1. Introduction
Artificial Intelligence (AI) is rapidly transforming numerous disciplines, and Traditional Chinese Medicine (TCM), a cornerstone of Chinese heritage, is no exception. The integration of AI with TCM presents a unique opportunity to revitalize and advance this ancient medical system. This article explores the burgeoning field of AI in TCM diagnosis, with a particular focus on the development and potential of Acupuncture Diagnosis Software. We will analyze the current landscape, identify existing challenges, and discuss the exciting future trajectory of AI-assisted TCM diagnostics in enhancing healthcare practices for an English-speaking audience.
TCM, rooted in millennia of empirical observation and practice [3], operates on a holistic framework, diagnosing and treating ailments by understanding the body’s intricate balance of Qi, blood, Yin, and Yang. Traditional TCM diagnosis relies heavily on the “Four Examinations”: Inspection, Listening and Smelling, Inquiry, and Palpation. These methods, while rich in qualitative data, are inherently subjective and can be influenced by environmental factors and the clinician’s experience. AI offers a pathway to objectify and quantify these diagnostic processes, paving the way for more standardized and accessible TCM practices.
The application of AI in TCM diagnosis dates back to the 1970s [5], coinciding with the broader AI revolution. Early attempts faced limitations in computational power and data availability. However, recent advancements in machine learning, deep learning [4, 9, 10], sensor technology [7], and image and speech recognition [8] have reignited the potential of AI in TCM. This progress is crucial for developing sophisticated acupuncture diagnosis software that can assist practitioners in making more precise and data-driven decisions.
This article aims to provide a comprehensive overview of AI’s role in auxiliary TCM diagnosis. We will delve into the applications of AI within each of the Four Examinations, highlighting the development of acupuncture diagnosis software and related technologies. Furthermore, we will address the current limitations and explore the future possibilities of AI in shaping the evolution of TCM diagnostics and its integration into global healthcare systems.
2. The Rise of AI in Medical Diagnostics: A Foundation for Acupuncture Diagnosis Software
The advent of computers in the mid-20th century sparked the vision of using machines to augment human cognitive abilities [4]. The field of AI, formally defined in 1955 [13], has since matured, with deep learning, particularly Convolutional Neural Networks (CNNs), becoming a dominant force in medical image analysis [17, 18]. These advancements are directly relevant to the development of acupuncture diagnosis software, particularly in areas like facial and tongue diagnosis.
Early medical expert systems emerged in the 1970s and 80s [14, 15], setting the stage for AI’s broader application in modern medicine. Today, AI is employed in diagnosing common diseases [16], analyzing medical images for cancer detection [19], and even assisting in surgical procedures [21, 22]. The DeepGestalt framework, for example, demonstrates AI’s capacity to surpass clinicians in identifying genetic syndromes through facial analysis [24]. These successes in modern medicine underscore the potential for similar AI-driven innovations in TCM, specifically in the realm of acupuncture diagnosis software.
TCM began exploring AI integration in the 1970s, marked by the “TCM Guan Youbo Hepatitis Diagnostic and Treatment Procedures” expert system [26]. Recent developments, such as the AI-based TCM auxiliary diagnosis system proposed by Zhang et al. [27], showcase impressive diagnostic accuracy for a wide range of TCM diseases. Furthermore, AI is contributing to various aspects of TCM, including drug development, prescription modeling, and even acupuncture point selection [28, 29]. These applications highlight the growing momentum and diverse potential for AI to enhance and modernize TCM practices, with acupuncture diagnosis software poised to play a central role.
3. AI Applications in TCM Diagnostic Methods: Building Blocks for Acupuncture Diagnosis Software
3.1. Inspection: Visual Cues for AI-Powered Diagnosis
Inspection, the first of the Four Examinations, relies on visual observation of the patient’s appearance to discern health status. Facial and tongue diagnosis are particularly prominent in TCM inspection, and AI is proving invaluable in analyzing these visual cues for acupuncture diagnosis software applications.
3.1.1. Facial Diagnosis: Unlocking Patterns in Facial Features with AI
Facial color and complexion are considered indicators of the body’s Qi and blood circulation and the health of internal organs in TCM. AI-powered facial analysis tools are being developed to objectively quantify these subtle variations. Dong [31] demonstrated the use of a TCM face digital detector to differentiate facial color characteristics among patients with various diseases. Similarly, Liu et al. [32] explored facial features in patients with chronic nephritis, finding correlations between facial color parameters and kidney function. Guo et al. [33] further validated the link between facial color changes and the progression of chronic renal failure. These studies lay the groundwork for integrating facial recognition and analysis into acupuncture diagnosis software, allowing for automated assessment of facial indicators relevant to TCM syndromes and informing acupuncture treatment strategies.
Facial recognition technology can be integrated into acupuncture diagnosis software to analyze subtle changes in facial color and features.
3.1.2. Tongue Diagnosis: AI Deciphers the Tongue’s Map of Health
Tongue diagnosis, another crucial aspect of TCM inspection, involves analyzing the tongue body and coating to understand the patient’s internal condition. The tongue is considered a microcosm reflecting the state of the viscera and bowels. AI-driven tongue diagnosis systems are emerging to standardize and objectify tongue assessment, offering valuable components for acupuncture diagnosis software.
Early systems like the “TCM tongue diagnosis automatic identification system” [34] developed in the 1990s, achieved quantitative analysis of tongue features. Jiang et al. [35] designed a computerized tongue diagnosis system using fuzzy theory to analyze tongue coating characteristics. Cui et al. [36] utilized a “TCM tongue diagnosis expert system” to study tongue manifestations in stroke patients. Zhang et al. [37] developed a Bayesian network-based tongue diagnosis system with promising accuracy. Lo et al. [38] explored tongue feature extraction for early breast cancer screening. Han et al. [39] found correlations between tongue coating thickness and colorectal cancer.
Modern tongue manifestation analyzers [40] offer advanced capabilities for multidimensional tongue image acquisition and analysis. Xu et al. [41] used a tongue and pulse condition analyzer to study tongue manifestations in chronic renal failure patients. Zhang et al. [42] analyzed tongue color in femoral head necrosis patients across different stages. These advancements in AI-powered tongue diagnosis are crucial for developing comprehensive acupuncture diagnosis software, providing objective data on tongue indicators to support TCM practitioners.
3.2. Listening and Smelling Examination: AI Captures Auditory and Olfactory Signals
Listening and smelling examination involves discerning health information from the patient’s sounds and odors. AI applications in this area, particularly in voice and odor analysis, are contributing to the development of more holistic acupuncture diagnosis software.
3.2.1. Listening Examination: Voice Analysis as a Diagnostic Tool
In TCM, voice characteristics are linked to different health conditions and constitutional types. AI-powered voice analysis is being explored for physique identification, disease diagnosis, and treatment efficacy evaluation. Wang [46] studied voice changes with age and Qi status using computer voice analysis. Qian et al. [47] used audio characteristics to identify constitutions in Parkinson’s patients. Dong et al. [48] analyzed speech signals in chronic pharyngitis patients, finding distinct energy features across different syndrome types. Li et al. [49] used acoustic parameters to evaluate the efficacy of bronchial asthma treatment.
Chen et al. [50] applied AI to study the “five visceral phonemes” theory, analyzing sound signals from patients with different organ diseases. Li and Zhang [51] investigated voice frequencies in women with different constitutions. Furthermore, AI is being applied to recognize other pathological sounds like bowel sounds [52] and cough sounds for pneumonia diagnosis [53]. Integrating voice analysis into acupuncture diagnosis software can provide another layer of objective diagnostic information, particularly useful in assessing internal organ imbalances and informing acupuncture point selection.
3.2.2. Smelling Examination: E-Nose Technology for Odor Profiling
Smelling examination in TCM involves discerning subtle body odors for diagnostic insights. Electronic nose (e-nose) technology, utilizing advanced gas sensors [54], offers a way to objectively analyze and map odor profiles. Lin et al. [55] used e-nose to differentiate odor characteristics between type 2 diabetes patients and healthy individuals. Liu [56] developed an oral odor detection system for TCM diagnosis based on e-nose principles and AI algorithms. While still in early stages, the integration of e-nose technology into acupuncture diagnosis software could potentially provide a unique and objective dimension to TCM diagnosis, particularly in identifying specific pathogenic factors and guiding treatment strategies, including acupuncture.
3.3. Inquiry: AI-Powered Patient History and Symptom Analysis
Inquiry, or asking questions about the patient’s condition, is crucial for gathering comprehensive diagnostic information in TCM. AI is being applied to develop intelligent inquiry systems that can streamline and standardize this process, enhancing the capabilities of acupuncture diagnosis software.
He et al. [59] developed a computerized TCM inquiry system that collects patient information and symptoms, providing preliminary diagnostic judgments. Zheng et al. [60] designed a standardized inquiry system for TCM splenic diseases. Liu et al. [61] created a TCM heart disease inquiry system. These AI-powered inquiry systems can be integrated into acupuncture diagnosis software to automate patient history taking, analyze symptom patterns, and provide valuable insights for syndrome differentiation and acupuncture treatment planning. Such systems can ensure comprehensive data collection, reduce subjective bias, and improve the efficiency of TCM consultations.
3.4. Palpation: Digital Pulse Diagnosis for Objective Assessment
Palpation, especially pulse diagnosis, is a cornerstone of TCM. AI is revolutionizing pulse diagnosis by providing objective methods for pulse data acquisition and analysis, contributing significantly to the development of acupuncture diagnosis software.
AI applications in pulse diagnosis focus on analyzing pulse waveform information using time-domain, frequency domain, time-frequency, and wavelet analysis [63]. Yang et al. [64] used a digital pulse analyzer to study pulse parameters in IgA nephropathy patients. Yan et al. [65] compared pulse characteristics in pregnant women and normal subjects using time-frequency analysis. Mo and Wang [66] designed a wavelet analysis-based pulse detection system for subhealth identification.
Mobile pulse diagnosis systems [67], wearable pulse detection devices [68, 72], and portable three-position pulse acquisition systems [69] are emerging. Remote pulse diagnosis systems are also under development [70, 71]. These advancements in digital pulse diagnosis are essential for creating robust acupuncture diagnosis software. By providing objective, quantifiable pulse data, AI can assist practitioners in accurately assessing pulse qualities, identifying underlying imbalances, and tailoring acupuncture treatments to individual patient needs.
4. Discussion: The Future of Acupuncture Diagnosis Software and AI in TCM
AI holds immense promise for modernizing TCM diagnostics and enhancing the effectiveness of treatments like acupuncture. The development of intelligent TCM diagnostic instruments, expert systems, and mobile platforms represents significant progress. Acupuncture diagnosis software, incorporating AI-driven analysis of the Four Examinations, can mitigate the subjectivity inherent in traditional TCM diagnosis and improve diagnostic accuracy. Early disease screening and diagnosis facilitated by AI can lead to timely interventions and improved patient outcomes.
However, challenges remain in the widespread adoption of AI in TCM diagnosis and the refinement of acupuncture diagnosis software. Currently, AI applications are unevenly distributed across the Four Examinations. For example, inspection is primarily focused on facial and tongue diagnosis, while other areas of inspection are less explored. While tongue diagnosis has seen progress in objective analysis of color and coating, capturing tongue motility remains challenging. Facial analysis is largely limited to color analysis, with less focus on facial expressions and their TCM significance. In listening and smelling examination, voice analysis is primarily applied to respiratory conditions, with limited research on other disease types and pathological sounds. E-nose technology in smelling examination is still in early stages of translating odor profiles into specific TCM syndromes. Inquiry systems need further refinement to standardize language and handle complex disease scenarios, integrating both TCM and modern medical perspectives. Pulse diagnosis is advancing with wearable and remote technologies, but replicating the nuanced tactile information of manual pulse-taking remains a technological hurdle.
Furthermore, the complexity of TCM information, the lack of unified diagnostic standards, and the limitations of current expert systems necessitate further development. Building comprehensive, standardized four-diagnosis databases and establishing international diagnostic criteria are crucial steps. Leveraging advanced deep learning algorithms like CNNs, RNNs, and DNNs is essential for creating smarter and more adaptable acupuncture diagnosis software.
The future of TCM lies in intelligent diagnostics. For AI to achieve universal application in clinical TCM and for acupuncture diagnosis software to reach its full potential, several key steps are necessary. These include fostering multidisciplinary expertise, securing funding for research and development of intelligent diagnostic tools, establishing robust data privacy policies, and creating reasonable cost structures for AI-driven diagnostic services. With continued innovation in AI and data science, breakthroughs in TCM diagnosis are anticipated, bringing the vision of accessible, accurate, and AI-assisted TCM healthcare, including acupuncture diagnosis software, closer to reality. This will not only enhance TCM practice but also contribute significantly to global healthcare advancements.
Acknowledgments
This study was supported by National Natural Science Foundation (81704170, 82074539), Heilongjiang Natural Science Foundation (LH2020H092), Postdoctoral Initiation Fund of Heilongjiang Province (LBH—Q18117), and Key Laboratory Project of Ministry of Education for Myocardial Ischemia (KF201614).
Contributor Information
Yang Li, Email: [email protected].
Tiansong Yang, Email: [email protected].
Data Availability
This study has no laboratory data, but the review study process and reference records are corrected and put in the data center of Heilongjiang University of Chinese Medicine, and these data will be kept for 8 years.
Conflicts of Interest
The authors declare that there are no conflicts of interest regarding the publication of the paper.
Authors’ Contributions
Chuwen Feng, Yuming Shao, and Bing Wang contributed equally to this work. Yang Li and Tiansong Yang made critical revision of the manuscript. All authors read and approved the final manuscript.
References
[References from the original article are included here and should be preserved in the rewritten article.]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
This study has no laboratory data, but the review study process and reference records are corrected and put in the data center of Heilongjiang University of Chinese Medicine, and these data will be kept for 8 years.
