Home About Browse Search
Svenska


Silventoinen, Alina, 2019. Clustering of facial action units in horses with pain. Second cycle, A2E. Uppsala: SLU, Dept. of Clinical Sciences

[img] PDF
Full text restricted to: Library staff only
Full text under embargo, will be available 1 March 2021.

1MB

Abstract

Pain assessment in horses is a relatively new topic in veterinary medicine and there is no universally accepted pain scale for assessing pain in horses. Assessing and managing pain is essential for the welfare of the animal.

Several different pain scale systems have been made for both humans and animals. An example of this is Composite Pain Scale (CPS), where several parameters, such as behaviour and physiological parameters, are being assessed at the same time. Studies of pain assessment through facial expressions are several in human medicine but only a few studies have been made on horses. The Horse Grimace Scale and Equine Pain Scale are examples of studies on facial expressions and pain in horses.

Equine Facial Action Coding System (EquiFACS) is a method of identifying and coding facial expressions on horses on the basis of underlying facial musculature and muscle movement. The system is based on Action Units and Action Descriptors which each represent a separate movement.
Facial actions have been used together with artificial intelligence in order to evaluate pain in humans. This is probably possible to do with horses as well, but more training data is needed.

This study included 28 films of horses with and without pain, both in experimental and clinical setting. The films were coded using Equine Facial Action Coding System in order to analyse how the different facial action units cluster in horses with and without pain. Relevant facial actions for the equine pain face were selected in order to visualise changes in duration, frequency and occurrence between painful and non-painful horses, as well as upper and lower facial actions. According to the two-sample t test, only one facial action unit (AD38, nostril dilator) had statistical significance. However, the linear discriminant analysis showed a 92% correct classification in duration in pain/no pain.

Stress may have influenced pain behaviour due to stress-induced analgesia or hyperalgesia. A larger sample size is needed in order to investigate this further, but there is potential for EquiFACS to be used as a tool for objective pain scoring.

Main title:Clustering of facial action units in horses with pain
Authors:Silventoinen, Alina
Supervisor:Haubro Andersen, Pia
Examiner:Rhodin, Marie
Series:UNSPECIFIED
Volume/Sequential designation:UNSPECIFIED
Year of Publication:2019
Level and depth descriptor:Second cycle, A2E
Student's programme affiliation:VY002 Veterinary Medicine Programme 330 HEC
Supervising department:(VH) > Dept. of Clinical Sciences
Keywords:pain assessment, equine pain face, EquiFACS
URN:NBN:urn:nbn:se:slu:epsilon-s-10997
Permanent URL:
http://urn.kb.se/resolve?urn=urn:nbn:se:slu:epsilon-s-10997
Subjects:Veterinary science and hygiene - General aspects
Language:English
Deposited On:04 Oct 2019 11:06
Metadata Last Modified:04 Jun 2020 13:55

Repository Staff Only: item control page

Downloads

Downloads per year (since September 2012)

View more statistics

Downloads
Hits