Method and apparatus for monitoring neuromuscular blockade, electronic device and storage medium
The ultrasound-based method for neuromuscular blockade monitoring addresses the inaccuracies and complexity of existing techniques by calculating a train-of-N ratio from muscle architectural variations, enhancing accuracy and usability in clinical settings.
Patent Information
- Application Number
- PCT/CN2025/112446
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2025-08-04
- Publication Date
- 2026-02-12
AI Technical Summary
Current quantitative neuromuscular monitoring methods for neuromuscular blockade are inaccurate and complex, requiring calibration and being sensitive to hand position and muscle temperature, with limited applicability in clinical settings.
A method using ultrasound imaging to determine temporal variations in muscle architectural parameters through consecutive electrical stimulations, calculating a train-of-N ratio without direct contact, which is self-calibrating and unaffected by hand position or muscle temperature, improving accuracy and simplifying the monitoring process.
The method provides high accuracy and consistency with traditional methods like AMG, with no need for calibration, and offers a non-invasive, user-friendly solution for monitoring neuromuscular blockade, reducing the risk of residual neuromuscular blockade complications.
Smart Images

Figure CN2025112446_12022026_PF_FP_ABST
Abstract
Description
METHOD AND APPARATUS FOR MONITORING NEUROMUSCULAR BLOCKADE, ELECTRONIC DEVICE AND STORAGE MEDIUMTECHNICAL FIELD
[0001] Embodiments of the present disclosure relate to the field of ultrasound imaging and, particularly, relate to a method and an apparatus for monitoring neuromuscular blockade, an electronic device, and a storage medium.BACKGROUND
[0002] In the post-anesthesia care unit, there is a high occurrence of residual neuromuscular blockade (rNMB) which puts patients at risk of negative consequences like hypoxia, upper airway obstruction, and aspiration.
[0003] The most commonly used quantitative neuromuscular monitoring (QNM) methods monitor neuromuscular blockade through approaches including acceleromyography (AMG) and electromyography (EMG) .
[0004] However, the current QNM methods have the drawbacks of low accuracy and a complex monitoring process.SUMMARY
[0005] Embodiments of the present disclosure provide a method and an apparatus for monitoring neuromuscular blockade, an electronic device, and a storage medium.
[0006] A first aspect of the embodiments of the present disclosure provides a method for monitoring neuromuscular blockade, including: determining, based on an ultrasound image sequence of a target muscle in response to N consecutive electrical stimulations at equal intervals, temporal variation information of an architectural parameter of the target muscle , where N is an integer greater than 1; determining, based on the temporal variation information and in response to N consecutive electrical stimulations at equal intervals, a first change amount in the architectural parameter of the target muscle after an Nth electrical stimulation and a second change amount in the architectural parameter after a first electrical stimulation; determining a train-of-N ratio based on a ratio of the first change amount to the second change amount; and determining a degree of neuromuscular blockade based on a value of the train-of-N ratio.
[0007] A second aspect of the embodiments of the present disclosure provides an apparatus for monitoring neuromuscular blockade, including: a first determining module, configured to determine, based on an ultrasound image sequence of a target muscle in response to N consecutive electrical stimulations at equal intervals, temporal variation information of an architectural parameter of the target muscle , where N is an integer greater than 1; a second determining module, configured to determine, based on the temporal variation information and in response to N consecutive electrical stimulations at equal intervals, a first change amount in the architectural parameter of the target muscle after an Nth electrical stimulation and a second change amount in the architectural parameter after a first electrical stimulation; a third determining module, configured to determine a train-of-N ratio based on a ratio of the first change amount to the second change amount; and a fourth determining module, configured to determine a degree of neuromuscular blockade based on a value of the train-of-N ratio.
[0008] A third aspect of the embodiments of the present disclosure provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable by the processor, where the processor, upon executing the program, is used to implement the method according to any one of the preceding embodiments.
[0009] A fourth aspect of the embodiments of the present disclosure provides a storage medium storing an instruction thereon, where the instruction, upon being executed by a processor, is used for implementing the method as described in any of the preceding embodiments.
[0010] The present disclosure provides a method and an apparatus for monitoring neuromuscular blockade, an electronic device, and a storage medium. The ultrasound image sequence is determined by the echo reflection of the target muscle without direct contact with the target muscle, unaffected by hand position and muscle temperature, which improves the accuracy of the method for monitoring neuromuscular blockade and simplifies the monitoring method. Moreover, the above monitoring method determines the degree of neuromuscular block by the size of the N consecutive stimulation ratios, where the first and second variations are the differences in the architectural parameter of the target muscle before and after the corresponding nerve electrical stimulation, without the need for calibration of the monitoring equipment, further improving the accuracy of the method for monitoring neuromuscular blockade and simplifying the monitoring method. The results of monitoring neuromuscular blockade by this method show no significant difference between the left and right hands, with a high mean accuracy, and have good consistency with the traditional method of monitoring neuromuscular blockade using AMG.
[0011] Additional aspects and advantages of the embodiments of the present disclosure will be partially given in the description below, which will become apparent from the description below, or will be learned through practice of the present disclosure.BRIEF DESCRIPTION OF THE DRAWINGS
[0012] In order to more clearly illustrate the technical solution of the embodiments of the present disclosure, the accompanying drawings required for describing of the embodiments of the present disclosure are briefly introduced as follows. Obviously, the accompanying drawings in the following description are only some embodiments of the present disclosure. For those skilled in the art, other accompanying drawings can be obtained based on these accompanying drawings without creative labor.
[0013] FIG. 1 is a flowchart of a method for monitoring neuromuscular blockade according to an embodiment of the present disclosure.
[0014] FIG. 2 is a schematic diagram of the structure of a target muscle according to an embodiment of the present disclosure.
[0015] FIG. 3 shows the variation of the mean thickness of APM with the number of times under four stimuli. TOFR = T4 / T1.
[0016] FIG. 4 shows the variation of the mean thickness of APM with the number of frames under four stimuli. TOFR = T4 / T1.
[0017] FIG. 5 shows the measurement results of TOFR in the left and right hands using the AMG and SMMG methods, respectively. *p > 0.05 (paired t-test) .
[0018] FIG. 6 shows the measurement results of TOFR using AMG and SMMG methods respectively as according to an embodiment of the present disclosure.
[0019] FIG. 7 shows a Bland-Altman plot of the differences in TOFR measured by AMG and SMMG against the corresponding means (n = 20) according to an embodiment of the present disclosure. UL = upper limit, LL = lower limit.
[0020] FIG. 8 is a flowchart illustrating another method for monitoring neuromuscular blockade according to by an embodiment of the present disclosure.
[0021] FIG. 9 is a flowchart according to an embodiment of the present disclosure, illustrating the process of generating a mask image corresponding to subsequent frame images through the XMem model and obtaining the mean thickness of the target muscle.
[0022] FIG. 10 is a flowchart of the process included in step 203 of FIG. 8.
[0023] FIG. 11 is a flowchart of the process included in step 206 of FIG. 8.
[0024] FIG. 12 is a flowchart of the process included in step 204 of FIG. 8.
[0025] FIG. 13 is a block diagram of an apparatus for monitoring neuromuscular blockade according to an exemplary embodiment.
[0026] FIG. 14 is a block diagram of an electronic device according to an exemplary embodiment.DETAILED DESCRIPTION
[0027] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. Unless otherwise indicated, when the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are merely examples of devices and methods consistent with some aspects of the present invention as detailed in the appended claims.
[0028] In the embodiments of the present disclosure, the terms used are only for the purpose of describing specific embodiments and are not intended to limit the present disclosure. The singular forms of "a" , "said" and "the" used in the present disclosure and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used in this article refers to and includes any or all possible combinations of one or more associated listed items. For example, A and / or B may represent: A exists alone, both A and B exist, and B exists alone. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. The term "multiple" refers to two or more. In view of this, "multiple" can also be understood as "at least two" in the embodiments of the present disclosure.
[0029] It should be understood that although the terms first, second, third, and the like may be used in the present disclosure to describe various information, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present disclosure, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, for example, the word "if" as used herein may be interpreted as "when" or "upon" or "in response to determining…"
[0030] The following will be combined with the drawings in the embodiments of the present disclosure to clearly and completely describe the technical solutions according to some embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by those skilled in the art without making creative work shall fall within the protection scope of the present disclosure.
[0031] The embodiments of the present disclosure provide a method and an apparatus for monitoring neuromuscular blockade, an electronic device, and a storage medium.
[0032] Herein, the methods and the devices are based on the same application concept. Since the methods and the devices solve the problem in a similar principle, the embodiments of the devices and the methods can refer to each other, and the repeated parts will not be elaborated.
[0033] As shown in FIG. 1, an embodiment of the present disclosure provides a method for monitoring neuromuscular blockade.
[0034] The method may include the following step (s) .
[0035] In step 101, temporal variation information of an architectural parameter of the target muscle is determined based on an ultrasound image sequence of a target muscle in response to N consecutive electrical stimulations at equal intervals, where N is an integer greater than 1.
[0036] When the target muscle is subjected to electrical stimulation, the amplitude of muscle twitch changes, thereby causing a change in the architectural parameter of the muscle. In some embodiments, the ultrasound image sequence is the image of the target muscle undergoing changes in the amplitude of muscle twitch under electrical stimulation. Each frame of the image can reflect the information of the architectural parameter of the target muscle changing over time. Since the muscle is a three-dimensional structure, its ultrasound image is not a regular shape. In some embodiments, the change in the amplitude of muscle twitch is measured based on the change in the architectural parameter of the target muscle.
[0037] In some embodiments, the architectural parameter includes at least one of muscle thickness, cross-sectional area, muscle fiber length, pennation angle, or location deformation. It should be noted that, the following embodiments are described by taking the muscle thickness (mean thickness) as an example, but those skilled in the art should understand that these embodiments are not intended to limit the protection scope of the invention. Based on the same concept disclosed in the embodiments related to muscle thickness, embodiments of deriving other parameters and utilizing their changes during the electrical stimulations are also conceivable by those skilled in the art and, thus, should fall within protection scope of the invention.
[0038] As shown in FIG. 2, an embodiment of the present disclosure provides a schematic diagram of the structure of a target muscle. The process of confirming the mean thickness of the target muscle is as follows: the mean vertical distance D0 between the first muscle boundary 101 and the second muscle boundary 102 of the target muscle 100 is taken as the mean thickness of the target muscle 100, and this is used to measure the degree of change in the amplitude of muscle twitch of the target muscle 100. Here, the direction from the first muscle boundary 101 to the second muscle boundary 102 is the thickness direction of the target muscle 100, and the thickness direction of the target muscle 100 is set parallel to the Y direction in FIG. 2. The first muscle boundary 101 and the second muscle boundary 102 appear as linear high-echo in each frame of the image.
[0039] The information on the mean thickness of the target muscle over time records the mean thickness values of the target muscle during the sampling time. Here, the mean thickness of the target muscle is characterized by the mean vertical distance between the first muscle boundary and the second muscle boundary in the target muscle, corresponding to the pixel distance. The sampling time can be represented by the frame number sequence or the time corresponding to the sampling frame number. The product of the pixel distance and the spatial resolution can be converted into physical dimensions. The sampling frame number divided by the sampling rate can be converted into time.
[0040] In some embodiments, the target muscles include but are not limited to the abductor pollicis minimi (APM) with relatively high sensitivity and the orbicularis oculi muscle. When the target muscles are different, the corresponding nerves for the target muscles may also be different. In some embodiments, the abductor pollicis minimi, which is innervated by the ulnar nerve, is recommended for quantitative neuromuscular blockade monitoring because it is easily accessible on the hand and is closely related to the recovery of most sensitive muscles, which can improve the accuracy of the method for monitoring neuromuscular block.
[0041] In step 102, a first change amount in the architectural parameter of the target muscle after an Nth electrical stimulation and a second change amount in the mean thickness after a first electrical stimulation is determined based on the temporal variation information and in response to N consecutive electrical stimulations at equal intervals.
[0042] In some embodiments, still taking the muscle thickness as an example of the architectural parameter, the change in the average thickness of the target muscle during the Nth muscle twitch response after the Nth electrical stimulation is the first change. The change in the average thickness of the target muscle during the first muscle twitch response after the first electrical stimulation is the second change. Specifically, in the information about the average thickness of the target muscle changing over time, within a time interval, the change in the average thickness of the target muscle between the initial and end moments of the Nth muscle twitch response is the first change. Within a time interval, the change in the average thickness of the target muscle between the initial and end moments of the first muscle twitch response is the second change. The aforementioned time interval is the time interval between any two adjacent electrical stimulations in N consecutive and equally spaced electrical stimulations. Since the first change and the second change are the differences in the average thickness of the target muscle after corresponding nerve electrical stimulation, there is no need to calibrate the monitoring equipment, which further improves the accuracy of the method for monitoring neuromuscular blockade and simplifies the monitoring method.
[0043] It should be noted that in some embodiments, the continuous N electrical stimulations felt by the target muscle are achieved by applying N pulses at equal intervals to the corresponding nerve.
[0044] In step 103, a train-of-N ratio is determined based on a ratio of the first change amount to the second change amount.
[0045] In step 104, a value of the train-of-N ratio is determined based on a degree of neuromuscular blockade.
[0046] By applying continuous N electrical stimulations to the nerve corresponding to the target muscle, N muscle twitches from T1 to TN are measured. The larger the value of the train-of-N ratio, the better the muscle relaxation recovery, the lower the degree of neuromuscular blockade, and the lower the incidence of residual neuromuscular blockade (rNMB) ; the smaller the value of the train-of-N ratio, the more obvious the muscle relaxation residue, the higher the degree of neuromuscular blockade, and the higher the incidence of residual neuromuscular blockade (rNMB) .
[0047] When the value of the train-of-N ratio is greater than or equal to the preset value, it indicates that the neuromuscular function can fully recover after continuous N electrical stimulations of the corresponding nerve, and the degree of neuromuscular blockade is low, with a very low incidence of residual neuromuscular blockade (rNMB) . When the value of the train-of-N ratio is less than the preset value, it indicates that the neuromuscular function does not recover well after continuous N electrical stimulations of the corresponding nerve, the degree of neuromuscular blockade is high, and the incidence of residual neuromuscular blockade (rNMB) is very high, which puts patients at risk of adverse consequences such as hypoxia, upper airway obstruction, and aspiration.
[0048] For example, when the target muscle is the APM, the number of N can be four. When the value of the train-of-four ratio (TOFR) is greater than or equal to 0.9, it indicates that the neuromuscular function can fully recover after continuous 4 times of electrical stimulation of the corresponding nerve, the degree of neuromuscular blockade is low, and the incidence of residual neuromuscular blockade (rNMB) is very low. When the value of TOFR is less than 0.9, it indicates that the neuromuscular function does not recover well after continuous 4 times of electrical stimulation of the corresponding nerve, the degree of neuromuscular blockade is high, and the incidence of residual neuromuscular blockade (rNMB) is very high, which puts the patient at risk of adverse consequences such as hypoxia, upper airway obstruction and aspiration.
[0049] It should be noted that some commonly used QNM modes, including but not limited to AMG and EMG, have the drawbacks of low accuracy and complex monitoring processes, as detailed below.
[0050] Clinical assessment has traditionally been the primary method for evaluating a patient's recovery of neuromuscular strength; however, its reliability remains questionable. The peripheral nerve stimulator (PNS) has emerged as a valuable tool, enabling nerve stimulation and allowing clinicians to assess muscle response through visual or tactile observation. Due to its ease of use and user-friendly design, PNS has gained widespread acceptance in clinical practice. Nevertheless, the subjective nature of nerve block evaluation using PNS constitutes a major limitation, particularly when compared to the higher precision offered by QNM, especially at low levels of neuromuscular blockade. To address the risks associated with residual neuromuscular blockade (rNMB) , measurement of the train-of-four ratio (TOFR) is recommended. Neuromuscular recovery is considered adequate when the TOFR is ≥ 0.9, particularly when measured at the APM. The TOFR stimulation protocol involves delivering four electrical stimuli at a frequency of 2 Hz, with an inter-stimulus interval of 0.5 seconds. The TOFR is calculated as the ratio of the amplitude of the fourth muscle twitch (T4) to that of the first twitch (T1) . Therefore, to enhance the accuracy of neuromuscular function assessment in clinical settings, it is recommended to employ TOFR in conjunction with quantitative monitoring devices. This approach not only improves the precision of neuromuscular assessment but also significantly reduces the risk of complications associated with residual neuromuscular blockade.
[0051] Currently, various QNM techniques are available for the assessment of TOFR. Mechanomyography (MMG) is recognized for its high sensitivity and is often considered a potential "gold standard" due to its accurate and reliable measurements. However, the application of MMG in clinical settings is limited by its complex setup, including the requirement for calibration prior to use, as well as the lack of commercially available systems tailored for clinical environments. Kinemyography (KMG) offers a simpler setup process that does not require an external display or calibration; however, its applicability is restricted to the APM. AMG provides greater flexibility, as it can be applied to any muscle capable of free movement, such as those located in the hand, foot, or face. Nevertheless, it is not suitable for immobilized muscles, and to ensure measurement accuracy, the device must be calibrated before the administration of neuromuscular blocking agents (NMBAs) and normalized against baseline values. EMG is regarded as equally accurate and reliable as MMG and can be utilized at a wide range of anatomical sites, including immobilized muscles. However, EMG is prone to electrical interference, and its accuracy may be affected by low muscle temperatures. Similar to AMG, EMG also requires device calibration before NMBA administration to maintain high measurement precision. Therefore, the primary barrier to the widespread adoption of QNM is not the cost of the technology, but rather the challenges related to usability and consistency. To overcome these limitations, there is a growing need for innovative monitoring techniques that are insensitive to patient hand positioning, capable of self-calibration, able to produce reliable results, easy to set up, and capable of generating reproducible responses. The development and implementation of such technologies would significantly promote the integration of QNM into routine clinical practice, thereby enhancing patient safety and improving the overall quality of anesthesia care.
[0052] To solve the above technical problems, the present disclosure provides a method for monitoring neuromuscular blockade. The ultrasound image sequence is determined by the echo reflection of the target muscle without direct contact with the target muscle, unaffected by hand position and muscle temperature, which improves the accuracy of the method for monitoring neuromuscular blockade and simplifies the monitoring method. Moreover, the above monitoring method determines the degree of neuromuscular block by the size of the N consecutive stimulation ratios, where the first and second variations are the differences in the architectural parameter of the target muscle before and after the corresponding nerve electrical stimulation, without the need for calibration of the monitoring equipment, further improving the accuracy of the method for monitoring neuromuscular blockade and simplifying the monitoring method. The results of monitoring neuromuscular blockade by this method show no significant difference between the left and right hands, with a high mean accuracy, and have good consistency with the traditional method of monitoring neuromuscular blockade using AMG.
[0053] Optionally, a sampling rate of the ultrasound image sequence is greater than or equal to 200 frames per second (fps) .
[0054] Ultrasound imaging presents a superior alternative for muscle monitoring, offering a non-invasive method to evaluate muscle architecture. Through B-mode ultrasound imaging, sonomyography (SMG) assesses dynamic muscle morphology, including thickness, pennation angle, fascicle length, and cross-sectional area. SMG is derived from B-mode ultrasound images, which are traditionally limited to a sampling rate of approximately 30 frames per second. However, advancements in ultrafast ultrasound imaging now permit B-mode imaging at significantly higher frame rates, often reaching several thousand frames per second (fps) . In some embodiments, the sampling rate of the ultrasound image sequence is greater than or equal to 2000 fps. Ultrafast ultrasound imaging with a frame rate exceeding 2000 fps is utilized to generate a two-dimensional representation of transient muscle motion, termed SMMG. The cutting-edge SMMG technology has recently demonstrated considerable potential in enhancing the comprehension of neuromuscular disorders, for example, using this method, it is found that the use of transverse friction massage could decrease the active muscle stiffness. Additionally, it offers valuable insights that aid in pinpointing changes in muscle activation timing and in evaluating the effectiveness of therapeutic interventions. SMMG's unique ability to measure muscle vibration from different locations within the muscle, coupled with its complementary features to existing muscle-related signals like EMG and MMG, makes it a versatile and valuable tool for various applications. This embodiment aims to develop a novel approach and protocol that leverages SMMG to map out muscle movements in two dimensions to achieve precise and reliable QNM during anesthesia and recovery, making it a promising technology for the future.
[0055] To verify that the monitoring results of the method for monitoring neuromuscular blockade show no significant difference between the left and right hands, have a high average accuracy, and have a good consistency with the method of monitoring neuromuscular blockade through traditional accelerated myography (AMG) , the following experimental process was carried out.
[0056] In some embodiments, SMMG of the APM and AMG of the thumb were employed to measure the train-of-four ratio (TOFR) in a cohort of 20 healthy adult subjects. The results revealed no statistically significant difference between the left and right hands for both AMG-and SMMG-derived TOFR values, with p-values from paired t-tests exceeding 0.05. Furthermore, the mean accuracy of SMMG-based TOFR measurements, expressed as relative error (0.6%) , was superior to that of AMG (1.4%relative error) . The Bland-Altman analysis demonstrated that all data points fell within the limits of agreement, with a mean bias of 0.02, indicating a high degree of agreement between the two methods. Notably, SMMG did not require any additional calibration prior to measurement, offering a practical advantage over AMG. Collectively, these findings suggest that the proposed method holds promise as a novel QNM technique and warrants further clinical investigation to evaluate its potential benefits for patient care. Future studies are necessary to assess the performance of this method in patients undergoing and recovering from anesthesia in order to validate its clinical applicability.
[0057] The following section provides a detailed description of the experimental procedure and the subsequent analysis of the obtained results.
[0058] In some embodiments, a convenience sampling approach was utilized to recruit a total of 20 healthy participants (7 females and 13 males) , aged between 18 and 52 years. All participants were required to be free from implantable electronic devices and to have no history of metabolic, neurological, or muscular disorders that could potentially affect the study outcomes. The inclusion criteria permitted the participation of individuals who engage in occasional physical activity or maintain a sedentary lifestyle. However, professional athletes were excluded from the study to minimize potential confounding effects associated with superior neuromuscular performance. Ethical approval for this study was obtained from the Human Subjects Ethics Sub-committee (HSESC) of The Hong Kong Polytechnic University (Reference No. HSEARS20240220002) , and all subjects provided written informed consent prior to participation. As presented in Table 1, data were collected from the 20 participants, who had a mean height of 171.5 ± 8.0 cm and a mean weight of 66.5 ± 13.6 kg. TABLE 1. Subjects characteristics
[0059] Neuromuscular stimulation was delivered using a biphasic constant current stimulator (DS8R, Digitimer Ltd., Hertfordshire, UK) . Concurrently, ultrasound images were captured with an advanced open ultrasound research platform (Verasonics Vantage 128, Verasonics, Inc., Kirkland, WA, USA) . A dedicated computer managed the image acquisition process and data storage. The custom imaging script was divided into two segments: the initial segment facilitated real-time imaging by accurately positioning the broadband linear ultrasound transducer L11-4v on the APM to ensure proper location and orientation. The subsequent segment managed the trigger output to the stimulator and the data acquisition process. The transducer operated at a frame rate of 2000 Hz, capturing a sequence of 5,000 frames. Upon initiating the recording, a trigger-out signal was dispatched from the Verasonics system to a function generator (AFG3052C, Tektronix, Beaverton, USA) , which in turn transmitted a signal to the current stimulator (1 MHz, 1-cycle burst, 5 Vpp, 2.5 V offset, 50%duty cycle) . Two self-adhesive stimulation electrodes, each with a diameter of 3.2 cm (ValuTrode, Axelgaard Manufacturing Co. Ltd., USA) , were employed to stimulate the ulnar nerve. The ultrasound system's radio-frequency raw data were preserved for subsequent image reconstruction and analysis.
[0060] The AMG measurements were conducted using a commercial anesthesia monitoring device (JS-100, Beijing Sigo Medical Technology Co., LTD, China) . AMG technology, grounded in Newton’s Second Law of Motion (force = mass × acceleration) , utilizes a piezoelectric sensor to quantify the acceleration of the thumb during muscle contractions. Notably, the minimum stimulation current threshold of the commercial AMG device is 35 mA, whereas the minimum current used in this study was 20 mA. To reconcile this discrepancy, the stimulation current amplitude of the AMG device was attenuated to 20mA by integrating a 100-ohm resistor with the external current stimulator. This adjustment allowed the AMG device's stimulation current output to serve as the input signal for the current stimulator.
[0061] Participants were seated comfortably in a chair with their arms and hands secured using a custom-designed plate and straps, ensuring stability during the testing procedure. The skin was cleansed with medical alcohol to lower skin impedance to an effective range, thereby enhancing the efficacy of the electrical stimulation. Electrodes were positioned on either side of the ulnar nerve to minimize potential errors due to nerve location variability. Subsequently, the thumb accelerometer and the SMMG probe were affixed to the thumb and palm, respectively. A specialized bracket was utilized to stabilize the ultrasound probe, ensuring that the measurement orientation remained consistently perpendicular to the APM. Ultrasound gel was then applied to the APM to facilitate clear imaging. Following the instructions of the AMG anesthesia monitoring device, the accelerometer sensor for AMG was secured to the thumb using the provided hand adapter and medical adhesive tape. Participants were instructed to remain relaxed and avoid any muscle activity throughout the experiment. Each stimulation trial consisted of four identical pulses, each with a duration of 0.2 ms and an amplitude of 20 mA, spaced 0.5 seconds apart, in line with protocols established by previous research. Measurements were taken five times for both the left and right hands of each subject under normal conditions, with a two-minute rest interval between each trial to prevent muscle fatigue. SMMG of the APM and AMG of the thumb were employed to determine the TOFR in the 20 healthy adults upon stimulation of the ulnar nerve. Participants might experience very mild discomfort during stimulation, comparable to sensations produced by commercially available electrical massage apparatuses. Upon completion of all stimulation trials, subjects were carefully disconnected from the sensors and released from the study.
[0062] The custom script written in MATLAB (MATLAB R2020b, The MathWorks, Inc., USA) was employed to reconstruct images from the raw radio-frequency ultrasound data. The data derived from image processing were subsequently analyzed using GraphPad Prism (GraphPad Prism 9.1.2, GraphPad Software, USA) . All values were presented as mean ±standard deviation (SD) . A paired sample t-test was utilized to evaluate the differences between measurements obtained from the left and right hand experiments. A p-value threshold of > 0.05 was established to indicate a lack of statistically significant difference between the measurements of the left and right hands. Furthermore, a TOFR of 1 was adopted as the normative standard under normal physiological conditions, and the accuracy (relative error) of measurements obtained by the two devices was calculated. Additionally, the TOFR measured by AMG and SMMG were compared using Bland–Altman analysis for repeated measurements. This analysis is designed to assess the agreement between two measurement methods by evaluating the mean difference (bias) and the 95%limits of agreement. A small bias and narrow limits of agreement are indicative of a strong agreement between the devices.
[0063] A representative B-mode ultrafast ultrasound image of the APM was capture a representative B-mode ultrafast ultrasound image of the APM by SMMG. The mean distance between the upper and lower boundaries of the muscle was selected as the measurement parameter. One of the key advantages of SMMG is its capacity for non-invasive visualization of transient muscle movements or vibrations, achieved with high spatial and temporal precision. FIG. 3 shows the variation of the mean thickness of APM with the number of times under four stimuli. TOFR =T4 / T1. FIG. 4 shows the variation of the mean thickness of APM with the number of frames under four stimuli. TOFR = T4 / T1. Consequently, as illustrated in FIG. 3 and FIG. 4, the variations in the mean thickness of the APM during a series of four identical stimuli (with 1 pixel equivalent to approximately 0.137 mm) were documented. The SMMG TOFR was then determined by dividing the change in muscle thickness observed in the fourth response (T4) by that of the first response (T1) , incorporating a self-calibration process. This method allows for a detailed and accurate assessment of neuromuscular function, leveraging the high frame rate capabilities of SMMG.
[0064] In the initial experiment, AMG was employed to measure the TOFR of the left and right hands of participants, as illustrated in FIG. 5. The mean AMG TOFR values for the left and right hands of the 20 subjects were 1.016 ± 0.016 and 1.011 ± 0.035, respectively. Statistical analysis revealed no significant difference between the left and right hands in terms of AMG TOFR, with a p-value of 0.50 (paired t-test) , which exceeds the threshold of 0.05. Concurrently, SMMG was utilized to determine the TOFR values for the subjects' left and right hands. The average SMMG TOFR values were 0.996 ± 0.023 for the left hand and 1.002 ±0.033 for the right hand. Similarly, the p-value calculated for the comparison between the left and right hands was 0.55. Therefore, it can be inferred that there was no significant difference in TOFR values between the left and right hands for both AMG and SMMG, with p-values (paired t-test) exceeding 0.05. After these findings, the next phase of the experiment involved a comparative analysis of the efficacy of the AMG and SMMG methods in measuring TOFR.
[0065] FIG. 6 displays the measurement outcomes of the TOFR utilizing both AMG and SMMG across 20 participants. The results reveal that the average TOFR values obtained through AMG and SMMG were 1.014 ± 0.021 and 0.994 ± 0.018, respectively. Given that a TOFR of 1 was established as the normative benchmark for subjects in a normal physiological state, the mean accuracy of the SMMG TOFR was calculated to be 0.6%relative error. This accuracy surpasses that of AMG, which demonstrated a 1.4%relative error. These findings underscore the superior precision of the SMMG technique in quantifying the TOFR of subjects under normal conditions.
[0066] FIG. 7 presents a Bland–Altman plot illustrating the differences in TOFR measurements obtained by AMG and SMMG plotted against their corresponding means. The plot reveals that all difference values fell within the defined narrow limits of agreement, ranging from -0.026 to 0.066. Additionally, a mean difference (bias) of 0.020 between the TOFR measurements by AMG and SMMG was observed, indicating that the TOFR values measured by AMG were generally higher than those measured by SMMG. When compared with biases and limits of agreement reported in other studies, as summarized in Table 2, both the bias and the agreement parameters for SMMG and AMG were superior to those of other devices with a similar number of tested subjects. Consequently, it can be inferred that these two methods demonstrate good agreement when assessing subjects in a normal, non-anesthetized state. It is important to note that the aforementioned experiments were conducted with subjects in a non-anesthetized condition. To draw more comprehensive conclusions, clinical tests under anesthesia are required, which will form the focus of the subsequent research endeavors. This future work will aim to validate the effectiveness and reliability of this monitoring method in a clinical anesthetic context. TABLE 2. Agreement between different devices
[0067] In this research, a novel QNM technique utilizing SMMG has been successfully developed. This method involves calculating the TOFR by dividing the change in thickness of the APM at the fourth stimulus by that at the first stimulus, with the process being self-calibrated. SMMG was applied to the APM and AMG to the thumb to determine the TOFR in 20 healthy adults upon ulnar nerve stimulation. The experimental findings indicated no significant differences in TOFR measurements between the left and right hands, whether using AMG or SMMG. Furthermore, the SMMG method was compared under normal conditions with traditional anesthesia monitoring techniques, i.e. AMG, and discovered that SMMG offers superior accuracy in measuring TOFR. Additionally, Bland–Altman plot analysis revealed a strong agreement between the SMMG device and the commercial AMG device.
[0068] Accurate and user-friendly monitoring of anesthesia levels and the recovery process is of paramount importance in clinical settings. The outcomes of this study imply that the system been developed holds promise as a medical device that could enhance patient care. Nonetheless, additional research is required to assess this new method in patients during and post-anesthesia to fully ascertain its clinical applicability and efficacy.
[0069] As shown in FIG. 8, an embodiment of the present disclosure provides another method for monitoring neuromuscular blockade.
[0070] The method may include the following step (s) .
[0071] In step 201, the ultrasound image sequence of the target muscle is acquired in response to N consecutive electrical stimulations on a nerve corresponding to the target muscle.
[0072] In some embodiments, the ultrasound imaging equipment emits ultrasound waves to the target muscle. When the ultrasound waves encounter the interface of different acoustic impedances of tissues in the target muscle, reflected echoes are generated. The reflection intensity is positively correlated with the difference in tissue density. The ultrasound imaging equipment obtains the ultrasound image sequence of the target muscle under the condition of continuous N times of electrical stimulation of the corresponding nerve based on the intensity and reception time of the reflected ultrasound waves received.
[0073] In step 202, normalization is performed on the ultrasound image sequence of the target muscle .
[0074] Specifically, the size of the ultrasound image sequences of the target muscle was set to 512 × 512, and the images were normalized to a pixel intensity range of 0 to 1 prior to being input into the model.
[0075] In step 203, a first frame image of the ultrasound image sequence and a mask image corresponding to the first frame image as initial information are acquired, where in the mask image corresponding to the first frame image, the target muscle and muscle background are segmented by a muscle boundary to determine a position of the target muscle in the first frame image.
[0076] In some embodiments, the annotations of muscle boundaries in the mask image corresponding to the first frame can be provided by the user.
[0077] In step 204, the target muscle in subsequent frame images of the ultrasound image sequence based on the initial information using a segmentation algorithm is tracked, and mask images corresponding to the subsequent frame images are generated , where in the mask images corresponding to the subsequent frames, the target muscle and the muscle background are segmented by the muscle boundaries to determine positions of the target muscle in the subsequent frame images.
[0078] In some embodiments, an anonymized dataset is prepared including sequential ultrasound images of muscle IH×W×3 and annotations of muscle borders MH×W×1, i.e., aponeurosis and tendon. A total of 1, 251 images from 35 sequences are collected using different devices from the gastrocnemius, diaphragm, tibialis anterior, and SMMG of the APM to enhance the variability of the dataset. The images were annotated by an experienced sonographer.
[0079] This step is a dynamic ultrasound segmentation technology based on machine learning. Its core idea is to automatically segment the target muscle and the background of the target muscle in subsequent frame images guided by the initial information. An example is given as follows.
[0080] The first frame image of the ultrasound image sequence and the corresponding mask image are used as the training starting point. The segmentation algorithm is adopted to learn the annotations of the muscle boundaries on the anonymous dataset, segments the target muscle and the background of the target muscle, and obtain the mask images corresponding to the subsequent frame images.
[0081] In step 205, a target muscle contour is extracted from the mask images corresponding to respective frame images of the ultrasound image sequence; where the target muscle contour includes a first muscle boundary and a second muscle boundary, a direction from the first muscle boundary pointing to the second muscle boundary is a thickness direction of the target muscle, and the first muscle boundary and the second muscle boundary appear as linear high echoes in the respective frame images.
[0082] FIG. 9 is a flowchart according to an embodiment of the present disclosure, illustrating the process of generating a mask image corresponding to subsequent frame images through the XMem model and obtaining the mean thickness of the target muscle. Specifically, as shown in FIG. 9, through a preset algorithm, the first muscle boundary and the second muscle boundary of the target muscle contour are extracted from the mask images corresponding to each frame of the ultrasonic image sequence.
[0083] In step 206, architectural parameters of the target muscle in the mask images corresponding to the respective frame images is determined based on the target muscle contour.
[0084] Still taking the muscle thickness as an example of the architectural parameter, as the muscle is a three-dimensional shape, the mean thickness of the target muscle in each frame of the ultrasound image sequence corresponding to the mask image can be determined based on the first muscle boundary and the second muscle boundary.
[0085] In step 207, the temporal variation information is determined based on the architectural parameters of the target muscle in the mask images corresponding to the respective frame images.
[0086] The mean thickness value of the target muscle can be characterized by the mean distance between the first muscle boundary and the second muscle boundary in the mask images corresponding to each frame of the ultrasound image sequence, and the sampling time can be characterized by the frame number sequence or the time corresponding to the sampling frame number. The product of the pixel distance and the spatial resolution can be converted into physical dimensions. The sampling frame number divided by the sampling rate can be converted into time.
[0087] In step 208, a first change amount in the architectural parameter of the target muscle after an Nth electrical stimulation and a second change amount in the architectural parameter after a first electrical stimulation is determined based on the temporal variation information and in response to N consecutive electrical stimulations at equal intervals.
[0088] For the specific execution and intended effects of step 208, the foregoing step 102 can be referred to, which is not elaborated here.
[0089] In step 209, a train-of-N ratio is determined based on a ratio of the first change amount to the second change amount.
[0090] For the specific execution and intended effects of step 209, the foregoing step 103 can be referred to, which is not elaborated here.
[0091] In step 210, a value of the train-of-N ratio is determined based on a degree of neuromuscular blockade.
[0092] For the specific execution and intended effects of step 210, the foregoing step 104 can be referred to, which is not elaborated here.
[0093] Based on the above description, the specific steps for determining the information that changes over time are described in detail as follows. In some embodiments, a semi-automatic algorithm for quantifying changes in architectural parameter in ultrafast B-mode ultrasound is proposed.
[0094] FIG. 10 is a flowchart of the process included in step 203 of FIG. 8. As shown in FIG. 10, the method may include the following step (s) .
[0095] Optionally, obtaining the first frame image of the ultrasound image sequence and the corresponding mask image includes step 2031, step 2032 and step 2033.
[0096] In step 2031, the first frame image of the ultrasound image sequence is acquired.
[0097] In step 2032, the muscle boundary in the first frame image of the ultrasound image sequence is acquired.
[0098] In step 2033, the mask image corresponding to the first frame image is generated based on the muscle boundary in the first frame image.
[0099] The above steps achieve the process of obtaining the initial information, in which the annotations of the muscle boundaries in the mask image corresponding to the first frame can be provided by the user, which can segment the target muscle from its background.
[0100] FIG. 11 is a flowchart of the process included in step 206 of FIG. 8. As shown in FIG. 11, The method may include the following step (s) .
[0101] Optionally, the determination of the architectural parameter of the target muscle in the mask image corresponding to each frame image based on the target muscle contour includes step 2061, step 2062 and step 2063.
[0102] In step 2061, a first fitted polynomial is obtained by performing linear fitting on a boundary point set of the first muscle boundary in the mask images.
[0103] Optionally, before obtaining a first fitted polynomial, burrs and falsely segmented contours within the mask images corresponding to the respective frame images are removed, using a morphological image processing algorithm.
[0104] As shown in FIG. 9, in the resulting binary mask the open transform was used to remove the burrs and the falsely segmented contours, presenting the main two boundaries. The above scheme improves the accuracy of the first fitted polynomial and the second fitted polynomial.
[0105] In step 2062, a second fitted polynomial is obtained by performing linear fitting on a boundary point set of the second muscle boundary in the mask images.
[0106] In step 2063, the architectural parameter of the target muscle is determined based on a vertical distance between the first fitted polynomial and the second fitted polynomial.
[0107] As shown in FIG. 9, taking the muscle thickness as an example of the architectural parameter and assuming that the first muscle boundary and the second muscle boundary appear as linear high-echo structures in the ultrasound image, the boundary points are fitted using a first-order polynomial. Subsequently, the mean muscle thickness is determined by calculating the mean vertical distance between the two fitted polynomials.
[0108] Optionally, the segmentation algorithm includes a long-term video object segmentation model.
[0109] In some embodiments, the long video object segmentation models include XMem model, Associating Objects with Transformers (AOT) model, Space-Time Memory (STM) model, Recurrent Dynamic Embedding (RDE) , and the like. Based on the initial information, the target muscles in the subsequent frame images of the ultrasound image sequence are tracked through the long video object segmentation model, and the mask images corresponding to the subsequent frame images are generated. Compared with obtaining the mask images corresponding to each frame image through manual annotation, the segmentation efficiency is improved.
[0110] Optionally, the long-term video object segmentation model includes XMem model.
[0111] In some embodiments, the XMem model adopts a unified feature memory for long videos, inspired by the Atkinson-Shiffrin memory model. Previous research on video object segmentation typically utilized only one type of feature memory. For videos exceeding one minute, a single feature memory model would closely associate memory consumption with accuracy. In contrast, the XMem model follows the idea of the Atkinson-Shiffrin model and designs a structure that includes multiple independent but deeply connected feature memories: a rapidly updated sensory memory, a high-resolution working memory, and a compact and persistent long-term memory. Crucially, the XMem model has developed a memory potential algorithm that periodically integrates active working memory elements into long-term memory, thereby avoiding memory explosion and minimizing performance degradation in long-term predictions.
[0112] The XMem model adopts a unified feature memory for long videos, inspired by the Atkinson-Shiffrin memory model. Previous research on video object segmentation typically utilized only one type of feature memory. For videos exceeding one minute, a single feature memory model closely associates memory consumption with accuracy. In contrast, the XMem model follows the idea of the Atkinson-Shiffrin model and designs a structure that includes multiple independent but deeply connected feature memories: a rapidly updated sensory memory, a high-resolution working memory, and a compact and persistent long-term memory. Crucially, the XMem model has developed a memory potential algorithm that periodically integrates active working memory elements into long-term memory, thereby avoiding memory explosion and minimizing performance degradation in long-term predictions.
[0113] FIG. 12 is a flowchart of the process included in step 204 of FIG. 8. As shown in FIG. 12, The method may include the following step (s) .
[0114] Optionally, generating the mask image corresponding to the subsequent frame images includes step 2041 and step 2042.
[0115] In step 2041, a human memory model is initialized using the initial information.
[0116] In step 2042, memory features of the subsequent frame images are read via the human memory mode, and the mask images corresponding to the subsequent frame images are generated, where the memory features include a memory key and a memory value.
[0117] As shown in FIG. 9, the process of generating mask images corresponding to subsequent frame images through the XMem model may be as follows.
[0118] XMem, a state-of-the-art long-term video tracking algorithm, was fine-tuned with the ultrasound dataset. Inspired by the Atkinson-Shiffrin human memory model, XMem consists of three memory modules, namely sensory, working, and long-term memory, which is suitable for ultrafast ultrasound images with high frame rates and long sequences. For t-th ultrasound frame IH×W×3 from the sequence, memory reading was performed to obtain memory features F representing memory key k and value v of memory. Together with the sensory memory from (t-1) -th frame, features F was fed into the decoder to generate a mask Memory was updated after each inference. In our task, the model was pre-trained on YouTubeVOS and DAVIS, and finetuned with 15,000 iterations in a PC with an RTX 3080 GPU card. The sample H × W size was set to 512 × 512, and the ultrasound images were scaled to a range of 0~1 before being input into the model.
[0119] Referring to FIG. 13, based on the same principle as the methods according to the embodiments of the present disclosure, embodiments of the present disclosure further provide an apparatus for monitoring neuromuscular blockade, where the apparatus includes: a first determining module 1301, configured to determine, based on an ultrasound image sequence of a target muscle in response to N consecutive electrical stimulations at equal intervals, temporal variation information of an architectural parameter of the target muscle, where N is an integer greater than 1; a second determining module 1302, configured to determine, based on the temporal variation information and in response to N consecutive electrical stimulations at equal intervals, a first change amount in the architectural parameter of the target muscle after an Nth electrical stimulation and a second change amount in the architectural parameter after a first electrical stimulation; a third determining module 1303, configured to determine a train-of-N ratio based on a ratio of the first change amount to the second change amount; and a fourth determining module 1304, configured to determine a degree of neuromuscular blockade based on a value of the train-of-N ratio.
[0120] Optionally, the architectural parameter includes at least one of muscle thickness, cross-sectional area, muscle fiber length, pennation angle, or location deformation.
[0121] Optionally, the apparatus further includes an acquiring module, configured to acquire the ultrasound image sequence of the target muscle in response to the N consecutive electrical stimulations on a nerve corresponding to the target muscle; and a performing module, configured to perform normalization on the ultrasound image sequence of the target muscle.
[0122] Optionally, a sampling rate of the ultrasound image sequence is greater than or equal to 2000 frames per second (fps) .
[0123] Optionally, the first determining module is specifically configured to: acquire a first frame image of the ultrasound image sequence and a mask image corresponding to the first frame image as initial information; where in the mask image corresponding to the first frame image, the target muscle and muscle background are segmented by a muscle boundary to determine a position of the target muscle in the first frame image; track the target muscle in subsequent frame images of the ultrasound image sequence based on the initial information using a segmentation algorithm, and generate mask images corresponding to the subsequent frame images; where in the mask images corresponding to the subsequent frames, the target muscle and the muscle background are segmented by the muscle boundaries to determine positions of the target muscle in the subsequent frame images; extract a target muscle contour from the mask images corresponding to respective frame images of the ultrasound image sequence; where the target muscle contour includes a first muscle boundary and a second muscle boundary, a direction from the first muscle boundary pointing to the second muscle boundary is a thickness direction of the target muscle, and the first muscle boundary and the second muscle boundary appear as linear high echoes in the respective frame images; determine, based on the target muscle contour, architectural parameters of the target muscle in the mask images corresponding to the respective frame images; and determine the temporal variation information based on the architectural parameters of the target muscle in the mask images corresponding to the respective frame images.
[0124] Optionally, the first determining module is specifically configured to: acquire the first frame image of the ultrasound image sequence; acquire the muscle boundary in the first frame image of the ultrasound image sequence; and generate the mask image corresponding to the first frame image based on the muscle boundary in the first frame image.
[0125] Optionally, the first determining module is specifically configured to: obtain a first fitted polynomial by performing linear fitting on a boundary point set of the first muscle boundary in the mask images; obtain a second fitted polynomial by performing linear fitting on a boundary point set of the second muscle boundary in the mask images; and determine the architectural parameter of the target muscle based on a vertical distance between the first fitted polynomial and the second fitted polynomial.
[0126] Optionally, the segmentation algorithm includes a long-term video object segmentation model.
[0127] Optionally, the long-term video object segmentation model includes XMem model.
[0128] Optionally, the first determining module is specifically configured to: initialize a human memory model using the initial information; and read memory features of the subsequent frame images via the human memory model, and generating the mask images corresponding to the subsequent frame images, where the memory features include a memory key and a memory value.
[0129] Optionally, the apparatus further includes a processing module, configured to remove burrs and falsely segmented contours within the mask images corresponding to the respective frame images, using a morphological image processing algorithm.
[0130] In an optional embodiment, embodiments of the present disclosure further provide an electronic device, as shown in FIG. 14, the electronic device 800 shown in FIG. 14 may be a server, including: a processor 801 and a memory 803. The processor 801 and the memory 803 are connected, such as through a bus 802. Optionally, the electronic device 800 may further include a transceiver 804. It should be noted that in actual applications, the transceiver 804 is not limited to one, and the structure of the electronic device 800 does not constitute a limitation on the embodiments of the present disclosure.
[0131] Processor 801 may be a CPU (Central Processing Unit) , a general-purpose processor, a DSP (Digital Signal Processor) , an ASIC (Application Specific Integrated Circuit) , an FPGA (Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It can implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the present disclosure. Processor 801 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, DSPs, microprocessors, or the like.
[0132] Bus 802 may include a path to transmit information between the above components. Bus 802 may be PCI (Peripheral Component Interconnect) bus, EISA (Extended Industry Standard Architecture) bus, or the like. Bus 802 may be divided into address bus, data bus, control bus, and the like. For ease of representation, only one thick line is used in FIG. 14, but it does not mean that there is only one bus or one type of bus.
[0133] Memory 803 may be ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, EEPROM (Electrically Erasable Programmable Read Only Memory) , CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, and the like) , magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto.
[0134] Memory 803 is configured to store application code for executing the solution of the present disclosure, and the execution is controlled by the processor 801. The processor 801 is configured to execute the application code stored in the memory 803 to implement the content shown in the above method embodiments.
[0135] The electronic device includes, but is not limited to, mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants) , PADs (tablet computers) , PMPs (portable multimedia players) , vehicle-mounted terminals (such as vehicle-mounted navigation terminals) , or the like, and fixed terminals such as digital TVs, desktop computers, or the like. The electronic device shown in FIG. 14 is only an example and should not bring any limitation to the functions and usage scope of the embodiments of the present disclosure.
[0136] The server provided by the present disclosure may be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The terminal may be a smart phone, tablet computer, laptop computer, desktop computer, smart speaker, smart watch, or the like, but is not limited thereto. The terminal and the server may be directly or indirectly connected via wired or wireless communication, which is not limited by the present disclosure.
[0137] The present disclosure provides a computer-readable storage medium. The computer program is essentially stored therein, and when it is run on a computer, the computer can execute the corresponding contents in the aforementioned method embodiments.
[0138] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times; and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.
[0139] It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus or device, or any combination of the above. More specific examples of computer-readable storage medium may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM) , a read-only memory (ROM) , an erasable programmable read-only memory (EPROM or flash memory) , an optical fiber, a portable compact disk read-only memory (CD-ROM) , an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, apparatus or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency) , or the like, or any suitable combination of the above.
[0140] The computer readable medium may be included in the electronic device; or it may be present separately without being not incorporated into the electronic device.
[0141] The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device executes the method shown in the above embodiments.
[0142] According to one aspect of the present disclosure, a computer program product or a computer program is provided, where the computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, thereby causing the computer device to implement the methods according to the various optional implementations as mentioned above.
[0143] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer; or partially on the user's computer; or as a separate software package; or partially on the user's computer and partially on a remote computer; or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN) , or may be connected to an external computer (e.g., through the Internet using an Internet service provider) .
[0144] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a partial code, and the module, the program segment or the partial code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box may also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.
[0145] The modules involved in the embodiments of the present disclosure can be implemented by software. It can be implemented by hardware. The name of a module does not limit the module itself in some cases. For example, module A can also be described as "module A configured to perform operation B" .
[0146] The above description includes only some preferred embodiments of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other.
[0147] In an optional embodiment, embodiments of the present disclosure further provide storage medium storing an instruction thereon, where the instruction, upon being executed by a processor, is used for implementing the method as described in any of the preceding embodiments.
[0148] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the application disclosed herein. This application is intended to cover any modifications, uses or adaptations of the present disclosure, which follow the general principles of the present disclosure and include common knowledge or customary technical means in the art that are not disclosed in the present disclosure.
[0149] It should be understood that the embodiments of the present disclosure are not limited to the precise structures described above and shown in the drawings, and various modifications and changes may be made without departing from the scope thereof. The scope of the embodiments of the present disclosure is limited only by the appended claims.
Claims
A method for monitoring neuromuscular blockade, comprising:determining, based on an ultrasound image sequence of a target muscle in response to N consecutive electrical stimulations at equal intervals, temporal variation information of an architectural parameter of the target muscle, where N is an integer greater than 1;determining, based on the temporal variation information and in response to N consecutive electrical stimulations at equal intervals, a first change amount in the architectural parameter of the target muscle after an Nth electrical stimulation and a second change amount in the architectural parameter after a first electrical stimulation;determining a train-of-N ratio based on a ratio of the first change amount to the second change amount; anddetermining a degree of neuromuscular blockade based on a value of the train-of-N ratio.The method according to claim 1, wherein the architectural parameter comprises at least one of muscle thickness, cross-sectional area, muscle fiber length, pennation angle, or location deformation.The method according to claim 1, wherein prior to determining the temporal variation information, the method further comprises:acquiring the ultrasound image sequence of the target muscle in response to the N consecutive electrical stimulations on a nerve corresponding to the target muscle; andperforming normalization on the ultrasound image sequence of the target muscle.The method according to claim 1, wherein a sampling rate of the ultrasound image sequence is greater than or equal to 200 frames per second (fps) .The method according to claim 1, wherein determining the temporal variation information comprises:acquiring a first frame image of the ultrasound image sequence and a mask image corresponding to the first frame image as initial information; wherein in the mask image corresponding to the first frame image, the target muscle and muscle background are segmented by a muscle boundary to determine a position of the target muscle in the first frame image;tracking the target muscle in subsequent frame images of the ultrasound image sequence based on the initial information using a segmentation algorithm, and generating mask images corresponding to the subsequent frame images; wherein in the mask images corresponding to the subsequent frames, the target muscle and the muscle background are segmented by the muscle boundaries to determine positions of the target muscle in the subsequent frame images;extracting a target muscle contour from the mask images corresponding to respective frame images of the ultrasound image sequence; wherein the target muscle contour comprises a first muscle boundary and a second muscle boundary, a direction from the first muscle boundary pointing to the second muscle boundary is a thickness direction of the target muscle, and the first muscle boundary and the second muscle boundary appear as linear high echoes in the respective frame images;determining, based on the target muscle contour, architectural parameters of the target muscle in the mask images corresponding to the respective frame images; anddetermining the temporal variation information based on the architectural parameters of the target muscle in the mask images corresponding to the respective frame images.The method according to claim 5, wherein acquiring the first frame image of the ultrasound image sequence and the mask image corresponding to the first frame image comprises:acquiring the first frame image of the ultrasound image sequence;acquiring the muscle boundary in the first frame image of the ultrasound image sequence; andgenerating the mask image corresponding to the first frame image based on the muscle boundary in the first frame image.The method according to claim 5, wherein determining, based on the target muscle contour, the architectural parameters of the target muscle in the mask images corresponding to the respective frame images comprises:obtaining a first fitted polynomial by performing linear fitting on a boundary point set of the first muscle boundary in the mask images;obtaining a second fitted polynomial by performing linear fitting on a boundary point set of the second muscle boundary in the mask images; anddetermining the architectural parameter of the target muscle based on a vertical distance between the first fitted polynomial and the second fitted polynomial.The method according to claim 5, wherein the segmentation algorithm comprises a long-term video object segmentation model.The method according to claim 8, wherein the long-term video object segmentation model comprises XMem model.The method according to claim 9, wherein generating the mask images corresponding to the subsequent frame images comprises:initializing a human memory model using the initial information; andreading memory features of the subsequent frame images via the human memory model, and generating the mask images corresponding to the subsequent frame images, wherein the memory features comprise a memory key and a memory value.The method according to claim 7, wherein prior to obtaining the first fitted polynomial, the method further comprises:removing burrs and falsely segmented contours within the mask images corresponding to the respective frame images, using a morphological image processing algorithm.An apparatus for monitoring neuromuscular blockade, comprising:a first determining module, configured to determine, based on an ultrasound image sequence of a target muscle in response to N consecutive electrical stimulations at equal intervals, temporal variation information of an architectural parameter of the target muscle , where N is an integer greater than 1;a second determining module, configured to determine, based on the temporal variation information and in response to N consecutive electrical stimulations at equal intervals, a first change amount in the architectural parameter of the target muscle after an Nth electrical stimulation and a second change amount in the architectural parameter after a first electrical stimulation;a third determining module, configured to determine a train-of-N ratio based on a ratio of the first change amount to the second change amount; anda fourth determining module, configured to determine a degree of neuromuscular blockade based on a value of the train-of-N ratio.The apparatus according to claim 12, wherein the architectural parameter comprises at least one of muscle thickness, cross-sectional area, muscle fiber length, pennation angle, or location deformation.The apparatus according to claim 12, further comprising:an acquiring module, configured to acquire the ultrasound image sequence of the target muscle in response to the N consecutive electrical stimulations on a nerve corresponding to the target muscle; anda performing module, configured to perform normalization on the ultrasound image sequence of the target muscle.The apparatus according to claim 12, wherein a sampling rate of the ultrasound image sequence is greater than or equal to 200 frames per second (fps) .The apparatus according to claim 12, wherein the first determining module is specifically configured to:acquire a first frame image of the ultrasound image sequence and a mask image corresponding to the first frame image as initial information; wherein in the mask image corresponding to the first frame image, the target muscle and muscle background are segmented by a muscle boundary to determine a position of the target muscle in the first frame image;track the target muscle in subsequent frame images of the ultrasound image sequence based on the initial information using a segmentation algorithm, and generate mask images corresponding to the subsequent frame images; wherein in the mask images corresponding to the subsequent frames, the target muscle and the muscle background are segmented by the muscle boundaries to determine positions of the target muscle in the subsequent frame images;extract a target muscle contour from the mask images corresponding to respective frame images of the ultrasound image sequence; wherein the target muscle contour comprises a first muscle boundary and a second muscle boundary, a direction from the first muscle boundary pointing to the second muscle boundary is a thickness direction of the target muscle, and the first muscle boundary and the second muscle boundary appear as linear high echoes in the respective frame images;determine, based on the target muscle contour, architectural parameters of the target muscle in the mask images corresponding to the respective frame images; anddetermine the temporal variation information based on the architectural parameters of the target muscle in the mask images corresponding to the respective frame images.The apparatus according to claim 16, wherein the first determining module is specifically configured to:acquire the first frame image of the ultrasound image sequence;acquire the muscle boundary in the first frame image of the ultrasound image sequence; andgenerate the mask image corresponding to the first frame image based on the muscle boundary in the first frame image.The apparatus according to claim 16, wherein the first determining module is specifically configured to:obtain a first fitted polynomial by performing linear fitting on a boundary point set of the first muscle boundary in the mask images;obtain a second fitted polynomial by performing linear fitting on a boundary point set of the second muscle boundary in the mask images; anddetermine the architectural parameter of the target muscle based on a vertical distance between the first fitted polynomial and the second fitted polynomial.The apparatus according to claim 16, wherein the segmentation algorithm comprises a long-term video object segmentation model.The apparatus according to claim 19, wherein the long-term video object segmentation model comprises XMem model.The apparatus according to claim 20, wherein the first determining module is specifically configured to:initialize a human memory model using the initial information; andread memory features of the subsequent frame images via the human memory model, and generating the mask images corresponding to the subsequent frame images, wherein the memory features comprise a memory key and a memory value.The apparatus according to claim 18, further comprising:a processing module, configured to remove burrs and falsely segmented contours within the mask images corresponding to the respective frame images, using a morphological image processing algorithm.An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable by the processor, wherein the processor, upon executing the program, is configured to implement the method according to any one of claims 1 to 11.A storage medium storing an instruction thereon, wherein the instruction, upon being executed by a processor, is configured for implementing the method as described in any of claims 1 to 11.
Citation Information
Patent Citations
Monitoring a neuromuscular blockade status
CN105873510A
Neuromuscular sensing device with multi-sensor array
CN111035363A
Muscle training method and system for providing visual feedback through ultrasonic imaging
CN112089442A
Neuromuscular monitoring display system
US20140107524A1
A device and means of assessing neuromuscular junction status with higher fidelity
US20180333569A1