Abstract
Given the ongoing technological advances on biosensing technologies it is not farfetched to envision biosensors that monitor the welfare status of fishes becoming one day commonplace items within the aquaculture industry. Yet first, the applicability of currently available biosensors must be validated, and the breadth of their potential uses investigated. Here, we explored how ECG and accelerometry-derived parameters measured using bio-loggers, such as heart rate (HR), external acceleration (EA) and variance of the external acceleration (VAR), relate to parameters measured using video analyses such as mouth opening and tail beat frequencies and amplitudes, O2 consumption rate (MO2) and blood variables related to stress and tissue damage. We show that EA is a good proxy for swimming activity as it explains 60 and 29% of the variability in tail beat frequency and amplitude, respectively. Additionally, statistical models including bio-logger-derived data could be used to predict aerobic metabolism-related parameters with models including HR, EA and VAR explaining 59, 74 and 91%, respectively, of the variability in MO2 and 17, 31 and 47% of the variability in PVO2. On the other hand, parameters more closely related to exhaustion and stress where better predicted by models which include VAR, with VAR being negatively correlated with blood pH and mean corpuscular hemoglobin content and positively correlated with plasma cortisol concentration. These correlations are explained by VAR being a good indicator of the onset of burst-swimming activity, which is often followed by acid-base and metabolic imbalances. The results herein indicate that bio-logger data, especially VAR, can be used to extrapolate a range of stress-related physiological events when these are accompanied by increases in activity and highlight the great potential of biosensors for monitoring fish welfare.