Data from: Body condition reveals hidden correlations between co-infection and behavior in sunfish
Data files
Jan 12, 2026 version files 49.30 KB
-
complete_final.csv
9.93 KB
-
complete2.csv
6.84 KB
-
README.md
9.55 KB
-
script_final.R
22.98 KB
Abstract
The role of parasites in maintaining consistent inter-individual differences in behavior (i.e., personality) is the subject of increasing study and debate. While behavioral differences may expose individuals differently to parasites, parasite infection can itself change host behavior, sometimes favoring the parasite’s own transmission. Furthermore, parasites can alter the functioning of energetically costly organs like the brain, thus affecting cognitive performance. However, relationships among infection, cognition, and behavior can be complex and difficult to interpret, especially in wild populations where individual health status is unknown. The inclusion of body condition as a fitness proxy may help reveal relationships between parasites and host traits that are otherwise masked. We examined relationships among host body condition, personality (i.e., exploration, boldness), cognition (i.e., aversive learning), and parasite density in wild-caught pumpkinseed sunfish (Lepomis gibbosus) naturally infected with endoparasites. We measured exploration (i.e,. surface explored in cm), boldness (i.e,. latency to emerge from shelter in seconds) and cognition (i.e. aversive learning; difference between trial 1 and 2 in latency time to enter the black side). Exploration and boldness were measured twice (trial 1 and 2). Cognitive performance had one learning trial, and two test trials (24-hour delay between trials). Exploration in an open field test was repeatable in sunfish. Boldness, assessed using a shelter test, was not repeatable, but was correlated with exploration. Host exploration decreased with both increasing parasite density and decreasing body condition. Only individuals in relatively lower body condition displayed a negative relationship between parasite density and exploration, suggesting a pathologic effect of the parasites on the sunfish. Aversive learning was not influenced by co-infection. Our results show that body condition is important to consider when studying wild populations as some patterns observed between parasite density and host behavior were only revealed when body condition was taken into consideration.
Dataset DOI: 10.5061/dryad.s4mw6m9j8
Description of the data and file structure
We measured exploration (i.e. surface explored in cm for 10 minutes using an open-field test), boldness (i.e. latency to emergence from the shelter in seconds for a maximum of 300 seconds) and cognitive performance (i.e. difference between trial 1 and 2 in the latency to enter the black side) in wild pumpkinseed sunfish (Lepomis gibbosus, n = 65). We selected individuals from Lake Cromwell (Saint-Hippolyte, Canada) with a range of infection (i.e. blackspots, cestodes, yellow grub). We tested exploration and boldness for two trials and measured repeatability of these traits. We tested cognition using one learning trial, then two test trials. We measured body condition using the Fulton index (K) using the adjusted fish mass (i.e. fish mass - parasite mass) (mass/standard length3 in cm).
We had n = 21 fish that were accidentally defrosted, and thus, we only considered n = 65 to measure repeatability and correlations between traits, but n = 44 for models looking at the impact of parasites since we could not quantify parasite loads in defrosted fish.
We looked at the impact of co-infection (i.e. blackspots and internal parasite density) on these behavioral and cognitive traits). We included trial, body condition, an interaction between blackspot density and body condition, an interaction between internal parasite density and body condition and an interaction between both parasite densities in our models for exploration, boldness and cognition.
Files and variables
Data file "complete_final.csv": this datafile was used for our models looking at the impact of parasite infection on exploration and boldness. Each fish ID appears twice (trial 1 and trial 2). We also used this datafile to measure repeatability and correlation between traits, while accounting for trials as a random factor.
| VARIABLE | DESCRIPTION |
|---|---|
| ID | Fish number |
| ES | Test trial number (1 or 2) |
| EX | Exploration trials, proportion of surface explored in 10 minutes |
| BO | Boldness trials, latency to emerge from a shelter after being chased with a net. (max 300s) |
| SUR | Boldness trials, binary variable. Whether the fish emerged (1) or not (0) from the shelter after 5 minutes |
| L.B | Logarithm of BO |
| BS | Density of blackspots (n/standard length in cm), scaled and centered |
| FUL | Fulton index, scaled and entered |
| OT | Density of internal parasites (n/standard length), scaled and centered |
| log.d.INT | Logarithm of OT |
| TOT | Total density of parasites (BS+OT), scaled and centered |
| GR | Fish group (4 groups), 1 and 2 from July, 3 and 4 from September. |
| MO | Month of testing (2 months, (1): July, (2): September) |
| defrost | Fish that were defrosted accidentally (1) (not used in analysis for internal parasite density) (n = 21) |
| adj.M | Adjusted mass (g), total mass minus parasite mass |
| INT_num | Number of internal parasites |
| BS_num | Number of blackspots on the right side of the fish |
| cog | Inhibitory avoidance learning, latency to go to the dark compartment on trial 2 |
| anxi | Anxiety, proportion of time spend in the dark side for 5 minutes |
Data file "complete2.csv": this datafile was used for our models looking at the impact of parasite infection on cognition. Each fish ID appears only once.
| VARIABLE | DESCRIPTION |
|---|---|
| ID | Fish number |
| GR | Fish group (4 groups), 1 and 2 from July, 3 and 4 from September. |
| ex.1 and ex.2 | Exploration trials (2 trials), percentage of surface explored in 10 minutes |
| es1, es2, es3 | Inhibitory avoidance learning trials (3 trials), latency to go to the dark compartment |
| lat.1, lat.2 | Boldness trials (2 trials), latency to emerge from a shelter after being chased with a net. (max 300s) |
| cog_performance | Difference between es2 and es3 (trial 1 and 2) in the inhibitory avoidance learning in the latency to enter the black side as a measure of cognitive performance |
| anxi | Anxiety, proportion of time spend in the dark side for 5 minutes. |
| M | Fish mass in grams |
| SL | Standard length in mm |
| SL.cm | Standard length in cm |
| adj.M | Adjusted mass (i.e. total fish mass minus parasite mass) |
| FUL | Fulton index: masse(g)/taille(cm)^3 |
| adj.FUL | Fulton index calculated with the adjusted mass |
| BS | Number of blackspots on the right side of the fish |
| d.BS | Blackspot density: BS/SL.cm |
| log.d.BS | Logarithm of d.BS |
| INT | Number of internal parasites |
| d.INT | Internal parasites density: INT/SL.cm |
| d.TOTAL | Total density of parasite: (INT+BS)/SL.cm |
| defrost | Fish that were accidentally defrosted (1) and not used in models looking at internal parasite density (n=21) |
Code/software
We used R statistical software to analyse the data version 4.0.5 as shown below:
R version 4.0.5 (2021-03-31) -- "Shake and Throw"
Copyright (C) 2021 The R Foundation for Statistical Computing
Platform: x86_64-w64-mingw32/x64 (64-bit)
The packages needed are tidyverse, MASS, glmmTMB, gt, latex2exp, posterior, gt, glue, dplyr,magrittr, ggplot2, cowplot, jpeg, magick, rptR, webshot2, car, stats, gridExtra, AICcmodavg, lmer4, vegan, lmerTest, rsq, ggpubr, MuMIn, jtools.
No other software is needed to access/visualize the data.
The file "script_final.R" is the script used in R to analyse the data.