Using AI to Detect Pain through Facial Expressions: A Review

Abstract

Pain assessment is a complex task largely dependent on the patient’s self-report. Artificial intelligence (AI) has emerged as a promising tool for automating and objectifying pain assessment through the identification of pain-related facial expressions. However, the capabilities and potential of AI in clinical settings are still largely unknown to many medical professionals. In this literature review, we present a conceptual understanding of the application of AI to detect pain through facial expressions. We provide an overview of the current state of the art as well as the technical foundations of AI/ML techniques used in pain detection. We highlight the ethical challenges and the limitations associated with the use of AI in pain detection, such as the scarcity of databases, confounding factors, and medical conditions that affect the shape and mobility of the face. The review also highlights the potential impact of AI on pain assessment in clinical practice and lays the groundwork for further study in this area.

Incorporation of “Artificial Intelligence” for Objective Pain Assessment: A Comprehensive Review

“…Some studies suggest that AI/ML can differentiate true pain versus faked pain and identify malingering [ 126 ]. This is of great importance in different situations, such as detecting patients seeking compensation [ 127 ], requesting unjustifiable sick leaves, and preventing unnecessary narcotics prescriptions for drug abusers and people with a substance-use disorder; this has the potential to reduce health care costs [ 26 , 88 , 107 ].…”

“…For example, patients should provide informed consent beforehand, as they may refuse facial analysis. Furthermore, algorithms might be trained for particular demographics, further marginalizing already vulnerable groups [ 26 , 135 , 136 ]. Additionally, using AI/ML algorithms to detect pain through facial expressions has several limitations [ 26 , 135 , 136 ].…”

“…Furthermore, algorithms might be trained for particular demographics, further marginalizing already vulnerable groups [ 26 , 135 , 136 ]. Additionally, using AI/ML algorithms to detect pain through facial expressions has several limitations [ 26 , 135 , 136 ]. Their application may be limited by many factors, e.g., the patient’s age, gender, behavior, as well as the head position and movements.…”

“…Moreover, some medical conditions can affect facial shape and mobility, such as Parkinson’s, stroke, facial injury, or deformity. All these factors can reduce the accuracy and sensitivity of the AI/ML to automatically detect pain experiences [ 26 , 135 , 136 ]. The “WHO” reiterates the importance of applying ethical principles and appropriate governance, as enumerated in the WHO guidance on the ethics and governance of AI for health, when designing, developing, and deploying AI for health.…”

Pain Assessment in Animals: human experts, make way for AI!

“…Yet even validated pain assessment instruments are limited by the observer's previous training and ability to interpret the pain responses accurately [39], as well as by observers' various biases, like gender, fatigue, experience and time consumption [40,41]. Adami et al [42] recently evaluated the inter-observer reliability of three feline pain scales commonly used in clinical practice: the Glasgow Feline Composite Measure Pain Scale (CMPS-Feline [43]), the Colorado State University Feline Acute Pain Scale (CSU-FAPS [44]) and the Feline Grimace Scale (FGS [30]).…”