Dissertation, Otto-Friedrich-Universität Bamberg, 2026

As dementia progresses, those affected lose their ability to think, speak, and express their inner states verbally. As a result, they are alone with their inner experiences and perceptions. Often, painful conditions go unnoticed by others, and sufferers remain unheard. And although the verbal ability to express pain may be lacking, non-verbal expressions of pain often remain preserved. For instance, an increased nociceptive pain response or facial response to pain is more likely to be observed in individuals with dementia. Additionally, there are specific indicators to measure non-verbal pain responses. Though some indicators, such as facial expressions and observer-based methods, are widely recognized and implemented, the use of autonomic responses to describe pain still lacks empirical robustness.

Thus, this dissertation aimed to describe and evaluate pain assessments that could be implemented in various populations, including individuals with dementia, while considering individual factors such as cognition and aging. Therefore, three empirical studies were conducted. Chapter 1 summarizes the theoretical background, and Chapter 2 describes Study 1, which focused on a practical challenge: How can we efficiently and validly observe what patients cannot say? With the development of a shortened version of the Pain Assessment in Impaired Cognition, an observer-based pain assessment tool was designed that combines clinical relevance with psychometric quality and focuses on facial expressions, body movements, and vocalizations as indicators of pain.

Until now, pain research has primarily been conducted in laboratory settings using standardized pain protocols. In Studies 2 and 3, we applied prior findings to real-world contexts and evaluated pain during motor challenges (box and kettlebell lift), activities of daily living (e.g., movements in bed or walking), and pain stimulation (pressure and heat) in a Living Lab environment.

With Study 2, addressed in Chapter 3, we explored multimodal indicators of pain (facial expressions, autonomic responses, and self-reports) during muscular pre-activation in healthy adults. For this, muscular pre-activation was induced by using an age simulation suit. The results demonstrate the selective sensitivity of specific indicators of pain: facial expressions and heart rate increased with muscle pre-activation, while heart rate variability and self-reports of pain remained surprisingly unaffected. Electrodermal activity increased from round 1 to round 2 only in the group with muscle pre-activation, indicating a delayed electrodermal activity response to muscle pre-activation.

Study 3, described in Chapter 4, extended this approach to older adults with and without cognitive impairment, assessing autonomic responses and self-reports of pain during both pain stimulation and activities of daily living. Surprisingly, cognitive impairment did not influence autonomic responses or self-reports of pain, while age itself proved to be a predictor of both.

Overall, these findings emphasize that integrative pain assessments are particularly useful for accurately describing pain when considering individual factors, which is addressed in Chapter 5. While observer-based assessment tools or facial expressions could be interpreted in isolation, autonomic responses such as heart rate variability and electrodermal activity should be considered as complementary indicators of arousal and not as stand-alone measures of pain, based on the results of this dissertation. Interestingly, even in individuals with mild forms of cognitive impairment, self-reports of pain could be reliably interpreted. These results highlight the importance of selecting context-sensitive pain assessment to reliably capture pain across various populations.