How can we know how salmon experience their own welfare beneath the water surface when a fish’s subjective experience of its condition cannot be measured directly?
The biological definition of animal welfare is: the animal's own subjective perception of its condition as it attempts to cope with the environment in which it lives. Although this subjective perception cannot be measured directly, it arises from and influences processes in the fish's brain, specifically communication between nerve cells within neural networks.
“We can say something about that because we can measure the processes that enable this communication and how effectively these processes in the brain respond to stress. From that, we can assess how well the fish is functioning and coping with its environment,” says Professor of Animal Behaviour Øyvind Øverli at the Norwegian University of Life Sciences (NMBU).
Professor Øverli researches fish welfare and the ability of fish to adapt to and cope with changing production environments, with a particular focus on methods for the early detection of biological conditions such as stress and poor welfare.
Assessing Fish Welfare in Practice
However, monitoring stress responses in a fish’s brain is not a practical method for assessing welfare in everyday aquaculture operations. Instead, fish health professionals use indicators, known as welfare indicators, which are intended to provide an indirect picture of the fish's condition based on informed assumptions.
Traditionally, fish health personnel inspect fish after they have been captured and removed from the water. Fish are assigned scores for various welfare indicators based on the extent of issues such as scale loss, fin damage, wounds, and other physical abnormalities. Weight and length are also recorded.
This method is time-consuming and limited to a small sample of the fish in a pen. The assessment results are also subjective, as scoring can vary between observers. Furthermore, the sampling process itself is stressful for the fish, making it difficult to distinguish between the fish's actual welfare state and its response to handling.
Biological Relevance: What Best Reflects the Fish’s Own Welfare State?
Returning to what happens inside the fish’s brain, how relevant are the visible external signs compared to the biological processes occurring internally? This relationship is known as biological relevance.
Biorelevance is the name of the research project in which Øverli and colleagues from NMBU, NIVA, and SINTEF Ocean investigated which welfare indicators most accurately reflect the fish’s actual self-experienced welfare state.
Measuring the Fish's Ability to Cope with Stress
To evaluate these indicators, researchers examined how fish responded when exposed to a standardized stress challenge in a confinement test. This approach turns traditional welfare assessment on its head by focusing on the fish’s ability to respond to stress rather than measuring stress levels directly.
The fish were held in a small volume of water for one hour to induce a maximum stress response. The experiments were conducted with approval from the Norwegian Food Safety Authority and followed established ethical guidelines for animal experimentation.
“If a fish does not show a normal response, meaning it is unable to respond to the stressor with a normal stress reaction, then it lacks coping capacity, which in turn indicates poor welfare,” says Øverli.
To assess the fish’s actual physiological condition, researchers measured the functioning of the nervous and immune systems in the brain. Through analyses of blood samples and brain tissue, they were able to determine the fish's true stress response.
Some external indicators proved more biologically relevant than others. The indicator that stood out most clearly was the condition factor, which reflects the relationship between a fish’s length and weight. Another indicator that contributed to welfare assessment was scale loss.
The Threshold on the Welfare Scale
Researchers identified a threshold value, or tipping point, at a condition factor of 1.0.
Fish that are long and thin, suggesting poor feed intake or inefficient feed utilization, tend to have condition factor values below 1.0. Fish with scores far below this threshold exhibit welfare so poor that normal physiological functioning begins to break down.
From a condition factor of 1.0 and above, no significant welfare-related changes were observed; the fish generally coped well with their environment.
The Future of Aquaculture Technology?
The goal of this research is to generate scientific knowledge that technology developers can use to measure welfare indicators more accurately, automatically, and in a biologically relevant way directly within sea cages.
The use of camera technology and machine vision has expanded rapidly within the aquaculture industry. Technologies capable of automatically detecting wounds, identifying scale loss, counting sea lice, measuring fish size, and recording other physiological signs have proven far more precise and efficient than manual visual inspections.
“Because machine vision can both detect and quantify with high precision, and can assess vastly more fish than is possible in a traditional observation programme, it is important to evaluate biological relevance rather than simply assigning numerical scores as is common practice today,” says Øverli.
Integrating Machine Vision with Biological Measurements
According to Øverli, the challenge is that we still know too little about how reliable camera-based and machine-vision measurements truly are, and how accurately they reflect fish health and welfare.
The next step in research is therefore to integrate machine vision with biological measurements to create a comprehensive welfare monitoring system. This means developing methods for the automatic analysis of observations and validating machine vision-generated data against biological processes and biological measurements.
“The field is developing rapidly with a high rate of innovation and will receive considerable attention in the years ahead. The Biorelevance project has already attracted substantial interest from the aquaculture industry, and several companies are requesting the integration of such methods into their systems,” says Øverli.
Transferring Behavioural Knowledge from Farm Animals to Fish
An emerging area of research involves using machine vision to interpret fish behaviour in sea cages and investigating how such observations can serve as indicators of welfare.
“This is an area that has received relatively little attention in aquaculture so far, but inspired by behavioural studies in terrestrial animals, researchers are now working to transfer behavioural knowledge to salmonids. The Ethology and Animal Welfare research group at NMBU, of which Øverli is a member, has extensive experience in behavioural studies of farm animals.”
Camera technology could potentially be used to analyse how fish move, position themselves within cages, and interact with one another. This could provide an entirely new understanding of fish welfare and contribute to more accurate welfare assessments.