Seasonal variation of behavior and brain size in a freshwater fish
Cite this dataset
Versteeg, Evan et al. (2023). Seasonal variation of behavior and brain size in a freshwater fish [Dataset]. Dryad. https://doi.org/10.5061/dryad.0cfxpnw35
Teleost fishes occupy a range of ecosystem and habitat types subject to large seasonal fluctuations. Temperate fishes in particular, survive large seasonal shifts in temperature, light availability, and access to certain habitats. Mobile species like lake trout (Salvelinus namaycush) can behaviorally respond to seasonal variation by shifting their habitat deeper and further offshore in response to warmer surface water temperatures during the summer. During cooler seasons, use of more structurally complex nearshore zones by lake trout could increase cognitive demands and potentially result in a larger relative brain size during those periods. Yet, there is limited understanding of how such behavioral responses to a seasonally shifting environment might shape, or be shaped by, the nervous system.
Here we quantified variation in relative brain size and the size of five externally visible brain regions in lake trout, across six consecutive seasons in two different lakes. Acoustic telemetry data from one of our study lakes was collected during the study period from a different subset of individuals and used to infer relationships between brain size and seasonal behaviors (habitat use and movement rate).
Our results indicated that lake trout relative brain size was larger in the fall and winter compared to the spring and summer in both lakes. Larger brains coincided with increased use of nearshore habitats and increased horizontal movement rates in the fall and winter based on acoustic telemetry. The telencephalon followed the same pattern as whole brain size, while the other brain regions (cerebellum, optic tectum, olfactory bulbs, hypothalamus) were only smaller in the spring.
These findings provide evidence that flexibility in brain size could underpin shifts in behavior, which could potentially subserve functions associated with differential habitat use during cold and warm seasons and allow fish to succeed in seasonally variable environments.
Study Sites and sample collection
Lake trout were collected seasonally from Lake of Two Rivers (hereafter referred to as “Two Rivers”), Ontario, Canada (45o34’42.6” N, 78o29’0.4” W; 274 ha surface area, 38 m maximum depth) and Lake Opeongo (hereafter referred to as “Opeongo”), Ontario, Canada (45o41’46.8” N, 78o22’27.8” W; 5800 ha surface area, 49.4 m maximum depth), each located within Algonquin Provincial Park. Unlike Two Rivers, Opeongo supports pelagic prey fish (e.g., lake cisco: Coregonus artedi). Mature lake trout were sampled seasonally from fall 2017 to winter 2019 (Table 1) using trap nets, gill nets, and angling equipment. The fish were euthanized immediately upon capture via severing of the spinal cord, and their lengths and weights collected (under approved University of Toronto Animal Use Protocols). Fish heads were then removed and placed in labelled containers with 10% neutral buffered formalin (Fisher Scientific Inc., New Jersey, USA). Fish that possessed undeveloped gonads were considered immature and removed from analysis.
Following previous conventions (Guzzo et al. 2017), the period after ice-off but before mean surface temperature (<6 m depth) exceeded 15˚C was denoted as spring, summer was defined as the period during which surface temperatures reached or surpassed 15 ˚C, fall began when lakes cooled to ≤ 15˚C and lasted until winter, defined as ice-on to ice-off. Water temperatures were measured throughout the upper 6 m of the water column in each lake using a string of data loggers (HOBO Temp Pro H20-001, Onset, Cape Cod, MA) deployed over the deepest point of each lake (Table 1).
Brain Mass and Region Volumes
Following procedures from Edmunds et al. (2016b), the brains were removed and trimmed of cranial nerves, and the spinal cord was cut at the obex. Each brain was blotted thoroughly to remove excess formalin and was then weighed on an analytical balance (Fisher Scientific Accu-124D), to a resolution of 0.0001 g to obtain brain mass, which was used to estimate brain size. Pictures were taken of the dorsal, left, and ventral sides of the brain using an Olympus SZ61 dissection microscope and a Canon Powershot G9 digital camera and PSREMOTE v1.7 software (Breeze and Breeze 2009). A calibration grid was included in each picture. The height, length, and width of brain regions visible on the resulting images (olfactory bulbs, telencephalon, optic tectum, cerebellum and hypothalamus) were measured using the measuring tool in ImageJ software (Rueden et al. 1997). Regional volumes (mm3) were estimated using the ellipsoid formula: V = (L x W x H) π/6, providing us with estimates of size for each brain region. Volume measurements were conducted by the same person (EJV) and measurement accuracy was assessed by repeating measurements of five randomly selected brains ten times (Appendix S1: Table S1).
Acoustic telemetry data were available from Lake of Two Rivers throughout the study period, albeit for a different subset of lake trout individuals (n=9). Fixed location reference tags within the array provided a measure of array performance throughout the study period. Estimated positions based on the Vemco (InnovaSea, WA, USA) positioning algorithm were compared with known positions of the tags to quantify positioning errors across seasons and years using methods proposed in Smith (2013). Reference tags at 5 m depth had a lower average error than tags placed at 18 m depth. Over 95% of detections at 5 m had a positional error of less than 6 m vs. 86% at the 18 m depth. Overall mean error was 2.41 m vs 5.24 m at 5 and 18 m depth, respectively. Additional details about the telemetry array setup, performance, and fish tagging can be found in (Appendix S2 of manuscript).
Metadata for the datasets are located in the README file, as well as written in the header of the associated R files.
R files were written in R studio, and require packages for usage.
Natural Sciences and Engineering Research Council, Award: 2017-06794
Natural Sciences and Engineering Research Council, Award: RGPIN-2020-04114