Population densities predict forebrain size variation in the cleaner fish Labroides dimidiatus

Abstract

The 'social brain hypothesis' proposes a causal link between social complexity and either brain size or the size of key brain parts known to be involved in cognitive processing and decision-making. While previous work has focused on comparisons between species, how social complexity affects plasticity in brain morphology at the intraspecific level remains mostly unexplored. A suitable study model is the mutualist 'cleaner' fish Labroides dimidiatus, a species that removes ectoparasites from a variety of 'client' fishes in iterative social interactions. Here, we report a positive relationship between the local density of cleaners, as a proxy of both intra- and interspecific sociality, and the size of the cleaner's brain parts suggested to be associated with cognitive functions, such as the diencephalon and telencephalon (that together form the forebrain). In contrast, the size of the mesencephalon, rhombencephalon, and brain stem, assumed more basal in function, were independent of local fish densities. Selective enlargement of brain parts, that is mosaic brain adjustment, appears to be driven by population density in cleaner fish.

Keywords: brain size; cleaning mutualism; mosaic brain; social brain hypothesis; sociality; strategic sophistication.

Figure 1.

The Labroides dimidiatus brain. (a) A representative magnetic resonance imaging (MRI) scan image of a sagittal section of L. dimidiatus brain inside its braincase, the lower panel shows the delineation of the five brain structures: telencephalon including the olfactory bulb (in red), diencephalon (in green), mesencephalon (in blue), rhombencephalon (in yellow), brain stem (in orange). Images generated with OsiriX Lite® v. 9.5.2. Cleaner fish photo credit to Z. T. (b) A 3D brain rendering showing the main brain parts from four different angles of the brain: top left image shows a partial frontal view, top right image shows a lateral view, bottom left image shows a superior view, and bottom right shows a dorsal view. The brain is subdivided into six major brain areas, depicted in different colours: telencephalon (in red), olfactory bulbs (in pink), diencephalon (in green), mesencephalon (in blue), rhombencephalon (in yellow), and brain stem (in light blue). Images were prepared with the AMIRA software by E. L. (Online version in colour.)

Figure 2.

Relationship of residuals from brain part sizes on body size and cleaner density. Box plots of median, interquartile range, upper and lower quartiles, and outliers. The raw data are depicted as dark red dots. (Online version in colour.)

Figure 3.

Relationship of the diencephalon size ratio to the rest of the brain and cleaner density. Box plots of median, interquartile range, upper and lower quartiles, and outliers. The raw data are depicted as dark red dots. p-values are estimated with post hoc analysis run on significant overall effect. (Online version in colour.)

Figure 4.

Pairwise comparisons of brain part sizes and their relationship with cleaner density. Box plots of median, interquartile range, upper and lower quartiles, and outliers of size ratio between diencephalon and (a) mesencephalon, (b) rhombencephalon, and (c) brain stem; and ratio between mesencephalon and (d) rhombencephalon, and (e) brain stem; and (f) ratio between brain stem and rhombencephalon. The raw data are depicted as dark red dots. p-values are estimated with post hoc analysis run on significant overall effects. (Online version in colour.)

Figure 5.

Relationship of brain measurements, involving the telencephalon data, and cleaner density. Box plots of median, interquartile range, upper and lower quartiles, and outliers of (a) residuals of telencephalon size on body size, (b) telencephalon size ratio to mesencephalon + rhombencephalon + brain stem combined, (c) ratio between telencephalon and diencephalon, and (d) forebrain size ratio as an estimate of telencephalon + diencephalon/mesencephalon + rhombencephalon + brain stem. The raw data are depicted as dark red dots. p-values are estimated with post hoc analysis run on significant overall effect. (Online version in colour.)