Encoding signs of orofacial neuropathic pain from facial expressions in mice.

Objective: To evaluate the effects of mental nerve injury in the facial reactions elicited by mechanical stimulation of different intensities and detect and quantify spontaneous facial pain-like expressions during a period free of stimuli, as signs of evoked and spontaneous pain in a mouse model for neuropathic orofacial pain.
Design: We recorded mouse heads in a fixed position during a stimulus-free period and with mechanical stimulation with 3 different Von Frey filaments. We extracted the Histograms of Oriented Gradients of each frame of the video recordings to be compared with a prototypical pain-like facial expression. The similarity score was then used to register and quantify the percentage of spontaneous pain-like facial reactions and evaluate the increased similarity to the prototypical pain-like face evoked by mechanical stimuli. The assessments were made one day before and four days after a unilateral mental nerve compression.
Results: Our findings show that mental nerve injury promotes an increase in spontaneous facial pain-like expressions and reduced mechanical threshold, reflected in a higher similarity to our pain-like face prototype, regardless of the intensity of the stimuli applied.
Conclusions: Machine vision encodes the facial expression associated with evoked and spontaneous pain after mental nerve injury for up to four days. Facial expression quantitatively reflects the increased mechanical sensitivity elicited by mental nerve injury. We also show that this technique can detect spontaneous pain-like responses from facial reactions. Artificial vision can be applied to evaluate signs of orofacial neuropathic pain to study the involved neural circuits.
(Copyright © 2022 Elsevier Ltd. All rights reserved.)