FACES: Facial Activities of Compound Emotional States of horses

Facial expressions (FE) of pain are widely used for assessment pain in horses.

However, FE of e.g. stress may go unnoticed in pain scorings, as shown in pilot studies of FE of stressed and of painful horses.

Using a facial action coding system (FACS) that exhaustively codes all facial activities, FE of stress and pain surprisingly showed both similarities and differences.

The main aim is to investigate FE of stress, pain and sedation in video of horses; and investigate FE resulting from combinations of these emotional challenges. We hypothesize that experience of combined emotional modalities will produce compound FE never described before.

The project will be conducted at SLU within an interdisciplinary group of researchers in horse pain, FACS and Machine Learning.

WP1 defines FACS-based expressions of single emotions of pain, stress and sedative drugs, based on re-use of videos from experimental studies of pain, stress and sedation. We use common and advanced machine learning principles to classify the facial expressions. WP2 characterizes the compound FE of horses experiencing two emotional modalities.

To do this, we perform a randomized, blinded cross-over trial of twelve horses challenged by stress and pain or pain and sedation. Video analysis is performed as in WP1. WP3 validates the findings of WP2 in clinical horse cases.

The outcome is original biological scientific papers which opens for practical revisions of current assessments of pain and stress.