mOB, medial olfactory bulb; Pa, pallium; rHb, right habenula; PoA, preoptic area; IPN + RN, interpeduncular nucleus and raphe nucleus; rHb1, rhombomere 1. Right: sketch of hypothesized synaptic connections based on activity maps and literature. (K and L) Proportion of ROIs from the different clusters for different anatomical regions. OB, olfactory bulb; Pa (Pallium); Hb, habenula (left + right habenula); PoA, preoptic area; IPN + RN (interpeduncular nucleus + superior raphe nucleus), rHb1 (rhombomere 1). See also Figure S4.

Figure 4

A small population of dorsal habenula, medial olfactory bulb, and pallial neurons respond similarly to temperature and salinity changes (A) Anatomical density distribution of motor-triggered ROIs. Selected cells can be either forward, left, or right-swim tuned. (B) Fraction of motor-triggered ROIs for each sensory-identified region. (C) Trial trigger average of all multimodal neurons found during temperature (left) and salt (right) sessions. (D) Mean of the multimodal neuron clusters (k = 2, mean ± standard error of the mean). (E) Raw activity of two example multimodal neurons (one from the olfactory bulb and one from the habenula) in black for both the temperature and the salt session. (F) Anatomical density distribution of multimodal neurons for the three projections. mOB, medial olfactory bulb; mPa, medial pallium; mr-dHb, medial nucleus of the right dorsal habenula. See also Figure S5.

Figure 5

PoA and dorsal Hb ablations impair different aspects of homeostatic navigation (A) Sketch of 2-photon ablations targeting the PoA (depicted in green). Stars are the chosen ablation sites. The control ablation is in the optic tectum (OT) (see text). (B) Increase in turn fraction depending on sensory context. Left: control OT ablation, right: PoA ablation. (median ± standard error of the median, Mann-Whitney nonparametric test.) (C) Turn correlation upon WCxt for the control and ablation groups (mean ± standard error of the mean, Mann-Whitney nonparametric test). (D) Sketch of chemogenetic ablations targeting the dHb (depicted in orange). (E) Increase in turn fraction depending on sensory context. Left: genetic control, middle: treatment control, right: dHb ablation (median ± standard error of the median, Mann-Whitney nonparametric test). (F) Turn correlation upon WCxt for the two controls and the ablation group (mean ± standard error of the mean, Mann-Whitney nonparametric test). See also Figure S6.

Figure 6

Homeostatic navigation (A) Behavioral principles underlying homeostatic navigation in larval zebrafish uncovered by this study. Fish use two complementary strategies: an increase in turning (reorientation probability) when experiencing a WCxt (being pushed away from the homeostatic setpoint, left panel) coupled with turning in a persistent direction during this WCxt (right panel). The first strategy is PoA dependent, whereas the second depends on both the PoA and the dorsal Hb. (B) Neural circuitry involved in homeostatic navigation. Neurons representing the three main different stimulus features are represented in different colors and the anatomical shading shows the fraction of response types in each region.