Characterization of hidden rules linking symptoms and selection of acupoint using an artificial neural network model

Discovering the key symptoms for identifying patterns in functional dyspepsia patients: Doctor's decision and machine learning

Background

Pattern identification is a crucial diagnostic process in Traditional East Asian Medicine, classifying patients with similar symptom patterns. This study aims to identify key symptoms for distinguishing patterns in patients with functional dyspepsia (FD) using explicit (doctor's decision-based) and implicit (computational model-based) approaches.

Methods

Data from twenty-one FD patients were collected from local clinics of traditional Korean Medicine and provided to three doctors in a standardized format. Each doctor identified patterns among three types: spleen-stomach weakness, spleen deficiency with qi stagnation/liver-stomach disharmony, and food retention. Doctors evaluated the importance of the symptoms indicated by items in the Standard Tool for Pattern Identification of Functional Dyspepsia questionnaire. Explicit importance was determined through doctors' survey by general evaluation and by selecting specific information used for the diagnosis of patient cases. Implicit importance was assessed by feature importance from the random forest classification models, which classify three types for general differentiation and perform binary classification for specific types.

Results

Key symptoms for distinguishing FD patterns were identified using two approaches. Explicit importance highlighted dietary and nausea-related symptoms, while implicit importance identified complexion or chest tightness as generally crucial. Specific symptoms important for particular pattern types were also identified, and significant correlation between implicit and explicit importance scores was observed for types 1 and 3.

Conclusion

This study showed important clinical information for differentiating FD patients using real patient data. Our findings suggest that these approaches can contribute to developing tools for pattern identification with enhanced accuracy and reliability.

A practical guide to implementing artificial intelligence in traditional East Asian medicine research

In this paper, we present a comprehensive guide for implementing artificial intelligence (AI) techniques in traditional East Asian medicine (TEAM) research. We cover essential aspects of the AI model development pipeline, including research objective establishment, data collection and preprocessing, model selection, evaluation, and interpretation. The unique considerations in applying AI to TEAM datasets, such as data scarcity, imbalance, and model interpretability, are discussed. We provide practical tips and recommendations based on best practices and our own experience. The potential of large language models in TEAM research is also highlighted. Finally, we discuss the challenges and future directions of AI application in TEAM, emphasizing the need for standardized data collection and sharing platforms.

Acupuncture indication knowledge bases: meridian entity recognition and classification based on ACUBERT

In acupuncture diagnosis and treatment, non-quantitative clinical descriptions have limited the development of standardized treatment methods. This study explores the effectiveness and the reasons for discrepancies in the entity recognition and classification of meridians in acupuncture indication using the Acupuncture Bidirectional Encoder Representations from Transformers (ACUBERT) model. During the research process, we selected 54 593 different entities from 82 acupuncture medical books as the pretraining corpus for medical literature, conducting classification research on Chinese medical literature using the BERT model. Additionally, we employed the support vector machine and Random Forest models as comparative benchmarks and optimized them through parameter tuning, ultimately leading to the development of the ACUBERT model. The results show that the ACUBERT model outperforms other baseline models in classification effectiveness, achieving the best performance at Epoch = 5. The model’s “precision,” “recall,” and F1 scores reached above 0.8. Moreover, our study has a unique feature: it trains the meridian differentiation model based on the eight principles of differentiation and zang-fu differentiation as foundational labels. It establishes an acupuncture-indication knowledge base (ACU-IKD) and ACUBERT model with traditional Chinese medicine characteristics. In summary, the ACUBERT model significantly enhances the classification effectiveness of meridian attribution in the acupuncture indication database and also demonstrates the classification advantages of deep learning methods based on BERT in multi-category, large-scale training sets.

Database URL: http://acuai.njucm.edu.cn:8081/#/user/login?tenantUrl=default

Acupuncture indication knowledge bases: meridian entity recognition and classification based on ACUBERT

In acupuncture diagnosis and treatment, non-quantitative clinical descriptions have limited the development of standardized treatment methods. This study explores the effectiveness and the reasons for discrepancies in the entity recognition and classification of meridians in acupuncture indication using the Acupuncture Bidirectional Encoder Representations from Transformers (ACUBERT) model. During the research process, we selected 54 593 different entities from 82 acupuncture medical books as the pretraining corpus for medical literature, conducting classification research on Chinese medical literature using the BERT model. Additionally, we employed the support vector machine and Random Forest models as comparative benchmarks and optimized them through parameter tuning, ultimately leading to the development of the ACUBERT model. The results show that the ACUBERT model outperforms other baseline models in classification effectiveness, achieving the best performance at Epoch = 5. The model's "precision," "recall," and F1 scores reached above 0.8. Moreover, our study has a unique feature: it trains the meridian differentiation model based on the eight principles of differentiation and zang-fu differentiation as foundational labels. It establishes an acupuncture-indication knowledge base (ACU-IKD) and ACUBERT model with traditional Chinese medicine characteristics. In summary, the ACUBERT model significantly enhances the classification effectiveness of meridian attribution in the acupuncture indication database and also demonstrates the classification advantages of deep learning methods based on BERT in multi-category, large-scale training sets. Database URL: http://acuai.njucm.edu.cn:8081/#/user/login?tenantUrl=default.

Deep autoencoder-powered pattern identification of sleep disturbance using multi-site cross-sectional survey data

Pattern identification (PI) is a diagnostic method used in Traditional East Asian medicine (TEAM) to select appropriate and personalized acupuncture points and herbal medicines for individual patients. Developing a reproducible PI model using clinical information is important as it would reflect the actual clinical setting and improve the effectiveness of TEAM treatment. In this paper, we suggest a novel deep learning-based PI model with feature extraction using a deep autoencoder and k-means clustering through a cross-sectional study of sleep disturbance patient data. The data were obtained from an anonymous electronic survey in the Republic of Korea Army (ROKA) members from August 16, 2021, to September 20, 2021. The survey instrument consisted of six sections: demographics, medical history, military duty, sleep-related assessments (Pittsburgh sleep quality index (PSQI), Berlin questionnaire, and sleeping environment), diet/nutrition-related assessments [dietary habit survey questionnaire and nutrition quotient (NQ)], and gastrointestinal-related assessments [gastrointestinal symptom rating scale (GSRS) and Bristol stool scale]. Principal component analysis (PCA) and a deep autoencoder were used to extract features, which were then clustered using the k-means clustering method. The Calinski-Harabasz index, silhouette coefficient, and within-cluster sum of squares were used for internal cluster validation and the final PSQI, Berlin questionnaire, GSRS, and NQ scores were used for external cluster validation. One-way analysis of variance followed by the Tukey test and chi-squared test were used for between-cluster comparisons. Among 4,869 survey responders, 2,579 patients with sleep disturbances were obtained after filtering using a PSQI score of >5. When comparing clustering performance using raw data and extracted features by PCA and the deep autoencoder, the best feature extraction method for clustering was the deep autoencoder (16 nodes for the first and third hidden layers, and two nodes for the second hidden layer). Our model could cluster three different PI types because the optimal number of clusters was determined to be three via the elbow method. After external cluster validation, three PI types were differentiated by changes in sleep quality, dietary habits, and concomitant gastrointestinal symptoms. This model may be applied to the development of artificial intelligence-based clinical decision support systems through electronic medical records and clinical trial protocols for evaluating the effectiveness of TEAM treatment.

Exploring Acupuncture Actions in the Body and Brain

Acupuncture's actions have been explained by biomedical research. However, the meridian system used in acupuncture needs further clarification. This review describes how acupuncture affects the body and brain. From the perspective of traditional East Asian medicine, the meridian system is closely connected with acupuncture's treatment effects. In the body, the indications of acupoints, primarily established based on the meridian system, have spatial symptom patterns. Spatial patterns of acupoint indications are distant from the stimulated sites and strongly associated with the corresponding meridian's route. Understanding how acupuncture works based on the original meridian system is important. From a neuroscience perspective, an acupuncture-induced sensation originates from the bottom-up action of simple needling in the peripheral receptor and the reciprocal interaction with top-down brain modulation. In the brain, enhanced bodily attention triggered by acupuncture stimulation can activate the salience network and deactivate the default mode network regardless of the actual stimulation. The application of data science technology to acupuncture research may provide new tools to uncover the principles of acupoint selection and enhance the clinical efficacy of acupuncture treatment in various diseases.

The Use of Artificial Intelligence in Complementary and Alternative Medicine: A Systematic Scoping Review

Background: The development of artificial intelligence (AI) in the medical field has been growing rapidly. As AI models have been introduced in complementary and alternative medicine (CAM), a systematized review must be performed to understand its current status. Objective: To categorize and seek the current usage of AI in CAM. Method: A systematic scoping review was conducted based on the method proposed by the Joanna Briggs Institute. The three databases, PubMed, Embase, and Cochrane Library, were used to find studies regarding AI and CAM. Only English studies from 2000 were included. Studies without mentioning either AI techniques or CAM modalities were excluded along with the non-peer-reviewed studies. A broad-range search strategy was applied to locate all relevant studies. Results: A total of 32 studies were identified, and three main categories were revealed: 1) acupuncture treatment, 2) tongue and lip diagnoses, and 3) herbal medicine. Other CAM modalities were music therapy, meditation, pulse diagnosis, and TCM syndromes. The majority of the studies utilized AI models to predict certain patterns and find reliable computerized models to assist physicians. Conclusion: Although the results from this review have shown the potential use of AI models in CAM, future research ought to focus on verifying and validating the models by performing a large-scale clinical trial to better promote AI in CAM in the era of digital health.

Special Issue: State of the Art in Research on Acupuncture Treatment

Acupuncture is a medical treatment that involves inserting a needle into the body [...].

Exploring traditional acupuncture point selection patterns for pain control: data mining of randomised controlled clinical trials

BackgroundThe underlying principles of traditional acupuncture point selection for pain control are complex. Analysis of acupuncture treatments from clinical studies may provide us with a potential rule when selecting traditional acupuncture points (hereafter abbreviated as "points") in treatment protocols for pain control. The aim of this study was to investigate which points were most commonly used to treat pain in randomised controlled clinical trials (RCTs).MethodsWe searched acupuncture treatment regimens in RCTs included in the Cochrane Database of Systematic Reviews for pain management. We analysed information on point selection (more than 10 RCTs included) from seven eligible systematic reviews on pain control. The frequency of the points used was calculated and visualised using a human body template.ResultsThe points most commonly used across diseases were SP6, ST36, LI4 and LR3. However, the most frequently used points varied across individual conditions. For example, the most frequently used points to treat migraine were GB20, LR3, GV20, Taiyang, LI4 and TE5, while the most frequently used points to manage dysmenorrhoea were SP6, CV4, SP8, LR3 and BL32. Both regional and distal points were used for pain management with acupuncture.ConclusionsOur findings suggest that local and segmental/extra-segmental neuromodulation appear to be the most common phenomena for pain control in acupuncture research. Analysis of information on point selection using a data-driven approach may unveil the hidden patterns of traditional acupuncture point utilisation in clinical practice.

Commonality and Specificity of Acupuncture Point Selections

OBJECTIVE

Because individual acupoints have a wide variety of indications, it is difficult to accurately identify the associations between acupoints and specific diseases. Thus, the present study aimed at revealing the commonality and specificity of acupoint selections using virtual medical diagnoses based on several cases.

METHODS

Eighty currently practicing Korean Medicine doctors were asked to prescribe acupoints for virtual acupuncture treatment after being presented with medical information extracted from 10 case reports. The acupoints prescribed for each case were quantified; the data were normalised and compared among the 10 cases using z-scores. A hierarchical cluster analysis was conducted to categorise diseases treated based on the acupoint prescription patterns. Additionally, network analyses were performed on the acupoint prescriptions, at the individual case and cluster level.

RESULTS

Acupoints ST36, LI4, and LR3 were most commonly prescribed across all diseases. Regarding the specific acupoints prescribed in each cluster, acupoints around the disease site (knee and lower back) were frequently used in cluster A (musculoskeletal symptoms), acupoints LI4, LR3, PC6, and KI3 were frequently used in cluster B (psychiatric symptoms), and acupoints ST36, LI4, LR3, PC6, CV12, and SP6 were frequently used in cluster C (several symptoms of diseases of internal medicine).

CONCLUSIONS

The present study identified the commonality and specificity of acupoint selections based on virtual acupuncture treatments prescribed by practicing clinicians. Acupoint selection patterns, which were defined using a top-down approach in previous studies and classical medical texts, may be further elucidated using a bottom-up approach based on patient medical records.

A Key Node Mining Method Based on Acupoint-Disease Network (ADN): A New Perspective for Exploring Acupoint Specificity

In the process of treating pro-diseases with acupuncture, traditional Chinese medicine (TCM) doctors may fine-tune acupuncture prescriptions according to different prior experiences. Different prescriptions will affect the efficiency and effect of acupuncture treatment, and even excessive acupoint selection may cause psychological pressure on patients. We still lack an effective means to analyze the meridian system and acupoint specificity to clarify the mapping relationship between acupoints and diseases. Given the inability of modern medical technology to provide effective evidence support for meridians and acupoints, we combined acupuncture theory with network science for an interdisciplinary discussion. In this paper, we constructed a weighted undirected acupoint-disease network (ADN) based on clinical acupuncture prescription literature and proposed a high-specificity key node mining method based on ADN. Combined with the principle of acupoint selection in TCM, the proposed method balanced the contribution of local areas to the network based on the distribution characteristics of meridians and selected 30 key acupoints with high influence on the global topology according to the evaluation index of key nodes. Finally, we compared the proposed method with the other six classical node importance evaluation algorithms in terms of resolution, network loss, and accuracy. The comprehensive results show that the marked key acupoint nodes make outstanding contributions to the connectivity, topological structure, and weighted benefits of the network, and the stability and specificity of the algorithm guarantee the reliability of the key acupoint nodes. We consider that these key acupoints with high centrality in ADN can be used as core acupoints to help researchers explore targeted and high-impact acupoint combinations under resource constraints and optimize existing acupuncture prescriptions.

Statistical inference of acupoint specificity: forward and reverse inference

Background

The acupoint specificity has been considered important issue in acupuncture research. In clinical aspects, it is essential to identify which acupoints are associated specifically with a particular disease. The present study aimed to identify the specificity of acupoint selection (forward inference) and the specificity of acupoint indication (reverse inference) from the online virtual diagnosis experiment.

Methods

Eighty Korean Medicine doctors conducted the virtual medical diagnoses provided for 10 different case reports. For forward inference, the acupoints prescribed for each disease were quantified and the data were normalised among 30 frequently used acupoints using Z-scores. For reverse inference, diseases for each acupoint were quantified and the data were normalized among 10 disease using Z-scores.

Results

Using forward inference we demonstrated the specificity of acupoint selection in each disease. Using reverse inference we identified the specificity of acupoint indication in each acupoint. In general, a certain acupoint can be selected specifically for a particular disease, and it has a specific indication for the disease. However, the specificity of acupoint indication and the specificity of acupoint selection are not always identical.

Conclusions

The selection of an acupoint for a particular disease does not imply that the acupoint has specific indications for that disease. Inferring the specificity of acupoint indication from clinical observations should be considered.

Revealing Associations between Diagnosis Patterns and Acupoint Prescriptions Using Medical Data Extracted from Case Reports

OBJECTIVE

The optimal acupoints for a particular disease can be determined by analysis of diagnosis patterns. The objective of this study was to reveal the association between such patterns and the acupoints prescribed in clinical practice using medical data extracted from case reports.

METHODS

This study evaluated online virtual diagnoses made by currently practicing Korean medical doctors (N = 80). The doctors were presented with 10 case reports published in Korean medical journals and were asked to diagnose the patients and prescribe acupoints accordingly. A network analysis and the term frequency-inverse document frequency (tf-idf) method were used to analyse and quantify the relationship between diagnosis patterns and prescribed acupoints.

RESULTS

The network analysis showed that ST36, LI4, LR3, and SP6 were the most frequently used acupoints across all diagnoses. The tf-idf values showed the acupoints used for specific diseases, such as BL40 for bladder disease and LU9 for lung disease.

CONCLUSIONS

The associations between diagnosis patterns and prescribed acupoints were identified using an online virtual diagnosis modality. Network and text mining analyses revealed commonly applied and disease-specific acupoints in both qualitative and quantitative terms.