Introduction: Pulsology is one of the most important clinical diagnostic methods in Persian medicine (PM) and has been used by traditional medicine specialists for centuries to assess clients' health conditions. Additionally, photoplethysmogram (PPG) is a simple and non-invasive method for measuring changes in blood volume with each pulse, and the amount of peripheral resistance (PR), which is crucial for understanding patients' clinical conditions, can be extracted from it. Moreover, a system based on fuzzy rules serves as one of the methods for prediction and estimation. This research establishes the relationship between PR in PPG and PM pulsology through a fuzzy system..
Method: Fuzzy theory underlies many studies aimed at diagnosing diseases. To design a stable fuzzy system, pulse information and PPG signals from 35 individuals, recorded by a PM expert, were used. Complete, normal, consistent, and symmetric fuzzy sets sensitive to the range of input variables were defined to encompass the input and output spaces and to generate rules using the available data. Stability conditions for the system were also established.
Results: A fuzzy system comprising 35 rules, triangular and trapezoidal membership functions, a singleton fuzzifier, a product inference engine, and a centroid defuzzifier was designed using MATLAB R2021b. This stable fuzzy system, with pulse frequency and pulse strength inputs, as well as PR output, operates within the defined range. The stability of the system is assured by the membership function definitions, with a maximum error of 0.01.
Conclusion: This intelligent stable fuzzy system provides a reliable estimate of the PR from the PPG signal by using PM pulse variables. It was observed that an increase in pulse frequency correlates with an increase in PR, while an increase in pulse strength correlates with a decrease in PR in the PPG. Therefore, when the photoplethysmography signal is unavailable, trained physicians can derive a suitable estimate of the patient's peripheral vascular resistance status based on the frequency and strength of the patient's pulse, which is valuable for the clinical evaluation and diagnosis of the patient's medical issues.
Type of Study:
Original Article |
Subject:
Artificial Intelligence in Healthcare Received: 2024/05/31 | Accepted: 2025/03/1