The Accuracy and Efficacy of AI-driven tVNS Algorithm

Plain-language summary

Pain, including somatic and visceral pain, is a common symptom. Persistent pain can lead to repetitive visits to hospitals and can limit patients' daily activities, which can result in tremendous medical cost and lower quality of life. For example, the prevalence rates of 25% are reported only for abdominal pain among adults (3), and it costs $10.2 billion each year in the US. Pain is usually treated according to the World Health Organisation (WHO) 3 steps analgesic ladder. Non-Steroidal Anti-Inflammatory Drugs (NSAIDs) are mainly used in step 1, which can cause serious side effects such as GI bleeding, renal failure and cardiovascular disease. In step 2 \& 3, opioids are used and are also associated with serious side effects (e.g., psychological addiction, dizziness, nausea, vomiting, constipation, physical dependence, tolerance, and respiratory depression). Therefore, a new effective non-pharmacological treatment is beneficial for patients. One such method is transcutaneous vagal nerve stimulation (tVNS). The auricular or cervical branch of the vagal nerve runs just under the skin and can be electrically stimulated through the skin by tVNS devices, which have shown the analgesic effects on various pain conditions. The autonomic activity, including parasympathetic tone, can be estimated from the beat to beat intervals in the electrocardiogram, which is called heart rate variability (HRV). To date, we have shown that visceral and somatic pain triggered the autonomic response with the change in HRV, and HRV could be a biomarker of pain. We hypothesised that the development of pain, including somatic pain and visceral pain, could be predicted by analysing heart rate pattern by artificial intelligence (AI). In this proof of concept study, we evaluate the detection rate of pain by the AI analysis of heart rate pattern. We also evaluate the effect of tVNS on the pain threshold.

Who can participate

What they're measuring

Trial details