System and method for quantitative evaluation of Parkinson's disease hand movement symptoms

Abstract

The invention provides a system for quantitative evaluation of Parkinson's disease hand motion symptoms, and the system comprises a glove base which is worn on the hand of a patient to be evaluated; the triboelectric sensors are located at finger joints on the back side of the glove base and used for collecting bending signals of the finger joints when the patient executes specified actions; the inertial measurement unit is positioned in a hand back area on the glove base and is used for collecting a movement signal of the whole hand when the patient executes a specified action; the data acquisition and processing unit is used for calculating a tremor evaluation index value and a muscular tension evaluation index value based on the motion signal of the whole hand, and calculating a motion delay evaluation index value based on the bending signal of the finger joint and the motion signal of the whole hand; and inputting the three types of evaluation index values into a pre-constructed mapping relationship between each evaluation index and a clinical score to obtain respective scores, and summing the scores to obtain an overall score. According to the invention, fine actions of the hand of the Parkinson's disease patient can be reflected, and accurate quantitative indexes of hand movement symptoms can be provided.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of sensing devices and medical rehabilitation technology, and particularly relates to a system and method for quantitatively evaluating hand movement symptoms of Parkinson's disease. BACKGROUND

[0002] Parkinson's disease is the second most common neurodegenerative disease in the world, often occurring in the elderly population, and is insidious in onset. Although there are many methods for treating Parkinson's disease that can improve symptoms or delay disease progression, they cannot cure it. Early detection and accurate diagnosis of Parkinson's disease is the primary task and the prerequisite for precise treatment.

[0003] Parkinson's disease is relatively difficult to diagnose and has a long course, and both diagnosis and treatment require motor function assessment. The clinical manifestations of Parkinson's disease include bradykinesia, tremor, muscle rigidity, and postural imbalance. The most authoritative method is the UPDRS-III Unified Parkinson's Disease Rating Scale method (3.1 to 3.17). However, this scale method has subjective and human factors, and the consistency and discriminability between assessors are poor, lacking objectivity and accuracy. In addition, UPDRS has poor sensitivity in the early stages of Parkinson's disease, and it is difficult for the human eye to accurately capture the subtle abnormalities of patient movement. The clinical heterogeneity of Parkinson's disease patients and the difficulty of accurately judging their motor signs increase the risk of misdiagnosis and missed diagnosis. Therefore, accurate evaluation and scoring of motor symptoms play a crucial role in avoiding misdiagnosis and missed diagnosis and early diagnosis of Parkinson's disease, and there is an urgent need for objective and good quantitative intelligent evaluation methods to accurately evaluate patients.

[0004] Wearable flexible sensors have important application value in health monitoring and disease diagnosis. Current flexible sensors mostly use batteries for power supply, which increases the system size and limits flexibility and extensibility. Friction-based sensors are a self-driven sensor that can output a higher electrical signal (volt level) from a small mechanical movement or deformation, facilitating subsequent signal processing. Flexible and stretchable friction-based sensors can better conform to the human skin, and by integrating wearable flexible bending sensors with multi-modal sensors such as inertial measurement units, it can help to comprehensively capture the hand movement characteristics of Parkinson's patients, providing an effective tool for accurately assessing disease progression.

[0005] A Chinese patent with publication number CN104127187A discloses a wearable system and method for quantitatively detecting main symptoms of Parkinson's disease. The system uses an inertial measurement unit located at a fingertip to detect finger movement. This method is prone to mutual influence of tremor signals and bradykinesia signals, and has the problem of signal drift, which cannot accurately reflect the fine motor condition of the patient's hand. A Chinese patent with publication number CN111513724A discloses a wearable device for detecting and evaluating the hand function rehabilitation of Parkinson's patients. The device uses a resistive bending sensor to capture the bending movement of the patient's fingers. This method has a long response time, high power consumption, and cannot well fit the human skin, and has poor integration and practicability.

[0006] Therefore, there is a need for a new wearable system and method to solve the problem of quantitatively evaluating the hand movement symptoms of Parkinson's disease. SUMMARY

[0007] The present application aims to at least partially solve one of the technical problems existing in the prior art.

[0008] To this end, the present application provides a system and method for quantitatively evaluating the hand movement symptoms of Parkinson's disease, which is low in cost, simple in manufacturing process and high in integration, and can reflect the fine action of the patient's hand, and provide accurate quantitative indicators of hand movement symptoms.

[0009] To achieve the above-mentioned purpose, the present application adopts the following technical solutions:

[0010] The present application provides a system for quantitatively evaluating the hand movement symptoms of Parkinson's disease in the first aspect, comprising:

[0011] A glove base worn on the hand of the patient to be evaluated;

[0012] A triboelectric sensor arranged at each finger joint on the back side of the glove base for collecting bending signals of the finger joints when the patient to be evaluated performs a specified action;

[0013] An inertial measurement unit arranged on the back of the hand on the glove base for collecting movement signals of the whole hand when the patient to be evaluated performs a specified action;

[0014] A data acquisition and processing unit for acquiring sensing signals from the triboelectric sensor and the inertial measurement unit, calculating tremor evaluation index values and muscle tension evaluation index values based on the movement signals of the whole hand, calculating bradykinesia evaluation index values based on the bending signals of the finger joints and the movement signals of the whole hand, inputting the three types of evaluation index values calculated respectively into the mapping relationship between each evaluation index and clinical score pre-constructed to obtain respective scores, and summing up the respective scores to obtain an overall score.

[0015] In some embodiments, the triboelectric sensor comprises a first encapsulation layer, a first electrode layer, a first triboelectric layer, a substrate layer, a second electrode layer and a second encapsulation layer arranged in sequence, a cavity is formed at one side of the first triboelectric layer opposite to the substrate layer, a second triboelectric layer is arranged on the sidewall of the substrate layer in the cavity, when the triboelectric sensor is bent with the glove substrate, the first triboelectric layer and the second triboelectric layer are in contact and separated from each other, a changing electromotive force is induced by the first electrode layer and the second electrode layer, and the first electrode layer and the second electrode layer are connected to the data acquisition and processing unit through wires.

[0016] In some embodiments, for tremor symptom evaluation, the specified action to be performed by the patient to be evaluated includes a static tremor evaluation action, a postural tremor evaluation action and a kinetic tremor evaluation action.

[0017] For muscle rigidity symptom evaluation, the specified action to be performed by the patient to be evaluated includes placing the patient to be evaluated in a relaxed state, gently rotating the patient's arm several times to naturally bend and stretch.

[0018] For bradykinesia symptom evaluation, the specified action to be performed by the patient to be evaluated includes a pinching action, a palm turning action and a fist clenching action.

[0019] In some embodiments, the data acquisition and processing unit comprises a signal acquisition module arranged at the center of the back of the glove substrate and a signal processing module arranged separately outside the glove substrate, the signal acquisition module is wiredly connected to the triboelectric sensor and the inertial measurement unit, and the signal acquisition module is wirelessly connected to the signal processing module.

[0020] In some embodiments, the signal acquisition module comprises a triboelectric signal retention circuit, an analog-to-digital converter, a master control chip and a power management circuit arranged on a PCB board, each triboelectric sensor is connected to a corresponding triboelectric signal retention circuit through a wire, the output signal of the triboelectric signal retention circuit is connected to the analog-to-digital converter, the analog signal is converted into a digital signal and then transmitted to the master control chip, and the master control chip is also connected to the inertial measurement unit and the signal processing module.

[0021] The triboelectric signal holding circuit comprises a resistor, a capacitor and an operational amplifier; the resistor is connected between the negative electrode of the triboelectric sensor and the inverting input terminal of the operational amplifier, the capacitor is connected in parallel between the output terminal and the inverting input terminal of the operational amplifier to form a feedback loop, and the non-inverting input terminal of the operational amplifier is connected to the negative electrode of the triboelectric sensor and is grounded together; the output voltage of the triboelectric signal holding circuit is proportional to the output charge of the triboelectric sensor and is inversely proportional to the capacitance in the feedback loop.

[0022] In some embodiments, the mapping relationship between each evaluation index and the clinical score is constructed in any of the following three ways:

[0023] The first way is to use interval judgment method to select a single quantitative index to evaluate the symptoms of Parkinson's patients; wherein the quantitative index used for tremor symptoms adopts the root mean square, standard deviation, kurtosis, skewness, dominant peak frequency, peak power, bandwidth or main frequency band energy proportion of the tremor signal collected when the patient performs a specified action; the quantitative index used for muscle tension enhancement symptoms adopts the root mean square, standard deviation or coefficient of variation of the angular velocity signal collected when the patient performs a specified action; the time evaluation index, amplitude evaluation index, speed evaluation index or overall evaluation index of the three is used for movement retardation symptoms, and the overall evaluation index adopts the movement amplitude-period standard deviation ellipse area index; the Parkinson's patients are divided into discrete groups according to the clinical score, the distribution characteristics of the quantitative index of each group are counted, the interval boundary of the quantitative index is determined, and then a corresponding rule table of the quantitative index interval and the clinical score is established;

[0024] The second way is to use continuous mapping method to select a single quantitative index to evaluate the symptoms of Parkinson's patients; wherein the quantitative index used for tremor symptoms adopts the root mean square, standard deviation, kurtosis, skewness, dominant peak frequency, peak power, bandwidth or main frequency band energy proportion of the tremor signal collected when the patient performs a specified action; the quantitative index used for muscle tension enhancement symptoms adopts the root mean square, standard deviation or coefficient of variation of the angular velocity signal collected when the patient performs a specified action; the time evaluation index, amplitude evaluation index, speed evaluation index or overall evaluation index of the three is used for movement retardation symptoms, and the overall evaluation index adopts the movement amplitude-period standard deviation ellipse area index; the linear correlation between the selected quantitative index and the clinical score is verified using Pearson test or Spearman correlation coefficient, then a regression model is established with the clinical score as the independent variable and the quantitative index as the dependent variable, and the quantitative index is converted into continuous clinical score through the regression model;

[0025] In a third mode, multi-feature fusion is used to extract multi-dimensional features from the sensing signals. For tremor symptoms, at least two combinations of the root mean square, standard deviation, kurtosis, skewness, dominant peak frequency, peak power, bandwidth, and main band energy proportion of the tremor signals collected when the patient performs a specified action are extracted. For muscle rigidity symptoms, at least two combinations of the root mean square, standard deviation, and coefficient of variation of the angular velocity signals collected when the patient performs a specified action are extracted. For bradykinesia symptoms, at least two combinations of the speed evaluation index, amplitude evaluation index, speed evaluation index, and overall evaluation index are extracted. After the features are obtained, a corresponding high-dimensional feature set is constructed, and a mapping model is constructed according to the following steps: 1) first, all features are standardized; subsequently, correlation analysis, principal component analysis, or recursive feature elimination is used to select a key feature subset with the highest correlation with clinical scores and the lowest redundancy from the high-dimensional feature set; 2) the key feature subset is used as an input variable, and the average of the scores of multiple physicians is used as a target variable to train a machine learning model; 3) the performance of the machine learning model is verified, and the parameters of the machine learning model are optimized to obtain a trained machine learning model as the mapping model.

[0026] In some embodiments, the signal processing module quantitatively evaluates the tremor symptoms of the patient to be evaluated, including:

[0027] Step S11, pre-processing the three-axis acceleration signals in the inertial measurement unit collected when the patient to be evaluated performs a specified action for tremor symptom evaluation to obtain a composite acceleration, and taking the composite acceleration as a tremor signal;

[0028] Step S12, calculating a variation index of the tremor signal using a sliding window, the variation index being the root mean square, standard deviation, or variance, identifying the time corresponding to the maximum value of the variation index of the tremor signal in all sliding windows, and extracting a window of a set time length before and after the center as a signal segment range for time domain feature and frequency domain feature calculation;

[0029] For the time domain features of the tremor signal:

[0030] calculating a first quantitative index of the tremor signal in the signal segment, the first quantitative index being a mixture of any one or more of the root mean square RMS, standard deviation , kurtosis Kurtosis, and skewness Skewness;

[0031] For the frequency domain features of the tremor signal:

[0032] calculating a second quantitative index of the tremor signal in the signal segment, the second quantitative index being a mixture of any one or more of the dominant peak frequency , peak power , bandwidth, and main band energy proportion Any one or more of the following; wherein, the main frequency band energy percentage The calculation formula is as follows:

[0033]

[0034] In the formula, , representing the energy of the main frequency band, is selected as the power spectral density integral value in the range of 3 Hz to 10 Hz; , which is the total energy of all frequencies of the flutter signal within the signal segment.

[0035] Step S13: Input the extracted time-domain and / or frequency-domain features of the tremor assessment into the pre-constructed mapping relationship between tremor assessment indicators and clinical scores to obtain a quantitative score for the severity of hand tremor symptoms in Parkinson's disease under different tremor assessment actions of the patient to be assessed.

[0036] In some embodiments, the signal processing module performs a quantitative assessment of the patient's increased muscle tone symptoms, including:

[0037] Step S21: Preprocess the triaxial angular velocity signal in the inertial measurement unit collected when the patient to be evaluated performs the specified action for assessing increased muscle tone symptoms to obtain a synthetic angular velocity, and use the synthetic angular velocity as a muscle tone assessment signal;

[0038] Step S22: Use threshold segmentation to identify and segment each complete flexion or extension movement segment, calculate the quantitative index of muscle tone assessment within each movement segment, the quantitative index is selected from one or more of the root mean square (RMS), standard deviation and coefficient of variation of the muscle tone assessment signal, and all movement segments use the same quantitative index; take the mean of the quantitative index calculated in all movement segments as the final quantitative index of this assessment.

[0039] Step S23: Input the final quantitative index into the pre-constructed mapping relationship between the muscle tone enhancement assessment index and the clinical score to obtain the quantitative score of the severity of hand muscle tone enhancement symptoms of Parkinson's disease under different muscle tone enhancement assessment actions of the patient to be assessed.

[0040] In some embodiments, the signal processing module performs a quantitative assessment of the patient's bradykinesia symptoms, including:

[0041] Step S31, pre-process the sensing signals of the inertial measurement unit and the triboelectric sensor collected during the execution of the specified action for bradykinesia symptom assessment by the patient to be evaluated; wherein, in the pinch action assessment and the fist action assessment, the sensing signal collected by the channel with the largest signal-to-noise ratio in each triboelectric sensor is used, and in the palm turning action assessment, the roll angle in the three-axis attitude angle calculated based on the sensing signal of the inertial measurement unit is used;

[0042] Step S32, period feature extraction is performed on the pre-processed sensing signals:

[0043] Step S321, multi-scale peak extraction is performed using a peak detection algorithm to obtain the peak and valley values of the sensing signal under periodic motion, and the calculated peak and valley values are matched;

[0044] Step S322, the period length and the period amplitude are calculated according to the matched peak and valley values;

[0045] Step S33, the bradykinesia symptom assessment indicators are calculated according to the period length and the period amplitude obtained under different bradykinesia assessment actions; the bradykinesia symptom assessment indicators include any one or a combination of multiple of the time assessment indicator, the amplitude assessment indicator, the speed assessment indicator, and the overall assessment indicator, wherein,

[0046] The time assessment indicator includes the average period time, the period time extreme value, the period time decay ratio, the period time standard deviation, and the period time coefficient of variation;

[0047] The amplitude assessment indicator includes the average amplitude, the amplitude extreme value, the amplitude decay ratio, the amplitude standard deviation, and the amplitude coefficient of variation;

[0048] The speed assessment indicator includes the average speed, the speed extreme value, the speed decay ratio, the speed standard deviation, and the speed coefficient of variation;

[0049] The overall assessment indicator uses the period-amplitude standard deviation ellipse area;

[0050] Step S34, the extracted time assessment indicator, amplitude assessment indicator, speed assessment indicator, and overall assessment indicator calculated under the bradykinesia assessment action are input into the pre-constructed mapping relationship between the bradykinesia assessment indicator and the clinical score to obtain the quantitative score for the severity of the hand bradykinesia symptom of Parkinson's disease for the patient to be evaluated under different bradykinesia assessment actions.

[0051] The second aspect of the present application provides a method for quantitatively assessing the hand movement symptoms of Parkinson's disease based on the above-mentioned system, comprising:

[0052] Step S100, the patient to be evaluated wears the wearable part of the system on the left and right hands respectively;

[0053] Step S200, the patient to be evaluated makes an upper limb movement ability evaluation action, including tremor evaluation action, movement retardation evaluation action and muscle tension evaluation action; real-time sensing signals output by the triboelectric sensor and the inertial measurement unit under different evaluation actions are collected;

[0054] Step S300, feature extraction and calculation are performed on the sensing signals, and corresponding symptom evaluation index values are converted;

[0055] Step S400, according to each symptom evaluation index, an overall score of the hand movement symptom of the patient to be evaluated is output.

[0056] The system and method for quantitatively evaluating the hand movement symptom of Parkinson's disease provided by the application have the following characteristics and beneficial effects:

[0057] 1. The system integrates a five-channel triboelectric sensor and an inertial measurement unit (IMU) to form an optimized combination for hand evaluation. The triboelectric sensor captures fine bending movements of the five fingers with high sensitivity and realizes stable measurement of sensing signals through a triboelectric signal holding circuit, which can comprehensively measure the bending amplitude and speed of the patient's fingers; the IMU accurately senses the three-dimensional posture, orientation change, tremor amplitude and frequency of the whole hand. The dual-mode signals are synchronously collected, covering the core dimensions of hand movement (fine action and overall posture / tremor).

[0058] 2. The triboelectric sensor, the IMU and the signal transmission module are integrated in the glove substrate to form an integrated wearable device. This design is non-invasive and light, allowing the patient to perform natural hand-specific actions (such as tremor evaluation, movement retardation evaluation and muscle tension evaluation actions) during evaluation, reducing the mutual interference of tremor signals and movement retardation signals.

[0059] 3. The method clearly requires the patient to perform standardized evaluation actions for core symptoms such as tremor, movement retardation and muscle tension enhancement. Based on this, the system uses a feature extraction method supported by clinical expert knowledge to calculate quantitatively evaluate indicators highly related to specific symptoms from dual-mode signals, realizing quantitative evaluation of each movement feature of the patient's hand, which is used to measure the severity of the patient's condition and monitor its progress. DETAILED DESCRIPTION

[0060] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0061] Figure 1is a whole structure diagram of a system for quantitatively evaluating hand movement symptoms of Parkinson's disease provided by an embodiment of the present application;

[0062] Figure 2 is a structure schematic diagram of a triboelectric sensor provided by an embodiment of the present application;

[0063] Figure 3 is a structure schematic diagram of a signal acquisition circuit board provided by an embodiment of the present application;

[0064] Figure 4 is a principle diagram of a triboelectric signal holding circuit provided by an embodiment of the present application;

[0065] Figure 5 is a flow chart of tremor evaluation by a signal processing module provided by an embodiment of the present application;

[0066] Figure 6 is a flow chart of muscle tension evaluation by a signal processing module provided by an embodiment of the present application;

[0067] Figure 7 is a flow chart of movement retardation evaluation by a signal processing module provided by an embodiment of the present application;

[0068] Figure 8 is a flow chart of a method for quantitatively evaluating hand movement symptoms of Parkinson's disease provided by an embodiment of the present application;

[0069] Figure 9 is a validity verification result diagram provided by an embodiment of the present application, wherein (a)~(e) are respectively turnover motion, fist clenching motion, pinching motion, static tremor evaluation action and postural tremor evaluation action;

[0070] In the figure: 1, glove base; 2, triboelectric sensor; 201, first friction layer; 202, second friction layer; 203, base layer; 204, second electrode layer; 205, first electrode layer; 206, first packaging layer; 207, second packaging layer; 3, inertial measurement unit; 4, data acquisition and processing unit; 41, signal acquisition module; 411, triboelectric signal holding circuit; 412, analog-to-digital converter; 413, main control chip; 414, power management circuit; 42, signal processing module. DETAILED DESCRIPTION

[0071] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application is further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0072] In contrast, the present application encompasses any alternatives, modifications, equivalent methods and solutions defined by the claims that are made on the essence and scope of the present application. Further, in order to make the public better understand the present application, in the following detailed description of the present application, some specific details are described in detail. The present application can also be fully understood without the description of these details by those skilled in the art.

[0073] The present application provides a system and method for quantitatively evaluating hand movement symptoms of Parkinson's disease, which are described in detail in combination with the drawings and specific embodiments as follows:

[0074] As shown in Figure 1 , the first aspect of the present application provides a system for quantitatively evaluating hand movement symptoms of Parkinson's disease, which comprises:

[0075] A glove base 1 is worn on the hand of a patient to be evaluated;

[0076] A triboelectric sensor 2 is arranged at each finger joint on the back side of the glove base 1, which is used to collect the bending signal of the finger joint with high sensitivity when the patient to be evaluated performs a specified action;

[0077] An inertial measurement unit 3 is arranged on the back of the glove base 1, which is used to collect the motion signal of the whole hand when the patient to be evaluated performs a specified action;

[0078] A data acquisition and processing unit 4 is used to collect the sensing signals from the triboelectric sensor 2 and the inertial measurement unit 3; calculate the tremor evaluation index value and the muscle tension evaluation index value based on the motion signal of the whole hand; calculate the movement retardation evaluation index value based on the bending signal of the finger joint and the motion signal of the whole hand; input the three types of evaluation index values calculated respectively into the mapping relationship between each evaluation index and the clinical score pre-constructed to obtain the respective score, and sum up the overall score.

[0079] Optionally, the glove base 1 is made of nylon, polyester and the like.

[0080] Optionally, referring to Figure 1 , Figure 2 , 5 triboelectric sensors 2 are respectively fixed at the 5 finger joints on the back side of the glove base 1 (the back side of the glove base is on the same side as the back side of the human hand) to capture the bending motion of each finger, and are connected to the data acquisition and processing unit 4 through wires.

[0081] Further, the structure of each triboelectric sensor 2 is the same and independent of each other, and one of them will be described in detail. The triboelectric sensor 2 includes a first encapsulation layer 206, a first electrode layer 205, a first friction layer 201, a substrate layer 203, a second electrode layer 204 and a second encapsulation layer 207 which are sequentially stacked, a cavity is formed at the side of the first friction layer 201 and the substrate layer 203, and a second friction layer 202 is arranged on the sidewall of the substrate layer 203 in the cavity. The substrate layer 203 can be made of flexible material, the first friction layer 201 and the second friction layer 202 are made of materials with different electron loss and gain capabilities, the first electrode layer 205 and the second electrode layer 204 can be made of the same or different conductive materials (preferably flexible conductive materials), and the first encapsulation layer 206 and the second encapsulation layer 207 can be made of the same or different flexible materials. When the triboelectric sensor 2 is bent with the glove substrate 1, the first friction layer 201 and the second friction layer 202 contact and separate each other, the first electrode layer 205 and the second electrode layer 204 act as electrodes to induce a changing electromotive force, and the first electrode layer 205 and the second electrode layer 204 are connected with wires to output the electrical signal of the finger joint bending movement in real time. In an embodiment of the present application, five triboelectric sensors 2 are provided, wherein the substrate layer 203, the first friction layer 201, the first encapsulation layer 206 and the second encapsulation layer 207 are all made of flexible silicone rubber material, the second friction layer 202 is made of glass fiber cloth, and the first electrode layer 205 and the second electrode layer 204 are both made of silicone rubber mixed with multi-walled carbon nanotubes. In a specific embodiment of the present application, the triboelectric sensor has a total length of 30 mm, a width of 12 mm and a thickness of 3.2 mm; wherein the substrate layer 203 made of flexible silicone rubber has a thickness of 1.1 mm at both ends and a thickness of 0.5 mm at the cavity; the first electrode layer 205 and the second electrode layer 204 made of flexible conductive silicone rubber have a thickness of 0.3 mm; and the first encapsulation layer 206 and the second encapsulation layer 207 made of flexible silicone rubber have a thickness of 0.2 mm.

[0082] Optionally, the inertial measurement unit 3 selects a commercially available inertial measurement unit IMU to collect the motion signals of the whole hand, specifically including 3-axis acceleration, 3-axis angular velocity and 3-axis magnetometer data. In a specific embodiment of the present application, the MPU9250 chip is selected.

[0083] Optionally, referring to Figure 1The data acquisition and processing unit 4 comprises a signal acquisition module 41 and a signal processing module 42, wherein the signal acquisition module 41 is arranged at the central part of the back of the glove base 1 for acquiring the sensing signals from the triboelectric sensor 2 and the inertial measurement unit 3, and transmitting the sensing signals to the signal processing module 42 arranged separately from the glove base 1 through wireless transmission, and then the signal processing module 42 performs subsequent sensing signal feature extraction and evaluation index calculation. Further, in order to improve the integration of the system, the inertial measurement unit 3 is integrated with the signal acquisition module 41, specifically, the inertial measurement unit 3 is integrated into the PCB board of the signal acquisition module 41; and the signal processing module 42 is arranged outside the glove base 1, that is, the glove base 1, the triboelectric sensor 2, the inertial measurement unit 3 and the signal acquisition module 41 together constitute the sensing glove in the system.

[0084] Optionally, as shown in Figure 3 , the signal acquisition module 41 comprises a triboelectric signal holding circuit 411, an analog-to-digital converter 412, a master control chip 413 and a power management circuit 414 arranged on the PCB board. Each triboelectric sensor 2 is connected to a corresponding triboelectric signal holding circuit 411 through a wire, the output signal of the triboelectric signal holding circuit 411 is connected to the analog-to-digital converter 412 to convert the analog signal into a digital signal, and then transmitted to the master control chip 413. The inertial measurement unit 3 is directly connected to the master control chip 413. After receiving the sensing signals from the analog-to-digital converter 412 and the inertial measurement unit 3, the master control chip 413 performs wireless transmission through a wireless transceiver. The signal acquisition module 41 further comprises a power management circuit 414 for powering the above circuits. In an embodiment of the present application, the analog-to-digital converter 412 preferably uses ADS131M04 chip, and the master control chip 413 preferably uses ESP32 chip, and wireless signal transmission is performed through Wi-Fi.

[0085] Optionally, as shown in Figure 4 , the triboelectric signal holding circuit 411 adopts the charge measurement principle, and the triboelectric signal holding circuit 411 uses the "virtual ground" characteristic of the operational amplifier to directly accumulate the charge generated by the triboelectric sensor 2 into the integration capacitor, preventing the loss of triboelectric charge. Specifically, the triboelectric signal holding circuit 411 comprises a resistor R, a capacitor C and an operational amplifier, the supply voltage of the operational amplifier is V C , the resistor R is connected between the negative electrode of the triboelectric sensor 2 and the inverting input terminal of the operational amplifier, the capacitor C is connected in parallel between the output terminal and the inverting input terminal of the operational amplifier to form a feedback loop, and the positive input terminal of the operational amplifier is connected to the negative electrode of the triboelectric sensor 2 and is commonly grounded. Let represent the impedance of the feedback loop, represent the internal impedance of the triboelectric sensor, The voltage V represents the voltage between the two electrodes of the triboelectric sensor, and the output voltage V of the triboelectric signal holding circuit 411 is... out It can be represented as:

[0086]

[0087] in, The output charge of triboelectric sensor 2, This refers to the internal capacitance of the triboelectric sensor 2. Considering... It is approximately 1 GΩ, which is significantly greater than the resistance R. Therefore, we can conclude that:

[0088]

[0089] Where C is the capacitor in the feedback loop. Therefore, the output voltage V of the triboelectric signal holding circuit 411 is... out With the output charge Q of triboelectric sensor 2 s It is directly proportional to the capacitance C in the feedback loop and inversely proportional to the capacitance C in the feedback loop, which can effectively measure the sensing signal of the triboelectric sensor 2.

[0090] In one specific embodiment of this application, the triboelectric signal holding circuit 411 preferably uses a precision operational amplifier LMP2234 to form a charge amplifier.

[0091] Optionally, the signal processing module 42 is responsible for extracting features from the sensor signal and calculating three types of evaluation indicators. Before the signal processing module 42 works, the patient needs to wear the sensor gloves of this embodiment and perform the specified actions with the left and right hands respectively.

[0092] For tremor symptom assessment, the designated actions that the patient to be assessed must perform include resting tremor assessment actions, postural tremor assessment actions, and action tremor assessment actions, among which:

[0093] Resting tremor assessment procedure: The patient to be assessed sits still in a chair for 10 seconds with both hands resting on the armrests of the chair and both feet comfortably on the floor. Observe the amplitude of the patient's hand tremor.

[0094] Postural tremor assessment maneuver: Instruct the patient to be assessed to extend their arms straight in front of their body with palms down, wrists straight and fingers spread apart without touching adjacent fingers, and hold this position for 10 seconds;

[0095] Action tremor assessment exercise: Instruct the patient to perform the action of repeatedly touching their nose with their finger and then extending it back. Specifically, starting with the arm extended, have the patient touch their nose with their index finger and then return to the extended arm position. This constitutes one repetition. Have the patient perform this exercise at least three times. When the patient's arm is extended, their fingers should be extended as far as possible to touch the assessor's fingertips. This action should be performed slowly to facilitate observation for any tremor.

[0096] For muscle tension symptom evaluation, the patient to be evaluated is in a relaxed state, and the patient's arm to be evaluated is rotated several times (usually three times) to naturally bend and stretch.

[0097] For bradykinesia symptom evaluation, the specified actions to be performed by the patient to be evaluated include finger pinching, palm turning, and fist clenching, wherein:

[0098] Finger pinching: the patient's thumb and index finger are opened as much as possible, and 10 taps are performed at the fastest speed;

[0099] Palm turning: the patient's palm is downward and the arm is straight in front of the body, and the palm is continuously turned upward and downward 10 times at the fastest speed;

[0100] Fist clenching: the patient's hand is clenched while the elbow is bent with the palm facing the tester, and the palm is opened as much as possible and then clenched at the fastest speed, and the palm clenching and opening is performed 10 times in succession.

[0101] In some embodiments, the working steps of the signal processing module 42 are as follows:

[0102] Step S10, receiving the sensing signals of the triboelectric sensor 2 and the inertial measurement unit 3 transmitted by the signal acquisition module 41, wherein the output signal size of the triboelectric sensor 2 represents the degree of finger bending; the sensing signals of the inertial measurement unit 3 include three-axis acceleration, three-axis angular velocity, and three-axis magnetic force, and the signal processing module 42 obtains three-axis attitude angles according to the sensing signals of the inertial measurement unit 3 (the specific solving process uses existing technical means in the art, which is not described here), the three-axis attitude angles are used to reflect the attitude direction of the whole hand, the three-axis angular velocity is used to reflect the rotation speed of the whole hand, and the three-axis acceleration is used to reflect the tremor amplitude of the whole hand.

[0103] Step S20, according to the different characteristics of tremor symptoms, muscle tension symptoms and bradykinesia symptoms, the motion signals of the patient's whole hand are calculated by a feature extraction method supported by clinical expert knowledge to obtain tremor evaluation index values and muscle tension evaluation index values, and the bending signals of the patient's finger joints and the motion signals of the whole hand are calculated by a feature extraction method supported by clinical expert knowledge to obtain bradykinesia evaluation index values.

[0104] Step S30, inputting each type of evaluation index value into the mapping relationship between each evaluation index and clinical score pre-constructed to obtain respective scores, and summing the scores to obtain an overall score.

[0105] Further, the mapping relationship between each evaluation index and clinical score is pre-constructed in the signal processing module 42. By collecting a large number of sensing signals of patients with Parkinson's disease of different severities when performing evaluation actions, extracting the features of the sensing signals, calculating the motion evaluation indexes, and at the same time, by a plurality of experienced physicians performing clinical scoring and taking the average value as the scoring "true value", the mapping relationship between each evaluation index and clinical score is established. More specifically, the mapping relationship between each evaluation index and clinical score is constructed by using any one of the following three methods:

[0106] Method one, interval judgment method is used to select a single quantitative index to evaluate the symptoms of patients with Parkinson's disease. The quantitative index used for tremor symptoms is the root mean square, standard deviation, kurtosis, skewness, dominant peak frequency, peak power, bandwidth or main band energy proportion of the tremor signal collected when the patient performs the specified action. The quantitative index used for muscle tension enhancement symptoms is the root mean square, standard deviation or coefficient of variation of the angular velocity signal collected when the patient performs the specified evaluation action. The time evaluation index, amplitude evaluation index, speed evaluation index or overall evaluation index of the motion retardation symptom is used, and the overall evaluation index is preferably used, which can specifically use the motion amplitude-period standard deviation ellipse area index. According to the clinical score, the patients with Parkinson's disease are divided into discrete groups (such as 0-4 points corresponding to 5 groups), the distribution characteristics (distribution characteristics can be measured by mean ± standard deviation) of the quantitative index of each group are counted, the interval boundary of the quantitative index is determined, and then the corresponding rule table of the quantitative index interval and the clinical score is established.

[0107] Method two, continuous mapping method is used to select a single quantitative index to evaluate the symptoms of patients with Parkinson's disease. The quantitative index used for tremor symptoms is the root mean square, standard deviation, kurtosis, skewness, dominant peak frequency, peak power, bandwidth or main band energy proportion of the tremor signal collected when the patient performs the specified evaluation action. The quantitative index used for muscle tension enhancement symptoms is the root mean square, standard deviation or coefficient of variation of the angular velocity signal collected when the patient performs the specified action. The time evaluation index, amplitude evaluation index, speed evaluation index or overall evaluation index of the motion retardation symptom is used, and the overall evaluation index is preferably used, which can specifically use the motion amplitude-period standard deviation ellipse area index. The linear correlation between the selected quantitative index and the clinical score is verified by using Pearson test or Spearman correlation coefficient, then the regression model is established by taking the clinical score as the independent variable X and the quantitative index as the dependent variable Y, and the goodness of fit of the regression model is evaluated. Thus, the quantitative index can be converted into continuous clinical score through the regression model.

[0108] Thirdly, multi-feature fusion method is used to extract multi-dimensional features from the sensing signals. For tremor symptoms, at least two combinations of the following features of the tremor signals collected from the patient performing the specified action are extracted: root mean square, standard deviation, kurtosis, skewness, dominant peak frequency, peak power, bandwidth, and main band energy proportion; for hypertonia symptoms, at least two combinations of the following features of the angular velocity signals collected from the patient performing the specified action are extracted: root mean square, standard deviation, and coefficient of variation; for bradykinesia symptoms, at least two combinations of the following evaluation indexes are extracted: speed evaluation index, amplitude evaluation index, speed evaluation index, and overall evaluation index. After obtaining the above features, a mapping model is constructed according to the following steps: 1) first, all features are standardized to eliminate the influence of dimensions; then, correlation analysis, principal component analysis, or recursive feature elimination is used to select a key feature subset with the highest correlation with clinical scores and the lowest redundancy from the high-dimensional feature set; 2) the selected key feature subset is used as the input variable, and the average value of the scores of multiple experienced physicians is used as the target variable to train a machine learning model, which can include but is not limited to support vector regression, random forest regression, gradient boosting decision tree, or artificial neural network; 3) the performance of the machine learning model is evaluated by cross-validation method, and the root mean square error, mean absolute error, and determination coefficient are used as the main evaluation indexes. The parameters of the machine learning model are optimized to prevent overfitting and improve the generalization ability, and the trained machine learning model is used as the mapping model.

[0109] In some embodiments, referring to the accompanying drawings Figure 5 The data processing module 42 evaluates the tremor symptoms of the patient to be evaluated based on the motion signals of the entire hand collected by the data acquisition module 41 when the patient to be evaluated performs the tremor evaluation action, specifically including:

[0110] Step S11, the collected sensing signals of the inertial measurement unit 3 are preprocessed, mainly using the three-axis accelerometer in the inertial measurement unit 3 to capture the micro-motions of the entire hand. The preprocessing includes: using resampling to adjust the sensing signal data to a uniform sampling rate of 100 Hz, and preferably using linear interpolation method to process data missing values; then filtering the resampled data using adjacent average filtering method; based on the filtered data, synthesizing three-axis acceleration to measure the amplitude of each axis tremor, obtaining the original synthesized acceleration a0, as shown in formula (1-1):

[0111]

[0112] wherein, , , represent the components of the three-axis acceleration signals in the x, y, and z axes of the inertial measurement unit, respectively.

[0113] The original synthesized acceleration a0 is band-pass filtered, preferably using a band-pass filter of 2-15 Hz, to remove low-frequency motion artifacts and high-frequency noise, to obtain a synthesized acceleration a, which is taken as the tremor signal.

[0114] In step S12, a variation index of the tremor signal is calculated using a sliding window, which can be root mean square (RMS), standard deviation, or variance, preferably with a sliding window length of 2 s and an overlap rate of 50%. The time point corresponding to the maximum value of the variation index of the tremor signal in all sliding windows (i.e., the time point at which the tremor amplitude is the largest) is identified, and a window of 10 s before and after this time point is extracted as the signal segment range for time-domain feature and frequency-domain feature calculation. This ensures that the analysis focuses on the most significant period of tremor, improving the sensitivity and representativeness of the features.

[0115] For the time-domain features of the tremor signal:

[0116] The first quantitative index of the tremor signal in the extracted signal segment is calculated, which can be any one or a combination of root mean square (RMS), standard deviation , kurtosis (Kurtosis), and skewness (Skewness). The calculation formulas of root mean square (RMS), standard deviation , kurtosis (Kurtosis), and skewness (Skewness) are as follows:

[0117]

[0118]

[0119]

[0120]

[0121] where n is the number of signal points extracted in the signal segment, is the tremor signal of the i-th signal point in the extracted signal segment; is the mean value of the tremor signal in the extracted signal segment; kurtosis (Kurtosis) is used to describe the difference between the distribution of the tremor signal in the extracted signal segment and the normal distribution; skewness (Skewness) is used to describe the asymmetry of the tremor signal distribution in the extracted signal segment, reflecting the skew direction of the waveform.

[0122] For the frequency-domain features of the tremor signal:

[0123] The second quantitative index of the tremor signal in the extracted signal segment is calculated, which can be the dominant peak frequency , peak power , bandwidth, and main band energy ratio Any one or more of the following. The Welch method is used to calculate the power spectral density of the flutter signal. , Let be the frequency of the tremor signal, and take . The frequency corresponding to the maximum value is the dominant peak frequency. Dominant peak frequency The dominant oscillation frequency represents the tremor signal. In Parkinson's patients, the tremor frequency is usually relatively stable, concentrated in the range of 4 Hz to 6 Hz. Peak power. Dominant peak frequency The power spectral density at a given frequency reflects the intensity of the flutter signal at its main frequency, with peak power being the highest. The higher the value, the stronger the tremor signal. Bandwidth is defined as the distance from the peak power of the tremor signal to its maximum power. half( The frequency range at which the peak power is measured is taken in this embodiment. The frequency width at -3 dB indicates that the tremor signal is frequency-stable; tremor signals from Parkinson's patients typically exhibit narrow bandwidth characteristics. (Dominant band energy percentage) The calculation formula is as follows:

[0124]

[0125] In the formula, , representing the energy of the main frequency band, preferably the power spectral density integral value in the range of 3Hz to 10Hz; , which is the total energy of all frequencies of the flutter signal within the extracted signal segment.

[0126] Step S13: Input the extracted time-domain and / or frequency-domain features of the tremor assessment into the pre-constructed mapping relationship between tremor assessment indicators and clinical scores to obtain a quantitative score for the severity of hand tremor symptoms in Parkinson's disease under different tremor assessment actions of the patient to be assessed.

[0127] In some embodiments, see Appendix Figure 6 The data processing module 42 assesses the increased muscle tone symptoms of the patient based on the overall hand movement signals acquired by the patient during the muscle tone assessment action transmitted by the data acquisition module 41. Specifically, this includes:

[0128] Step S21, the collected inertial measurement unit 3 sensor signal is preprocessed, mainly using the three-axis angular velocity in the inertial measurement unit 3 to capture the motion of the whole hand. The preprocessing includes: using resampling to adjust the sensor signal data to a uniform sampling rate of 100 Hz, and preferably using linear interpolation method to process data missing values; based on the resampled data, the three-axis angular velocity is synthesized to measure the rotation speed of the whole arm, and the original synthesized angular velocity is obtained As shown in formula (1-2):

[0129]

[0130] Among them, 、 、 respectively represent the components of the three-axis angular velocity signal in the inertial measurement unit in the x, y and z axes.

[0131] The original synthesized angular velocity is low-pass filtered, preferably using a 5 Hz low-pass filter, to remove high-frequency noise and retain the incoordination caused by rigidity in the patient to be evaluated in the pulling motion, and the synthesized angular velocity is obtained, which is used as the muscle tone evaluation signal.

[0132] Step S22, using threshold segmentation method to identify and segment each complete flexion or extension motion segment. Preferably, 10% of the maximum value of the synthesized angular velocity is used as the threshold value, the continuous region exceeding the threshold value is found, and the motion segment is extracted. In the provided embodiment, three arm rotations have three flexion motion segments and three extension motion segments. Calculate the quantitative indicators of muscle tone evaluation in each motion segment, and the quantitative indicators are selected from one or more of the mixed root mean square (RMS), standard deviation and coefficient of variation. All motion segments use the same quantitative indicators. Finally, the average of the quantitative indicators calculated in all motion segments is taken as the final quantitative indicator of the evaluation.

[0133] Step S23, the final quantitative indicator calculated under the extracted muscle tone evaluation action is input into the pre-constructed mapping relationship between the muscle tone evaluation indicator and the clinical score, and the quantitative score of the patient to be evaluated for the severity of the muscle tone enhancement symptom of Parkinson's disease under the muscle tone evaluation action can be obtained.

[0134] In some embodiments, referring to the attached Figure 7 , the data processing module 42 evaluates the bradykinesia symptom of the patient to be evaluated based on the bending signal of the finger joint and the motion signal of the whole hand collected by the data acquisition module 41 when the patient to be evaluated performs the bradykinesia evaluation action, specifically including:

[0135] Step S31, preprocessing the collected sensing signals: using resampling to adjust the sensing signal data to a uniform sampling rate of 100 Hz, preferably using linear interpolation method to handle missing values. Among them, the sensing signals of the triboelectric sensor 2 are used in the pinch action evaluation and fist action evaluation to characterize the degree of finger bending; the roll angle data in the calculated three-axis attitude angle are used in the palm turning action evaluation. Calculate the signal-to-noise ratio of each channel of the triboelectric sensor 2, and use the channel with the highest signal-to-noise ratio for subsequent analysis. The calculation method of the signal-to-noise ratio of each channel is shown in formula (7):

[0136]

[0137] wherein SNR is the signal-to-noise ratio, is the average power of the sensing signal of a certain channel in the triboelectric sensor 2, is the average power of the noise of a certain channel in the triboelectric sensor 2.

[0138] Subsequently, the roll angle data and the sensing signals of the selected channels of the triboelectric sensor 2 are subjected to low-pass filtering, preferably using a low-pass filter of 15 Hz to remove high-frequency noise; the roll angle data and the sensing signals of the selected channels of the triboelectric sensor 2 after low-pass filtering are subjected to smoothing processing, preferably using an 80 ms window to average the sliding window smoothing of the sensing signals, to obtain the preprocessed sensing signals.

[0139] Step S32, periodic feature extraction of sensing signals

[0140] Step S321, peak extraction and pairing, which needs to be performed for each sensing signal collected for different movement retardation evaluation actions: using the peak detection algorithm AMPD (Asymmetric Maximum Peak Detection) to perform multi-scale peak extraction to obtain the peak and valley values of the sensing signals under periodic motion, and matching the calculated peak and valley values to ensure the integrity of the period.

[0141] Step S322, calculating the period length and period amplitude according to the matched peak and valley values: let the time points of two adjacent valley values of the sensing signal be and , the corresponding valley values be and , and the peak value be p, then the calculation expressions of the period length T and the period amplitude A are shown in formula (8) and formula (9) respectively:

[0142]

[0143]

[0144] Step S33, according to different movement retardation evaluation action, the period length and the period amplitude obtained under the action, the movement retardation symptom evaluation index is calculated. The movement retardation symptom evaluation index includes one or more combinations of time evaluation index, amplitude evaluation index, speed evaluation index and overall evaluation index. Among them,

[0145] The time evaluation index includes: average cycle time, cycle time extreme value, cycle time attenuation ratio (the ratio of total time of the last 5 movement cycles to the total time of the first 5 movement cycles), cycle time standard deviation and cycle time coefficient of variation (the ratio of cycle time standard deviation to average cycle time).

[0146] The amplitude evaluation index includes: average amplitude, amplitude extreme value, amplitude attenuation ratio (the ratio of the average amplitude of the last 5 movements to the average amplitude of the first 5 movements), amplitude standard deviation and amplitude coefficient of variation (the ratio of amplitude standard deviation to average amplitude).

[0147] The speed feature is calculated from the amplitude and cycle time (i.e. speed = amplitude / cycle time), and the speed evaluation index includes: average speed, speed extreme value, speed attenuation ratio (the ratio of the average speed of the last 5 movements to the average speed of the first 5 movements), speed standard deviation and speed coefficient of variation (the ratio of speed standard deviation to average speed).

[0148] The overall evaluation uses the cycle-amplitude standard deviation ellipse area, the larger the area, the greater the movement retardation evaluation index value, the more retarded the movement, and the calculation method is shown in formula (10):

[0149]

[0150] Where S is the standard deviation ellipse area calculated, is the standard deviation of the cycle length, is the standard deviation of the cycle amplitude.

[0151] Step S34, the extracted time evaluation index, amplitude evaluation index, speed evaluation index and overall evaluation index calculated under the movement retardation evaluation action are input into the mapping relationship between the movement retardation evaluation index and the clinical score obtained by a large number of clinical data, and the quantitative score of the patient to be evaluated under different movement retardation evaluation actions for the severity of hand movement retardation symptoms of Parkinson's disease can be obtained.

[0152] The signal processing module 42 of the present application needs to calculate the comprehensive score after obtaining the above three types of evaluation indexes. Specifically, the quantitative scores obtained under each evaluation action are added, i.e. the comprehensive score of the patient's hand movement symptoms is obtained.

[0153] The embodiment of the second aspect of the present application provides a method for quantitatively evaluating hand movement symptoms of Parkinson's disease, referring to the accompanying Figure 8 , comprising the following steps:

[0154] Step S100, the patient to be evaluated wears the wearable part in the system provided by the first aspect of the present application on the left and right hands respectively;

[0155] Step S200, the patient to be evaluated performs an upper limb movement ability evaluation action, including tremor evaluation action, movement retardation evaluation action and muscle tension evaluation action; real-time gesture signals output by the triboelectric sensor 2 and the inertial measurement unit 3 under different evaluation actions are collected by the wearable part, and are sent to the signal processing module 42 through the signal transmission module 41;

[0156] Step S300, the signal processing module 42 extracts and calculates the features of the received sensing signals, and converts them into corresponding symptom evaluation indexes;

[0157] Step S400, according to the symptom evaluation indexes, the overall score of the hand movement symptoms of the patient to be evaluated is output.

[0158] It should be noted that the above explanation of the embodiment of the system for quantitatively evaluating hand movement symptoms of Parkinson's disease is also applicable to the method for quantitatively evaluating hand movement symptoms of Parkinson's disease of the present embodiment, which will not be described here in detail.

[0159] The present application will be further described below in combination with specific embodiments.

[0160] Embodiment 1:

[0161] The triboelectric sensor 2 is prepared and the signal acquisition circuit board is prepared, and the sensing glove is built. The patient to be evaluated sits on a chair, and wears the sensing glove on the left and right hands, respectively, and performs the static tremor evaluation action, the postural tremor evaluation action, the kinetic tremor evaluation action, the pinching action, the palm turning action, the fist clenching action and the muscle tension evaluation action.

[0162] The real-time gesture signals output by the triboelectric sensor 2 and the inertial measurement unit 3 under different gesture actions are collected by the sensing glove, and are sent to the host computer as the signal processing module through the circuit board as the signal transmission module. The sensing signals are preprocessed on the host computer, and then according to the clinical expert knowledge, the features are extracted and the corresponding evaluation indexes are calculated for different evaluation actions. The static tremor evaluation index, the postural tremor evaluation index, the kinetic tremor evaluation index, the pinching action evaluation index, the palm turning action evaluation index, the fist clenching action evaluation index and the muscle tension evaluation index are calculated respectively.

[0163] Based on the above calculated evaluation indicators, the clinical score is converted through the pre-established mapping relationship (reference UPDRS part III score standard). The mapping relationship is constructed by the following method:

[0164] 1. Data collection: 100 patients with different severity of Parkinson's disease (UPDRS score range 0-4) were collected when performing the same evaluation action, the features were extracted and the evaluation indicators were calculated, and at the same time, 3 experienced neurologists independently performed clinical scoring, and the average value was taken as the scoring true value.

[0165] 2. Mapping relationship establishment (based on method three): a nonlinear mapping between evaluation indicators and clinical scores is established using multi-feature fusion method. Specifically, the following procedures are included:

[0166] a) Feature vector construction: correlation analysis is used to remove highly correlated features between features (Pearson correlation coefficient > 0.9 is considered highly correlated), and then recursive feature elimination method is used to select the key feature subset with the most relevant clinical score and the lowest redundancy from the feature set.

[0167] b) Data preprocessing: Z-score standardization is used to eliminate the influence of different index dimensions and numerical ranges.

[0168] c) Model selection and training: for example, support vector regression model (SVR) is used as the fusion model. The core of this model is to find a nonlinear mapping function that can optimally map the input feature vector in high-dimensional feature space to continuous clinical score value. During training, the collected patient data is divided into training set and test set in the ratio of 8:2, and the training set data is used to train the SVR model, with the average value of physician score as the target value, and the kernel function type is selected as radial basis function (RBF). Through grid search and other conventional optimization techniques, the hyperparameters of the model are determined.

[0169] d) Model verification: the performance of the trained model is verified using test set data. The determination coefficient R² between the model predicted score and the average value of physician score is calculated to ensure the prediction accuracy of the model. For example, the model is required to have a determination coefficient greater than 0.85 on the test set to meet the reliability requirements of clinical application.

[0170] 3. Score calculation and report generation: input the evaluation indexes of the patient to be evaluated calculated in this embodiment into the mapping model to obtain the resting tremor score, postural tremor score, kinetic tremor score, pinching action score, palm turning action score, fist clenching action score, and muscle tone score, with the score range being 0-4 (0 normal, 4 severe); add the above scores to calculate the overall quantified score of the hand movement symptoms. Finally, a comprehensive evaluation report of the hand movement symptoms of the patient with Parkinson's disease is generated, including the sub-item scores, overall score, and textual description based on the severity of the score (such as "mild tremor" and "moderate movement retardation"), to assist the physician in diagnosis and efficacy evaluation.

[0171] Validation of effectiveness of the embodiment of the present application:

[0172] This validation example includes 13 patients with Parkinson's disease and 6 age- and gender-matched healthy controls, collects demographic characteristics such as gender and age of the patients and clinical characteristics such as disease duration, uses the Unified Parkinson's Disease Rating Scale part III (UPDRS III) score to evaluate the motor symptoms of the patients, and captures the transmission signals of each task action to calculate the evaluation indexes. The effectiveness of the embodiment of the present application is evaluated by the following method:

[0173] The subjects are divided into a patient group and a control group for statistical difference analysis. Exemplarily, the root mean square of the synthesized acceleration RMS is used as the evaluation index for the tremor evaluation action, and the cycle-amplitude standard deviation ellipse area is used as the evaluation index for the periodic action. Since the sample distribution is not normally distributed and the sample size is small (n < 30), the Mann-Whitney U test is used to compare the differences between the two independent samples. The results are shown in Table 1. Figure 9 As can be seen from the test results, there are significant statistical differences in the hand movement characteristic indexes of the patient group (Patient) and the control group (Control) in the five task actions (palm turning motion PS, fist clenching motion HM, pinching motion FT, resting tremor evaluation action RT, and postural tremor evaluation action PT): (1) In the periodic motion (palm turning motion, fist clenching motion, and pinching motion), the movement retardation indexes of the patient group have significant statistical differences compared with those of the control group (p < 0.05), which indicates that the index can effectively evaluate the subtle but measurable retardation symptom differences between the patient group and the control group during the task motion; (2) In the tremor evaluation action (resting tremor and postural tremor), the tremor index of the patient group is larger than that of the control group (p < 0.05), which is consistent with the trend of clinical evaluation, i.e., the proposed index can effectively distinguish between patients with Parkinson's disease and healthy people and quantitatively evaluate the severity of the symptoms.

[0174] It will be apparent to those skilled in the art that the application is not limited to the details of the above-exemplified embodiments and that the present application can be implemented in other particular forms without departing from the spirit or essential characteristics of the present application. The presently disclosed embodiments are, therefore, to be considered in all respects as illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. No feature of the application is to be construed as limiting the scope of the claims to the instant embodiment.

[0175] Furthermore, it should be understood that although the description above refers to particular embodiments, the description can include more than one independent technical solution, and the description is presented herein in a way of clarity only and should be considered as a whole, and the technical solutions in each embodiment can be combined appropriately to form other embodiments which can be understood by those skilled in the art.

Claims

1. A system for quantitative assessment of hand motor symptoms in Parkinson's disease, characterized in that, include: The glove base is worn on the hand of the patient to be evaluated; Triboelectric sensors are installed on the back side of the glove base at each finger joint to collect bending signals of the finger joints when the patient to be evaluated performs a specified action. An inertial measurement unit, disposed on the back of the hand area of ​​the glove base, is used to collect motion signals of the entire hand when the patient to be evaluated performs a specified action; The data acquisition and processing unit is used to acquire sensing signals from the triboelectric sensor and the inertial measurement unit, calculate tremor assessment index value and muscle tone assessment index value based on the overall motion signal of the hand, calculate bradykinesia assessment index value based on the bending signal of the finger joints and the overall motion signal of the hand, input the calculated three types of assessment index values ​​into the pre-constructed mapping relationship between each assessment index and clinical score to obtain their respective scores, and sum the individual scores to obtain the overall score.

2. The system according to claim 1, characterized in that, The triboelectric sensor includes a first encapsulation layer, a first electrode layer, a first friction layer, a base layer, a second electrode layer, and a second encapsulation layer stacked sequentially. A cavity is formed in a portion of the side of the first friction layer opposite to the base layer. The second friction layer is provided on the sidewall of the base layer within the cavity. When the triboelectric sensor bends with the glove base, the first friction layer and the second friction layer come into contact with each other and separate. A changing electromotive force is induced by the first electrode layer and the second electrode layer. The first electrode layer and the second electrode layer are connected to the data acquisition and processing unit through wires.

3. The system according to claim 1, characterized in that, For tremor symptom assessment, the specified actions that the patient to be assessed must perform include resting tremor assessment actions, postural tremor assessment actions, and action tremor assessment actions. For the assessment of increased muscle tone symptoms, the specified actions that the patient to be assessed needs to perform include: putting the patient to be assessed in a relaxed state, gently rotating the patient's arm several times to allow it to bend and stretch naturally; For the assessment of bradykinesia symptoms, the specified movements that the patient to be assessed needs to perform include pinching fingers, turning the palm over, and clenching the fist.

4. The system according to claim 1, characterized in that, The data acquisition and processing unit includes a signal acquisition module located at the center of the back of the hand on the glove base and a signal processing module separately located outside the glove base. The signal acquisition module is wiredly connected to the triboelectric sensor and the inertial measurement unit, and wirelessly connected to the signal processing module.

5. The system according to claim 4, characterized in that, The signal acquisition module includes a triboelectric signal holding circuit, an analog-to-digital converter, a main control chip, and a power management circuit mounted on a PCB board. Each triboelectric sensor is connected to a corresponding triboelectric signal holding circuit via a wire. The output signal of the triboelectric signal holding circuit is connected to the analog-to-digital converter to convert the analog signal into a digital signal, which is then transmitted to the main control chip. The main control chip is also connected to the inertial measurement unit and the signal processing module. The triboelectric signal holding circuit includes a resistor, a capacitor, and an operational amplifier. The resistor is connected between the negative terminal of the triboelectric sensor and the inverting input terminal of the operational amplifier. The capacitor is connected in parallel between the output terminal and the inverting input terminal of the operational amplifier to form a feedback loop. The positive input terminal of the operational amplifier is connected to the negative terminal of the triboelectric sensor and is grounded together. The output voltage of the triboelectric signal holding circuit is proportional to the output charge of the triboelectric sensor and inversely proportional to the capacitance in the feedback loop.

6. The system according to claim 4, characterized in that, Use any one of the following three methods to construct the mapping relationship between each assessment indicator and the clinical score: Method 1: Using the interval judgment method, a single quantitative indicator is selected to assess the symptoms of Parkinson's patients. Specifically, the quantitative indicator used for tremor symptoms is the root mean square, standard deviation, kurtosis, skewness, dominant peak frequency, peak power, bandwidth, or dominant frequency band energy percentage of the tremor signal collected when the patient performs a specified action. The quantitative indicator used for increased muscle tone symptoms is the root mean square, standard deviation, or coefficient of variation of the angular velocity signal collected when the patient performs a specified action. The quantitative indicator used for bradykinesia symptoms is a time assessment indicator, an amplitude assessment indicator, a velocity assessment indicator, or a combination of these three indicators. The overall assessment indicator is the area of ​​the ellipse representing the amplitude-period standard deviation. Parkinson's patients are divided into discrete groups according to clinical scores. The distribution characteristics of the quantitative indicators for each group are statistically analyzed to determine the interval boundaries of the quantitative indicators. Then, a correspondence table between the quantitative indicator intervals and clinical scores is established. Method 2: Using the continuous mapping method, a single quantitative indicator is selected to assess the symptoms of Parkinson's patients. Specifically, the quantitative indicator used for tremor symptoms is the root mean square, standard deviation, kurtosis, skewness, dominant peak frequency, peak power, bandwidth, or dominant frequency band energy percentage of the tremor signal collected when the patient performs a specified action. The quantitative indicator used for increased muscle tone symptoms is the root mean square, standard deviation, or coefficient of variation of the angular velocity signal collected when the patient performs a specified action. The quantitative indicator used for bradykinesia symptoms is a time assessment indicator, an amplitude assessment indicator, a velocity assessment indicator, or a combination of these three indicators. The overall assessment indicator is the area of ​​the ellipse representing the amplitude-period standard deviation. The linear correlation between the selected quantitative indicator and the clinical score is verified using the Pearson test or Spearman correlation coefficient. Then, a regression model is established with the clinical score as the independent variable and the quantitative indicator as the dependent variable. The regression model converts the quantitative indicator into a continuous clinical score. Method 3: Using a multi-feature fusion method, multi-dimensional features are extracted from the sensor signals. For tremor symptoms, at least two combinations of the root mean square, standard deviation, kurtosis, skewness, dominant peak frequency, peak power, bandwidth, and dominant band energy percentage of the tremor signal collected when the patient performs a specified action are extracted. For increased muscle tone symptoms, at least two combinations of the root mean square, standard deviation, and coefficient of variation of the angular velocity signal collected when the patient performs a specified action are extracted. For bradykinesia symptoms, at least two combinations of velocity assessment indicators, amplitude assessment indicators, velocity assessment indicators, and overall assessment indicators are extracted. In obtaining... Following the above features, corresponding high-dimensional feature sets are constructed, and a mapping model is built according to the following steps: 1) First, all features are standardized; then, correlation analysis, principal component analysis, or recursive feature elimination are used to select the key feature subset that is most relevant to the clinical score and has the lowest redundancy from the high-dimensional feature set; 2) Using the key feature subset as the input variable and the average score of multiple physicians as the target variable, the machine learning model is trained; 3) The performance of the machine learning model is verified, and the parameters of the machine learning model are optimized to obtain the trained machine learning model, which serves as the mapping model.

7. The system according to claim 4, characterized in that, The signal processing module includes a quantitative assessment of the tremor symptoms of the patient to be evaluated, comprising: Step S11: Preprocess the triaxial acceleration signal in the inertial measurement unit collected when the patient to be evaluated performs the specified action for tremor symptom assessment to obtain the synthetic acceleration, and use the synthetic acceleration as the tremor signal; Step S12: Calculate the variation index of the tremor signal using a sliding window. The variation index is expressed as root mean square, standard deviation, or variance. Identify the time corresponding to the maximum value of the variation index of the tremor signal in all sliding windows. Using this as the center, extract the window before and after a set duration as the signal segment range for time domain and frequency domain feature calculation. Regarding the time-domain characteristics of tremor signals: Calculate the first quantization index of the flutter signal within the signal segment, wherein the first quantization index is selected as root mean square (RMS) and standard deviation. A mixture of any one or more of kurtosis and skewness; Regarding the frequency domain characteristics of the tremor signal: Calculate the second quantization index of the flutter signal within the signal segment, using the dominant peak frequency as the second quantization index. Peak power bandwidth and main frequency band energy ratio Any one or more of the following; wherein, the main frequency band energy percentage The calculation formula is as follows: In the formula, , representing the energy of the main frequency band, is selected as the power spectral density integral value in the range of 3 Hz to 10 Hz; , which is the total energy of all frequencies of the flutter signal within the signal segment. Step S13: Input the extracted time-domain and / or frequency-domain features of the tremor assessment into the pre-constructed mapping relationship between tremor assessment indicators and clinical scores to obtain a quantitative score for the severity of hand tremor symptoms in Parkinson's disease under different tremor assessment actions of the patient to be assessed.

8. The system according to claim 4, characterized in that, The signal processing module performs a quantitative assessment of the increased muscle tone symptoms in the patient to be evaluated, including: Step S21: Preprocess the triaxial angular velocity signal in the inertial measurement unit collected when the patient to be evaluated performs the specified action for assessing increased muscle tone symptoms to obtain a synthetic angular velocity, and use the synthetic angular velocity as a muscle tone assessment signal; Step S22: Use threshold segmentation to identify and segment each complete flexion or extension movement segment, calculate the quantitative index of muscle tone assessment within each movement segment, the quantitative index is selected from one or more of the root mean square (RMS), standard deviation and coefficient of variation of the muscle tone assessment signal, and all movement segments use the same quantitative index; take the mean of the quantitative index calculated in all movement segments as the final quantitative index of this assessment. Step S23: Input the final quantitative index into the pre-constructed mapping relationship between the muscle tone enhancement assessment index and the clinical score to obtain the quantitative score of the severity of hand muscle tone enhancement symptoms of Parkinson's disease under different muscle tone enhancement assessment actions of the patient to be assessed.

9. The system according to claim 4, characterized in that, The signal processing module includes a quantitative assessment of the bradykinesia symptoms of the patient to be evaluated, comprising: Step S31: Preprocess the sensing signals from the inertial measurement unit and the triboelectric sensor collected when the patient to be evaluated performs the specified action for assessing bradykinesia symptoms; wherein, in the finger pinching action assessment and fist clenching action assessment, the sensing signal collected by the channel with the highest signal-to-noise ratio among the triboelectric sensors is used, and in the palm flipping action assessment, the roll angle in the three-axis attitude angle calculated based on the sensing signal from the inertial measurement unit is used. Step S32: Extract periodic features from the preprocessed sensor signal. Step S321: Use the peak detection algorithm to extract peak values ​​at multiple scales, obtain the peak and valley values ​​of the sensing signal under periodic motion, and match the calculated peak and valley values. Step S322: Calculate the cycle duration and cycle amplitude based on the matching peak and trough values; Step S33: Calculate the bradykinesia symptom assessment index based on the cycle duration and cycle amplitude obtained under different bradykinesia assessment actions; the bradykinesia symptom assessment index includes any one or more combinations of time assessment index, amplitude assessment index, speed assessment index, and overall assessment index, wherein... The time evaluation indicators include average cycle time, cycle time extreme value, cycle time decay ratio, cycle time standard deviation, and cycle time coefficient of variation. Amplitude evaluation metrics include average amplitude, amplitude extremes, amplitude attenuation ratio, amplitude standard deviation, and amplitude coefficient of variation. Speed ​​evaluation metrics include average speed, speed extremes, speed decay ratio, speed standard deviation, and speed coefficient of variation. The overall evaluation index uses the area of ​​the ellipse representing the period-amplitude standard deviation; Step S34: Input the time assessment index, amplitude assessment index, speed assessment index and overall assessment index calculated under the extracted bradykinesia assessment action into the pre-constructed mapping relationship between bradykinesia assessment index and clinical score to obtain a quantitative score for the severity of hand bradykinesia symptoms of Parkinson's disease under different bradykinesia assessment actions of the patient to be assessed.

10. A method for quantitative assessment of hand motor symptoms in Parkinson's disease based on the system described in any one of claims 1 to 9, characterized in that, include: Step S100: The patient to be evaluated wears the wearable part of the system on both hands; Step S200: The patient to be evaluated performs upper limb motor ability assessment actions, including tremor assessment actions, bradykinesia assessment actions, and muscle tone assessment actions; real-time sensing signals output by the triboelectric sensor and the inertial measurement unit under different assessment actions are collected; Step S300: Extract and calculate features from the sensor signal and convert them into corresponding symptom assessment index values; Step S400: Based on the various symptom assessment indicators, output the overall score of the hand movement symptoms of the patient to be assessed.

Images