V-QBA vs QBA – How do video and live analysis compare for Qualitative Behaviour Assessment?

A - Papers appearing in refereed journals

Cooke, A. S., Mullan, S. M., Morten, C., Hockenhull, J., Lee, M. R. F., Cardenas, L. M. and Rivero, M. J. 2022. V-QBA vs QBA – How do video and live analysis compare for Qualitative Behaviour Assessment? Frontiers in veterinary science. 9, p. 832239. https://doi.org/10.3389/fvets.2022.832239

AuthorsCooke, A. S., Mullan, S. M., Morten, C., Hockenhull, J., Lee, M. R. F., Cardenas, L. M. and Rivero, M. J.
Abstract

Animal welfare is an inextricable part of livestock production and sustainability. Assessing welfare, beyond physical indicators of health, is challenging and often relies on qualitative techniques. Behaviour is a key component of welfare to consider and Qualitative Behaviour Assessment (QBA) aims to achieve this by systematically scoring behaviour across specific terms. In recent years, numerous studies have conducted QBA by using video footage, however, the method was not originally developed using video and video QBA (V-QBA) requires validation. Forty live QBAs were conducted, by two assessors, on housed beef cattle to help fill this validation gap. Video was recorded over the assessment period and a second video assessment was conducted. Live and video scores for each term were compared for both correlation and significant difference. Principle component analysis (PCA) was then conducted and correlations and differences between QBA and V-QBA for the first two components were calculated. Of the 20 terms, three were removed due to an overwhelming majority of scores of zero. Of the remaining 17 terms, 12 correlated significantly, and a significant pairwise difference was found for one (‘Bored’). QBA and V-QBA results correlated across both PC1 (defined as ‘arousal’) and PC2 (defined as ‘mood’). Whilst there was no significant difference between the techniques for PC1, there was for PC2, with V-QBA generally yielding lower scores than QBA. Furthermore, based on PC1 and PC2, corresponding QBA and V-QBA scores were significantly closer than would be expected at random. Results found broad agreement between QBA and V-QBA at both univariate and multivariate levels. However, the lack of absolute agreement and muted V-QBA results for PC2 mean that caution should be taken when implementing V-QBA and that it should ideally be treated independently from live QBA until further evidence is published. Future research should focus on a greater variety of animals, environments, and assessors to address further validation of the method.

KeywordsAnimal welfare; Animal behavior; Animal behaviour; QBA; Qualitative behavior analysis; Qualitative behaviour analysis; Cattle ; Livestock; Agriculture; Zoology
Year of Publication2022
JournalFrontiers in veterinary science
Journal citation9, p. 832239
Digital Object Identifier (DOI)https://doi.org/10.3389/fvets.2022.832239
Web address (URL)https://www.frontiersin.org/articles/10.3389/fvets.2022.832239/full
Open accessPublished as ‘gold’ (paid) open access
FunderBiotechnology and Biological Sciences Research Council
Funder project or codeS2N - Soil to Nutrition - Work package 2 (WP2) - Adaptive management systems for improved efficiency and nutritional quality
The North Wyke Farm Platform- National Capability [2017-22]
Publisher's version
Output statusPublished
Publication dates
Online16 Mar 2022
Publication process dates
Accepted20 Jan 2022
PublisherFrontiers Media SA
ISSN2297-1769

Permalink - https://repository.rothamsted.ac.uk/item/987w0/v-qba-vs-qba-how-do-video-and-live-analysis-compare-for-qualitative-behaviour-assessment

10 total views
7 total downloads
8 views this month
2 downloads this month
Download files as zip