Acute pain is one of the most important protective reactions related to the restriction of physical functions. This is an important public health issue and one of the main complaints of patients seeking medical attention. As many as 95% of people experience acute or prolonged pain on their own or in their immediate environment. Pain management remains a complex issue nowadays because its assessment requires special training by professionals. Inadequate analgesia is associated with an increased stress response, anxiety, depression, a possible transition to chronic. As technology advances, there are more and more opportunities to assess pain objectively and quantitatively, but currently, the usual human vital parameters (respiratory rate, oxygen saturation, heart rate, arterial blood pressure, temperature) used in clinical practice do not always indicate the individual pain level objectively enough. In addition, data on molecular mechanisms of pain response are insufficient or performed only in the presence of chronic pain (e.g., cortisol, melatonin). Scientific studies are also limited by the lack of informative, complex physiological and biochemical databases. This shortcoming is expected to be filled by creating a highly informative physiological database that includes both dynamic multimodal physiological signals and biochemical markers during the pain of different intensities. The main goal of the project is to develop a prototype of the diagnostic system based on machine learning algorithms to assess the experience of the pain of different intensities using dynamic multimodal physiological signals and biochemical markers. The expected results of the project would not only take a step towards the development of objective methods for assessing the level of pain, but also simplify pain assessment and facilitate the work of medical staff in the application of individualized analgesia.
KTU Research and Innovation Fund
During the project, highly informative databases of dynamic multimodal (physiological and biomechanical) signals and saliva proteins and hormones were collected, algorithms for pre-processing, parameterization, feature extraction and machine learning-based pain classification of physiological and biomechanical signals were developed, and a prototype of a diagnostic system based on machine learning algorithms for pain assessment was developed.
Period of project implementation: 2021-04-01 - 2021-12-31
Project partners: Lithuanian University of Health Sciences