Consistent individual differences in behavior are prevalent throughout the animal kingdom and are thought to be important contributors to evolutionary adaptation. However, the biological basis for individual differences is poorly understood. One explanatory framework that has gained traction is the pace of life syndrome (POLS) hypothesis. The POLS hypothesis proposes that behavioral variation arises from variation in basic physiological functions like metabolism. In particular, the POLS predicts that individuals with higher baseline metabolic demands will be more willing to take risks to attain the requisite resources. To date, support for this hypothesis when applied within species has been mixed, leading to the tentative conclusion that the relationship between metabolism and behavior depends on factors like species, sex, and context. We sought to determine if the POLS hypothesis held in zebrafish, a widely used model organism with well-developed genetic and neurobiological tools that would allow us to better understand how biological and environmental factors influence predictions of the POLS hypothesis. We tested the POLS hypothesis in adult zebrafish by measuring the relationship between routine metabolic rate, as assessed by oxygen consumption, and exploration of a novel tank. We found no clear relationship between boldness-related behaviors in the novel tank and metabolism in female or male zebrafish of the TU or WIK strains. Our findings suggest that a key prediction of the POLS hypothesis does not hold in zebrafish.

Subjects

Subjects were TU (Tübingen) and WIK (Wild India Kolkata) zebrafish aged 4 to 8 months. The TU strain was established from a pet shop in the 1990’s and was bred to remove embryonic lethal mutations (Haffter et al., 1996; Mullins et al., 1994) and used to generate the zebrafish genome (Howe et al., 2013). In contrast, the WIK line was more recently established from wild caught fish. They are more genetically polymorphic than the TU line (Rauch et al., 1997), more closely resembling wild animals (Wilson et al., 2014). All fish were bred and raised at Wayne State University and within two generations of animals obtained from the Zebrafish International Resource Center at the University of Oregon. Animals were kept on high density racks (20-40 fish per 4 or 8 liter tanks at a density of ~5 fish per liter) under standard conditions (temperature 27.5 ± 0.5 °C; water conductivity 500 ± 10 μS, and a pH of 7.5 ± 0.2) with a 14:10 light:dark cycle (lights on at 08:00). Fish were fed twice a day with a dry feed in the morning (Gemma 300, Skretting, Westbrook, ME, USA) and brine shrimp (Artemia salina; Brine Shrimp Direct, Ogden, UT, USA) in the afternoon. Behavioral testing took place between 11:00 and 14:00. The sex of fish was determined using three secondary sexual characteristics: shape (prominent belly for females), color (with males being more red/pink), and presence of pectoral fin tubercles (exclusively found on males; (McMillan et al., 2015)). We dissected animals after experiments to confirm the presence or absence of eggs. Animals that were assigned the wrong sex were removed from analysis. One fish with negative oxygen consumption was removed from analysis as this was likely due to a data entry error. All procedures were approved by the Wayne State University Institutional Animal Care and Use Committee.

Novel tank test

The novel tanks consisted of five-sided tanks (15 × 15 × 15 cm) made from frosted acrylic (TAP Plastics, San Leandro, CA, USA). Tanks were placed in an enclosure of white corrugated plastic to diffuse light and prevent exposure to external stimuli. D435 Intel RealSense™ cameras (Intel, Santa Clara, CA, USA) were mounted 20 cm above tanks to capture three-dimensional videos (Kuroda, 2018; Rajput et al., 2022). These cameras capture three-dimensional videos using the synchronous capture of two video streams: a color stream (red / green / blue) and a depth stream. The depth stream is generated via stereoscopic imaging using the disparity between two infrared cameras. Firmware on the camera synchronizes the capture of the two streams. Cameras were connected to a Linux workstation via high-speed USB cables (NTC distributing, Santa Clara, CA, USA), and video capture was controlled via custom written Python scripts. Animals with videos that were not fully recorded due to malfunction were not analyzed. Experimental tanks were filled with 2.5 L of fish facility water. Individual fish were placed in the tanks for 6 minutes while video was recorded for offline analysis. Tanks were rinsed between animals and water was replaced.

One week prior to testing, fish were placed as male/female pairs into 2 L tanks to enable non-invasive identification across days. The tanks were divided in half using a transparent divider with two fish in each section and four fish in each tank. On days when behavior was assessed, fish were taken off housing racks and moved to the procedural space at least 1 hour prior to testing. Following testing, fish waited at least thirty minutes before being returned to the housing racks.

Animal tracking and behavioral analysis

Fish were tracked along five points on the body using DeepLabCut (Mathis et al., 2018). The model was trained as previously described (Rajput et al., 2022). Animals unable to be accurately tracked were removed from analysis (1 fish). After tracking, four parameters were extracted: distance from tank center, distance travelled, distance from bottom, and percentage explored, wherein the tank was divided into 1000 voxels, and the percentage of explored voxels was calculated.

We generated a boldness index that combined the z-scores for the percentage of the tank explored and the bottom distance. These were calculated within each sex and strain. The boldness index is based on prior work that found that these two measures best distinguished bold and shy fish (Beigloo et al., 2024; Rajput et al., 2022).

We also calculated the percentage of time fish spent immobile and their maximum velocity. For immobility, we first plotted a distribution of velocities and found a break in the data at ~1.5 mm/s, which we used as a cut-off for determining immobility. For maximum velocity, we used the median of the top 5% of velocity measurements for each fish.

Dissolved O₂ measurements

Dissolved O₂ measurements were taken while fish swam in custom built tanks (15.2 cm [L] x 8.9 cm [W] x 6.0 cm [H]) made from P95 clear and frosted acrylic (4.8 mm thick; TAP Plastics). Each tank was filled with 500 mL of facility water and placed in the same enclosure used for novel tank tests. The tank lid consisted of two layers: (1) a top layer of P99 non-glare acrylic (3.2 mm thick) and (2) a bottom layer of clear cast acrylic (4.8 mm), sealed with a bead of aquarium silicone (Aqueon, Franklin, WI, USA) along the rim to ensure an air-tight seal.

Dissolved O₂ was measured using an optical meter (DO850, Apera Instruments, Columbus, OH, USA) as mg of O₂ per liter. We measured the initial and final O₂ concentrations before and after fish were placed in the sealed tank for 30 minutes. To account for changes in dissolved oxygen unrelated to fish metabolism, such as degassing or temperature changes, oxygen levels in a ‘blank tank’ containing the same volume of water, but no fish, were simultaneously recorded. This change in dissolved oxygen of the ‘blank tank’ was subtracted from that of the tank containing the fish to obtain each fish’s overall oxygen consumption:

[O₂]_consumped = ([O₂]_f,t=1 - [O₂]_f,t=2) - ([O₂]_b,t=1 - [O₂]_b,t=2)

Where [O₂]_consumed is the amount of oxygen consumed by the fish, [O₂]_f is oxygen measured in the chamber containing the fish, [O₂]_b is the oxygen measured in the ‘blank tank’, and t = 1,2 corresponds to the beginning and end of the trial, respectively.

16-hour fasting

Male and female fish were randomly assigned to either the fasted (“unfed”) or control (“fed”) groups prior to the experiment. Fish were fasted for approximately 16 hours by withholding their morning feed. Control groups were fed their morning feed approximately 1 hour prior to the experiment. 15 minutes after feed administration, fish were transported to the experimental room and left to habituate for 1 hour. Individual fish of each group were then placed into the dissolved oxygen chamber for 30 minutes to capture their metabolic activity. Two days later, individuals were switched between groups and oxygen consumption was measured. A separate cohort of fish were divided into fed and unfed groups and tested in the novel tank as described above.

Coding and statistical analysis

Data analysis was performed using R version 4.5.0 (R Core Team, 2016). All graphs were generated using ggplot2 (Wickham, 2016). Normality was assessed using the Shapiro-Wilks test. For correlations, we used Pearson’s correlations if the normality assumption was not violated otherwise we used Spearman’s ρ. To determine repeatability, we used intraclass correlations (ICC) and calculated 95% confidence intervals (CI) by bootstrapping 1,000 times with replacement using the rptr package (version 0.9.23). For comparing four groups we used permutation ANOVAs (with 10,000 resamples). For comparing two groups, we used Welsch’s independent sample t-tests, unless the normality assumption was violated, then we used the Mann-Whitney U test for independent samples or Wilcoxon rank-sum tests for paired samples. Effect sizes were calculated using Cohen’s d for two groups or η² for ANOVAs. Interpretation of effect sizes were small (0.01 < η² < 0.06; 0.2 < d < 0.5), medium (0.06 ≤ η² < 0.14; 0.5 ≤ d < 0.8) or large (η² ≥ 0.14; d ≥ 0.8) based on Cohen (Cohen, 1988).

References

Beigloo, F., Davidson, C. J., Gjonaj, J., Perrine, S. A., & Kenney, J. W. (2024). Individual differences in the boldness of female zebrafish are associated with alterations in serotonin function. Journal of Experimental Biology, 227(12), jeb247483. https://doi.org/10.1242/jeb.247483

Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences (Rev. ed.). Lawrence Erlbaum Associates. https://doi.org/10.4324/9780203771587.

Haffter, P., Granato, M., Brand, M., Mullins, M. C., Hammerschmidt, M., Kane, D. A., Odenthal, J., J. M. van Eeden, F., Jiang, Y.-J., Heisenberg, C.-P., Kelsh, R. N., Furutani-Seiki, M., Vogelsang, E., Beuchle, D., Schach, U., Fabian, C., & Nüsslein-Volhard, C. (1996). The identification of genes with unique and essential functions in the development of the zebrafish, Danio rerio. Development, 123(1), 1–36. https://doi.org/10.1242/dev.123.1.1

Howe, K., Clark, M. D., Torroja, C. F., Torrance, J., Berthelot, C., Muffato, M., Collins, J. E., Humphray, S., McLaren, K., Matthews, L., McLaren, S., Sealy, I., Caccamo, M., Churcher, C., Scott, C., Barrett, J. C., Koch, R., Rauch, G. J., White, S., … Stemple, D. L. (2013). The zebrafish reference genome sequence and its relationship to the human genome. Nature, 496(7446), 498–503. https://doi.org/10.1038/nature12111

Kuroda, T. (2018). A system for the real-time tracking of operant behavior as an application of 3D camera. Journal of the Experimental Analysis of Behavior, 110(3), 522–544. https://doi.org/10.1002/jeab.471

Mathis, A., Mamidanna, P., Cury, K. M., Abe, T., Murthy, V. N., Mathis, M. W., & Bethge, M. (2018). DeepLabCut: Markerless pose estimation of user-defined body parts with deep learning. Nature Neuroscience, 21(9), 1281–1289. https://doi.org/10.1038/s41593-018-0209-y

McMillan, Stephanie. C., Géraudie, J., & Akimenko, M.-A. (2015). Pectoral Fin Breeding Tubercle Clusters: A Method to Determine Zebrafish Sex. Zebrafish, 12(1), 121–123. https://doi.org/10.1089/zeb.2014.1060

Mullins, M. C., Hammerschmidt, M., Haffter, P., & Nüsslein-Volhard, C. (1994). Large-scale mutagenesis in the zebrafish: In search of genes controlling development in a vertebrate. Current Biology, 4(3), 189–202.

R Core Team. (2016). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Http://Www.R-Project.Org/. https://cir.nii.ac.jp/crid/1574231874043578752

Rajput, N., Parikh, K., & Kenney, J. W. (2022). Beyond bold versus shy: Zebrafish exploratory behavior falls into several behavioral clusters and is influenced by strain and sex. Biology Open, 11(8), bio059443. https://doi.org/10.1242/bio.059443

Rauch, G.-J., Granato, M., & Haffter, P. (1997). A polymorphic zebrafish line for genetic mapping using SSLPs on high-percentage agarose gels. Technical Tips Online, 2(1), 148–150. https://doi.org/10.1016/S1366-2120(08)70068-0

Wickham, H. (2016). Data Analysis. In H. Wickham (Ed.), Ggplot2: Elegant Graphics for Data Analysis (pp. 189–201). Springer International Publishing. https://doi.org/10.1007/978-3-319-24277-4_9

Wilson, C. A., High, S. K., McCluskey, B. M., Amores, A., Yan, Y., Titus, T. A., Anderson, J. L., Batzel, P., Carvan, M. J., III, Schartl, M., & Postlethwait, J. H. (2014). Wild Sex in Zebrafish: Loss of the Natural Sex Determinant in Domesticated Strains. Genetics, 198(3), 1291–1308. https://doi.org/10.1534/genetics.114.169284