\received
DD MMMM YYYY \acceptedDD MMMM YYYYINTRODUCTION
Collectively, the processes of perception, learning and memory that permit decision making constitute cognition (Shettleworth, 2010a). Cognitive performance impacts fitness and welfare of animals by influencing their capacities to forage, find potential mates and avoid predators (Boogert et al., 2018; Fuss, 2021). To collect and process the multitude of sensory inputs from their surroundings, fish have evolved sophisticated brains that allow them to distinguish between predator and prey, assess competitors, discern their own movements through their environments, evaluate risk, and decide whether and where to hide (Fernö et al. 2020). Although not yet well studied, brain development has been found in some fish species to be reliant upon the quality of their diet during larval stages (Hou and Fuiman, 2020). In particular, omega-3 long-chain (≥ 20 C) polyunsaturated fatty acids (n‑3 LC‑PUFA), especially docosahexaenoic acid (DHA 22:6n‑3), are important for neural development, structure and function in vertebrates (Pilecky et al., 2021; Twining et al., 2021; Závorka et al., 2023). Because vertebrates have limited capacity to biosynthesise n‑3 LC‑PUFA, they must be acquired, at least in part, via the diet (Twining et al., 2016). Laboratory studies have shown a positive association between n‑3 LC-PUFA enrichment and brain growth in freshwater and marine fishes (Lund et al., 2016; Ishizaki et al., 2001). In many animals, including fishes, cognition has been linked to relative brain size (Kotrschal et al., 2013; Triki et al., 2021). Brain size is associated with fitness-influencing behaviours such as prey capture (Edmunds et al., 2016) and predator avoidance (Kondoh, 2010). However, fish brain morphology varies widely across species (Triki et al., 2021), and so relative brain size alone may not always be the best proxy for fish cognition (Marhounová et al., 2019; Závorka et al., 2022b). In addition, different cognitive skills are associated with the relative sizes of specific brain regions: for example, learning and cognitive flexibility in response to visual stimuli, which are often investigated in animal cognition studies, are regulated by the telencephalon and optic tectum in guppies (Poecilia reticulata ) (Triki et al., 2022). Therefore, there is a need for experimental work integrating the impacts of diet quality on brain size, morphology and biochemical composition, and the cognitive and behavioural abilities they promote.
For stream-dwelling fishes, the combination of prey from freshwater aquatic and terrestrial origins provides for a varied dietary intake of essential nutrients (Sánchez-Hernández et al., 2018; Syrjänen et al., 2011). Of the invertebrates typically consumed by juvenile salmonids, freshwater prey tends to be relatively rich in a n‑3 LC-PUFA, eicosapentaenoic acid (EPA 20:5n-3), which can be converted to DHA at relatively low metabolic cost, while prey of terrestrial origin tends to be richer in the short-chain precursor, α-linolenic acid (ALA 18:3n‑3), which can only be converted to DHA at high metabolic cost (Pilecky et al., 2021; Twining et al., 2021). The energetic costs of short- to long-chain conversion may constrain the rate of DHA biosynthesis and thus the ability to allocate DHA to the brain in consumers of diets low in LC-PUFA (Pilecky et al., 2021; Závorka et al., 2021). Furthermore, because n‑3 and n‑6 precursor fatty acids compete for the same enzymes, both synthesis pathways from α-linolenic acid (ALA 18:3n‑3) to docosahexaenoic acid (DHA 22:6n‑3) and linoleic acid (LIN 18:2n‑6) to arachidonic acid (ARA 20:4n-6) should be considered when investigating bioconversion of DHA precursors (Geiger et al., 1993): biosynthesis of ARA is, ipso facto , powerful evidence of even greater n‑3 LC-PUFA conversion than what may be deduced from DHA alone (Sprecher, 2000; Hastings et al., 2001).
Apart from diet, greater environmental complexity or enrichment can increase brain size and cognitive ability in fish (Ebbesson and Braithwaite, 2012; Salena et al., 2021): evidence of plasticity-inducing effects of habitat complexity on relative brain size exists for related salmonids, chinook salmon Oncorhynchus tschawytscha (Kihslinger et al., 2006) and Atlantic salmon Salmo salar (Näslund et al., 2012), as well as brown trout themselves (Závorka et al. 2022b). Most of the conversion of short-chain to long-chain fatty acids occurs in the liver, from which DHA is distributed to the brain (Rapoport et al., 2007; Wang et al., 2017), where it may be incorporated into polar lipids (PL – mainly phospholipids) that increase neuronal membrane fluidity and accelerate the formation of synapses and synaptic vesicles as vehicles for neurotransmitters (Pilecky et al. 2021). DHA and its precursors that are not currently needed for a specific physiological function are stored in form of neutral lipids (NL, mainly triacylglycerols) or allocated to other tissues, such as gonads or gametes for the next generation (Rigaud et al., 2023; Závorka et al., 2023; Zhu et al., 2019). Uneven distribution of DHA, and its precursors, whether biosynthesised or dietary, may be caused by activation of storage lipids and rerouting of the required fatty acids to the brain (Lacombe et al., 2018; Pifferi et al., 2021). The combination of biosynthesis and priority routing of DHA and its precursors may represent a compensatory mechanism for those animals with diets depleted in such nutrients.
Brown trout Salmo trutta are a widespread species throughout Europe, and introduced worldwide, that exhibit enormous genetic, morphological and life-history variation and sexual dimorphism (Klemetsen, 2013; Koene et al., 2020; Reyes-Gavilán et al., 1997). Most variants typically are born and spend their early years as juveniles in low-productivity streams (Ferguson et al., 2019), which offer both physical habitat complexity and instability (Guo et al. 2018) and the opportunity for social complexity in the form of dominance hierarchies in response to competition for preferred microhabitats (Sloman et al., 2000). Juvenile brown trout often lack morphological features that correspond to competitive ability that allows them to assess one another at the onset of each potential dyadic conflict (sensu Kodric-Brown and Brown, 1984), beyond mere differences in size (Jacob et al., 2007). So, recognition of individual rivals, memory of previous confrontations, and inference of competitive ability become important cognitive competencies when a potential opponent’s fighting capabilities can otherwise be assessed only through escalated contests (Drew, 1993).
To test the effect of dietary quality on memory and inference abilities that are crucial for juvenile brown trout during conflicts (Johnsson andÅkerman, 1998; Grosenick et al., 2007; Shettleworth, 2010b), we conducted an experiment in which juvenile brown trout raised in either complex or simple habitats and on either high or low n‑3 LC-PUFA diets were subjected to a series of agonistic dyadic trials. We hypothesised that larger brains, generally, larger optic tecta and telencephala, specifically, and greater DHA content of brain polar lipids are associated with greater cognitive performance; and that habitat complexity and dietary DHA deficiency stimulates increased compensatory DHA biosynthesis and routing to brain polar lipids, the energetic cost of which is reflected in lower somatic growth. We tested the predictions that: 1) trout deprived of n‑3 LC-PUFA and in a complex habitat would show lower somatic growth than those raised on a high n‑3 LC-PUFA diet and in a simple habitat; 2) trout raised on a high n‑3 LC-PUFA diet and in a complex habitat would show greater cognitive performance than those raised on a low n‑3 LC-PUFA diet and in a simple habitat; 3) trout raised on a high n‑3 LC-PUFA diet and in a complex habitat would have larger relative brain size, and larger optic tecta and telencephala, than those raised on a low n‑3 LC-PUFA diet and in a simple habitat; 4) that larger brains, and larger optic tecta and telencephala, would be associated with greater cognitive performance; and 5) trout deprived of n‑3 LC-PUFA and in a complex habitat would biosynthesise DHA, and route DHA to brain polar lipids, to a greater extent than those raised on a high n‑3 LC-PUFA diet and in a simple habitat.
There has been a suggestion that sex can play a role in routing and synthesis of n‑3 LC-PUFA in fish, with female Eurasian perch Perca fluviatilis showing significantly higher proportions of n‑3 LC‑PUFA in muscle tissue (Scharnweber and Gårdmark, 2020), and rates of n‑3 LC-PUFA biosynthesis before spawning (Rigaud et al., 2023), than males. Although sex differences are normally unexpected in brown trout until they reach maturity (Reyes-Gavilán et al., 1997), we considered the potential impact of sex in all of our analyses.