This experimental method describes a solution for the kinematic analysis of acupuncture manipulation with three-dimensional finger motion tracking technology.
Three-dimensional (3D) motion tracking has been used in many fields, such as the researches of sport and medical skills. This experiment aimed to use 3D motion tracking technology to measure the kinematic parameters of the joints of fingers during acupuncture manipulation (AM) and establish three technical indicators “amplitude, velocity and time”. This method can reflect the operation characteristics of AM and provide quantitative parameters along three axes of multiple finger joints. The current evidence shows that the method has great potential for future applications such as the study of the dose-effect relationship of acupuncture, teaching, and learning of AM, and the measurement and preservation of famous acupuncturists’ AM.
As a kind of the clinical skills of traditional Chinese medicine (TCM) and physical stimulation, acupuncture manipulation (AM) is often regarded as an important factor that affects the therapeutic effect of acupuncture1,2. Many studies have confirmed that different AM or different stimulation parameters (needling velocity, amplitude, frequency, etc.) of the same AM resulted in different therapeutic effects3,4,5,6,7. Therefore, the measurement of relevant kinematic parameters of AM and correlation analysis with the therapeutic effect can provide useful data support and reference for the clinical treatment with acupuncture8,9.
The measurement of kinematic parameters of the AM began in the 1980s10. In the early days, the electrical signal conversion technology based on variable resistance was mainly used to convert the displacement signal of the needle body into a voltage or current signal for displaying and recording the amplitude and frequency data of AM11. Moreover, the famous ATP-II Chinese medicine acupuncture technique tester II (ATP-II) with this technology currently has been used by many traditional Chinese medicine universities of China12. After that, with the continuous development and innovation of sensor technology, different types of sensors were used to collect kinematic parameters of AM. For example, the three axes electromagnetic motion sensor was attached to the needle handle to acquire needling amplitude and velocity13; the bioelectric signal sensor was placed on the dorsal horn of the animal's spinal cord to record needling frequency14, etc. Although the quantitative research of AM based on the above two types of technologies has completed the acquisition of relevant kinematic parameters during needling, its main disadvantages are the inability to perform the real-time non-invasive measurement and the change of operating feel caused by the modification of the needle body.
In recent years, motion tracking technology was gradually applied to the quantitative research of AM15,16. Because it is based on the frame-by-frame analysis of needling video, the measurement of acupuncture parameters can be acquired during in vivo operation without modifying the needle body. This technology has been used to measure the kinematic parameters such as amplitude, velocity, acceleration, and frequency of four tracking points of thumb and forefinger during needling in a two-dimensional (2D) plane and established the corresponding finger stick figure15. Some studies also measured the angle change range of interphalangeal (IP) joint of thumb and forefinger with similar technology9,17,18. However, the current studies on AM analysis are still mainly limited to the 2D motion plane, and the number of tracking points is relatively small. So far, there is no complete three-dimensional (3D) kinematics measurement and analysis method for AM, and no related data was published.
To solve the above problems, this study will use 3D motion tracking technology to measure the kinematic parameters of the seven tracking points of hand during needling. This protocol aims to provide a complete technical solution for the kinematic analysis on AM, as well as the further study on the dose-effect correlation of acupuncture.
This study was approved by the ethics committee of Yueyang Hospital, affiliated with Shanghai University of Traditional Chinese Medicine (reference no. 2021-062), and each participant signed an informed consent form.
1. Experiment preparations
2. Video shooting and editing
3. Project configuration of Simi Reality Motion System (motion capture and analysis software)
4. Video analysis
5. Data analysis
NOTE: An original PHP script is used to browse and analyze the data files exported by the motion capture and analysis software. All the source code has been shared in a GitHub repository20.
After establishing this experimental method, the lifting-thrusting and twirling skills of basic AM of nineteen acupuncture teachers from the School of Acupuncture-Moxibustion and Tuina of Shanghai University of TCM were measured using 3D motion tracking. According to the definition of a joint coordinate system (JCS) for the shoulder, elbow, wrist, and hand proposed by the Standardization and Terminology Committee (STC) of the International Society of Biomechanics21, seven finger tracking points have been selected. The stick view generated by the motion capture and analysis software based on the anatomical positions of these points is shown in Figure 2B. The typical coordinate-time curves along three axes of each point are shown in Figure 4, and two videos of lifting-thrusting and twirling skills with stick view (Video 1 and Video 2).
As shown in Figure 4C,E, because of the minimal movement amplitude along the main motion axes during different skills (the Z-axis of lifting-thrusting skill and the Y-axis of twirling skill) of the wrist joint (WJ) can be fixed, and the movement seems to occur from the thumb and index finger. Therefore, the data of the other six points were exported by the motion capture and analysis software for further kinematic analysis of AM. After data analysis, the average values of amplitude and velocity along three axes and the operating time of the action "lifting", "thrusting", "twirling left" and "twirling right" of each tracking point on fingers were calculated and shown in Table 2, Table 3 and Table 4.
In addition, the finger motion of participants was also tracked when they performed AM on ATP-II. The data derived from ATP-II was compared with the data exported by the motion capture and analysis software. The results show that the shape of the coordinate-time curve of TT along the Z-axis was similar to the voltage-time curve generated by ATP-II during the lifting-thrusting skill. Meanwhile, during the twirling skill, the shape of the amplitude-time curve along the Y-axis of TT was also similar to the voltage-time curve of ATP-II. Furthermore, after calculation, the average operating cycles of these two types of curves were basically the same (Figure 5).
Figure 1: Camera positions and the placement of 3D calibration frame. (A) The positions of three cameras. (B) Front view of 3D calibration frame. (C) Top view of 3D calibration frame. Please click here to view a larger version of this figure.
Figure 2: The positions of tracking markers and their stick view. (A) The positions of tracking markers on hand. (B) The stick view generated by the motion capture and analysis software based on the anatomical positions of these points. Please click here to view a larger version of this figure.
Figure 3: Schematic diagram of calculation method of kinematic parameters. The average amplitude and velocity can be calculated based on curve crest and trough positioning. Please click here to view a larger version of this figure.
Figure 4: Typical coordinate-time curves during lifting-thrusting and twirling skills. (A,B,C) The typical coordinate-time curves along the X-, Y-, Z-axis of each tracking point during the lifting-thrusting skill, respectively. (D,E,F) The curves with the same settings of lifting-thrusting skill during twirling skill. Please click here to view a larger version of this figure.
Figure 5: Comparison of the curves generated by ATP-II and motion capture and analysis software. (A) Finger motions of participants were tracked when they performed AM on ATP-II. (B) The voltage-time curve of ATP-II during the lifting-Thrusting skill. (C) The coordinate-time curve along the Z-axis of TT during the lifting-thrusting skill. (D) The voltage-time curve of ATP-II during twirling skill. (E) The coordinate-time curve along the Y-axis of TT during twirling skill. Please click here to view a larger version of this figure.
Table 1: Coordinate parameters of the calibration points. The coordinate values of three axes of eight calibration points. Please click here to download this Table.
Table 2: Kinematics data of each tracking point during the lifting-thrusting skill. The average values of amplitude and velocity along three axes of each tracking point on figures during the lifting-thrusting skill. Please click here to download this Table.
Table 3: Kinematics data of each tracking point during twirling skill. The average values of amplitude and velocity along three axes of each tracking point on figures during twirling skill. Please click here to download this Table.
Table 4: Operating time during lifting-thrusting and twirling skills The average values of operating time in the processes of lifting, thrusting, twirling left, and twirling right actions Please click here to download this Table.
Video 1: Lifting-thrusting skill. (Top left) The stick view of the hand. (Top right, Bottom left, Bottom right) The typical coordinate-time dynamic curve along the X-, Y-, Z-axis of each tracking point during the lifting-thrusting skill Please click here to download this Video.
Video 2: Twirling skill: The stick view of the hand and typical coordinate-time dynamic curves with the same settings as Video 1 during the twirling skill. Please click here to download this Video.
Supplementary File 1: Video synchronization instructions. Screenshots and steps of video synchronization instructions of the video editing software used in this study. Please click here to download this File.
This study established the measurement method of the kinematic parameters of AM in vivo and obtained the data of motion amplitude, velocity, and operation time of the six important tracking points on the thumb and forefinger along three axes. Meanwhile, based on the 3D calibration frame, a 3D stick view and corresponding animation of the thumb and forefinger during needling were generated. The thumb and forefinger movement of AM can be fully displayed with the synchronous playback of kinematic parameter curve and stick animation, which can help researchers to explore the movement characteristics and compare the similarities and differences of different AM skills.
Throughout the entire experimental process, some critical steps that affect the results of the analysis can be summarized-first, the experimental environment configuration. The recommended temperature of the experimental environment is constant 22-25 °C, and relative humidity is about 60% without obvious airflow in the room. Meanwhile, there is no strong noise and electromagnetic source interference in the surrounding environment. Second, the placements of the camera and tripod. In the process of motion tracking, all tracking points should be recorded by all cameras to obtain high-precision data. Therefore, a reasonable camera position is key to reducing experimental errors. Furthermore, the tripods should be adjusted to a proper height (higher than the table and ensure that the experimental devices on the table and the hand of the participant can be clearly recorded). Third, calibration and automatic motion tracking. All analysis data is calculated based on the position of each tracking point in the 3D calibration system in each frame of the motion video; therefore, successful calibration and automatic tracking of each point are prerequisites for performing calculations. Finally, identification of crests and troughs. The technical indicators of AM can be calculated by the positioning of the crests and troughs in each cycle. In this protocol, the steps of automatic identification and manual review are designed to ensure the accuracy of the experimental data.
In order to apply 3D motion tracking technology to the kinematic analysis of AM, two modifications were made to this technology commonly used in the large joints of human limbs. First, the customization of a small 3D calibration frame for fingers. A 15×15×15cm 3D calibration frame was customized for improving the measurement accuracy of finger movements. Through 3D laser scanning, the calibration accuracy of the frame is 0.01mm. Second, the establishment of technical indicators of AM and related calculation methods. According to the motion characteristics of AM and the raw data exported by the motion tracking system, three technical indicators, "amplitude, velocity, and time" along three axes were established for each finger tracking point. These parameters can be calculated by PHP script based on the inflection point recognition of the coordinate-time curve. The possible crests and troughs can be identified according to the logical expression (1) and (2), respectively.
(1)
(2)
Where dc, dt and dt2 are the differentiations of coordinate value, time and time squared, d2c is the quadratic differentiation of coordinate. According to the test results of experimental sample data, two types of thresholds were set for verifying the validities of these crests and troughs. The time threshold is 80% of the average operating cycle, the crest and trough thresholds are 75% and 25% of the maximum operating amplitude. After traversing all the crests and troughs, the crest whose interval time from the previous crest is greater than the time threshold and coordinate value is greater than the crest threshold is identified as the valid crest. The trough whose interval time from the previous crest is greater than the time threshold and coordinate value is less than the trough threshold is identified as the valid trough. Although, in most cases, the crests and troughs can be identified automatically, there are still a few cases that need to be adjusted manually. Therefore, as the main limitation of this solution, the recognition algorithm needs to be improved in future work. The preliminary analysis of the experimental data showed that the movement amplitude and velocity of MCP joints were the smallest, and the related parameters of IP or PIP joint and fingertips were larger and largest, respectively. Moreover, the needle body was driven by the vertical or tangential movement of the fingertips to move up and down or rotate on a fixed axis. In summary, AM is a kind of rhythmic movement performed by fingertips driven by MCP joints of the thumb and forefinger. Moreover, no matter which AM skill was used, a certain range of movement occurred along three axes at all tracking points, which suggests that during the operation of the lifting-thrusting skill, although the fingertips mainly move in the vertical direction, it is still accompanied by a tangential coupled movement, and the tangential-based twirling skill is also accompanied by a vertical coupled movement. These results indicate that the AM is not a simple single-axis movement.
Similar to other studies that use this technology to analyze finger motion, the motion tracking technology in this protocol also provides three-axis kinematics data of finger joints with high accuracy22. However, a secondary analysis on raw data according to the skill characteristics of AM was performed, and corresponding technical indicators were established in this protocol for further comparative analysis. Furthermore, compared with the portable, easy-to-use and low-cost hand motion tracking devices such as Leap Motion, standard marker-based motion tracking analysis has the advantages of higher accuracy and wider application range23,24. Compared with the traditional AM analysis device ATP-II, the amplitude-time curve along the main motion axis derived from motion tracking analysis and the voltage-time curve derived by ATP-II have significant conformity in the same AM skill. Moreover, the operating cycles calculated by the two measurement methods were also relatively consistent. These results showed that this experimental method can not only reflect similar skill characteristics to that of ATP-II but also provide more kinematics parameters along three axes of multiple tracking points, which cannot be measured by previous experimental technology.
This experimental method provides an efficient way for analyzing complicated movements of fingers involved in AM. It has great potential for future applications. First, the study of the dose-effect relationship of acupuncture. 3D finger motion tracking technology provides a solution for determining the stimulation amount of manual acupuncture and can be used to carry out studies such as the correlation analysis between needling velocity, amplitude and therapeutic effect, so as to provide more scientific data support for the clinical application of acupuncture. Second, the quantitative evaluation and feedback for the teaching and learning of AM. The results from data analysis combined with the teacher's verbal feedback can help learners adjust their finger actions and reduce the cognitive load24,25. Previous studies have used the data provided by 3D motion tracking technology to improve the effect of motor skills learning, such as repetitive overarm throwing26 and musical performance27,28. Some reports also showed that medical skills such as colonoscopy29, laparoscopic30, arthroscope31 and other endoscope32,33 could also be enhanced with this technology. And another study suggested that the video-based self-reflection and discussion with learners engaging at a higher cognitive level than the standard descriptive feedback34. Third, the measurement and preservation of famous acupuncturists' AM. Because all the AM is collected, recorded, and analyzed based on motion videos stored in the database, these videos and relevant data of AM can be browsed by researchers at any time for further learning and inheritance.
The establishment of this experimental method opens up a new way for the quantitative research of AM. In the future, more camera positions, higher-definition lenses, and Higher precision calibration frames can be applied to further improve data accuracy and dig out more meaningful technical indicators to provide more data reference for the clinical application, education, and promotion of acupuncture.
The authors have nothing to disclose.
This work was supported by the National Natural Science Foundation of China (Grant Number. 82174506).
3D calibration frame | Any brand | 15 x 15 x 15 cm | |
Acupuncture needles | Suzhou Medical Appliance Factory | 0.35 x 40 mm | |
Double-sided tape | Any brand | Round, 1 cm-diameter | |
Reflective balls | Simi Reality Motion Systems GmbH | 6.5 mm-diameter | |
SD card | Western Digital Corporation | SDXC UHS-I | |
SD card reader | UGREEN Group Limited | USB 3.0 | |
Simi Motion | Simi Reality Motion Systems GmbH | Ver.8.5.15 | |
Swab | Any brand | The volume fraction of ethanol is 70%-80% | |
Three cameras | Victor Company of Japan, Limited | JVC GC-PX100BAC | |
Three tripods | Any brand |