CN115620853A - Model training method for TMS strategy automatic selection, automatic selection method and system - Google Patents
Model training method for TMS strategy automatic selection, automatic selection method and system Download PDFInfo
- Publication number
- CN115620853A CN115620853A CN202211091331.1A CN202211091331A CN115620853A CN 115620853 A CN115620853 A CN 115620853A CN 202211091331 A CN202211091331 A CN 202211091331A CN 115620853 A CN115620853 A CN 115620853A
- Authority
- CN
- China
- Prior art keywords
- tms
- patient
- strategy
- index
- electrophysiological
- 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.)
- Pending
Links
- 238000012549 training Methods 0.000 title claims abstract description 117
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000010187 selection method Methods 0.000 title claims abstract description 14
- 230000003925 brain function Effects 0.000 claims abstract description 81
- 238000007637 random forest analysis Methods 0.000 claims abstract description 58
- 230000000694 effects Effects 0.000 claims abstract description 34
- 238000003066 decision tree Methods 0.000 claims abstract description 25
- 238000005070 sampling Methods 0.000 claims abstract description 11
- 239000000523 sample Substances 0.000 claims description 71
- 230000033001 locomotion Effects 0.000 claims description 49
- 210000004556 brain Anatomy 0.000 claims description 39
- 230000004913 activation Effects 0.000 claims description 27
- 230000003902 lesion Effects 0.000 claims description 24
- 230000000638 stimulation Effects 0.000 claims description 24
- 210000003710 cerebral cortex Anatomy 0.000 claims description 20
- 238000011156 evaluation Methods 0.000 claims description 18
- 230000006870 function Effects 0.000 claims description 17
- 108010054147 Hemoglobins Proteins 0.000 claims description 16
- 102000001554 Hemoglobins Human genes 0.000 claims description 16
- 230000000763 evoking effect Effects 0.000 claims description 12
- 206010019468 Hemiplegia Diseases 0.000 claims description 11
- 238000001514 detection method Methods 0.000 claims description 11
- 239000013610 patient sample Substances 0.000 claims description 10
- 210000004720 cerebrum Anatomy 0.000 claims description 8
- 230000001502 supplementing effect Effects 0.000 claims description 8
- 238000012360 testing method Methods 0.000 claims description 8
- 238000012935 Averaging Methods 0.000 claims description 6
- 210000001364 upper extremity Anatomy 0.000 claims description 6
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 claims description 4
- 239000008280 blood Substances 0.000 claims description 4
- 210000004369 blood Anatomy 0.000 claims description 4
- 238000002790 cross-validation Methods 0.000 claims description 4
- 229910052760 oxygen Inorganic materials 0.000 claims description 4
- 239000001301 oxygen Substances 0.000 claims description 4
- 230000009466 transformation Effects 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000011491 transcranial magnetic stimulation Methods 0.000 description 107
- 210000004092 somatosensory cortex Anatomy 0.000 description 16
- 210000000976 primary motor cortex Anatomy 0.000 description 15
- 230000001976 improved effect Effects 0.000 description 10
- 208000006011 Stroke Diseases 0.000 description 8
- 230000008901 benefit Effects 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 210000000337 motor cortex Anatomy 0.000 description 4
- 210000003813 thumb Anatomy 0.000 description 4
- 230000002146 bilateral effect Effects 0.000 description 3
- 238000004590 computer program Methods 0.000 description 3
- 230000006378 damage Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 201000010099 disease Diseases 0.000 description 3
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000007659 motor function Effects 0.000 description 3
- 210000003205 muscle Anatomy 0.000 description 3
- 238000007792 addition Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000002329 infrared spectrum Methods 0.000 description 2
- 210000005036 nerve Anatomy 0.000 description 2
- 230000001953 sensory effect Effects 0.000 description 2
- 208000012661 Dyskinesia Diseases 0.000 description 1
- 206010061296 Motor dysfunction Diseases 0.000 description 1
- 108010064719 Oxyhemoglobins Proteins 0.000 description 1
- 208000027418 Wounds and injury Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 210000005013 brain tissue Anatomy 0.000 description 1
- 230000002490 cerebral effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 210000003414 extremity Anatomy 0.000 description 1
- 210000004247 hand Anatomy 0.000 description 1
- 210000003128 head Anatomy 0.000 description 1
- 230000001939 inductive effect Effects 0.000 description 1
- 208000014674 injury Diseases 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 210000000578 peripheral nerve Anatomy 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000007634 remodeling Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 210000003625 skull Anatomy 0.000 description 1
- 230000004936 stimulating effect Effects 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
- G16H10/60—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61N—ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY
- A61N2/00—Magnetotherapy
- A61N2/004—Magnetotherapy specially adapted for a specific therapy
- A61N2/006—Magnetotherapy specially adapted for a specific therapy for magnetic stimulation of nerve tissue
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61N—ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY
- A61N2/00—Magnetotherapy
- A61N2/02—Magnetotherapy using magnetic fields produced by coils, including single turn loops or electromagnets
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
Landscapes
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Public Health (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Primary Health Care (AREA)
- Epidemiology (AREA)
- Radiology & Medical Imaging (AREA)
- Life Sciences & Earth Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Veterinary Medicine (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Neurology (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention provides a model training method, an automatic selection method and a system for TMS strategy automatic selection, comprising the following steps: acquiring a first training sample set, wherein each sample comprises basic information, brain function indexes, electrophysiological indexes of a single patient and a TMS strategy with a good rehabilitation effect; carrying out N times of putting back random sampling on the first training sample set to form N second training sample sets with the number consistent with that of the first training sample set; and training by using the N second training sample sets to generate N decision trees, and constructing a random forest model for automatically selecting the TMS strategy. And inputting the basic information, the brain function index and the electrophysiological index of the real patient into the obtained random forest model to obtain the TMS strategy of the patient. The invention can make an individualized TMS strategy based on the brain function index and the electrophysiological index of the patient according to the illness state of the patient.
Description
Technical Field
The invention relates to the field of nerve regulation and rehabilitation, in particular to a model training method, an automatic selection method and a system for TMS strategy automatic selection.
Background
In recent years, the incidence rate of stroke is higher and higher, the number of newly diagnosed stroke is basically up to 250 ten thousand people every year, and the number is rising year by year, wherein 55-75% of stroke patients show dyskinesia. The traditional rehabilitation training method can promote the motor function recovery to a certain extent, but has high requirements on the professional ability of a rehabilitation therapist and the active cooperation degree of a patient.
According to the current evidence-based medical data, transcranial Magnetic Stimulation (TMS) has the effect of improving the motor function of stroke patients. The transcranial magnetic stimulation technology is a painless and noninvasive green treatment method, magnetic signals can penetrate through the skull without attenuation to stimulate cerebral nerves, and in practical application, the magnetic stimulation technology is not limited to stimulation of the head, and peripheral nerve and muscles can also stimulate the magnetic stimulation technology, so the magnetic stimulation technology is called as magnetic stimulation.
Although TMS can be developed in the early stage of stroke and does not have the requirement on the active coordination ability of a patient, the selection of the stimulation strategy in the current clinic mainly depends on the experience of a doctor, is fuzzy in selection basis and single in stimulation strategy, and is influenced by multiple factors such as the cause of disease, the injury degree, the course of disease, the focus position and the like after stroke. More importantly, the improvement effect difference of the motor function of the patient receiving TMS is large, meanwhile, the root cause of the motor dysfunction left by the patient with stroke is the brain tissue function damage of the lesion part, the essence of limb rehabilitation is the brain function remodeling, the TMS strategy should be selected according to the brain function state of each patient, and therefore, the universal TMS strategy cannot meet the personalized requirement.
Disclosure of Invention
In view of this, embodiments of the present invention provide a model training method, an automatic selection method, and a system for TMS policy automatic selection, so as to eliminate or improve one or more defects in the prior art, and solve the problems that the selection of the conventional universal TMS policy is based on ambiguity, the TMS policy has a single type, and cannot meet personalized requirements.
In one aspect, the invention provides a model training method for TMS strategy automatic selection, which comprises the following steps:
acquiring a first training sample set, wherein the first training sample set comprises a plurality of patient samples, and each patient sample comprises basic information, a brain function index, an electrophysiological index and a corresponding TMS strategy of a corresponding patient; wherein the basic information comprises the lesion type, the degree of hemiplegia, the lesion position and Fugl-Meyer scale evaluation data of the patient; the brain function index comprises the activation degree and the lateral deviation index of each brain area of the patient and the function connection index inside and between every two partial brain areas; the electrophysiological indexes comprise cerebral cortex movement threshold, movement evoked potential and latency of the patient;
performing a set number of times of replaced random sampling on the first training sample set to form a set number of second training sample sets consistent with the number of samples in the first training sample set;
and respectively generating the set number of decision trees by using the set number of second training sample sets to construct a random forest model for automatically selecting the TMS strategy, and taking a final prediction result obtained by the decision result of each decision tree according to a voting principle as the output of the random forest model.
In some embodiments of the present invention, the method adopts a k-fold cross validation method to construct a random forest model and test, including:
and dividing the first training sample set into k parts in a form of not putting back random samples, wherein k-1 part is used as a second training sample set and is used for training the random forest model, and the rest 1 part is used for testing the random forest model.
In some embodiments of the present invention, before obtaining the first set of training samples, further comprising:
calculating a Pearson correlation coefficient of the near-infrared brain function index of the patient corresponding to the sample and evaluation data of the Fugl-Meyer scale, and taking the near-infrared brain function index of which the Pearson correlation coefficient of the upper limb function score in the evaluation data of the Fugl-Meyer scale is larger than a first preset value as the brain function index of the corresponding sample; wherein the brain function index is a subset of the near-infrared brain function index.
In some embodiments of the present invention, the brain function index is calculated according to the oxygenated hemoglobin signals of each area of cerebral cortex of a patient sample in a motion detection state collected by a near infrared spectrometer;
the activation degree of each brain region is obtained by averaging the amplitudes of the channels of each brain region after averaging the amplitudes of the acquisition channels in the time domain according to the oxygenated hemoglobin signal;
the laterality index is calculated by dividing the difference of the activation degrees of the dominant cerebral hemisphere and the affected cerebral hemisphere by the sum of the activation degrees of the dominant cerebral hemisphere and the affected cerebral hemisphere;
the functional connection index inside the brain region and between every two channels is represented by wavelet phase coherence between every two channels, which is obtained by phase information calculation after the oxygenated hemoglobin signals of all the acquisition channels are subjected to continuous complex wavelet transformation.
In some embodiments of the invention, the electrophysiological index is measured for each patient during an electrophysiological examination using a single pulse TMS from a magnetic stimulator.
In some embodiments of the invention, the TMS strategy includes a stimulation pulse frequency and a stimulation target.
On the other hand, the invention provides a TMS strategy automatic selection method, which comprises the following steps:
acquiring basic information, brain function indexes and electrophysiological indexes of a patient to be selected with the TMS strategy; wherein the basic information comprises the type of lesion, degree of hemiplegia, lesion location and initial Fugl-Meyer scale assessment data of the patient; the brain function index comprises the activation degree and the lateral deviation index of each brain area of the patient and the function connection index between every two brain areas; the electrophysiological indexes comprise cerebral cortex movement threshold, movement evoked potential and latency of the patient;
inputting the basic information, the brain function index and the electrophysiological index into a random forest model in a model training method of TMS strategy automatic selection according to any one of claims 1 to 6 to obtain the TMS strategy of the patient.
In some embodiments of the present invention, the TMS policy automatic selection method further comprises:
and supplementing the basic information, the brain function index, the electrophysiological index and the corresponding TMS strategy of the patient with the rehabilitation effect higher than the preset standard into the first training sample set, and repeating the step of the model training method automatically selected by the TMS strategy as mentioned above to construct a new random forest model.
In some embodiments of the present invention, supplementing the basic information, the brain function index, the electrophysiological index and the corresponding TMS strategy of the patient with a rehabilitation effect higher than a preset standard into the first training sample set further comprises:
when the number of the patients with the accumulated rehabilitation effect higher than the preset standard is larger than a second preset value, supplementing the basic information, the brain function index, the electrophysiological index and the corresponding TMS strategy of each patient into the first training sample set;
and when the number of the patients is less than a second preset value, only recording the basic information, the brain function index, the electrophysiological index and the corresponding TMS strategy of each patient.
On the other hand, the invention also provides a TMS strategy automatic selection system, which comprises:
the basic information input module is used for inputting basic information of a patient to be subjected to TMS strategy selection, including lesion type, hemiplegia degree, lesion position and Fugl-Meyer scale evaluation data of the patient, and transmitting the basic information to the intelligent learning module;
the near-infrared brain blood oxygen data acquisition module is used for acquiring data of oxygenated hemoglobin signals of each area of a cerebral cortex of the patient in a motion detection state and transmitting the data to the near-infrared brain function evaluation module;
the near-infrared brain function evaluation module is used for calculating the activation degree and the lateral deviation index of each brain area and the function connection index between every two brain areas according to the acquired oxygenated hemoglobin signals to obtain the brain function index of the patient and transmitting the brain function index to the intelligent learning module;
the electrophysiological examination module is used for carrying out electrophysiological examination on the patient by adopting the monopulse TMS, measuring and collecting a movement threshold, a movement evoked potential and a latency of the cerebral cortex of the patient to obtain electrophysiological indexes of the patient, and transmitting the electrophysiological indexes to the intelligent learning module;
an intelligent learning module for performing the steps of the TMS strategy automatic selection method as mentioned hereinbefore;
the TMS strategy display module is used for displaying the TMS strategy output by the intelligent learning module;
and the feedback module is used for feeding back the rehabilitation effect recommended by the random forest model after the TMS strategy is adopted and transmitting the patient information with the rehabilitation effect higher than the preset standard to the intelligent learning module.
The invention has the beneficial effects that:
the invention provides a model training method, an automatic selection method and a system for TMS strategy automatic selection. And inputting the basic information, the brain function index and the electrophysiological index of the real patient into the obtained random forest model to obtain the TMS strategy of the patient. The invention can make an individualized TMS strategy based on the brain function index and the electrophysiological index of the patient according to the illness state of the patient.
Furthermore, the patient with the TMS strategy obtained through the random forest model is tracked for a set time length, the rehabilitation effect of the patient treated by the TMS strategy is observed, basic information, brain function indexes and electrophysiological indexes of the patient with the rehabilitation effect higher than a preset standard and corresponding TMS strategies are supplemented into the first training sample set, the step of the model training method for automatically selecting the TMS strategy is carried out again, a new random forest model is built, the accuracy of random forest model selection is improved, the clinical rehabilitation curative effect of the patient is effectively improved, and the rehabilitation efficiency is improved.
Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and drawings.
It will be appreciated by those skilled in the art that the objects and advantages that can be achieved with the present invention are not limited to what has been particularly described hereinabove, and that the above and other objects that can be achieved with the present invention will be more clearly understood from the following detailed description.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the principles of the invention. In the drawings:
fig. 1 is a schematic diagram illustrating steps of a model training method for TMS strategy automatic selection according to an embodiment of the present invention.
Fig. 2 is a schematic structural diagram of a random forest model in an embodiment of the present invention.
Fig. 3 is a schematic structural diagram of a TMS policy automatic selection system in an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the following embodiments and accompanying drawings. The exemplary embodiments and descriptions of the present invention are provided to explain the present invention, but not to limit the present invention.
It should be noted that, in order to avoid obscuring the present invention with unnecessary details, only the structures and/or processing steps closely related to the scheme according to the present invention are shown in the drawings, and other details not so relevant to the present invention are omitted.
It should be emphasized that the term "comprises/comprising" when used herein, is taken to specify the presence of stated features, elements, steps or components, but does not preclude the presence or addition of one or more other features, elements, steps or components.
It is also noted herein that the term "coupled," if not specifically stated, may refer herein to not only a direct connection, but also an indirect connection in which an intermediate is present.
Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the drawings, the same reference numerals denote the same or similar parts, or the same or similar steps.
It should be emphasized that the step labels mentioned in the following are not limitations to the order of steps, but should be understood that the steps may be executed in the order mentioned in the embodiments, may be executed in a different order from the embodiments, or may be executed simultaneously.
In order to solve the problems that the selection of the conventional universal TMS strategy is fuzzy, the TMS strategy is single in type and cannot meet individual requirements, the invention provides a model training method for automatically selecting the TMS strategy, which comprises the following steps of S101-S103 as shown in FIG. 1:
step S101: acquiring a first training sample set, wherein the first training sample set comprises a plurality of patient samples, and each patient sample comprises basic information, a brain function index, an electrophysiological index and a corresponding TMS strategy of a corresponding patient; the basic information comprises the lesion type, the degree of hemiplegia, the lesion position and Fugl-Meyer scale evaluation data of the patient; the brain function index comprises the activation degree and the lateral deviation index of each brain region of the patient and the function connection index inside and between every two partial brain regions; the electrophysiological metrics include cerebral cortex motor threshold, motor evoked potential and latency of the patient.
Step S102: and randomly sampling the first training sample set for a set number of times with the set back to form a set number of second training sample sets consistent with the number of samples in the first training sample set.
Step S103: and respectively generating a set number of decision trees by using a set number of second training sample sets to construct a random forest model for automatically selecting the TMS strategy, and taking a final prediction result obtained by the decision result of each decision tree according to a voting principle as the output of the random forest model.
In step S101, a first training sample set for model training is obtained, where each sample includes basic information of a single patient, a brain function index, an electrophysiological index, and a corresponding TMS strategy.
Specifically, the method for constructing the first training sample set includes the following steps S1011 to S1013:
step S1011: basic information of a plurality of patients is recorded, wherein the basic information at least comprises the lesion type, the degree of hemiplegia, the lesion position and the initial Fugl-Meyer scale of each patient, and in some embodiments, the basic information of the patients also comprises the identity information of the age, the sex and the like of the patients.
Step S1012: the brain function indexes of each patient in the motion detection state are collected and calculated, and the electrophysiological indexes of each patient are collected by using a magnetic stimulator.
In some embodiments, the brain function index is calculated from the oxygenated hemoglobin signals of various regions of the cerebral cortex in the motion detection state of a patient sample collected by the NIR spectrometer.
Wherein Near Infrared (NIR) is an electromagnetic wave between visible light and mid-Infrared light, and the wavelength range of the Near Infrared spectrum defined by American Society for Testing and Materials (ASTM) is 780 to 2526nm.
In some embodiments, the motion detection state is a task-oriented training performed by each patient using the upper limb of the affected side, and a passive training mode, an assisted training mode, and an active training mode can be selected according to the functional state of the upper limb of each patient, wherein the affected refers to involvement, and in the present invention, the stroke-related disease involves the cerebral cortex.
The cerebral cortex is divided into 6 areas called brain areas, namely a left primary motor cortex, a right primary motor cortex, a left motor anterior area, a right motor anterior area, a left somatosensory cortex and a right somatosensory cortex. And acquiring local oxygenated hemoglobin concentration change signal data of each cerebral cortex area of each patient in a motion detection state by using a near infrared spectrum measuring instrument. And respectively calculating the activation degree and the laterality index of each brain area of the patient and the function connection index inside and between every two partial brain areas according to the oxygenated hemoglobin signals of each cerebral cortex area so as to obtain the required brain function index.
Specifically, the calculation method of the activation degree index of each brain region is as follows: and carrying out continuous complex wavelet transformation on the oxygenated hemoglobin signals of each acquisition channel to obtain corresponding time-frequency domain signals, averaging the signals in a time domain to obtain wavelet amplitudes of each acquisition channel, and averaging the wavelet amplitudes of the channels in each brain region to obtain the activation degree of each brain region.
The laterality index is calculated by dividing the difference between the activation degrees of the dominant and affected hemispheres by the sum of the activation degrees of the dominant and affected hemispheres.
The method for calculating the function connection index between every two brain areas comprises the following steps: phase information obtained after continuous complex wavelet transformation is carried out on oxygenated hemoglobin signals of all the acquisition channels is used for calculating wavelet phase coherence between every two channels, and brain function connection between every two channels is represented according to the wavelet phase coherence. The functional connection inside each brain region and between two brain regions is the mean value of the functional connection between every two channels. The functional connection between every two of the 6 brain areas comprises an affected side somatosensory cortex layer-an affected side somatosensory cortex layer, an affected side somatosensory cortex layer-a dominant side locomotion anterior area, an affected side somatosensory cortex layer-a dominant side somatosensory cortex layer, an affected side somatosensory cortex layer-a dominant side primary locomotion cortex layer, an affected side primary locomotion cortex layer-a dominant side locomotion anterior area, an affected side primary locomotion cortex layer-a dominant side primary locomotion cortex layer, a dominant side locomotion anterior area-a dominant side locomotion anterior area layer, and a dominant side somatosensory cortex layer-a dominant side primary locomotion cortex layer.
Wherein, the above mentioned primary motor cortex of left primary motion, primary motor cortex of right primary motion and dominant side primary motor cortex, affected side primary motor cortex are bilateral primary motor cortex of patient, but because each affected side of patient is different, it can't be determined whether the affected side is left side or right side directly, so use two kinds of description ways, similarly, the anterior region of left motion, anterior region of right motion and dominant side anterior region of motion, anterior region of affected side motion are bilateral anterior region of motion of patient; the left somatosensory cortex, the right somatosensory cortex, the dominant lateral somatosensory cortex and the affected lateral somatosensory cortex are the bilateral somatosensory cortex of the patient.
In some embodiments, the electrophysiological index is obtained by performing electrophysiological examination measurements on each patient using a single pulse TMS from a magnetic stimulator.
Specifically, the electrophysiological examination step is: taking a patient in a sitting position or a supine position; the patient is relaxed with the muscles of the hands, and the surface electrodes are used for recording the target muscle groups of the handedness; stimulating a brain affected side thumb movement cortex region using a single pulse mode of a magnetic stimulator, wherein the thumb movement cortex region corresponds to the above-mentioned primary movement region; when the motion of the thumb abductor is induced, the potential recorded by the surface electrode is taken as a single motion-induced potential; taking the minimum stimulation intensity quantity capable of inducing the motion of thumb abductor at least 5 times of 10 times of stimulation as a motion threshold, wherein the motion-induced potentials of the 10 times of stimulation reach more than 50 muV; the time from the pulse sent by the magnetic stimulator to the motion-evoked potential peak is taken as a single latency, and the motion-evoked potential and the latency recorded in the first training sample set are the mean values of the results of 5 times of stimulation.
Step S1013: the method comprises the steps of making a corresponding TMS strategy according to information such as lesion positions, damage degrees, cerebral cortex activation modes and the like of patients, carrying out TMS treatment for 5 times per week, wherein the TMS treatment is carried out for two weeks, the treatment effect of each patient is updated in a Fugl-Meyer scale, the score of the Fugl-Meyer scale of each patient is compared, the higher the score is, the better the rehabilitation effect of the patient is represented, the lower the score is, the worse the rehabilitation effect of the patient is represented, basic information, brain function indexes, electrophysiological indexes of the patient with the better rehabilitation effect and the corresponding TMS strategy are selected to construct a first training sample set, and the first training sample set is equivalent to a personalized TMS strategy knowledge base.
In some embodiments, before obtaining the first set of training samples, further comprising: and calculating the Pearson correlation coefficient of all near-infrared brain function indexes of the corresponding sample patients and Fugl-Meyer scale evaluation data, and selecting the near-infrared brain function indexes which are obviously related to Fugl-Meyer upper limb function scores by taking the near-infrared brain function indexes of which the Pearson correlation coefficient of the upper limb function scores in the Fugl-Meyer scale evaluation data is greater than a first preset value as the brain function indexes of the first training sample set. Wherein the brain function index is a subset of the near-infrared brain function index. Pearson product-moment correlation (PPMCC or PCCs) is used to measure the linear correlation between two variables, and its value is between-1 and 1.
In step S102, the first training sample set is randomly sampled to construct a second training sample set, so that the random sampling of the first training sample set is performed to ensure that training sets of sub-models in the random forest model are basically different to meet the requirement that the sub-models are independent of each other, wherein the sub-models are decision trees, and the random forest model is composed of a plurality of decision trees, so that the decision trees are called as sub-models of the random forest model. The constructed second training sample set, i.e. the training set of the sub-models, needs to be guaranteed to be basically different, because if the training set of each sub-model is consistent, the trained sub-models are the same, and the decision results obtained by the sub-models are also the same, then the random forest algorithm is completely equivalent to a single decision tree algorithm, and the advantage of using an integrated algorithm cannot be embodied.
The method includes the steps that the first training sample set is subjected to replaced random sampling for N times to construct N second training sample sets, although repeated samples can appear in the second training sample sets, and each submodel cannot be completely independent, in fact, the purpose of adopting a random forest algorithm is not to obtain a plurality of independent submodels, combined judgment is carried out according to output results of each submodel, and team cooperation efficiency of the random forest model is improved to a certain extent by the replaced random sampling.
In step S103, N decision trees are respectively generated by training using the N second training sample sets obtained in step S102, and a random forest model for automatically selecting the TMS strategy is constructed from the N decision trees.
In some embodiments, as shown in fig. 2, N is 5, that is, 5 times of random sampling with putting back is performed on the first training sample set, to obtain 5 second training sample sets with the same number of samples as the first training sample set, and 5 decision trees are generated by using the 5 second training sample sets through training respectively, where the random forest model is composed of 5 decision trees and a voting module.
In some embodiments, the model training method for TMS strategy automatic selection adopts a k-fold cross validation method to construct a random forest model and test, and divides a first training sample set into k parts in a form of non-return random sampling, wherein k-1 part is used as a second training sample set for training the random forest model, and the remaining 1 part is used for testing the random forest model.
Specifically, a 5-fold cross validation method is adopted, the first training sample set is divided into 5 parts, wherein 4 parts are used for training the random forest model, and the remaining 1 part is used for testing the random forest model.
For the random forest model obtained by training, the basic information, the brain function index and the electrophysiological index of each patient sample in the first training sample set are input into the random forest model, specifically, the input comprises 3 basic information characteristic values of the lesion type, the hemiplegia degree and the lesion position of the corresponding patient; the activation degree of the left primary motor cortex, the activation degree of the right primary motor cortex, the activation degree of the left motor cortex, the activation degree of the right motor cortex, the activation degree of the left somatosensory cortex, the activation degree of the right somatosensory cortex, the laterality index, the functional connection index of the affected side body cortex-the affected side body cortex, the functional connection index of the affected side body cortex-the dominant side motor cortex, the functional connection index of the affected side body cortex-the dominant side body cortex, the functional connection index of the affected side body cortex-the dominant side primary motor cortex functional connection indexes of an affected side primary motor cortex-dominant side motor cortex, functional connection indexes of the affected side primary motor cortex-dominant side primary motor cortex, functional connection indexes of the dominant side motor cortex-dominant side body sensory cortex, functional connection indexes of the dominant side motor cortex-dominant side primary motor cortex, and functional connection indexes of the dominant side body sensory cortex-dominant side primary motor cortex are 17 brain function index characteristic values; the motor threshold, the latency period and the motor evoked potential are 3 electrophysiological index characteristic values, and the total number is 23 input characteristic values.
And the TMS strategy obtained by the automatic model selection is the random forest model output. Specifically, after 23 input feature values of a patient are input into a random forest model, decision trees in the model respectively obtain respective decision results, the decision results are input into a voting module, a final prediction result obtained according to a voting principle is used as output of the random forest model, namely, a TMS strategy of the patient is output, wherein the voting principle is that a mode of the decision results of the decision trees is used as a final prediction result. Illustratively, the decision result of the decision tree 1 is a TMS strategy a, the decision result of the decision tree 2 is a TMS strategy a, the decision result of the decision tree 3 is a TMS strategy a, the decision result of the decision tree 4 is a TMS strategy B, the decision result of the decision tree 5 is a TMS strategy B, and the 5 decision trees respectively generate 5 decision results, wherein the decision results are 3 of the TMS strategies a and 2 of the TMS strategies B, so that the TMS strategy a is a mode, the TMS strategy a is taken as a final prediction result of a random forest model and is output, and the TMS strategy is finally selected as the TMS strategy a.
In some embodiments, the TMS strategy includes a stimulation pulse frequency and a stimulation target. Specifically, the stimulation pulse frequency is divided into 2 types of high frequency and low frequency, wherein the high frequency refers to the frequency with the frequency value larger than or equal to 5Hz, and the low frequency is 1Hz; the stimulation targets are divided into 3 types of dominant side primary motor cortex, affected side primary motor cortex and healthy side motor prozone.
The invention also provides a TMS strategy automatic selection method, which comprises the following steps S201-S202:
step S201: acquiring basic information, brain function indexes and electrophysiological indexes of a patient to be selected with the TMS strategy; wherein the basic information comprises the type of lesion, degree of hemiplegia, lesion location and initial Fugl-Meyer scale assessment data of the patient; the brain function index comprises the activation degree and the lateral deviation index of each brain region of the patient and the function connection index inside and between every two partial brain regions; the electrophysiological metrics include cerebral cortex motor threshold, motor evoked potential and latency of the patient.
Step S202: the basic information, brain function indices and electrophysiological indices are entered into a random forest model in a model training method as mentioned above for any of the above mentioned automatic selection of TMS strategies to obtain a TMS strategy that is suitable for the patient.
In step S201, the brain function index and electrophysiological index of the patient are obtained according to the method in step S1012.
In step S202, the TMS strategy is obtained for the basic information, brain function index and electrophysiological index of the patient, wherein the TMS strategy includes a stimulation pulse frequency and a stimulation target.
In some embodiments, the basic information, the brain function index, the electrophysiological index and the corresponding TMS strategy of the patient with a rehabilitation effect higher than a preset standard are supplemented into the first training sample set, and the steps of the model training method automatically selected by the TMS strategy as mentioned above are repeated to construct a new random forest model.
The method includes the steps that data with a good rehabilitation effect are supplemented into a first training sample set, the first training sample set is continuously updated and optimized, a model is built by means of the first training sample set, a decision tree obtained through training has better decision-making capacity, then a random forest model obtains a more accurate TMS strategy, treatment is conducted according to specific conditions of a patient in a targeted mode, the clinical rehabilitation curative effect of the patient is effectively improved, and rehabilitation efficiency is improved.
Specifically, the rehabilitation effect tracking is carried out on each patient for a set time, illustratively, the patient is treated by TMS for two weeks 5 times per week, the treatment result of each time is updated in a Fugl-Meyer scale, the score of the Fugl-Meyer scale and the actual rehabilitation condition of the patient are integrated, and a preset standard is formulated for judging the good and the poor rehabilitation effect.
In some embodiments, supplementing the basic information of the patient with a rehabilitation effect higher than a preset standard, the brain function index, the electrophysiological index and the corresponding TMS strategy into the first training sample set further comprises: when the number of the patients with the accumulated rehabilitation effect higher than the preset standard is larger than a second preset value, supplementing the basic information, the brain function index, the electrophysiological index and the corresponding TMS strategy of each patient into a first training sample set; and when the number of the patients is less than a second preset value, only recording the basic information, the brain function index, the electrophysiological index and the corresponding TMS strategy of each patient.
The second preset value can be modified and set according to actual conditions, and the model is updated after a certain number of effective samples are accumulated, so that repeated work is reduced, and the effect of the random forest model updated every time is improved greatly.
The invention also provides a TMS strategy automatic selection system, as shown in fig. 3, the TMS strategy automatic selection system includes: the system comprises a basic information input module 1, a near-infrared brain blood oxygen data acquisition module 2, an electrophysiological examination module 3, a near-infrared brain function evaluation module 4, an intelligent learning module 5, a TMS strategy display module 6 and a feedback module 7.
The basic information input module 1 includes an information entry module 11 and a display module 12. Basic information of a patient to be selected with the TMS strategy is input into the information entry module 11 and transmitted to the intelligent learning module 5. Wherein the basic information includes the type of lesion, degree of hemiplegia, lesion location and Fugl-Meyer scale assessment data of the patient.
The near-infrared brain blood oxygen data acquisition module 2 comprises a near-infrared light source 21, a probe 22 and a near-infrared host 23. Is used for acquiring data of oxygenated hemoglobin signals of each area of the cerebral cortex of the patient in a motion detection state and transmitting the data to the near-infrared brain function evaluation module 4.
The near-infrared brain function evaluation module 4 is configured to calculate, according to the acquired oxyhemoglobin signal, an activation degree and a laterality index of each brain region and a function connection index inside and between each two of partial brain regions to obtain a brain function index of the patient, and transmit the brain function index to the intelligent learning module 5, where the brain function index of the patient is obtained according to the method in step S1012.
The electrophysiological examination module 3 includes an electrophysiological examination host 31, a stimulation coil 32, and an electrophysiological index detection module 33. And (3) performing electrophysiological examination on the patient by using the monopulse TMS, measuring and collecting a movement threshold, a movement evoked potential and a latency of the cerebral cortex of the patient by using the electrophysiological index detection module 33 to obtain electrophysiological indexes of the patient, and transmitting the electrophysiological indexes to the intelligent learning module 5, wherein the electrophysiological indexes of the patient are obtained according to the method in the step S1012.
The intelligent learning module 5 includes a first training sample set 51 and a random forest model 52, wherein the random forest model is a model selected automatically by the TMS strategy, which is trained by the above-mentioned model training method for automatically selecting by the TMS strategy. The smart learning module 5 is used to perform the TMS strategy automatic selection method as mentioned above.
And the TMS strategy display module 6 is used for displaying the TMS strategy output by the intelligent learning module 5. The TMS strategy comprises stimulation pulse frequency and stimulation target points, and is convenient for doctors to perform targeted treatment.
The feedback module 7 is configured to feed back a rehabilitation effect recommended by the random forest model after the TMS strategy is adopted, supplement patient information with a rehabilitation effect higher than a preset standard to the first training sample set 51 of the intelligent learning module 5, continuously update and optimize the first training sample set 51, and reconstruct the random forest model 52 by using the first training sample set 51, so as to improve the accuracy of the random forest model 52.
The present invention also provides a computer readable storage medium having stored thereon a computer program that, when executed by a processor, performs the steps of the method for model training for TMS strategy automatic selection and the method for TMS strategy automatic selection.
In accordance with the above method, the present invention also provides an apparatus comprising a computer device including a processor and a memory, the memory having stored therein computer instructions, the processor being configured to execute the computer instructions stored in the memory, the apparatus implementing the steps of the method as described above when the computer instructions are executed by the processor.
Embodiments of the present invention further provide a computer-readable storage medium, on which a computer program is stored, where the computer program is executed by a processor to implement the foregoing steps of the edge computing server deployment method. The computer readable storage medium may be a tangible storage medium such as Random Access Memory (RAM), memory, read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, floppy disks, hard disks, removable storage disks, CD-ROMs, or any other form of storage medium known in the art.
In summary, the present invention provides a model training method, an automatic selection method and a system for TMS strategy automatic selection, which acquire a first training sample set by collecting basic information, brain function indexes, electrophysiological indexes of a plurality of patients and a TMS strategy with good rehabilitation effect, construct a plurality of second training sample sets in a random sampling manner, and acquire a plurality of decision trees by training the plurality of second patrol sample sets to construct a random forest model for TMS strategy automatic selection. And inputting the basic information, the brain function index and the electrophysiological index of the real patient into the obtained random forest model to obtain the TMS strategy of the patient. The invention can make an individualized TMS strategy based on the brain function index and the electrophysiological index of the patient according to the illness state of the patient.
Furthermore, the patient with the TMS strategy obtained through the random forest model is tracked for a set time length, the rehabilitation effect of the patient treated by the TMS strategy is observed, basic information, brain function indexes and electrophysiological indexes of the patient with the rehabilitation effect higher than a preset standard and corresponding TMS strategies are supplemented into the first training sample set, the step of the model training method for automatically selecting the TMS strategy is carried out again, a new random forest model is built, the accuracy of random forest model selection is improved, the clinical rehabilitation curative effect of the patient is effectively improved, and the rehabilitation efficiency is improved.
Those of ordinary skill in the art will appreciate that the various illustrative components, systems, and methods described in connection with the embodiments disclosed herein may be implemented as hardware, software, or combinations of both. Whether this is done in hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention. When implemented in hardware, it may be, for example, an electronic circuit, an Application Specific Integrated Circuit (ASIC), suitable firmware, plug-in, function card, or the like. When implemented in software, the elements of the invention are the programs or code segments used to perform the required tasks. The program or code segments can be stored in a machine-readable medium or transmitted by a data signal carried in a carrier wave over a transmission medium or a communication link.
It is to be understood that the invention is not limited to the specific arrangements and instrumentality described above and shown in the drawings. A detailed description of known methods is omitted herein for the sake of brevity. In the above embodiments, several specific steps are described and shown as examples. However, the method processes of the present invention are not limited to the specific steps described and illustrated, and those skilled in the art can make various changes, modifications and additions or change the order between the steps after comprehending the spirit of the present invention.
Features that are described and/or illustrated with respect to one embodiment may be used in the same way or in a similar way in one or more other embodiments and/or in combination with or instead of the features of the other embodiments in the present invention.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made to the embodiment of the present invention by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A TMS strategy automatic selection model training method is characterized by comprising the following steps:
acquiring a first training sample set, wherein the first training sample set comprises a plurality of patient samples, and each patient sample comprises basic information, a brain function index, an electrophysiological index and a corresponding TMS strategy of a corresponding patient; wherein the basic information comprises the lesion type, the degree of hemiplegia, the lesion position and Fugl-Meyer scale evaluation data of the patient; the brain function index comprises the activation degree and the lateral deviation index of each brain area of the patient and the function connection index inside and between every two partial brain areas; the electrophysiological indexes comprise cerebral cortex movement threshold, movement evoked potential and latency of the patient;
performing a set number of times of replaced random sampling on the first training sample set to form a set number of second training sample sets consistent with the number of samples in the first training sample set;
and respectively generating the set number of decision trees by using the set number of second training sample sets to construct a random forest model for automatically selecting the TMS strategy, and taking a final prediction result obtained by the decision result of each decision tree according to a voting principle as the output of the random forest model.
2. The TMS strategy automatic selection model training method of claim 1, wherein the method adopts a k-fold cross validation method to construct a random forest model and test, and comprises the following steps:
and dividing the first training sample set into k parts in a random sampling mode without returning, wherein k-1 part is used as a second training sample set and used for training the random forest model, and the rest 1 part is used for testing the random forest model.
3. The method of TMS strategy auto-selection model training of claim 1, further comprising, prior to acquiring the first set of training samples:
calculating a Pearson correlation coefficient of the near-infrared brain function index of the corresponding sample patient and the evaluation data of the Fugl-Meyer scale, and taking the near-infrared brain function index of which the Pearson correlation coefficient of the upper limb function score in the evaluation data of the Fugl-Meyer scale is larger than a first preset value as the brain function index of the corresponding sample; wherein the brain function index is a subset of the near-infrared brain function index.
4. The model training method for TMS strategy automatic selection according to claim 1, wherein the brain function index is calculated from the oxygenated hemoglobin signals of each area of cerebral cortex in motion detection state of patient sample collected by near infrared spectrometer;
the activation degree of the brain areas is obtained by averaging the amplitudes of the channels of each brain area after averaging the amplitudes of the acquisition channels in the time domain according to the oxygenated hemoglobin signals;
the laterality index is calculated by dividing the difference of the activation degrees of the dominant cerebral hemisphere and the affected cerebral hemisphere by the sum of the activation degrees of the dominant cerebral hemisphere and the affected cerebral hemisphere;
the functional connection index inside the brain region and between every two channels is represented by wavelet phase coherence between every two channels, which is obtained by phase information calculation after the oxygenated hemoglobin signals of all the acquisition channels are subjected to continuous complex wavelet transformation.
5. The method of claim 1, wherein the electrophysiological metric is measured from a single pulse TMS delivered by a magnetic stimulator for electrophysiological examination of each patient.
6. The method of model training for TMS strategy automatic selection of claim 1, wherein said TMS strategy comprises stimulation pulse frequency and stimulation target.
7. A TMS strategy automatic selection method is characterized by comprising the following steps:
acquiring basic information, brain function indexes and electrophysiological indexes of a patient to be subjected to TMS strategy selection; wherein the basic information comprises the type of lesion, degree of hemiplegia, lesion location and initial Fugl-Meyer scale assessment data of the patient; the brain function index comprises the activation degree and the lateral deviation index of each brain area of the patient and the function connection index between every two brain areas; the electrophysiological indexes comprise cerebral cortex movement threshold, movement evoked potential and latency of the patient;
inputting the basic information, the brain function index and the electrophysiological index into a random forest model in a model training method of TMS strategy automatic selection according to any one of claims 1 to 6 to obtain the TMS strategy of the patient.
8. The method of claim 7, further comprising:
and supplementing the basic information, the brain function index, the electrophysiological index and the corresponding TMS strategy of the patient with the rehabilitation effect higher than a preset standard into the first training sample set, and repeating the step of the model training method automatically selected by the TMS strategy according to any one of claims 1 to 6 to construct a new random forest model.
9. The method of claim 8, wherein supplementing the base information, the brain function index, the electrophysiological index of the patient with a rehabilitation effect above a predetermined criteria, and the corresponding TMS strategy into the first set of training samples further comprises:
when the number of the patients with the accumulated rehabilitation effect higher than the preset standard is larger than a second preset value, supplementing the basic information, the brain function index, the electrophysiological index and the corresponding TMS strategy of each patient into the first training sample set;
and when the number of the patients is less than a second preset value, only recording the basic information, the brain function index, the electrophysiological index and the corresponding TMS strategy of each patient.
10. A TMS strategy automatic selection system, said TMS strategy automatic selection system comprising:
the basic information input module is used for inputting basic information of a patient to be subjected to TMS strategy selection, including lesion type, hemiplegia degree, lesion position and Fugl-Meyer scale evaluation data of the patient, and transmitting the basic information to the intelligent learning module;
the near-infrared brain blood oxygen data acquisition module is used for acquiring data of oxygenated hemoglobin signals of each area of the cerebral cortex of the patient in a motion detection state and transmitting the data to the near-infrared brain function evaluation module;
the near-infrared brain function evaluation module is used for calculating the activation degree and the laterality index of each brain area and the function connection index between every two brain areas according to the acquired oxygenated hemoglobin signals to obtain the brain function index of the patient and transmitting the brain function index to the intelligent learning module;
the electrophysiological examination module is used for carrying out electrophysiological examination on the patient by adopting the monopulse TMS, measuring and collecting a movement threshold, a movement evoked potential and a latency of the cerebral cortex of the patient to obtain electrophysiological indexes of the patient, and transmitting the electrophysiological indexes to the intelligent learning module;
an intelligent learning module for executing the steps of the TMS strategy automatic selection method according to any of claims 7 to 9;
the TMS strategy display module is used for displaying the TMS strategy output by the intelligent learning module;
and the feedback module is used for feeding back the rehabilitation effect recommended by the random forest model after the TMS strategy is adopted and transmitting the patient information with the rehabilitation effect higher than the preset standard to the intelligent learning module.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211091331.1A CN115620853A (en) | 2022-09-07 | 2022-09-07 | Model training method for TMS strategy automatic selection, automatic selection method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211091331.1A CN115620853A (en) | 2022-09-07 | 2022-09-07 | Model training method for TMS strategy automatic selection, automatic selection method and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115620853A true CN115620853A (en) | 2023-01-17 |
Family
ID=84858330
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211091331.1A Pending CN115620853A (en) | 2022-09-07 | 2022-09-07 | Model training method for TMS strategy automatic selection, automatic selection method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115620853A (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111631731A (en) * | 2020-05-09 | 2020-09-08 | 国家康复辅具研究中心 | Near-infrared brain function and touch force/motion information fusion assessment method and system |
CN112527958A (en) * | 2020-12-11 | 2021-03-19 | 平安科技(深圳)有限公司 | User behavior tendency identification method, device, equipment and storage medium |
US20210353205A1 (en) * | 2018-09-13 | 2021-11-18 | Quantalx Neuroscience Ltd | A reliable tool for evaluating brain health |
US20210393955A1 (en) * | 2010-01-06 | 2021-12-23 | Evoke Neuroscience, Inc. | Transcranial stimulation device and method based on electrophysiological testing |
US20220168585A1 (en) * | 2019-03-04 | 2022-06-02 | The Methodist Hospital | Dynamically Controlling Self-Directed Magnetic Stimulation |
-
2022
- 2022-09-07 CN CN202211091331.1A patent/CN115620853A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20210393955A1 (en) * | 2010-01-06 | 2021-12-23 | Evoke Neuroscience, Inc. | Transcranial stimulation device and method based on electrophysiological testing |
US20210353205A1 (en) * | 2018-09-13 | 2021-11-18 | Quantalx Neuroscience Ltd | A reliable tool for evaluating brain health |
US20220168585A1 (en) * | 2019-03-04 | 2022-06-02 | The Methodist Hospital | Dynamically Controlling Self-Directed Magnetic Stimulation |
CN111631731A (en) * | 2020-05-09 | 2020-09-08 | 国家康复辅具研究中心 | Near-infrared brain function and touch force/motion information fusion assessment method and system |
CN112527958A (en) * | 2020-12-11 | 2021-03-19 | 平安科技(深圳)有限公司 | User behavior tendency identification method, device, equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Papagiannis et al. | Methodology of surface electromyography in gait analysis: review of the literature | |
WO2018045623A1 (en) | Emotion classification method and device based on pulse wave time sequence analysis | |
Doheny et al. | Feature-based evaluation of a wearable surface EMG sensor against laboratory standard EMG during force-varying and fatiguing contractions | |
KR20170086056A (en) | Scoring method based on improved signals analysis | |
Djamal et al. | Identification of post-stroke EEG signal using wavelet and convolutional neural networks | |
Greco et al. | Assessment of muscle fatigue during isometric contraction using autonomic nervous system correlates | |
Pun et al. | Brain-computer interaction research at the Computer Vision and Multimedia Laboratory, University of Geneva | |
Herter et al. | Vision does not always help stroke survivors compensate for impaired limb position sense | |
Chauhan et al. | Real-time stress assessment through PPG sensor for VR biofeedback | |
Alamudun et al. | Removal of subject-dependent and activity-dependent variation in physiological measures of stress | |
US20220183611A1 (en) | Identifying and extracting electroencephalogram signals | |
CN117133404B (en) | Intelligent rehabilitation nursing device to thorax export syndrome | |
CN116312951B (en) | Exercise function assessment method and system based on multi-modal coupling analysis | |
Zhou et al. | Upper-limb functional assessment after stroke using mirror contraction: A pilot study | |
CN115969395A (en) | System and method for non-invasively evaluating excitability of cerebral cortex | |
CN115620853A (en) | Model training method for TMS strategy automatic selection, automatic selection method and system | |
Suviseshamuthu et al. | EEG-based spectral analysis showing brainwave changes related to modulating progressive fatigue during a prolonged intermittent motor task | |
Sun et al. | [Retracted] Optimization of Surface Electromyography‐Based Neurofeedback Rehabilitation Intervention System | |
Celka et al. | Pulse wave harmony: ancient wisdoms for modern age | |
JP6749777B2 (en) | Rehabilitation equipment | |
Litvak | Analysis of the effects of transcranial magnetic stimulation on functional states and connectivity of the human cerebral cortex using electroencephalography | |
Golmohammadishouraki | Prediction of Fatigue in Lower Extremity using EMG Sensor and Machine Learning | |
Henshaw | Improved brain-computer interface methods with application to gaming | |
KR101530957B1 (en) | Method of displaying a plurality of simulated points and system thereof | |
Liu et al. | The human primary somatosensory cortex response contains components related to stimulus frequency and perception in a frequency discrimination task |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20230117 |
|
RJ01 | Rejection of invention patent application after publication |