Decoding the effect of expectation in pain from mice's facial expressions using a visual machine analysis

An environmental context can modulate pain perception by generating expectations based on previous experiences. A significant issue has been the limitations in objectively evaluating pain and its modulation in rodents. This work aimed to utilize machine vision to analyze the characteristics of facial pain-like expressions and their modulation in C57BL/6 mice. We employed a conditioned sensitivity model and subsequently generated an expectation induced by the environmental context. Our findings indicate that machine vision can detect facial pain-like expressions following thermal nociceptive stimulation, resulting in an exacerbated facial response upon re-exposure to a conditioning context, even in the absence of thermal stimulation. Female mice exhibited a stronger facial expression in response to thermal stimulation compared to males. Conversely, by analyzing facial responses, these results affirm male-specific conditioned pain and female-conditioned responses in a non-context-dependent manner. Thus, we propose that analyzing facial expressions through machine vision is a novel and useful tool for studying pain and its modulation in mice.