CN117298449B - Closed-loop DBS regulation and control method and system based on wearable equipment - Google Patents
Closed-loop DBS regulation and control method and system based on wearable equipment Download PDFInfo
- Publication number
- CN117298449B CN117298449B CN202311428526.5A CN202311428526A CN117298449B CN 117298449 B CN117298449 B CN 117298449B CN 202311428526 A CN202311428526 A CN 202311428526A CN 117298449 B CN117298449 B CN 117298449B
- Authority
- CN
- China
- Prior art keywords
- index
- data
- foot
- acceleration
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000033228 biological regulation Effects 0.000 title claims abstract description 100
- 238000000034 method Methods 0.000 title claims abstract description 91
- 210000001364 upper extremity Anatomy 0.000 claims abstract description 78
- 238000012549 training Methods 0.000 claims abstract description 63
- 230000002159 abnormal effect Effects 0.000 claims abstract description 61
- 210000003141 lower extremity Anatomy 0.000 claims abstract description 47
- 230000005021 gait Effects 0.000 claims abstract description 39
- 206010044565 Tremor Diseases 0.000 claims abstract description 38
- 206010006100 Bradykinesia Diseases 0.000 claims abstract description 37
- 208000006083 Hypokinesia Diseases 0.000 claims abstract description 37
- 230000008014 freezing Effects 0.000 claims abstract description 32
- 238000007710 freezing Methods 0.000 claims abstract description 32
- 230000033001 locomotion Effects 0.000 claims abstract description 26
- 208000012661 Dyskinesia Diseases 0.000 claims abstract description 24
- 230000003183 myoelectrical effect Effects 0.000 claims abstract description 18
- 239000002131 composite material Substances 0.000 claims abstract description 14
- 230000001133 acceleration Effects 0.000 claims description 205
- 238000000605 extraction Methods 0.000 claims description 128
- 239000013598 vector Substances 0.000 claims description 36
- 230000005856 abnormality Effects 0.000 claims description 27
- 238000001914 filtration Methods 0.000 claims description 20
- 230000008569 process Effects 0.000 claims description 20
- 238000011282 treatment Methods 0.000 claims description 18
- 230000015654 memory Effects 0.000 claims description 17
- 238000005070 sampling Methods 0.000 claims description 12
- 230000007787 long-term memory Effects 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 10
- 230000006403 short-term memory Effects 0.000 claims description 10
- 208000027089 Parkinsonian disease Diseases 0.000 claims description 9
- 206010034010 Parkinsonism Diseases 0.000 claims description 9
- 238000009499 grossing Methods 0.000 claims description 9
- 238000001228 spectrum Methods 0.000 claims description 6
- 230000008859 change Effects 0.000 claims description 3
- 238000013480 data collection Methods 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 230000001131 transforming effect Effects 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 description 14
- 208000024891 symptom Diseases 0.000 description 13
- 230000002829 reductive effect Effects 0.000 description 10
- 210000004556 brain Anatomy 0.000 description 8
- 230000007547 defect Effects 0.000 description 7
- 238000004891 communication Methods 0.000 description 5
- 238000012544 monitoring process Methods 0.000 description 5
- 208000018737 Parkinson disease Diseases 0.000 description 4
- 230000001105 regulatory effect Effects 0.000 description 4
- 230000000638 stimulation Effects 0.000 description 4
- 206010056242 Parkinsonian gait Diseases 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 238000004590 computer program Methods 0.000 description 2
- 238000013075 data extraction Methods 0.000 description 2
- 201000010099 disease Diseases 0.000 description 2
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 2
- 210000005036 nerve Anatomy 0.000 description 2
- 238000011002 quantification Methods 0.000 description 2
- 238000002560 therapeutic procedure Methods 0.000 description 2
- 206010017577 Gait disturbance Diseases 0.000 description 1
- 206010061533 Myotonia Diseases 0.000 description 1
- 206010071390 Resting tremor Diseases 0.000 description 1
- 230000003321 amplification Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000036772 blood pressure Effects 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 238000004883 computer application Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 210000003414 extremity Anatomy 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 239000007943 implant Substances 0.000 description 1
- 230000002401 inhibitory effect Effects 0.000 description 1
- 230000002427 irreversible effect Effects 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- 230000004770 neurodegeneration Effects 0.000 description 1
- 208000015122 neurodegenerative disease Diseases 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000001144 postural effect Effects 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 238000001959 radiotherapy Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000029865 regulation of blood pressure Effects 0.000 description 1
- 230000035807 sensation Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61N—ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY
- A61N1/00—Electrotherapy; Circuits therefor
- A61N1/18—Applying electric currents by contact electrodes
- A61N1/32—Applying electric currents by contact electrodes alternating or intermittent currents
- A61N1/36—Applying electric currents by contact electrodes alternating or intermittent currents for stimulation
- A61N1/3605—Implantable neurostimulators for stimulating central or peripheral nerve system
- A61N1/3606—Implantable neurostimulators for stimulating central or peripheral nerve system adapted for a particular treatment
- A61N1/36067—Movement disorders, e.g. tremor or Parkinson disease
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
- A61B5/1036—Measuring load distribution, e.g. podologic studies
- A61B5/1038—Measuring plantar pressure during gait
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
- A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
- A61B5/1101—Detecting tremor
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
- A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
- A61B5/112—Gait analysis
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/389—Electromyography [EMG]
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/40—Detecting, measuring or recording for evaluating the nervous system
- A61B5/4076—Diagnosing or monitoring particular conditions of the nervous system
- A61B5/4082—Diagnosing or monitoring movement diseases, e.g. Parkinson, Huntington or Tourette
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
- A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
- A61B5/6802—Sensor mounted on worn items
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
- A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61N—ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY
- A61N1/00—Electrotherapy; Circuits therefor
- A61N1/18—Applying electric currents by contact electrodes
- A61N1/32—Applying electric currents by contact electrodes alternating or intermittent currents
- A61N1/36—Applying electric currents by contact electrodes alternating or intermittent currents for stimulation
- A61N1/3605—Implantable neurostimulators for stimulating central or peripheral nerve system
- A61N1/36128—Control systems
- A61N1/36135—Control systems using physiological parameters
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Veterinary Medicine (AREA)
- Public Health (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Physics & Mathematics (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Surgery (AREA)
- Molecular Biology (AREA)
- Medical Informatics (AREA)
- Heart & Thoracic Surgery (AREA)
- Neurology (AREA)
- Physiology (AREA)
- Neurosurgery (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Artificial Intelligence (AREA)
- Dentistry (AREA)
- Radiology & Medical Imaging (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Evolutionary Computation (AREA)
- Hospice & Palliative Care (AREA)
- Developmental Disabilities (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Psychiatry (AREA)
- Signal Processing (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention provides a closed-loop DBS regulation and control method and system based on wearable equipment, comprising the following steps: collecting physiological sign data collected by the intelligent shoe, the myoelectric armring and the wearable patch; extracting features based on physiological sign data to obtain gait variation index, tremor index, bradykinesia index, balance index, freezing index, stiffness index, upper limb fluctuation index, trunk fluctuation index, and lower limb fluctuation index; taking the gait variation index, the tremor index, the bradykinesia index, the balance index, the freezing index and the stiffness index as training characteristics, and taking the open-phase or closed-phase state as a training label to obtain a Parkinson model; taking the upper limb abnormal index, the trunk abnormal index and the lower limb movement index as training characteristics, and taking whether abnormal movement exists as a training label to obtain a regulation and control model; and adjusting the parameters of the DBS by utilizing the composite judgment result of the Parkinson model and the regulation model. The invention can detect the abnormal state in real time, and find the abnormal state to accurately regulate and control the DBS in time.
Description
Technical Field
The invention relates to the technical field of medical computer application, in particular to a closed-loop DBS regulation and control method and system based on wearable equipment.
Background
Parkinson's disease (hereinafter referred to as PD patient) is the second most common neurodegenerative disease, whose clinical features are mainly resting tremor, bradykinesia, myotonia and four major motor symptoms of postural gait disorder. In middle and late stages of the disease, the patient is easy to have exercise complications, including exercise symptom fluctuation and catastrophe, which are difficulties in clinical treatment and seriously affect the working capacity and life quality of the patient. At present, the evaluation of the motor symptom fluctuation of the PD patient mainly depends on the evaluation of a clinical scale, the scale has the defects of long time consumption, complex evaluation and the like, and the patient diary is subjective record and has the defects of recall deviation, omission and the like. Thus, objective collection of motor symptom changes in PD patients has become a strong need for clinicians and researchers. The wearable device based on the motion sensor has the characteristics of convenience in operation, objectivity, quantification, fineness and accuracy, so that the accurate treatment of the Parkinson is realized by objectivity, quantification and digitization of the clinical symptoms of the Parkinson.
Deep brain electro-stimulation (deep brain stimulation, DBS) is a treatment means currently used for treating PD patients, and can significantly improve the disease. DBS therapy is mainly to implant electrodes into the brain of a patient, stimulate certain nuclei in the deep brain with a pulse generator, correct abnormal brain electrical loops, and thereby alleviate symptoms in these nerves. Unlike some treatments (cautery or radiotherapy) that permanently unadjustable and irreversible damage the brain, DBS does not disrupt brain structure, allowing future further treatment.
In the application process of DBS therapy, DBS must be regulated and controlled so as to achieve the aim of inhibiting abnormal states of PD patients, so that the regulation and control of DBS equipment is important, and a dynamic optimized nerve sensing method disclosed in Chinese patent application publication CN114949595A, for example, is applied to DBS regulation and control, and the specific method is as follows: determining, by an Implantable Medical Device (IMD), an electrode of a plurality of electrodes of a lead for delivering electrical stimulation to a patient at a particular time; selecting, by the IMD and based on the determined electrodes, an electrode set of the plurality of electrodes; and sensing, by the IMD and via the selected electrode set, an electrical signal of the patient at a particular time. During its application, the clinician is required to control the implantable medical device by means of the programmer to deliver stimulation to the brain, and to query the patient for patient sensation during stimulation for further adjustment. That is, the clinical doctor performs empirical control on DBS control basis by subjective judgment of a patient in a query mode and the like, and the patient has the defects of unclear description of the state of the patient, increased control difficulty and the like. When symptoms appear, the patient needs to reach a hospital to finish DBS regulation, and abnormal states of the patient cannot be well monitored in real time and stimulated in an auxiliary mode, namely, the patient can relieve the abnormal symptoms only through the regulation of the hospital, so that the duration of symptoms of the patient is prolonged.
In addition, the current evaluation of the motion symptom fluctuation of the PD patient by the hospital mainly depends on clinical scale evaluation and patient diaries, the scales have the defects of long time consumption, complex evaluation and the like, the patient diaries are subjective records, the actual situation of the PD patient cannot be evaluated due to the problems of recall deviation, omission and the like, the evaluation of the PD patient by different people is different, and the evaluation can only be performed by a method of taking an average value through multi-person evaluation, and the method has a series of defects of long time consumption, complex evaluation and the like.
Disclosure of Invention
In view of the above problems, the embodiment of the invention provides a closed-loop DBS regulation and control method and system based on wearable equipment, which solve the existing technical problems.
In order to solve the technical problems, the invention provides the following technical scheme:
in a first aspect, the present invention provides a closed-loop DBS regulation method based on a wearable device, the method comprising:
collecting three-axis acceleration data, three-axis angular velocity data and pressure data of the foot, which are collected by the intelligent shoes and related to the left foot and the right foot; collecting myoelectricity data related to the upper limb and triaxial angular velocity data of the upper limb, which are collected by a myoelectricity arm ring; collecting trunk triaxial acceleration data which are collected by the wearable patch and related to the trunk;
Performing feature extraction based on acceleration of the foot triaxial acceleration data in the Y-axis direction to obtain gait variation indexes; performing feature extraction based on the foot triaxial acceleration data and the foot triaxial angular velocity data to obtain a tremor index and a bradykinesia index; extracting features based on foot pressure data to obtain a balance index; performing feature extraction on the basis of the angular speed of the foot pressure data and the foot triaxial angular speed data in the Z-axis direction to obtain a freezing index; extracting features based on myoelectricity data to obtain a stiffness index; extracting features based on the triaxial angular velocity data of the upper limb to obtain an upper limb abnormal movement index; feature extraction is carried out based on the trunk triaxial acceleration data to obtain a trunk fluctuation index; extracting features based on the triaxial acceleration data of the foot to obtain a lower limb abnormal movement index;
taking gait variation index, tremor index, bradykinesia index, balance index, freezing index and stiffness index as training characteristics, taking an artificial tag in an open period or closed period state as a training tag, and inputting the training tag into a long-period and short-period memory network for training to obtain a Parkinson model;
the upper limb abnormal movement index, the trunk abnormal movement index and the lower limb movement index are used as training characteristics, and artificial labels with abnormal movement are used as training labels and are input into a long-term and short-term memory network for training to obtain a regulation and control model;
And adjusting the parameters of the DBS by utilizing the composite judgment result of the Parkinson model and the regulation model.
In a second aspect, the present invention provides a closed-loop DBS regulation system based on a wearable device, the system comprising:
and a data collection module: the intelligent shoe is used for collecting three-axis acceleration data, three-axis angular velocity data and pressure data of the foot, which are collected by the intelligent shoe and are related to the left foot and the right foot; collecting myoelectricity data related to the upper limb and triaxial angular velocity data of the upper limb, which are collected by a myoelectricity arm ring; collecting trunk triaxial acceleration data which are collected by the wearable patch and related to the trunk;
and the feature extraction module is used for: the method comprises the steps of performing feature extraction on acceleration in the Y-axis direction based on three-axis acceleration data of the foot to obtain gait variation indexes; performing feature extraction based on the foot triaxial acceleration data and the foot triaxial angular velocity data to obtain a tremor index and a bradykinesia index; extracting features based on foot pressure data to obtain a balance index; performing feature extraction on the basis of the angular speed of the foot pressure data and the foot triaxial angular speed data in the Z-axis direction to obtain a freezing index; extracting features based on myoelectricity data to obtain a stiffness index; extracting features based on the triaxial angular velocity data of the upper limb to obtain an upper limb abnormal movement index; feature extraction is carried out based on the trunk triaxial acceleration data to obtain a trunk fluctuation index; extracting features based on the triaxial acceleration data of the foot to obtain a lower limb abnormal movement index;
A first model training module: the method comprises the steps of taking gait variation index, tremor index, bradykinesia index, balance index, freezing index and stiffness index as training characteristics, taking an artificial tag in an open phase or closed phase state as a training tag, and inputting the training tag into a long-term and short-term memory network for training to obtain a Parkinson model;
and a second model training module: the method comprises the steps of taking an upper limb abnormal index, a trunk abnormal index and a lower limb movement index as training characteristics, taking an artificial tag with abnormal state as a training tag, and inputting the training tag into a long-term and short-term memory network for training to obtain a regulation and control model;
DBS regulation and control module: and the method is used for adjusting the parameters of the DBS by utilizing the composite judgment result of the Parkinson model and the regulation model.
In a third aspect, the present invention provides an electronic device comprising:
a processor, a memory, an interface in communication with the gateway;
the memory is used for storing programs and data, and the processor calls the programs stored in the memory to execute the closed-loop DBS regulation method based on the wearable device.
In a fourth aspect, the present invention provides a computer readable storage medium comprising a program, which when executed by a processor is adapted to carry out a wearable device-based closed-loop DBS regulation method provided in any of the first aspects.
Compared with the prior art, the technical scheme provided by the embodiment of the application has the following advantages: the invention innovatively provides a closed-loop DBS regulation and control realization method, and the BDS equipment is controlled jointly through different combination modes of the two models. The invention can provide real and effective objective data through daily state monitoring and abnormal state monitoring, helps doctors to verify subjective judgment, reduces the time consumption of evaluation, has complex evaluation and other defects, records the state of a patient in real time, and avoids the problems of recall deviation, omission and the like; the comprehensive judgment is carried out on the description of the self state through the characteristic combination of objective data extraction collected by the wearable equipment, so that the operability is provided, the regulation difficulty is reduced, and the judgment by means of subjective experience in the past is reduced; the intelligent regulation and control at home can be realized, the regulation and control are timely, the PD patient does not need to go to a hospital for regulation and control, and accurate advice can be given to the PD patient through a model; and can provide corresponding objective data for doctors, and the doctor can conveniently know the problem that the abnormal state of the PD patient is not solved after basic regulation by utilizing the objective data. In general, the invention aims at realizing closed-loop DBS regulation and control, and utilizes the wearable device to quantitatively drive the closed-loop accurate regulation and control of DBS through deeply analyzing the physiological sign data of a patient.
Drawings
Fig. 1 is a schematic flow chart of a closed-loop DBS regulation method based on a wearable device according to an embodiment of the present invention;
fig. 2 is a schematic diagram of an execution relationship of a closed-loop DBS regulation method based on a wearable device according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of a closed-loop DBS regulation system based on a wearable device according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the invention.
Detailed Description
The present invention will be further described with reference to the drawings and the detailed description below, in order to make the objects, technical solutions and advantages of the present invention more apparent. It will be apparent that the described embodiments are only some, but not all, embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Based on the shortcomings of the prior art, the embodiment of the invention provides a specific implementation manner of a closed-loop DBS regulation method based on wearable equipment, as shown in fig. 1 and in combination with fig. 2, the method specifically comprises the following steps:
S110, collecting three-axis acceleration data, three-axis angular velocity data and pressure data of the foot, which are collected by the intelligent shoe and related to the left foot and the right foot; collecting myoelectricity data related to the upper limb and triaxial angular velocity data of the upper limb, which are collected by a myoelectricity arm ring; torso triaxial acceleration data related to the torso collected by the wearable patch is collected.
Specifically, the wearable device is a device worn on and in contact with the human body, and the wearable device used in the present invention includes a smart shoe, an myoelectric arm ring, and a wearable patch. The intelligent shoe is worn on the left foot and the right foot of a PD patient, a triaxial acceleration sensor, a triaxial angular velocity sensor and a pressure sensor are arranged in the intelligent shoe, triaxial acceleration data of the foot are acquired by the triaxial acceleration sensor, triaxial angular velocity data of the foot are acquired by the triaxial angular velocity sensor, foot pressure data are acquired by the pressure sensor (see an auxiliary shoe for relieving frozen gait, which is disclosed by the applicant in 2022, 5 and 3 days and disclosed by publication number CN 216416202U); the myoelectric arm ring is worn on the upper limb of a human body, a triaxial angular velocity sensor and a plurality of myoelectric acquisition electrodes (namely a multi-channel acquisition electrode) are arranged on the myoelectric arm ring, the number of the multi-channel myoelectric acquisition electrodes is at least two, that is, the myoelectric arm ring is provided with a plurality of myoelectric acquisition electrodes, so that the number of channels corresponds, myoelectric data are acquired by the myoelectric acquisition electrodes, triaxial angular velocity data of the upper limb are acquired by the triaxial angular velocity sensor (see a portable patch type motion monitoring device with the publication number of CN216258992U disclosed by applicant in 2022, 10 and 14 months and a portable motion detection patch with the publication number of CN307129270S disclosed by 2022, 2 and 25 months); the wearable patch is attached to the trunk (namely, the head and the rest of limbs of a human body are removed), a triaxial acceleration sensor is arranged in the wearable patch, and trunk triaxial acceleration data is acquired by utilizing the triaxial acceleration sensor (see an intelligent armlet (2) with a publication number of CN307592055S, which is published by the applicant in 2022, 4 and 12, and a circuit structure for acquiring electromyographic signals, which is disclosed by the applicant in 2022, 10 and 14, and has a publication number of CN 217588045U).
S120, carrying out feature extraction on the acceleration of the foot in the Y-axis direction based on the three-axis acceleration data of the foot to obtain gait variation indexes; performing feature extraction based on the foot triaxial acceleration data and the foot triaxial angular velocity data to obtain a tremor index and a bradykinesia index; extracting features based on foot pressure data to obtain a balance index; performing feature extraction on the basis of the angular speed of the foot pressure data and the foot triaxial angular speed data in the Z-axis direction to obtain a freezing index; extracting features based on myoelectricity data to obtain a stiffness index; extracting features based on the triaxial angular velocity data of the upper limb to obtain an upper limb abnormal movement index; feature extraction is carried out based on the trunk triaxial acceleration data to obtain a trunk fluctuation index; and extracting features based on the triaxial acceleration data of the foot to obtain the lower limb abnormal movement index.
Specifically, the step is to obtain the characteristic data of objective physiological sign data in all directions, obtain gait variation index, tremor index, bradykinesia index, balance index, freezing index and rigidity index related to the period of movement and the period of opening by extracting the physiological sign data, and then obtain upper limb abnormal index, trunk abnormal index and lower limb abnormal index related to movement abnormality.
S130, taking gait variation index, tremor index, bradykinesia index, balance index, freezing index and stiffness index as training characteristics, taking an artificial tag in an open phase or closed phase state as a training tag, and inputting the training tag into a long-term and short-term memory network for training to obtain a Parkinson Model (PM).
Specifically, after wearing intelligent shoes, myoelectric armors and wearable patches, the PD patient and healthy person extract gait variation index, tremor index, bradykinesia index, balance index, freezing index and stiffness index through step S120, meanwhile, a doctor judges the off-period state and the on-period state according to the behavior of each person in each time period, marks corresponding data with artificial tags, takes the on-period state and the off-period state as training tags, takes the state variation index, tremor index, bradykinesia index, balance index, freezing index and stiffness index extracted in step S120 as training characteristics, and inputs the training signals into a long-short-period memory network (LSTM) for training, so that a Parkinson model is obtained. At this time, the parkinsonism model is trained, and when the state variation index, tremor index, bradykinesia index, balance index, freezing index and rigidity index from the PD patient are input, the parkinsonism model automatically judges the off-phase state or the off-phase state and outputs the result.
S140, taking the upper limb abnormal index, the trunk abnormal index and the lower limb movement index as training characteristics, taking the artificial label with abnormal state as a training label, and inputting the training label into a long-term and short-term memory network for training to obtain a regulation and Control Model (CM).
Specifically, after the PD patient and the healthy person wear intelligent shoes, myoelectric arm rings and wearable patches, the upper limb abnormality index, the trunk abnormality index and the lower limb movement index are extracted through the step S120, meanwhile, a doctor judges whether abnormality exists or not according to the behavior of each person in each time period, marks corresponding data with artificial labels, takes the abnormality and the abnormality as training labels, takes the upper limb abnormality index, the trunk abnormality index and the lower limb movement index extracted through the step S120 as training characteristics, and inputs the upper limb abnormality index, the trunk abnormality index and the lower limb movement index into a long-short-term memory network (LSTM) for training, thereby obtaining a regulation model. At this time, the regulation model is trained, and when the upper limb movement index, the trunk movement index and the lower limb movement index from the PD patient are input, the regulation model automatically judges whether the movement exists or not and outputs the result.
And S150, adjusting the parameters of the DBS by utilizing the composite judgment result of the Parkinson model and the regulation model.
Specifically, according to the fact that the output result of the parkinsonism model is in an open-phase state or in a closed-phase state, the output result of the regulation model is abnormal or not, the parkinsonism model and the output result of the regulation model are combined to be used as a composite judgment result, the problem that the result is wrong due to single judgment can be solved through the composite strategy, the accuracy of executing the regulation strategy is improved, and meanwhile DBS misoperation caused by misjudgment is avoided.
In this embodiment, a method for implementing closed-loop control of DBS is innovatively provided, and control over BDS devices is completed jointly by different combinations of two models. In the embodiment, real and effective objective data can be provided through daily state monitoring and abnormal state monitoring, a doctor is helped to verify subjective judgment, the defects of long evaluation time consumption, complex evaluation and the like are reduced, the state of a patient is recorded in real time, and the problems of recall deviation, omission and the like are avoided; the comprehensive judgment is carried out on the description of the self state through the characteristic combination of objective data extraction collected by the wearable equipment, so that the operability is provided, the regulation difficulty is reduced, and the judgment by means of subjective experience in the past is reduced; the intelligent regulation and control at home can be realized, the regulation and control are timely, the PD patient does not need to go to a hospital for regulation and control, and accurate advice can be given to the PD patient through a model; and can provide corresponding objective data for doctors, and the doctor can conveniently know the problem that the abnormal state of the PD patient is not solved after basic regulation by utilizing the objective data. In general, the embodiment aims at realizing closed-loop DBS regulation and control, and the wearable device is utilized to quantitatively drive the closed-loop accurate regulation and control of DBS through deeply analyzing the physiological sign data of a patient.
In an embodiment of the present invention, in step S120, performing feature extraction based on acceleration of the three-axis acceleration data of the foot in the Y-axis direction, to obtain gait variability index includes:
s1.1: y-axis acceleration acc for respectively extracting three-axis acceleration data of foot relevant to each foot y ；
Acceleration acc of Y axis y The data segment with the value of the first threshold value which is continuously appeared in the data sequence is used as an abnormal data segment, and the first threshold value is preferably 2;
s1.2, all values in the abnormal data segment are zeroed to be removed, the removed data is filtered by adopting a moving average method, the abnormal data segment is removed by adopting a data zeroing method, and the width of a sliding window adopted in the filtering of the moving average method is preferably 10, so that interference is reduced;
S1.3Y-axis acceleration acc after screening and filtering y All valid acceleration data segments greater than the second threshold and calculate eachThe length of each effective acceleration data segment is recorded as L i1 The method comprises the steps of carrying out a first treatment on the surface of the The second threshold is preferably 0.1;
s1.4, extracting the maximum value of the acceleration index in each effective acceleration data segment and marking as acc max,i ；
S1.5, obtaining gait variation indexes by using a gait variation characteristic extraction model;
the gait variation characteristic extraction model is as follows:
wherein GVS represents gait variability index; w (w) 1 、w 2 And w 3 Representing the weight value; n is n L1 Representing the number of left foot effective acceleration data segments; n is n R1 Representing the number of right foot effective acceleration data segments; s is S m A standard deviation representing the product of the length of the effective acceleration data segment and the maximum value; l (L) m A standard deviation representing the length of each effective acceleration data segment; acc (acc) max,m A standard deviation representing a maximum value in each of the effective acceleration data segments; n (N) 1 Representing the number of valid acceleration data segments;
wherein,
where n represents the number of valid acceleration data segments.
The gait variation characteristic extraction model comprehensively considers the gait variation conditions of the left foot and the right foot, the left foot and the right foot are processed by using S1.1-S1.4, abnormal data are removed in the data processing process, data filtering processing is carried out, and after the data acquisition is completed, the gait variation characteristic extraction model is used for obtaining the final gait variation index, so that the gait variation index is more objective and accurate, the gait variation is quantized, and the gait variation is visually presented in a numerical mode.
In step S120, feature extraction is performed based on the foot triaxial acceleration data and the foot triaxial angular velocity data, and obtaining the tremor index includes:
s2.1: calculating component acceleration acc in triaxial acceleration data of foot x 、acc y And acc (sic) z Vector sum in X, Y and Z axis direction, get vector acceleration ACC;
the vector acceleration ACC is obtained specifically by:
ACC＝axx x 2 +axx y 2 +axx z 2 ；
s2.2: calculating component angular velocity gyr in foot three-axis angular velocity data x 、gyr y And gyr z Vector sum in X, Y and Z direction to obtain vector angular velocity GYR;
the vector angular velocity GYR is obtained specifically by:
GYR＝gyr x 2 +gyr y 2 +gyr z 2 ；
s2.3: average value filtering processing is carried out on the vector acceleration ACC and the vector angular velocity GYR to obtain an acceleration signal ACC0 and an angular velocity signal GYR0;
the processing procedure of the vector acceleration ACC is as follows:
wherein ACC0 i Representing an ith acceleration signal; ACC (ACC) i Representing the i-th vector acceleration; fs represents the sampling rate; ACC (ACC) k A baseline (i.e., average) of vector accelerations;
the acceleration signal ACC0 is ACC0 i Is a set of (3).
The vector angular velocity GYR is processed as follows:
in the formula, GYR0 i Representing an ith angular velocity signal; GYR i Represents the i-th vector angular velocity; fs represents the sampling rate; GYR k A baseline (i.e., average) of vector angular velocity;
the angular velocity signal GYR0 is GYR0 i Is a set of (3).
S2.4: counting the number of zero crossings related to the left foot in the acceleration signal ACC0 and the angular velocity signal GYR0 to obtain the number N of the acceleration zero crossings of the left foot respectively L_acc_0 And the number of zero crossing points of angular velocity N L_gyr_0 ；
S2.5: counting the number of zero crossings related to the right foot in the acceleration signal ACC0 and the angular velocity signal GYR0 to obtain the number N of the acceleration zero crossings of the right foot respectively R_acc_0 And the number of zero crossing points of angular velocity N R_gyr_0 ；
S2.6: obtaining tremor index by using a tremor characteristic extraction model;
the tremor feature extraction model is as follows:
wherein TRS represents tremor index; max represents the extraction maximum value; fs denotes the sampling rate.
The tremor index obtained by the method can quantify tremors in the walking process of PD patients, and the tremors are visually presented in a numerical mode.
In step S120, performing feature extraction based on the foot triaxial acceleration data and the foot triaxial angular velocity data, and obtaining the bradykinesia index includes:
s3.1: component acceleration acc in foot triaxial acceleration data x 、acc y And acc (sic) z Adding to obtain total acceleration;
s3.2: triaxial footComponent angular velocity gyr in angular velocity data x 、gyr y And gyr z Adding to obtain a total angular velocity;
s3.3: calculating first-order differential signals of the total acceleration and the total angular velocity, converting all values of the first-order differential signals into positive values, sequencing, recording data point positions of one fourth of the front ranking, calculating the average value of the total acceleration and the total angular velocity at the corresponding positions as a base line, and calculating the difference value between the total acceleration and the base line and the total angular velocity to obtain the total acceleration and the total angular velocity after primary filtering;
S3.4: smoothing the primary filtered total acceleration and total angular velocity with a moving average window of a predetermined length to obtain a secondary filtered total acceleration signal acc s And a total angular velocity signal gyr x The method comprises the steps of carrying out a first treatment on the surface of the The length of the moving average window is preferably 10;
s3.5: extracting the total acceleration signal acc s And a total angular velocity signal gyr s The data length of the data segment which is larger than the third threshold value is used for obtaining a total acceleration signal acc s Data length L of data segment greater than third threshold au And a total angular velocity signal gyr s Data length L of data segment greater than third threshold gu The method comprises the steps of carrying out a first treatment on the surface of the The third threshold is preferably 0.1;
s3.6: extracting the total acceleration signal acc s And a total angular velocity signal gyr s The data length of the data segment smaller than the third threshold value is used for obtaining a total acceleration signal acc s Data length L of data segment smaller than third threshold al And a total angular velocity signal gyr s Data length L of data segment smaller than third threshold gl ；
S3.7: extracting the total acceleration signal acc s Maximum value in the data segment larger than the third threshold value is obtained to obtain the maximum value A of the acceleration m ；
S3.8: extracting the total angular velocity signal gyr s Maximum value in data segment greater than the third threshold value, obtaining maximum value G of angular velocity m ；
S3.9: obtaining a bradykinesia index by using a bradykinesia feature extraction model;
The bradykinesia feature extraction model is:
wherein BRS represents the bradykinesia index; ln represents taking the logarithm.
The bradykinesia index obtained by the method can quantify the bradykinesia of the PD patient in the walking process and visually present in a numerical mode.
In step S120, feature extraction is performed based on the foot pressure data, and obtaining the balance index includes:
s4.1: first pressure threshold TH for fitting foot pressure data using pressure threshold configuration model 1 ；
The pressure threshold configuration model is:
wherein N is 2 The amount of foot pressure data; p (P) i Data representing an ith foot pressure; p (P) min Representing a minimum value in the foot pressure data;
s4.2: rejecting foot pressure data greater than a first pressure threshold TH 1 The length of the data segment is larger than twice the sampling frequency, the effective foot pressure data segment is obtained, and the length of the effective foot pressure data segment is recorded as L i2 ；
S4.3: extracting the maximum pressure value of the foot pressure data in each effective foot pressure data segment and marking as P max,i ；
S4.4: obtaining a balance index by using a balance characteristic extraction model;
the balance characteristic extraction model is as follows:
wherein BAS represents a balance index; ln represents taking the logarithm;n L2 representing the number of effective foot pressure data segments of the left foot; n is n R2 The number of effective foot pressure data segments for the right foot is indicated.
The balance index obtained by the method can quantify the balance of the PD patient in the walking process and visually present the balance in a numerical mode.
In step S120, feature extraction is performed based on the angular velocity of the foot pressure data and the foot triaxial angular velocity data in the Z-axis direction, and obtaining the freezing index includes:
s5.1: configuring the second pressure threshold TH using a sliding average method based on foot pressure data 2 The method comprises the steps of carrying out a first treatment on the surface of the The sliding window width is preferably twice the sampling rate fs;
s5.2: the foot pressure data is greater than the second pressure threshold TH 2 Is marked as 0 for the walk state and the remaining data points are marked as 1 for the stop state;
s5.3: extracting angular velocity gyr of foot three-axis angular velocity data corresponding to a data segment marked as 0 in the Z-axis direction z Finally, the angle change quantity at the moment is used for obtaining the turning angle EU;
the turning angle EU is obtained specifically by:
wherein EU represents a turning angle; n (N) 4 Representing the number of data segments marked 0;indicating the angular velocity in the i-th Z-axis direction in the data segment marked 0;
s5.4: judging the angle which is larger than a fourth threshold value and the number of times of continuous occurrence exceeds three in the turning angle as a turning process, and connecting the turning angles of the left foot and the right foot in the same turning process into a new angle vector EULR;
Specifically, the fourth threshold is preferably 10, and the degree vector EULR is a set of turning angles of the left foot and the right foot during turning;
s5.5: obtaining a freezing index by utilizing a freezing characteristic extraction model;
the freezing characteristic extraction model is as follows:
wherein FOS is a freeze index; w (w) 4 And w 5 Is a weight value; n (N) 3 Is the number of turns; EULR (EULR) i Angle vector of the ith turn; EULR (EULR) m Is the average of the angle vectors of all turns.
Wherein,
the freezing index obtained by the method can quantify the frozen gait of the PD patient in the walking process, and the frozen gait is visually presented in a numerical mode.
In step S120, performing feature extraction based on myoelectricity data, where obtaining a stiffness index includes:
s6.1: carrying out band-pass filtering on myoelectric data by adopting a Butterworth filter to obtain pure myoelectric signals EMG; myoelectricity data can be acquired by a plurality of channels, that is, a plurality of groups of myoelectricity signals can be obtained by a plurality of channels;
s6.2: all the values of the electromyographic signals EMG are converted into positive values, and average filtering is carried out by using a sliding window with a preset length to obtain an electromyographic signal envelope curve ENV;
s6.3: extracting data segments which are larger than half of average value of electromyographic signal envelope curve ENV in electromyographic signal EMG, and obtaining the number n of the data segments p Length L of each data segment n The mean value m of each data segment n ；
S6.4: fourier transforming the electromyographic signal EMG to obtain an electromyographic spectrum signal F;
s6.5: the myoelectricity spectrum signals of 100-200Hz and 20-100Hz are summed respectively to correspondingly obtain first energy values of the signals of two frequency bandsE 1 And a second energy value E 2 The method comprises the steps of carrying out a first treatment on the surface of the Namely, feature extraction is carried out by taking 100Hz as a boundary line;
s6.6: obtaining a rigidity index by using a rigidity characteristic extraction model;
the rigidity characteristic extraction model is as follows:
wherein RIS represents a stiffness index; k represents the number of channels for collecting electromyographic signals; fs denotes the sampling rate.
The stiffness index obtained by the method can quantify the stiffness state of the PD patient in the walking process and visually present the stiffness state in a numerical mode.
In step S120, feature extraction is performed based on the upper limb triaxial angular velocity data, and obtaining the upper limb abnormal index includes:
s7.1: respectively calculating the angular speed integral of the upper limb triaxial angular speed data in three directions to obtain angle signals in the three directions;
s7.2: extracting the maximum value EU of each direction max And an angle minimum value EU min Number of corresponding angle maxima n max And an angle minimum number n min ；
S7.3: calculating adjacent two angle maxima EU in time sequence max Or an angle minimum value EU min The absolute value of the difference value between the two directions is obtained and the average value is obtained to obtain the average value EUM of the maximum angle value corresponding to each direction max Sum angle minimum mean EUM min ；
S7.4: obtaining an upper limb abnormality index by using an upper limb abnormality extraction model;
the upper limb abnormal extraction model is as follows:
wherein AMS u Indicating an upper limb movement index;representing the average value of the maximum value of the angle of the X axis direction of the upper limb;representing the average value of the minimum value of the X-axis angle of the upper limb; />Representing the average value of the maximum value of the angle in the Y-axis direction of the upper limb;representing the average value of the minimum value of the Y-axis angle of the upper limb; />Representing the average value of the maximum value of the Z-axis direction angle of the upper limb;representing the average value of the minimum value of the Z-axis angle of the upper limb;
the upper limb abnormal state obtained by the method can be quantified and visually presented in a numerical mode in the walking process of the PD patient.
In step S120, performing feature extraction based on the trunk triaxial acceleration data, and obtaining the trunk fluctuation index includes:
s8.1: smoothing the acceleration signals of each direction of the trunk triaxial acceleration data by using a moving average window with a preset length, and then calculating acceleration integrals in three directions to obtain speed signals in three directions; the width of the sliding average window is 10;
S8.2: extracting the maximum value VE of the speed signal in each direction max Number of corresponding maximum speed values n max ；
S8.3: calculating adjacent velocity maxima VE in time sequence max The absolute value of the difference value between the two is taken and the average value is obtained to obtain the average value VEM of the maximum speed value corresponding to each direction max ；
S8.4: obtaining a trunk fluctuation index by using a trunk fluctuation extraction model;
the trunk transaction extraction model is as follows:
wherein AMS m The index of the abnormal movement of the trunk is expressed,representing the average value of the maximum speed value of the trunk in the X-axis direction;representing the average value of the maximum speed value of the trunk in the Y-axis direction; />Representing the average value of the maximum speed value of the trunk in the Z-axis direction;
the trunk abnormal state obtained by the method can quantify the trunk abnormal state of the PD patient in the walking process, and the trunk abnormal state is visually presented in a numerical mode.
In step S120, performing feature extraction based on the triaxial acceleration data of the foot, and obtaining the lower limb movement index includes:
s9.1: smoothing the acceleration signals of the foot triaxial acceleration data in each direction by using a moving average window with a preset length, and then calculating acceleration integrals in three directions to obtain speed signals in the three directions;
s9.2: extracting the maximum value VE of the speed signal in each direction max Number of corresponding maximum speed values n max ；
S9.3: calculating adjacent velocity maxima VE in time sequence max The absolute value of the difference value between the two is taken and the average value is obtained to obtain the average value VEM of the maximum speed value corresponding to each direction max ；
S1.4: obtaining a lower limb abnormality index by using a lower limb abnormality extraction model;
the lower limb abnormal extraction model is as follows:
wherein AMS l The index of the lower limb movement is represented,representing the average value of the maximum speed values of the lower limb in the X-axis direction;representing the average value of the maximum speed values of the lower limbs in the Y-axis direction; />The average value of the maximum speed value in the Z-axis direction of the lower limb is shown.
The characteristic extraction mode of the lower limb abnormal index and the characteristic extraction mode of the trunk abnormal index are the same (only the data sources are different), and the trunk abnormal state of the PD patient in the walking process can be quantified by the lower limb abnormal index obtained by adopting the mode, and the lower limb abnormal state can be visually presented in a numerical mode.
In this embodiment, specific extraction processes of gait variability index, tremor index, bradykinesia index, balance index, freezing index, stiffness index, upper limb abnormality index, trunk abnormality index, lower limb abnormality index are provided, and filtering processing is performed on the data in each index rejection process, so that noise is removed and random errors are reduced, and each movement abnormality can be objectively and accurately expressed by the finally extracted index. In addition, the feature extraction of each index in the embodiment is a deep analysis mainly of symptoms of a patient suffering from Parkinson's disease, rather than a common extraction mode mainly of conventional statistical methods such as median mode, average number, frequency and the like, so that the deep analysis of clinical symptoms of Parkinson's disease is achieved, the data is more objective and more suitable for clinical requirements, and the method is a medical cross extraction mode and more suitable for clinical diagnosis and treatment routine of Parkinson's disease.
In an embodiment of the present invention, adjusting parameters of the DBS in step S150 using the composite determination result of the parkinson model and the regulation model includes:
defining the open-phase state of the Parkinson model as 1 and the closed-phase state as 0;
defining the regulation model to be 1 without variation and 0 with variation;
if the parkinsonism model output is 1 and the regulation model output is 1, no regulation instruction is required to be sent to the DBS;
if the output of the Parkinson model is 1 and the output of the regulation model is 0, sending a step-down instruction to the DBS;
if the Parkinson model output is 0 and the regulation model output is 1, a boosting instruction is sent to the DBS;
if the output of the Parkinson model is 0 and the output of the regulation model is 0, sending prompt information of timely medical treatment of the patient.
For a clearer description of the above regulation process, the following formula may be used:
more specifically, after the DBS receives the adjustment instruction, corresponding adjustment is performed according to the obtained current parameters of the DBS, such as voltage (V), frequency (f), and Pulse Width (PW), and the parkinson model: when boosting, the amplification is +0.2; regulation model: the voltage is reduced, and the amplitude is reduced to-0.1. The voltage adjustment interval is: current parameters of DBS ± 50%.
And judging whether to perform adjustment when the off-period state is found currently, if yes, starting to perform adjustment, receiving feedback information of whether the patient has abnormal feeling currently after the adjustment is completed, if the feedback information is abnormal, maintaining the adjusted voltage, continuously monitoring the off-period state of the patient, if the off-period state still exists, repeating the operation until the off-period state disappears, and maintaining the adjusted parameter state to operate by the DBS after the off-period state disappears.
In the regulation and control process, if the current closing state is found, receiving feedback information of whether the patient is regulated, and if the feedback information is not, prompting to seek medical attention in time;
after the adjustment is finished, receiving feedback information of whether the patient has abnormal feeling currently, if the feedback information is yes, stopping boosting, performing depressurization according to the depressurization amplitude, repeating the operation until the feedback information of the patient is no, and maintaining the adjusted parameter state by the DBS.
In the boost regulation, if the monitored off-period state still exists after the boost is 1.5 times of the DBS parameter, prompt for medical attention in time.
In the regulation of blood pressure reduction, if the blood pressure is reduced to 0.5 times of DBS parameters before regulation, feedback information of a patient is thrown as abnormal, and prompt of medical treatment is given timely.
In this embodiment, a specific regulation and control process of DBS is provided, and through objective data, the operation parameters of DBS can be controlled and regulated according to the composite result of the parkinson model and the regulation and control model, and intervention is timely performed when abnormality occurs, so as to achieve the purpose of relieving symptoms of a patient, reduce the pain time of the patient, realize helping the PD patient monitor the abnormal state in real time, find out that the abnormal state of the PD patient is subjected to targeted home regulation and control, and provide corresponding advice.
Based on the same inventive concept, the embodiment of the present application further provides a closed-loop DBS regulation system based on the wearable device, which can be used to implement a closed-loop DBS regulation method based on the wearable device described in the above embodiment, as described in the following embodiments. Because the principle of solving the problem of the closed-loop DBS regulation and control system based on the wearable device is similar to that of the closed-loop DBS regulation and control method based on the wearable device, the implementation of the system can be implemented by referring to the method, and the repetition is omitted. As used below, the term "unit" or "module" may be a combination of software and/or hardware that implements the intended function. While the system described in the following embodiments is preferably implemented in software, implementation in hardware, or a combination of software and hardware, is also possible and contemplated.
The invention provides a closed-loop DBS regulation and control system based on wearable equipment, as shown in fig. 3, the system comprises:
the data collection module 210: the intelligent shoe is used for collecting three-axis acceleration data, three-axis angular velocity data and pressure data of the foot, which are collected by the intelligent shoe and are related to the left foot and the right foot; collecting myoelectricity data related to the upper limb and triaxial angular velocity data of the upper limb, which are collected by a myoelectricity arm ring; collecting trunk triaxial acceleration data which are collected by the wearable patch and related to the trunk;
The feature extraction module 220: the method comprises the steps of performing feature extraction on acceleration in the Y-axis direction based on three-axis acceleration data of the foot to obtain gait variation indexes; performing feature extraction based on the foot triaxial acceleration data and the foot triaxial angular velocity data to obtain a tremor index and a bradykinesia index; extracting features based on foot pressure data to obtain a balance index; performing feature extraction on the basis of the angular speed of the foot pressure data and the foot triaxial angular speed data in the Z-axis direction to obtain a freezing index; extracting features based on myoelectricity data to obtain a stiffness index; extracting features based on the triaxial angular velocity data of the upper limb to obtain an upper limb abnormal movement index; feature extraction is carried out based on the trunk triaxial acceleration data to obtain a trunk fluctuation index; extracting features based on the triaxial acceleration data of the foot to obtain a lower limb abnormal movement index;
the first model training module 230: the method comprises the steps of taking gait variation index, tremor index, bradykinesia index, balance index, freezing index and stiffness index as training characteristics, taking an artificial tag in an open phase or closed phase state as a training tag, and inputting the training tag into a long-term and short-term memory network for training to obtain a Parkinson model;
The second model training module 240: the method comprises the steps of taking an upper limb abnormal index, a trunk abnormal index and a lower limb movement index as training characteristics, taking an artificial tag with abnormal state as a training tag, and inputting the training tag into a long-term and short-term memory network for training to obtain a regulation and control model;
DBS regulation module 250: parameters for adjusting DBS using composite decisions of a parkinsonism model and a regulatory model
In an embodiment of the present invention, the feature extraction module 220 performs feature extraction based on acceleration of the three-axis acceleration data of the foot in the Y-axis direction, and the obtaining of the gait variability index includes:
y-axis acceleration acc for respectively extracting three-axis acceleration data of foot relevant to each foot y ；
Acceleration acc of Y axis y The data segment with the value of the first threshold value continuously appearing in the data sequence is used as an abnormal data segment;
all values in the abnormal data segment are zeroed to be removed, and the removed data is subjected to filtering treatment by adopting a moving average method;
y-axis acceleration acc after screening and filtering y All the effective acceleration data segments larger than the second threshold value, calculating the length of each effective acceleration data segment and recording as L i1 ；
Extracting the maximum value of the acceleration index in each effective acceleration data segment and marking as acc max,i ；
Obtaining gait variation indexes by using a gait variation feature extraction model;
the gait variation characteristic extraction model is as follows:
wherein GVS represents gait variability index; w (w) 1 、w 2 And w 3 Representing the weight value; n is n L1 Representing the number of left foot effective acceleration data segments; n is n R1 Representing the number of right foot effective acceleration data segments; s is S m A standard deviation representing the product of the length of the effective acceleration data segment and the maximum value; l (L) m A standard deviation representing the length of each effective acceleration data segment; acc (acc) max,m A standard deviation representing a maximum value in each of the effective acceleration data segments; n (N) 1 Representing the number of valid acceleration data segments.
In an embodiment of the present invention, the feature extraction module 220 performs feature extraction based on the three-axis acceleration data and the three-axis angular velocity data of the foot, and the obtaining of the tremor index includes:
calculating component acceleration acc in triaxial acceleration data of foot x 、acc y And acc (sic) z Vector sum in X, Y and Z axis direction, get vector acceleration ACC;
calculating the angular velocity of foot three axesComponent angular velocity gyr in the degree data x 、gyr y And gyr z Vector sum in X, Y and Z direction to obtain vector angular velocity GYR;
average value filtering processing is carried out on the vector acceleration ACC and the vector angular velocity GYR to obtain an acceleration signal ACC0 and an angular velocity signal GYR0;
Counting the number of zero crossings related to the left foot in the acceleration signal ACC0 and the angular velocity signal GYR0 to obtain the number N of the acceleration zero crossings of the left foot respectively L_acc_0 And the number of zero crossing points of angular velocity N L_gyr_0 ；
Counting the number of zero crossings related to the right foot in the acceleration signal ACC0 and the angular velocity signal GYR0 to obtain the number N of the acceleration zero crossings of the right foot respectively R_acc_0 And the number of zero crossing points of angular velocity N R_gyr_0 ；
Obtaining tremor index by using a tremor characteristic extraction model;
the tremor characteristic extraction model is as follows:
wherein TRS represents tremor index; max represents the extraction maximum value; fs denotes the sampling rate.
In an embodiment of the present invention, the feature extraction based on the foot triaxial acceleration data and the foot triaxial angular velocity data to obtain the bradykinesia index includes:
component acceleration acc in foot triaxial acceleration data x 、acc y And acc (sic) z Adding to obtain total acceleration;
component angular velocity gyr in foot triaxial angular velocity data x 、gyr y And gyr z Adding to obtain a total angular velocity;
calculating first-order differential signals of the total acceleration and the total angular velocity, converting all values of the first-order differential signals into positive values, sequencing, recording data point positions of one fourth of the front ranking, calculating the average value of the total acceleration and the total angular velocity at the corresponding positions as a base line, and calculating the difference value between the total acceleration and the base line and the total angular velocity to obtain the total acceleration and the total angular velocity after primary filtering;
Smoothing the primary filtered total acceleration and total angular velocity with a moving average window of a predetermined length to obtain a secondary filtered total acceleration signal acc s And a total angular velocity signal gyr s ；
Extracting the total acceleration signal acc s And a total angular velocity signal gyr s The data length of the data segment which is larger than the third threshold value is used for obtaining a total acceleration signal acc s Data length L of data segment greater than third threshold au And a total angular velocity signal gyr z Data length L of data segment greater than third threshold gu ；
Extracting the total acceleration signal acc s And a total angular velocity signal gyr s The data length of the data segment smaller than the third threshold value is used for obtaining a total acceleration signal acc s Data length L of data segment smaller than third threshold al And a total angular velocity signal gyr s Data length L of data segment smaller than third threshold gl ；
Extracting the total acceleration signal acc s Maximum value in the data segment larger than the third threshold value is obtained to obtain the maximum value A of the acceleration m ；
Extracting the total angular velocity signal gyr s Maximum value in data segment greater than the third threshold value, obtaining maximum value G of angular velocity m ；
Obtaining a bradykinesia index by using a bradykinesia feature extraction model;
the bradykinesia feature extraction model is:
wherein BRS represents the bradykinesia index; ln represents taking the logarithm.
In an embodiment of the present invention, the feature extraction module 220 performs feature extraction based on the foot pressure data, and the obtaining the balance index includes:
by means of pressure thresholdFirst pressure threshold TH of foot pressure data of model fitting 1 ；
The pressure threshold configuration model is:
wherein N is 2 The amount of foot pressure data; p (P) i Data representing an ith foot pressure; p (P) min Representing a minimum value in the foot pressure data;
rejecting foot pressure data greater than a first pressure threshold TH 1 The length of the data segment is larger than twice the sampling frequency, the effective foot pressure data segment is obtained, and the length of the effective foot pressure data segment is recorded as L i2 ；
Extracting the maximum pressure value of the foot pressure data in each effective foot pressure data segment and marking as P max,i ；
Obtaining a balance index by using a balance characteristic extraction model;
the balance characteristic extraction model is as follows:
wherein BAS represents a balance index; ln represents taking the logarithm; n is n L2 Representing the number of effective foot pressure data segments of the left foot; n is n R2 The number of effective foot pressure data segments for the right foot is indicated.
In an embodiment of the present invention, the feature extraction module 220 performs feature extraction based on the angular velocity of the foot pressure data and the foot triaxial angular velocity data in the Z-axis direction, and the obtaining the freezing index includes:
Configuring the second pressure threshold TH using a sliding average method based on foot pressure data 2 ；
The foot pressure data is greater than the second pressure threshold TH 2 Is marked as 0 for the walk state and the remaining data points are marked as 1 for the stop state;
extracting angular velocity gyr of foot three-axis angular velocity data corresponding to a data segment marked as 0 in the Z-axis direction z Finally, the angle change quantity at the moment is used for obtaining the turning angle;
judging the angle which is larger than a fourth threshold value and the number of times of continuous occurrence exceeds three in the turning angle as a turning process, and connecting the turning angles of the left foot and the right foot in the same turning process into a new angle vector EULR;
obtaining a freezing index by utilizing a freezing characteristic extraction model;
the freezing characteristic extraction model is as follows:
wherein FOS is a freeze index; w (w) 4 And w 5 Is a weight value; n (N) 3 Is the number of turns; EULR (EULR) i Angle vector of the ith turn; EULR (EULR) m Is the average of the angle vectors of all turns.
In an embodiment of the present invention, the feature extraction module 220 performs feature extraction based on myoelectricity data, and the obtaining of the stiffness index includes:
carrying out band-pass filtering on myoelectric data by adopting a Butterworth filter to obtain pure myoelectric signals EMG;
All the values of the electromyographic signals EMG are converted into positive values, and average filtering is carried out by using a sliding window with a preset length to obtain an electromyographic signal envelope curve ENV;
extracting data segments which are larger than half of average value of electromyographic signal envelope curve ENV in electromyographic signal EMG, and obtaining the number n of the data segments p Length L of each data segment n The mean value m of each data segment n ；
Fourier transforming the electromyographic signal EMG to obtain an electromyographic spectrum signal F;
the myoelectricity spectrum signals of 100-200Hz and 20-100Hz are summed respectively to correspondingly obtain a first energy value E of the signals of two frequency bands 1 And a second energy value E 2 ；
Obtaining a rigidity index by using a rigidity characteristic extraction model;
the rigidity characteristic extraction model is as follows:
wherein RIS represents a stiffness index; k represents the number of channels for collecting electromyographic signals; fs denotes the sampling rate.
In an embodiment of the present invention, the feature extraction module 220 performs feature extraction based on the upper limb triaxial angular velocity data, and the obtaining of the upper limb abnormal movement index includes:
respectively calculating the angular speed integral of the upper limb triaxial angular speed data in three directions to obtain angle signals in the three directions;
extracting the maximum value EU of each direction max And an angle minimum value EU min Number of corresponding angle maxima n max And an angle minimum number n min ；
Calculating adjacent two angle maxima EU in time sequence max Or an angle minimum value EU min The absolute value of the difference value between the two directions is obtained and the average value is obtained to obtain the average value EUM of the maximum angle value corresponding to each direction max Sum angle minimum mean EUM min ；
Obtaining an upper limb abnormality index by using an upper limb abnormality extraction model;
the upper limb abnormal extraction model is as follows:
wherein AMS u Indicating an upper limb movement index;representing the average value of the maximum value of the angle of the X axis direction of the upper limb;representing the X axis of the upper limbAverage value of angle minimum value; />Representing the average value of the maximum value of the angle in the Y-axis direction of the upper limb;representing the average value of the minimum value of the Y-axis angle of the upper limb; />Representing the average value of the maximum value of the Z-axis direction angle of the upper limb;representing the average value of the minimum value of the Z-axis angle of the upper limb;
feature extraction based on the trunk triaxial acceleration data in the feature extraction module 220, the obtaining the trunk transaction index includes:
smoothing the acceleration signals of each direction of the trunk triaxial acceleration data by using a moving average window with a preset length, and then calculating acceleration integrals in three directions to obtain speed signals in three directions;
extracting the maximum value VE of the speed signal in each direction max Number of corresponding maximum speed values n max ；
Calculating adjacent velocity maxima VE in time sequence max The absolute value of the difference value between the two is taken and the average value is obtained to obtain the average value VEM of the maximum speed value corresponding to each direction max ；
Obtaining a trunk fluctuation index by using a trunk fluctuation extraction model;
the trunk transaction extraction model is as follows:
wherein AMS m The index of the abnormal movement of the trunk is expressed,representing the average value of the maximum speed value of the trunk in the X-axis direction;representing the average value of the maximum speed value of the trunk in the Y-axis direction; />Representing the average value of the maximum speed value of the trunk in the Z-axis direction;
the feature extraction module 220 performs feature extraction based on the triaxial acceleration data of the foot, and the obtaining of the lower limb abnormal index includes:
smoothing the acceleration signals of the foot triaxial acceleration data in each direction by using a moving average window with a preset length, and then calculating acceleration integrals in three directions to obtain speed signals in the three directions;
extracting the maximum value VE of the speed signal in each direction max Number of corresponding maximum speed values n max ；
Calculating adjacent velocity maxima VE in time sequence max The absolute value of the difference value between the two is taken and the average value is obtained to obtain the average value VEM of the maximum speed value corresponding to each direction max ；
Obtaining a lower limb abnormality index by using a lower limb abnormality extraction model;
the lower limb abnormal extraction model is as follows:
wherein AMS l The index of the lower limb movement is represented, Representing the average value of the maximum speed values of the lower limb in the X-axis direction;representing the average value of the maximum speed values of the lower limbs in the Y-axis direction; />The average value of the maximum speed value in the Z-axis direction of the lower limb is shown.
In an embodiment of the present invention, the adjusting the parameters of the DBS in the DBS adjusting module 250 according to the composite determination result of the parkinson model and the adjusting model includes:
defining the open-phase state of the Parkinson model as 1 and the closed-phase state as 0;
defining the regulation model to be 1 without variation and 0 with variation;
if the parkinsonism model output is 1 and the regulation model output is 1, no regulation instruction is required to be sent to the DBS;
if the output of the Parkinson model is 1 and the output of the regulation model is 0, sending a step-down instruction to the DBS;
if the Parkinson model output is 0 and the regulation model output is 1, a boosting instruction is sent to the DBS;
if the output of the Parkinson model is 0 and the output of the regulation model is 0, sending prompt information of timely medical treatment of the patient.
The embodiment of the present application further provides a specific implementation manner of an electronic device capable of implementing all the steps in the method in the foregoing embodiment, and referring to fig. 4, the electronic device 300 specifically includes the following:
a processor 310, a memory 320, a communication unit 330, and a bus 340;
wherein the processor 310, the memory 320, and the communication unit 330 perform communication with each other through the bus 340; the communication unit 330 is configured to implement information transmission between the server-side device and the terminal device.
The processor 310 is configured to invoke a computer program in the memory 320, and when the processor executes the computer program, implement all the steps in a closed-loop DBS regulation method based on a wearable device in one of the above embodiments.
Those of ordinary skill in the art will appreciate that: the Memory may be, but is not limited to, random access Memory (Random Access Memory; RAM; ROM; programmable Read-Only Memory; PROM; erasable ROM; erasable Programmable Read-Only Memory; EPROM; electrically erasable ROM; electric Erasable Programmable Read-Only Memory; EEPROM; etc.). The memory is used for storing a program, and the processor executes the program after receiving the execution instruction. Further, the software programs and modules within the memory may also include an operating system, which may include various software components and/or drivers for managing system tasks (e.g., memory management, storage device control, power management, etc.), and may communicate with various hardware or software components to provide an operating environment for other software components.
The processor may be an integrated circuit chip with an exponential processing capability. The processor may be a general-purpose processor, including a central processing unit (Central Process ing Unit, abbreviated as CPU), a Network Processor (NP), and the like. The disclosed methods, steps, and logic blocks in the embodiments of the present application may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The present application also provides a computer readable storage medium comprising a program for performing a closed-loop DBS regulation method based on a wearable device provided by any of the foregoing method embodiments when executed by a processor.
Those of ordinary skill in the art will appreciate that: all or part of the steps for implementing the method embodiments described above may be performed by hardware associated with program instructions. The foregoing program may be stored in a computer readable storage medium. The program, when executed, performs steps including the method embodiments described above; and the aforementioned storage medium includes: various media may store program code, such as ROM, RAM, magnetic or optical disks, and the specific type of media is not limiting in this application.
The present invention is not limited to the above-mentioned embodiments, and any changes or substitutions that can be easily understood by those skilled in the art within the technical scope of the present invention are intended to be included in the scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.
Claims (9)
1. A method for closed-loop DBS regulation and control based on a wearable device, the method comprising:
collecting three-axis acceleration data, three-axis angular velocity data and pressure data of the foot, which are collected by the intelligent shoes and related to the left foot and the right foot; collecting myoelectricity data related to the upper limb and triaxial angular velocity data of the upper limb, which are collected by a myoelectricity arm ring; collecting trunk triaxial acceleration data which are collected by the wearable patch and related to the trunk;
performing feature extraction based on acceleration of the foot triaxial acceleration data in the Y-axis direction to obtain gait variation indexes; performing feature extraction based on the foot triaxial acceleration data and the foot triaxial angular velocity data to obtain a tremor index and a bradykinesia index; extracting features based on foot pressure data to obtain a balance index; performing feature extraction on the basis of the angular speed of the foot pressure data and the foot triaxial angular speed data in the Z-axis direction to obtain a freezing index; extracting features based on myoelectricity data to obtain a stiffness index; extracting features based on the triaxial angular velocity data of the upper limb to obtain an upper limb abnormal movement index; feature extraction is carried out based on the trunk triaxial acceleration data to obtain a trunk fluctuation index; extracting features based on the triaxial acceleration data of the foot to obtain a lower limb abnormal movement index;
Taking gait variation index, tremor index, bradykinesia index, balance index, freezing index and stiffness index as training characteristics, taking an artificial tag in an open period or closed period state as a training tag, and inputting the training tag into a long-period and short-period memory network for training to obtain a Parkinson model;
the upper limb abnormal movement index, the trunk abnormal movement index and the lower limb abnormal movement index are used as training characteristics, and artificial labels with abnormal movement are used as training labels and are input into a long-term and short-term memory network for training to obtain a regulation and control model;
adjusting the parameters of the DBS by utilizing the composite judgment result of the Parkinson model and the regulation model;
the adjusting the parameters of the DBS by utilizing the composite judgment result of the Parkinson model and the regulation model comprises the following steps:
defining the open-phase state of the Parkinson model as 1 and the closed-phase state as 0;
defining the regulation model to be 1 without variation and 0 with variation;
if the parkinsonism model output is 1 and the regulation model output is 1, no regulation instruction is required to be sent to the DBS;
if the output of the Parkinson model is 1 and the output of the regulation model is 0, sending a step-down instruction to the DBS;
if the Parkinson model output is 0 and the regulation model output is 1, a boosting instruction is sent to the DBS;
If the output of the Parkinson model is 0 and the output of the regulation model is 0, sending prompt information of timely medical treatment of the patient.
2. The method for closed-loop DBS regulation and control based on wearable equipment according to claim 1, wherein the feature extraction of acceleration in the Y-axis direction based on the three-axis acceleration data of the foot, and the obtaining of gait variation index comprises:
y-axis acceleration acc for respectively extracting three-axis acceleration data of foot relevant to each foot y ；
Acceleration acc of Y axis y The data segment with the value of the first threshold value continuously appearing in the data sequence is used as an abnormal data segment;
all values in the abnormal data segment are zeroed to be removed, and the removed data is subjected to filtering treatment by adopting a moving average method;
y-axis acceleration acc after screening and filtering y All the effective acceleration data segments larger than the second threshold value, calculating the length of each effective acceleration data segment and recording as L i1 ；
Extracting the maximum value of the acceleration index in each effective acceleration data segment and marking as acc max,i ；
Obtaining gait variation indexes by using a gait variation feature extraction model;
the gait variation characteristic extraction model is as follows:
wherein GVS represents gait variability index; w (w) 1 、w 2 And w 3 Representing the weight value; n is n L1 Representing the number of left foot effective acceleration data segments; n is n R1 Representing the number of right foot effective acceleration data segments; s is S m A standard deviation representing the product of the length of the effective acceleration data segment and the maximum value; l (L) m A standard deviation representing the length of each effective acceleration data segment; acc (acc) max,m A standard deviation representing a maximum value in each of the effective acceleration data segments; n (N) 1 Representing the number of valid acceleration data segments.
3. The method of claim 2, wherein the feature extraction based on the foot triaxial acceleration data and the foot triaxial angular velocity data to obtain the tremor index comprises:
calculating component acceleration acc in triaxial acceleration data of foot x 、acc y And acc (sic) z Vector sum in X, Y and Z axis direction, get vector acceleration ACC;
calculating component angular velocity gyr in foot three-axis angular velocity data x 、gyr y And gyr z Vector sum in X, Y and Z direction to obtain vector angular velocity GYR;
average value filtering processing is carried out on the vector acceleration ACC and the vector angular velocity GYR to obtain an acceleration signal ACC0 and an angular velocity signal GYR0;
counting the number of zero crossings related to the left foot in the acceleration signal ACC0 and the angular velocity signal GYR0 to obtain the number N of the acceleration zero crossings of the left foot respectively L_acc_0 And the number of zero crossing points of angular velocity N L_gyr_0 ；
Counting the number of zero crossings related to the right foot in the acceleration signal ACC0 and the angular velocity signal GYR0 to obtain the number N of the acceleration zero crossings of the right foot respectively R_acc_0 And the number of zero crossing points of angular velocity N R_gyr_0 ；
Obtaining tremor index by using a tremor characteristic extraction model;
the tremor characteristic extraction model is as follows:
wherein TRS represents tremor index; max represents the extraction maximum value; fs denotes the sampling rate.
4. The method of claim 1, wherein the performing feature extraction based on the foot triaxial acceleration data and the foot triaxial angular velocity data to obtain the bradykinesia index comprises:
component acceleration acc in foot triaxial acceleration data x 、acc y And acc (sic) z Adding to obtain total acceleration;
component angular velocity gyr in foot triaxial angular velocity data x 、gyr y And gyr z Adding to obtain a total angular velocity;
calculating first-order differential signals of the total acceleration and the total angular velocity, converting all values of the first-order differential signals into positive values, sequencing, recording data point positions of one fourth of the front ranking, calculating the average value of the total acceleration and the total angular velocity at the corresponding positions as a base line, and calculating the difference value between the total acceleration and the base line and the total angular velocity to obtain the total acceleration and the total angular velocity after primary filtering;
Smoothing the primary filtered total acceleration and total angular velocity with a moving average window of a predetermined length to obtain a secondary filtered total acceleration signal acc s And a total angular velocity signal gyr s ；
Extracting the total acceleration signal acc s And a total angular velocity signal gyr s The data length of the data segment which is larger than the third threshold value is used for obtaining a total acceleration signal acc s Data length L of data segment greater than third threshold au Sum total ofAngular velocity signal gyr s Data length L of data segment greater than third threshold gu ；
Extracting the total acceleration signal acc s And a total angular velocity signal gyr s The data length of the data segment smaller than the third threshold value is used for obtaining a total acceleration signal acc s Data length L of data segment smaller than third threshold al And a total angular velocity signal gyr s Data length L of data segment smaller than third threshold gl ；
Extracting the total acceleration signal acc s Maximum value in the data segment larger than the third threshold value is obtained to obtain the maximum value A of the acceleration m ；
Extracting the total angular velocity signal gyr s Maximum value in data segment greater than the third threshold value, obtaining maximum value G of angular velocity m ；
Obtaining a bradykinesia index by using a bradykinesia feature extraction model;
the bradykinesia feature extraction model is:
wherein BRS represents the bradykinesia index; ln represents taking the logarithm.
5. The method of claim 1, wherein the performing feature extraction based on foot pressure data to obtain a balance index comprises:
first pressure threshold TH for fitting foot pressure data using pressure threshold configuration model 1 ；
The pressure threshold configuration model is:
wherein N is 2 The amount of foot pressure data; p (P) i Data representing an ith foot pressure; p (P) min Representing a minimum value in the foot pressure data;
rejecting foot pressure data greater than a first pressure threshold TH 1 The length of the data segment is larger than twice the sampling frequency, the effective foot pressure data segment is obtained, and the length of the effective foot pressure data segment is recorded as L i2 ；
Extracting the maximum pressure value of the foot pressure data in each effective foot pressure data segment and marking as P max,i ；
Obtaining a balance index by using a balance characteristic extraction model;
the balance characteristic extraction model is as follows:
wherein BAS represents a balance index; ln represents taking the logarithm; n is n L2 Representing the number of effective foot pressure data segments of the left foot; n is n R2 The number of effective foot pressure data segments for the right foot is indicated.
6. The method of claim 1, wherein the feature extraction based on the angular velocity of the foot pressure data and the foot triaxial angular velocity data in the Z-axis direction, to obtain the freezing index, comprises:
Configuring the second pressure threshold TH using a sliding average method based on foot pressure data 2 ；
The foot pressure data is greater than the second pressure threshold TH 2 Is marked as 0 for the walk state and the remaining data points are marked as 1 for the stop state;
extracting angular velocity gyr of foot three-axis angular velocity data corresponding to a data segment marked as 0 in the Z-axis direction z Finally, the angle change quantity at the moment is used for obtaining the turning angle;
judging the angle which is larger than a fourth threshold value and the number of times of continuous occurrence exceeds three in the turning angle as a turning process, and connecting the turning angles of the left foot and the right foot in the same turning process into a new angle vector EULR;
obtaining a freezing index by utilizing a freezing characteristic extraction model;
the freezing characteristic extraction model is as follows:
wherein FOS is a freeze index; w (w) 4 And w 5 Is a weight value; n (N) 3 Is the number of turns; EULR (EULR) i Angle vector of the ith turn; EULR (EULR) m Is the average of the angle vectors of all turns.
7. The method of claim 1, wherein the performing feature extraction based on myoelectricity data to obtain a stiffness index comprises:
carrying out band-pass filtering on myoelectric data by adopting a Butterworth filter to obtain pure myoelectric signals EMG;
All the values of the electromyographic signals EMG are converted into positive values, and average filtering is carried out by using a sliding window with a preset length to obtain an electromyographic signal envelope curve ENV;
extracting data segments which are larger than half of average value of electromyographic signal envelope curve ENV in electromyographic signal EMG, and obtaining the number n of the data segments p Length L of each data segment n The mean value m of each data segment n ；
Fourier transforming the electromyographic signal EMG to obtain an electromyographic spectrum signal F;
the myoelectricity spectrum signals of 100-200Hz and 20-100Hz are summed respectively to correspondingly obtain a first energy value E of the signals of two frequency bands 1 And a second energy value E 2 ；
Obtaining a rigidity index by using a rigidity characteristic extraction model;
the rigidity characteristic extraction model is as follows:
wherein RIS represents a stiffness index; k represents the number of channels for collecting electromyographic signals; fs denotes the sampling rate.
8. The method for closed-loop DBS regulation and control based on wearable equipment according to claim 1, wherein the feature extraction based on the upper limb triaxial angular velocity data comprises:
respectively calculating the angular speed integral of the upper limb triaxial angular speed data in three directions to obtain angle signals in the three directions;
extracting the maximum value EU of each direction max And an angle minimum value EU min Number of corresponding angle maxima n max And an angle minimum number n min ；
Calculating adjacent two angle maxima EU in time sequence max Or an angle minimum value EU min The absolute value of the difference value between the two directions is obtained and the average value is obtained to obtain the average value EUM of the maximum angle value corresponding to each direction max Sum angle minimum mean EUM min ；
Obtaining an upper limb abnormality index by using an upper limb abnormality extraction model;
the upper limb abnormal extraction model is as follows:
wherein AMS u Indicating an upper limb movement index;representing the average value of the maximum value of the angle of the X axis direction of the upper limb; />Representing the average value of the minimum value of the X-axis angle of the upper limb; />Representing the average value of the maximum value of the angle in the Y-axis direction of the upper limb; />Representing the average value of the minimum value of the Y-axis angle of the upper limb; />Representing the average value of the maximum value of the Z-axis direction angle of the upper limb; />Representing the average value of the minimum value of the Z-axis angle of the upper limb;
the feature extraction is performed based on the trunk triaxial acceleration data, and the trunk fluctuation index obtaining comprises the following steps:
smoothing the acceleration signals of each direction of the trunk triaxial acceleration data by using a moving average window with a preset length, and then calculating acceleration integrals in three directions to obtain speed signals in three directions;
extracting the maximum value VE of the speed signal in each direction max Number of corresponding maximum speed values n max ；
Calculating adjacent velocity maxima VE in time sequence max The absolute value of the difference value between the two is taken and the average value is obtained to obtain the average value VEM of the maximum speed value corresponding to each direction max ；
Obtaining a trunk fluctuation index by using a trunk fluctuation extraction model;
the trunk transaction extraction model is as follows:
wherein AMS m The index of the abnormal movement of the trunk is expressed,representing the average value of the maximum speed value of the trunk in the X-axis direction; />Representing the average value of the maximum speed value of the trunk in the Y-axis direction; />Representing the average value of the maximum speed value of the trunk in the Z-axis direction;
the feature extraction is performed based on the triaxial acceleration data of the foot, and the obtaining of the lower limb abnormal index comprises the following steps:
smoothing the acceleration signals of the foot triaxial acceleration data in each direction by using a moving average window with a preset length, and then calculating acceleration integrals in three directions to obtain speed signals in the three directions;
extracting the maximum value VE of the speed signal in each direction max Number of corresponding maximum speed values n max ；
Calculating adjacent velocity maxima VE in time sequence max The absolute value of the difference value between the two is taken and the average value is obtained to obtain the average value VEM of the maximum speed value corresponding to each direction max ；
Obtaining a lower limb abnormality index by using a lower limb abnormality extraction model;
The lower limb abnormal extraction model is as follows:
wherein AMS l The index of the lower limb movement is represented,representing the average value of the maximum speed values of the lower limb in the X-axis direction; />Representing the average value of the maximum speed values of the lower limbs in the Y-axis direction; />The average value of the maximum speed value in the Z-axis direction of the lower limb is shown.
9. A closed-loop DBS regulation and control system based on a wearable device, the system comprising:
and a data collection module: the intelligent shoe is used for collecting three-axis acceleration data, three-axis angular velocity data and pressure data of the foot, which are collected by the intelligent shoe and are related to the left foot and the right foot; collecting myoelectricity data related to the upper limb and triaxial angular velocity data of the upper limb, which are collected by a myoelectricity arm ring; collecting trunk triaxial acceleration data which are collected by the wearable patch and related to the trunk;
and the feature extraction module is used for: the method comprises the steps of performing feature extraction on acceleration in the Y-axis direction based on three-axis acceleration data of the foot to obtain gait variation indexes; performing feature extraction based on the foot triaxial acceleration data and the foot triaxial angular velocity data to obtain a tremor index and a bradykinesia index; extracting features based on foot pressure data to obtain a balance index; performing feature extraction on the basis of the angular speed of the foot pressure data and the foot triaxial angular speed data in the Z-axis direction to obtain a freezing index; extracting features based on myoelectricity data to obtain a stiffness index; extracting features based on the triaxial angular velocity data of the upper limb to obtain an upper limb abnormal movement index; feature extraction is carried out based on the trunk triaxial acceleration data to obtain a trunk fluctuation index; extracting features based on the triaxial acceleration data of the foot to obtain a lower limb abnormal movement index;
A first model training module: the method comprises the steps of taking gait variation index, tremor index, bradykinesia index, balance index, freezing index and stiffness index as training characteristics, taking an artificial tag in an open phase or closed phase state as a training tag, and inputting the training tag into a long-term and short-term memory network for training to obtain a Parkinson model;
and a second model training module: the method comprises the steps of taking an upper limb abnormal index, a trunk abnormal index and a lower limb abnormal index as training characteristics, taking an artificial tag with abnormal conditions as a training tag, and inputting the training tag into a long-term and short-term memory network for training to obtain a regulation and control model;
DBS regulation and control module: the method is used for adjusting the parameters of the DBS by utilizing the composite judgment result of the Parkinson model and the regulation model; the adjusting the parameters of the DBS by utilizing the composite judgment result of the Parkinson model and the regulation model comprises the following steps:
defining the open-phase state of the Parkinson model as 1 and the closed-phase state as 0;
defining the regulation model to be 1 without variation and 0 with variation;
if the parkinsonism model output is 1 and the regulation model output is 1, no regulation instruction is required to be sent to the DBS;
if the output of the Parkinson model is 1 and the output of the regulation model is 0, sending a step-down instruction to the DBS;
If the Parkinson model output is 0 and the regulation model output is 1, a boosting instruction is sent to the DBS;
if the output of the Parkinson model is 0 and the output of the regulation model is 0, sending prompt information of timely medical treatment of the patient.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311428526.5A CN117298449B (en) | 2023-10-31 | 2023-10-31 | Closed-loop DBS regulation and control method and system based on wearable equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311428526.5A CN117298449B (en) | 2023-10-31 | 2023-10-31 | Closed-loop DBS regulation and control method and system based on wearable equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117298449A CN117298449A (en) | 2023-12-29 |
| CN117298449B true CN117298449B (en) | 2024-04-09 |
Family
ID=89246424
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311428526.5A Active CN117298449B (en) | 2023-10-31 | 2023-10-31 | Closed-loop DBS regulation and control method and system based on wearable equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117298449B (en) |
Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106377837A (en) * | 2016-09-19 | 2017-02-08 | 天津大学 | Functional muscle electrical stimulation walk-assisting device based on gait recognition and control method |
| CN206197948U (en) * | 2016-07-19 | 2017-05-31 | 泉州装备制造研究所 | Quantify the wearable smart machine of detection Parkinson's motor symptoms |
| CN106725504A (en) * | 2016-11-24 | 2017-05-31 | 苏州大学附属第二医院 | The wearable device and its monitoring method of a kind of multinode motion monitoring |
| CN108629304A (en) * | 2018-04-26 | 2018-10-09 | 深圳市臻络科技有限公司 | A kind of freezing of gait online test method |
| CN108742612A (en) * | 2018-05-24 | 2018-11-06 | 杭州航弈生物科技有限责任公司 | DBS validation checking equipment based on myoelectricity marker |
| CN110946556A (en) * | 2019-12-27 | 2020-04-03 | 南京信息工程大学 | Parkinson resting state tremor evaluation method based on wearable somatosensory network |
| CN111528842A (en) * | 2020-05-26 | 2020-08-14 | 复嶂环洲生物科技(上海)有限公司 | Quantitative assessment method for Parkinson disease symptoms based on physiological and behavioral indexes |
| CN112617807A (en) * | 2020-12-31 | 2021-04-09 | 中国科学院合肥物质科学研究院 | Device and method for preventing and relieving frozen gait of Parkinson disease patient |
| CN112656406A (en) * | 2021-01-27 | 2021-04-16 | 山东农业大学 | Wearable sensor-based lower limb movement detection method for Parkinson's disease |
| CN112691292A (en) * | 2021-01-12 | 2021-04-23 | 天津大学 | Parkinson closed-loop deep brain stimulation system based on wearable intelligent equipment |
| CN215651118U (en) * | 2021-06-01 | 2022-01-28 | 江苏省人民医院（南京医科大学第一附属医院） | Parkinson's disease patient dyskinesia recognition device |
| CN113975634A (en) * | 2021-12-08 | 2022-01-28 | 杭州范斯凯科技有限公司 | Left and right hand wearing identification method and device, wrist strap medical equipment and storage medium |
| CN114246596A (en) * | 2020-09-11 | 2022-03-29 | 首都医科大学附属北京天坛医院 | Evaluation method applied to DBS for treating curative effect of Parkinson disease patient |
| CN114259651A (en) * | 2022-01-17 | 2022-04-01 | 天津大学 | Active real-time closed-loop electrical stimulation system for Parkinson's disease |
| CN115486815A (en) * | 2022-06-27 | 2022-12-20 | 青岛大学 | Real-time analysis system and method for three-dimensional posture of Parkinson's disease hand tremor |
| WO2023108315A1 (en) * | 2021-12-13 | 2023-06-22 | 中国科学院深圳先进技术研究院 | Gait analysis and anomaly detection method based on wearable step sensor |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8702629B2 (en) * | 2005-03-17 | 2014-04-22 | Great Lakes Neuro Technologies Inc. | Movement disorder recovery system and method for continuous monitoring |
| US20080053253A1 (en) * | 2006-09-06 | 2008-03-06 | Individual Monitoring Systems, Inc | Fully ambulatory, self-contained gait monitor |
| US9238142B2 (en) * | 2012-09-10 | 2016-01-19 | Great Lakes Neurotechnologies Inc. | Movement disorder therapy system and methods of tuning remotely, intelligently and/or automatically |
| WO2021234228A1 (en) * | 2020-05-20 | 2021-11-25 | Feetme | Insole and stimulation method |
-
2023
- 2023-10-31 CN CN202311428526.5A patent/CN117298449B/en active Active
Patent Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN206197948U (en) * | 2016-07-19 | 2017-05-31 | 泉州装备制造研究所 | Quantify the wearable smart machine of detection Parkinson's motor symptoms |
| CN106377837A (en) * | 2016-09-19 | 2017-02-08 | 天津大学 | Functional muscle electrical stimulation walk-assisting device based on gait recognition and control method |
| CN106725504A (en) * | 2016-11-24 | 2017-05-31 | 苏州大学附属第二医院 | The wearable device and its monitoring method of a kind of multinode motion monitoring |
| CN108629304A (en) * | 2018-04-26 | 2018-10-09 | 深圳市臻络科技有限公司 | A kind of freezing of gait online test method |
| CN108742612A (en) * | 2018-05-24 | 2018-11-06 | 杭州航弈生物科技有限责任公司 | DBS validation checking equipment based on myoelectricity marker |
| CN110946556A (en) * | 2019-12-27 | 2020-04-03 | 南京信息工程大学 | Parkinson resting state tremor evaluation method based on wearable somatosensory network |
| CN111528842A (en) * | 2020-05-26 | 2020-08-14 | 复嶂环洲生物科技(上海)有限公司 | Quantitative assessment method for Parkinson disease symptoms based on physiological and behavioral indexes |
| CN114246596A (en) * | 2020-09-11 | 2022-03-29 | 首都医科大学附属北京天坛医院 | Evaluation method applied to DBS for treating curative effect of Parkinson disease patient |
| CN112617807A (en) * | 2020-12-31 | 2021-04-09 | 中国科学院合肥物质科学研究院 | Device and method for preventing and relieving frozen gait of Parkinson disease patient |
| CN112691292A (en) * | 2021-01-12 | 2021-04-23 | 天津大学 | Parkinson closed-loop deep brain stimulation system based on wearable intelligent equipment |
| CN112656406A (en) * | 2021-01-27 | 2021-04-16 | 山东农业大学 | Wearable sensor-based lower limb movement detection method for Parkinson's disease |
| CN215651118U (en) * | 2021-06-01 | 2022-01-28 | 江苏省人民医院（南京医科大学第一附属医院） | Parkinson's disease patient dyskinesia recognition device |
| CN113975634A (en) * | 2021-12-08 | 2022-01-28 | 杭州范斯凯科技有限公司 | Left and right hand wearing identification method and device, wrist strap medical equipment and storage medium |
| WO2023108315A1 (en) * | 2021-12-13 | 2023-06-22 | 中国科学院深圳先进技术研究院 | Gait analysis and anomaly detection method based on wearable step sensor |
| CN114259651A (en) * | 2022-01-17 | 2022-04-01 | 天津大学 | Active real-time closed-loop electrical stimulation system for Parkinson's disease |
| CN115486815A (en) * | 2022-06-27 | 2022-12-20 | 青岛大学 | Real-time analysis system and method for three-dimensional posture of Parkinson's disease hand tremor |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117298449A (en) | 2023-12-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11160979B1 (en) | Method and system for tuning of movement disorder therapy devices | |
| EP3856330B1 (en) | Predictive therapy neurostimulation systems | |
| US20210077032A1 (en) | Classifying seizures as epileptic or non-epileptic using extra-cerebral body data | |
| Lauer et al. | Application of a neuro-fuzzy network for gait event detection using electromyography in the child with cerebral palsy | |
| CN102626538B (en) | Sleep detection-based integrated closed loop deep brain stimulator | |
| CN112353368B (en) | Multi-input signal epileptic seizure detection system based on feedback adjustment | |
| WO2010126625A1 (en) | Patient state detection based on support vector machine based algorithm | |
| WO2004034881A2 (en) | Signal quality monitoring and control for a medical device system | |
| CN115463344A (en) | Computer readable storage medium | |
| Patel et al. | Analysis of the severity of dyskinesia in patients with Parkinson's disease via wearable sensors | |
| CN112691292B (en) | Parkinson closed-loop deep brain stimulation system based on wearable intelligent equipment | |
| CN113577559B (en) | Closed-loop deep brain stimulation device, system and equipment based on multiple signals | |
| CN105125186B (en) | A kind of method and system of definite therapeutic intervention mode | |
| CN114259651A (en) | Active real-time closed-loop electrical stimulation system for Parkinson's disease | |
| CN117298449B (en) | Closed-loop DBS regulation and control method and system based on wearable equipment | |
| CN116312951B (en) | Exercise function assessment method and system based on multi-modal coupling analysis | |
| Powell et al. | Parkinson's disease is associated with greater regularity of repetitive voluntary movements. | |
| CN116313029B (en) | Method, system and device for dynamic control optimization of digital acupuncture | |
| US11918810B1 (en) | Method and system for tuning of movement disorder therapy devices | |
| CN105310695A (en) | Dyskinesia assessment equipment | |
| CN117017234B (en) | Multi-parameter integrated anesthesia monitoring and analyzing system | |
| Hortal et al. | Starting and finishing gait detection using a BMI for spinal cord injury rehabilitation | |
| US20230181094A1 (en) | Classifying seizures as epileptic or non-epileptic using extra-cerebral body data | |
| CN110432888B (en) | Screening method of vestibular nerve weight information coding neurons | |
| CN117323537A (en) | Helmet for mental state adjustment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |