WO2024114834A1 - 癫痫检测方法及系统 - Google Patents

癫痫检测方法及系统 Download PDF

Info

Publication number
WO2024114834A1
WO2024114834A1 PCT/CN2024/073897 CN2024073897W WO2024114834A1 WO 2024114834 A1 WO2024114834 A1 WO 2024114834A1 CN 2024073897 W CN2024073897 W CN 2024073897W WO 2024114834 A1 WO2024114834 A1 WO 2024114834A1
Authority
WO
WIPO (PCT)
Prior art keywords
signal
module
data analysis
physiological
epilepsy
Prior art date
Application number
PCT/CN2024/073897
Other languages
English (en)
French (fr)
Inventor
彭希
王小沙
冉雨杭
王亮
苟践
蒋伟
Original Assignee
重庆医科大学附属第二医院
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 重庆医科大学附属第二医院 filed Critical 重庆医科大学附属第二医院
Publication of WO2024114834A1 publication Critical patent/WO2024114834A1/zh

Links

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • 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
    • 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
    • A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
    • 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
    • A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
    • A61B5/346—Analysis of electrocardiograms
    • 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
    • A61B5/369—Electroencephalography [EEG]
    • 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
    • A61B5/369—Electroencephalography [EEG]
    • A61B5/372—Analysis of electroencephalograms
    • 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
    • A61B5/24—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]
    • A61B5/397—Analysis of electromyograms

Definitions

  • the present invention relates to the technical field of medical devices, and specifically to a portable epilepsy detection method and system.
  • Epilepsy is a chronic neurological disease caused by sudden abnormal discharges of brain neurons, which can lead to short-term brain dysfunction, limb rigidity, abnormal convulsions of the limbs, loss of consciousness and other symptoms.
  • epileptic seizures patients are often accidentally injured due to absence of consciousness, uncontrolled body, respiratory arrest and other reasons. If the seizure is not treated in time, the inflammatory response of the brain may aggravate the damage to the nervous system, resulting in more serious consequences.
  • Epilepsy seizures are sudden and random, affecting the normal work and life of patients and causing anxiety in patients.
  • Epilepsy seizures are accompanied by almost imperceptible short-term absence of consciousness or long-term violent convulsions. The situation is complex and diverse, and there is no obvious pattern.
  • the current automatic epilepsy detection system mainly distinguishes epileptic seizures from normal states based on the differences in some characteristics of abnormal physiological activities during epileptic seizures and normal physiological activities in EEG, ECG, limb movements, etc., and mostly uses EEG signals, acceleration signals, ECG signals, EMG signals, etc. as input.
  • Wearable devices are portable devices that can be worn directly on the body or integrated into the patient's clothes or accessories. Based on hardware devices, powerful functions can be achieved through software support, data interaction, and cloud interaction. Implementing epilepsy alarms based on wearable devices can greatly reduce the harm of epileptic seizures to patients and improve the quality of life of patients. On the one hand, it can meet the needs of epilepsy monitoring and alarm, reduce patient damage and improve the quality of life. On the other hand, due to the commonness and concealment of the device, the patient's sense of stigma is completely eliminated.
  • the prior art has proposed a technical solution for epilepsy identification based on multiple physiological signals.
  • the patent number is ZL 202011240677.4
  • the invention name is an invention patent for a multi-input signal epilepsy attack detection system based on feedback regulation. It obtains acceleration, angular velocity, skin electrical signal, myoelectric signal and temperature through sensors, and combines the signals, and then performs signal processing and analysis on each signal combination, that is, extracts features from the pre-processed signals, and then analyzes the extracted features to obtain the final detection results, which can overcome the problem of low accuracy of epilepsy detection with a single signal.
  • this detection device collects multiple physiological signals, it does not collect EEG signals, which undoubtedly greatly reduces the accuracy of detection.
  • the object of the present invention is to provide an epilepsy detection method and system to partially solve or alleviate the above-mentioned deficiencies in the prior art and to more accurately predict epilepsy.
  • the first aspect of the present invention is to provide an epilepsy detection system, which includes: a signal acquisition module, which is used to collect multiple physiological signals of a monitored user; the multiple physiological signals include: brain wave signals and electrocardiogram signals, acceleration, and electromyography signals; a preprocessing module, which is connected to the signal acquisition module and is used to preprocess the signal combination signal of the physiological signals collected by the signal acquisition module; a data analysis module, which is connected to the preprocessing module and is used to perform data analysis on the time-frequency physiological signal combination signal obtained after preprocessing through a neural network model to identify whether the monitored user has epilepsy; an early warning module, which is connected to the data analysis module and is used to issue an early warning when the data analysis module identifies that the monitored user has an epileptic seizure; wherein the preprocessing module includes: a bandpass filter, which is used to filter out out-of-band noise signals to obtain the physiological signal combination signal that only retains in-band information; a median filter, which is used to filter out the denoised The
  • the signal acquisition module includes: at least four EEG signal acquisition microelectrodes that can be attached to the brain of the monitored user and are used to acquire the EEG signals of the monitored user; a wireless communication unit integrated with the EEG signal acquisition microelectrodes and used to send the EEG signals acquired by the EEG signal acquisition microelectrodes to the preprocessing module; and a heart rate sensor integrated with the preprocessing module, the data analysis module, and the early warning module and used to acquire the electrocardiogram signals of the monitored user.
  • the signal acquisition module also includes: an electromyographic signal acquisition electrode for collecting electromyographic signals, an acceleration sensor for collecting acceleration, and a gyroscope sensor for collecting angular velocity, which are integrated with the preprocessing module, the data analysis module, and the early warning module and connected to the data analysis module.
  • the heart rate sensor, the preprocessing module, the data analysis module, and the early warning module are integrated on a wearable device, and the wearable device includes a bracelet.
  • the early warning module includes: a voice unit connected to the data analysis module, used to automatically broadcast a pre-stored voice message for help when the data analysis module identifies that the user has an epileptic seizure; and/or an early warning notification unit connected to the data analysis module, used to send an early warning message to a mobile terminal of a pre-associated guardian via the Internet of Things when the data analysis module identifies that the user has an epileptic seizure.
  • the epilepsy detection system further includes: a wireless communication module connected to the data analysis module and used for data communication with a pre-bound mobile terminal of the monitored user or guardian.
  • the data analysis module includes: a first data analysis unit, which is used to match the corresponding signal combination mode according to the seizure type when the seizure type of the monitored user has been specified, and The signal acquisition module is triggered to collect corresponding physiological signals according to the matched signal combination mode; or, when the seizure type of the monitored user is not specified, a preset default signal combination mode is obtained, and the signal acquisition module is triggered to collect corresponding physiological signals according to the default signal combination mode; a second data analysis unit is used to perform data analysis on the physiological signals preprocessed by the preprocessing module to identify whether the monitored user has epilepsy.
  • the signal combination methods include: a first combination method: EEG, ECG + acceleration; a second combination method: EEG + acceleration; a third combination method: EEG + ECG + EMG + acceleration; a fourth combination method: EEG + EMG; a fifth combination method: EEG + EMG + acceleration.
  • the second aspect of the present invention is to provide an epilepsy detection method, which is based on the above-mentioned epilepsy detection system, and the epilepsy detection system includes: a signal acquisition module for collecting multiple physiological signals of a monitored user; the multiple physiological signals include: brain wave signals, electrocardiogram signals, electromyography signals, acceleration and angular velocity; a preprocessing module for preprocessing the physiological signals collected by the signal acquisition module; a data analysis module for performing data analysis on the time-frequency physiological signals obtained after preprocessing through a neural network model to identify whether the monitored user has epilepsy; an early warning module for issuing an early warning when the data analysis module identifies that the monitored user has an epileptic seizure; accordingly, the epilepsy detection method specifically includes the steps of: obtaining the epileptic seizure type of the monitored user through the data analysis module, and matching it in a database according to the epileptic seizure type Corresponding signal combination methods; wherein the signal combination methods include: a first combination method: electroencephalogram signal,
  • the third aspect of the present invention is to provide an epilepsy early warning system, which includes the above-mentioned epilepsy detection system, a server and a user terminal, wherein the epilepsy detection system and the user terminal communicate data with the server via a wireless network; wherein the epilepsy detection system is used to collect physiological signals of the monitored user, and use the built-in processing module to pre-process the signal combination signal of the physiological signal before performing data analysis, and then send the collected physiological signal combination signal and/or data analysis results to the server and/or the user terminal via the wireless network; wherein the pre-processing module includes: a bandpass filter, used to filter out out-of-band noise signals to obtain the physiological signal combination signal that only retains the in-band information; a median filter, used to eliminate the baseline offset of the denoised physiological signal combination signal; The smoothing filter is used to perform denoising on the physiological signal combination signal after the baseline offset is eliminated, so as to eliminate the noise in the signal band and obtain the true original physiological signal combination signal; the differential filter is used to filter
  • the collected original physiological signals are preprocessed by passing them through a bandpass filter, a median filter, a balance filter, and a differential filter in sequence to obtain real time-frequency physiological signals, and then input into a pre-trained neural network model for identification.
  • a bandpass filter a median filter
  • a balance filter a differential filter
  • the input to the neural network model is a real physiological signal, no data will be omitted or lost due to the use of feature extraction, so its detection accuracy is higher.
  • a variety of physiological signals are collected, and each signal combination includes an EEG signal that most directly reflects the discharge pattern of the monitored user during an epileptic seizure.
  • brain waves are collected by four microelectrodes set at fixed positions on the brain of the monitored user, which makes it easy to carry and does not require the monitored user to go to the hospital or use large equipment for detection, so that the user can monitor at any time.
  • FIG1 is a functional module diagram of an epilepsy detection system according to an exemplary embodiment of the present invention.
  • FIGS. 2a and 2b are schematic diagrams of four electroencephalogram signal microelectrodes pasted at fixed positions on the head in an epilepsy detection system according to an exemplary embodiment of the present invention
  • FIG3a is a collected original brain wave signal of the monitored user
  • FIG3b is an EEG signal after the original EEG signal shown in FIG3a is bandpass filtered
  • FIG3c is an EEG signal after the EEG signal shown in FIG3b is processed by median filtering
  • FIG3d is a time-frequency signal of the brain wave after the brain wave signal shown in FIG3c is processed by smoothing filtering;
  • FIG. 4a to FIG. 4g are the recognition results of 30 monitored users detected by using two types of epilepsy detection bracelets respectively.
  • orientation or position relationship indicated by the terms “upper”, “lower”, “inner”, “outer”, “front”, “back”, “one end”, “other end” and the like is based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention.
  • first and “second” are used only for descriptive purposes and cannot be understood as indicating or implying relative importance.
  • the terms “installed”, “provided with”, “connected”, etc. should be understood in a broad sense.
  • connection can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, a direct connection, an indirect connection through an intermediate medium, or a connection between the two components.
  • “And/or” in this document includes any and all combinations of one or more listed related items.
  • “Multiple” in this document means two or more, that is, it includes two, three, four, five, etc.
  • Embodiment 1 it is a functional module diagram of an exemplary embodiment of an epilepsy detection system of the present invention.
  • the epilepsy detection system comprises: a signal acquisition module, which is used to acquire a variety of physiological signals of a monitored user; wherein the physiological signals include: brain wave signals and electrocardiogram signals; a preprocessing module, which is used to preprocess the physiological signals acquired by the signal acquisition module; specifically, the preprocessing module comprises: a bandpass filter, which is used to filter out-of-band noise signals to obtain the physiological signals that only retain the in-band information; a median filter, which is used to perform baseline offset elimination on the denoised physiological signals; and a smoothing filter, which is used to perform denoising on the physiological signals after baseline offset elimination to eliminate the signal bandpass.
  • the invention relates to a method for eliminating the noise in the signal band and obtaining the real original physiological signal; a differential filter is used to filter the physiological signal that eliminates the noise in the signal band and obtain the time-frequency physiological signal to be analyzed; a data analysis module is used to perform data analysis on the pre-processed signal through a neural network model to identify whether the monitored user has epilepsy; specifically, a large number of physiological signals of epilepsy patients (including different types of epilepsy patients) during epileptic seizures are collected in advance, and the collected physiological signals are pre-processed by the above-mentioned pre-processing module and used as training samples for training the neural network module; specifically, a CNN neural network model can be used; an early warning module is used to make early warning when the above-mentioned data analysis module identifies that the monitored user has an epileptic seizure. police.
  • the out-of-band noise signal is filtered out by a bandpass filter, and only the in-band information of the physiological signal is retained.
  • the in-band signal if there is a baseline offset, the signal feature detection will be affected by the baseline drift, and there is a high probability of misjudgment, so the baseline offset must be eliminated; after eliminating the baseline offset, there is still in-band noise in the signal band.
  • the in-band noise is filtered out by an adaptive smoothing filter to restore the original signal more realistically; in order to make the signal change characteristics of the physiological signal more obvious, it is filtered by a differential filter, and the amplitude change of the filtered signal is more obvious. After being sent to the neural network CNN, the detection accuracy is higher.
  • the brain wave signal processed by the bandpass filter eliminates the out-of-band interference noise compared to the original signal, as shown in Figure 3c, the brain wave signal processed by the median filter eliminates the baseline drift, as shown in Figure 3d, the brain wave signal processed by the smoothing filter eliminates the burrs in the signal.
  • the signal is amplified by a differential filter. That is, the signal quality is improved through the above four filters, so that high-quality physiological signals can be input into the data analysis module for machine learning or automatic recognition, which greatly improves the recognition accuracy.
  • the time-frequency signal obtained by directly preprocessing the collected physiological signals is input into the pre-trained CNN neural network for epilepsy recognition, which greatly improves the detection accuracy.
  • epilepsy detection was performed on 30 monitored users using the feature extraction method and the method of this embodiment. Specifically, the 30 patients wore two types of epilepsy detection systems on their left and right hands, respectively. The detection results are shown in Figures 4a to 4g.
  • the epilepsy detection system of this embodiment identified that all 30 monitored users had epileptic seizures, which was consistent with the hospital diagnosis results, and its accuracy was 100%.
  • the device for epilepsy identification using the feature extraction method identified that 28 of the 30 monitored users had epileptic seizures, and 2 were inconsistent with the hospital diagnosis results, and its accuracy was 93.33%.
  • the signal acquisition module specifically includes: at least four electroencephalogram signal acquisition microelectrodes that can be attached to the brain of the monitored user and are used to collect the electroencephalogram signal of the monitored user; and integrated with the electroencephalogram signal acquisition microelectrodes.
  • a wireless communication unit for sending the EEG signals collected by the EEG signal collection microelectrodes to the preprocessing module; a heart rate sensor integrated with the preprocessing module, the data analysis module, and the early warning module for collecting the ECG signals of the monitored user.
  • four microelectrodes for collecting EEG signals can be respectively attached to four fixed positions on the brain of the monitored user, see Figure 2a and Figure 2b.
  • the four microelectrodes and the wireless communication unit are portable accessories of the detection system, so that the monitored user does not need to go to a hospital or other medical institution to collect EEG signals with special equipment, and it is also easy to carry and has good concealment, which is also convenient for the monitored user to better monitor at any time.
  • an analog-to-digital converter can be further set between the two.
  • the signal acquisition module further includes: an electromyographic signal acquisition electrode for acquiring electromyographic signals, an acceleration sensor for acquiring acceleration, and a gyroscope sensor for acquiring angular velocity, which are integrated with the preprocessing module, the data analysis module, and the warning module and connected to the data analysis module.
  • the heart rate sensor, the preprocessing module, the data analysis module, and the warning module are integrated into a wearable device, and the wearable device includes a wristband.
  • the above-mentioned early warning module includes: a voice unit, connected to the above-mentioned data analysis module, used to automatically broadcast a pre-stored voice message for help when the data analysis module identifies that the monitored user has an epileptic seizure; and/or, an early warning notification unit, connected to the above-mentioned data analysis module, used to send an early warning message to a mobile terminal of a pre-associated guardian through the Internet of Things when the data analysis module identifies that the monitored user has an epileptic seizure.
  • a voice unit is set up to play a pre-stored voice message for help, for example, "The patient has an epileptic seizure, please dial XXXXX", "The patient has an epileptic seizure, please put XXX in the XXX bag carried by the patient into the patient's mouth”..., not only can the people around know the specific situation of the monitored user, but they can also provide immediate assistance based on the voice information.
  • the early warning notification unit can also be used to immediately notify the guardian of the monitored user's epileptic seizure, so that the guardian can quickly take appropriate measures. For example, when the guardian is close to the monitored user, he or she can quickly come to the monitored user to provide assistance.
  • the warning module further includes: a positioning unit connected to the data analysis module, for feeding back the current positioning information of the monitored user to the data analysis module when the data identifies that the user has an epileptic seizure, and obtaining the medical institution information of the nearest medical institution; an automatic help unit connected to the data analysis module
  • the data analysis module is connected to the monitoring device, and is used to send the positioning information and the physiological signal to the medical institution when the data analysis module identifies that the user has an epileptic seizure.
  • the current location information of the monitored user can be accurately located through the positioning unit (for example, GPS, etc.), so that while sending an early warning notification to the guardian, the positioning information can also be sent to the guardian, so that even if the guardian is far away from the monitored user, the guardian can still grasp the dynamics of the monitored user; at the same time, the nearest medical institution can be automatically searched according to the current positioning information of the monitored user, so as to send the corresponding distress information (including positioning information, physiological signals and basic information of the monitored user, etc.) to it (for example, the emergency clinic of the medical institution).
  • the positioning unit for example, GPS, etc.
  • the epilepsy detection system further includes: a wireless communication module connected to the data analysis module and used for data communication with a pre-bound mobile terminal of the monitored user or guardian.
  • the above-mentioned data analysis module includes: a first data analysis unit, which is used to match the corresponding signal combination method according to the seizure type of the monitored user when the seizure type has been specified, and trigger the signal acquisition module to collect corresponding physiological signals according to the matched signal combination method; or, when the seizure type of the monitored user is not specified, obtain a preset default signal combination method, and trigger the signal acquisition module to collect corresponding physiological signals according to the default signal combination method; a second data analysis unit, which is used to use a pre-trained neural network model to perform data analysis on the physiological signals preprocessed by the preprocessing module to identify whether the monitored user has epilepsy.
  • the above-mentioned signal combinations include: the first combination method: EEG signal, ECG signal + acceleration; the second combination method: EEG signal + acceleration; the third combination method: EEG signal + ECG signal + EMG signal + acceleration; the fourth combination method: EEG signal + EMG signal; the fifth combination method: EEG signal + EMG signal + acceleration.
  • the epilepsy detection system of this embodiment is described below in conjunction with its working principle.
  • the data analysis module triggers the signal acquisition module to collect the corresponding physiological signal according to the epilepsy attack type selected by the monitored user; wherein the data analysis module pre-stores the physiological signal combination mode corresponding to each epilepsy attack type, so once an epilepsy attack type is selected, it will automatically match the corresponding physiological signal combination according to the epilepsy attack type, and then trigger the corresponding sensor in the signal acquisition module to collect the corresponding physiological signal; under the triggering action of the data analysis module, the signal acquisition module adopts the corresponding physiological signal and sends it to the preprocessing module for preprocessing, and then sends the preprocessed physiological secondary signal to the data analysis module for data analysis to identify whether the monitored user has epilepsy; if so, the early warning module automatically voice broadcasts the alarm signal, so that the monitored user When the monitored user has an
  • the warning module automatically searches for the nearest medical institution based on the current location information of the monitored user, dials the corresponding emergency call, and sends the location information to the medical institution, thereby improving the efficiency of rescue.
  • EEG signals ECG signals
  • EMG signals EEG signals
  • data analysis is performed according to all the above signal combinations for identification.
  • a signal combination is dynamically determined based on the feedback from the monitored user or his/her guardian. For the monitored user, only the physiological signals in this signal combination will be collected in the later stage, and other signal combinations will no longer be considered to reduce power consumption.
  • Embodiment 2 Based on the above-mentioned epilepsy detection system, the present invention also provides an epilepsy early warning system, which comprises the above-mentioned epilepsy detection system, a server and a user terminal, wherein the epilepsy detection system and the user terminal communicate data with the server via a wireless network; specifically, the epilepsy detection system is used to collect physiological signals of the monitored user, and use a built-in preprocessing module to preprocess the physiological signals before performing data analysis, and then send the recognition results and/or the collected physiological signals and/or the data analysis results to the server and/or the user terminal via a wireless network.
  • the user terminal includes a mobile terminal of the monitored user, and/or a mobile terminal of a guardian, so that the monitored user and/or the guardian can understand the monitoring situation at any time.
  • Embodiment 3 Based on the epilepsy detection system of the above-mentioned embodiment 1, the present invention further provides an epilepsy detection method.
  • the epilepsy detection method comprises the steps of: obtaining the epilepsy attack type of the monitored user through a data analysis module, and matching the corresponding signal combination mode in the database according to the epilepsy attack type; usually, when the monitored user uses the epilepsy detection system for the first time, for example, when wearing the epilepsy detection system in the form of a bracelet for the first time, the user will set his or her own epilepsy attack type in the epilepsy detection system, thereby realizing the initialization of the epilepsy detection system, that is, automatically matching the corresponding signal combination mode according to the set epilepsy attack type, so that in the subsequent detection process, the corresponding signal acquisition unit in the signal acquisition module is triggered according to the matched signal combination mode to collect the corresponding physiological Signal;
  • the signal combination method includes: the first combination method: EEG signal, ECG signal + acceleration; the second combination method: E
  • the method of triggering the early warning module to issue an early warning includes any one or more of the following: 1) sending an early warning message to a mobile terminal of a pre-associated guardian; 2) automatically broadcasting a voice message according to a pre-stored voice message; 3) obtaining the current location information of the monitored user and nearby medical institutions, and sending the location information and basic information of the monitored user and the collected physiological signals to the medical institution, so that the medical institution can prepare for rescue in advance and accurately obtain the location of the monitored user.
  • the specific early warning method can be pre-set according to the type and severity of the attack of the monitored user.

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Medical Informatics (AREA)
  • Animal Behavior & Ethology (AREA)
  • Veterinary Medicine (AREA)
  • Biophysics (AREA)
  • Pathology (AREA)
  • Engineering & Computer Science (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Public Health (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Cardiology (AREA)
  • Psychiatry (AREA)
  • Psychology (AREA)
  • Physiology (AREA)
  • Dentistry (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
  • Measuring And Recording Apparatus For Diagnosis (AREA)

Abstract

一种癫痫检测方法及系统，癫痫检测系统包括信号采集模块，用于采集被监测用户的多种生理信号；预处理模块，用于对所采集的生理信号组合信号进行预处理；数据分析模块，用于通过神经网络模型对经过预处理的生理信号组合信号进行数据分析，以识别被监测用户是否发生癫痫；预警模块，用于当识别出被监测用户癫痫发作时，进行预警；其中，预处理模块包括：带通滤波器、中值滤波器、平滑滤波器和差分滤波器，该四个滤波器依次对采集的生理信号组合信号进行预处理，使得经过预处理的生理信号组合信号直接输入数据分析模块，提高了癫痫识别的精确率。

Description

癫痫检测方法及系统
优先权申请
本申请要求2022年12月02日提交的中国发明专利申请【2022115404286】“【癫痫检测方法及系统】”的优先权，该优先权发明专利申请以引用方式全文并入。
技术领域
本发明涉及医疗装置技术领域，具体设计一种便携式的癫痫检测方法及系统。
背景技术
癫痫是由大脑神经元突发性异常放电所引起一类慢性神经系统疾病，会导致短暂的大脑功能障碍，产生肢体僵直、四肢异常抽搐、意识丧失等症状。癫痫发作时常常由于失神、躯体不受控制、呼吸停止等原因导致患者受到意外伤害，且发作时如果得不到及时治疗可能会因脑部炎症反应而加重神经系统损伤，造成更加严重的后果。癫痫发作具有突发性和随机性，影响患者的正常工作与生活，使患者产生焦虑情绪。癫痫发作时伴有几乎无法察觉的短暂失神或长时间的剧烈阵挛，其情况复杂多样，没有明显规律。若患者发病时不在公众场合或无人看护，就很难被发现，事后也很难回忆自己的发作史。抽搐发作是癫痫相关的损伤和癫痫致死情况的主要因素。除此之外，癫痫发作与病耻感、头痛以及诸如注意力缺陷、多动障碍等精神疾病有很大的关系。如果不正常的神经活动限制在大脑的一个特定区域，称为局灶性癫痫发作；当传播到大脑的其他区域时，则称为全身性发作。对癫痫患者和他们的看护者来说，对发作的恐惧是一直存在的。他们的生活也会被癫痫可能发作的恐惧一直支配着，严重降低了他们的生活质量。基于上述癫痫发作检测中的困难与癫痫发作对患者造成的严重影响，癫痫发作自动检测方法是当今医学界与医疗电子领域的重要研究课题之一。
目前的癫痫自动检测系统主要基于癫痫发作时的异常生理活动与正常生理活动在脑电、心电、肢体运动等方面的一些特征的差异来区分癫痫发作与正常状态，多以脑电信号、加速度信号、心电信号、肌电信号等作为输入，主要有以下几种实现方式：(1)基于脑电信号实现癫痫自动检测：龚光红等于2019年申请的专利《基于监督梯度提升器的多级癫痫脑电信号自动识别方法》(专利号CN109934089A)通过梯度提升分类器进行癫痫信号的检查。(2)基于心电信号实现癫痫自动检测：宋晓宇等于2015年申请的《癫痫病人心跳异常智能预警癫痫发作系统》(专利号CN 104997499 A)通过患者胸部周围的数个电极采集信号并通过心率变化检测心跳信号的异常；Wangcai Liao等于2016年申请的专利《IDENTIFYING SEIZURES USING HEART DATA FROM TWO OR MORE WINDOWS》(专利号US9498162B2)也通过统计心率变化来检 测异常。(3)基于人体躯干或头部的加速度信号实现癫痫自动检测：陈蕾等于2016年申请的《癫痫检测装置及癫痫检测方法》(专利号CN105232000A)在含有三轴无线加速度传感器的手环中使用癫痫检测方法进行癫痫发作的检测。然而，上述各种检测该类癫痫报警装置的不足在于对癫痫发作判断的准确性不足，使用单一生理传感器的数据进行分析，容易出现假阳性的情况；该类装置还存在携带不便，增加病耻感的缺点，无法时刻对患者进行监测。可穿戴装置是一种可以直接穿在身上，或者集成到患者的衣服或配件中的一种便携装置，基于硬件装置可以通过软件支持、数据交互、云端交互来实现强大的功能。基于可穿戴装置实现癫痫的报警，能在很大程度上减小癫痫发作对患者的伤害，改善患者的生活质量。一方面可以满足癫痫监测与报警的需求，减少病人损伤的同时提升生活质量。另一方面由于装置的常见性和隐蔽性，完全消除了患者的病耻感。
针对上述问题，现有技术提出了基于多种生理信号进行癫痫识别的技术方案，例如，专利号为ZL 202011240677.4，发明名称为一种基于反馈调节的多输入信号癫痫发作检测系统的发明专利，其通过传感器获得加速度、角速度、皮肤电信号、肌电信号和温度，并把信号进行组合，再对每种信号组合进行信号处理和分析，即对预处理后的信号进行特征提取，然后基于提取的特征进行分析得到最终检测结果，可以克服单一信号检测癫痫准确度较低的问题。这种检测装置虽然采集了多种生理信号，却没有采集脑电信号，这无疑大大降低了检测的精确度。
发明内容
本发明的目的在于提供一种癫痫检测方法及系统，部分地解决或缓解现有技术中的上述不足，能够更加精确地预测癫痫。
为了解决上述所提到的技术问题，本发明具体采用以下技术方案：
本发明的第一方面，在于提供一种癫痫检测系统，其包括：信号采集模块，用于采集被监测用户的多种生理信号；所述多种生理信号包括：脑电波信号和心电信号、加速度、肌电信号；预处理模块，与所述信号采集模块相连，用于对所述信号采集模块所采集的生理信号的信号组合信号进行预处理；数据分析模块，与所述预处理模块相连，用于通过神经网络模型对经过预处理得到的时频生理信号组合信号进行数据分析，以识别所述被监测用户是否发生癫痫；预警模块，与所述数据分析模块相连，用于当所述数据分析模块识别出所述被监测用户癫痫发作时，进行预警；其中，所述预处理模块包括：带通滤波器，用于滤除带外噪声信号，得到仅保留了带内信息的所述生理信号组合信号；中值滤波器，用于对去噪后的 所述生理信号组合信号进行基线偏移消除；平滑滤波器，用于对进行基线偏移消除后的所述生理信号组合信号进行去噪处理，以消除信号带内的噪声，得到真实的原始生理信号组合信号；差分滤波器，用于对消除信号带内噪声的所述原始生理信号组合信号进行滤波，得到待分析的所述时频生理信号组合信号。
在本发明的一些实施例中，所述信号采集模块包括：至少四个可贴合于所述被监测用户脑部，用于采集所述被监测用户的脑电信号的脑电信号采集微电极；与所述脑电信号采集微电极集成在一起的，用于将所述脑电信号采集微电极所采集的脑电信号发送至所述预处理模块的无线通信单元；与所述预处理模块、所述数据分析模块、所述预警模块集成在一起，用于采集所述被监测用户的心电信号的心率传感器。
在本发明的一些实施例中，所述信号采集模块还包括：与所述预处理模块、所述数据分析模块、所述预警模块集成在一起，且与所述数据分析模块相连的，分别用于采集肌电信号的肌电信号采集电极、用于采集加速度的加速度传感器，以及用于采集角速度的陀螺仪传感器。
在本发明的一些实施例中，所述心率传感器、所述预处理模块、所述数据分析模块、所述预警模块集成于一可穿戴装置上，所述可穿戴装置包括手环。
在本发明的一些实施例中，所述预警模块包括：语音单元，与所述数据分析模块相连，用于当所述数据分析模块识别出所述用户癫痫发作时，自动播报预存的求救语音信息；和/或，预警通知单元，与所述数据分析模块相连，用于当所述数据分析模块识别出所述用户癫痫发作时，通过物联网向预先关联的监护人的移动终端发送预警消息。
在本发明的一些实施例中，所述预警模块还包括：定位单元，与所述数据分析模块相连，用于当所述数据识别出所述用户癫痫发作时，向所述数据分析模块反馈所述被监测用户当前的定位信息，并获取最近的医疗机构的医疗机构信息；自动求助单元，与所述数据分析模块相连，用于当所述数据分析模块识别出所述用户癫痫发作时，将所述定位信息和所述生理信号发送至所述医疗机构。
在本发明的一些实施例中，所述一种癫痫检测系统还包括：无线通信模块，与所述数据分析模块相连，用于与预先绑定的所述被监测用户或监护人的移动终端进行数据通信。
在本发明的一些实施例中，所述数据分析模块包括：第一数据分析单元，用于当已指定所述被监测用户的发作类型时，根据所述发作类型匹配到相应的信号组合方式，并根 据所匹配到的信号组合方式触发所述信号采集模块采集相应的生理信号；或者，当未指定所述被监测用户的发作类型时，获取预设的默认信号组合方式，并根据所述默认信号组合方式触发所述信号采集模块采集相应的生理信号；第二数据分析单元，用于对经过所述预处理模块预处理后的所述生理信号进行数据分析，以识别所述被监测用户是否发生癫痫。
在本发明的一些实施例中，所述信号组合方式包括：第一组合方式：脑电、心电+加速度；第二组合方式：脑电+加速度；第三组合方式：脑电+心电+肌电+加速度；第四组合方式：脑电+肌电；第五组合方式：脑电+肌电+加速度。
本发明的第二方面，在于提供一种癫痫检测方法，其基于上述的癫痫检测系统，所述癫痫检测系统包括：用于采集被监测用户的多种生理信号的信号采集模块；所述多种生理信号包括：脑电波信号、心电信号、肌电信号、加速度和角速度；用于对所述信号采集模块所采集的生理信号进行预处理的预处理模块；用于通过神经网络模型对经过预处理得到的时频生理信号进行数据分析，以识别所述被监测用户是否发生癫痫的数据分析模块；用于当所述数据分析模块识别出所述被监测用户癫痫发作时，进行预警的预警模块；相应地，所述癫痫检测方法具体包括步骤：通过所述数据分析模块获取所述被监测用户的癫痫发作类型，并根据所述癫痫发作类型在数据库中匹配得到相应的信号组合方式；其中，所述信号组合方式包括：第一组合方式：脑电信号、心电信号+加速度；第二组合方式：脑电信号+加速度；第三组合方式：脑电信号+心电信号+肌电信号+加速度；第四组合方式：脑电信号+肌电信号；第五组合方式：脑电信号+肌电信号+加速度；通过所述数据分析模块触发所述信号采集模块按照所匹配到的信号组合方式采集相应的生理信号；通过所述预处理模块对所述信号采集模块所采集到的所述生理信号进行预处理，得到时频生理信号；通过所述数据分析模块对经过所述预处理模块进行预处理的所述视频生理信号进行数据分析，以识别所述被监测用户是否发生癫痫；若是，通过所述数据分析模块触发所述预警模块进行预警。
本发明的第三方面，在于提供一种癫痫预警系统，其包括上述的癫痫检测系统，服务器和用户终端，所述癫痫检测系统和所述用户终端通过无线网络与所述服务器进行数据通信；其中，癫痫检测系统，用于采集被监测用户的生理信号，并利用内置的与处理模块对所述生理信号的信号组合信号进行预处理后再进行数据分析，然后通过所述无线网络将所采集的生理信号组合信号和/或数据分析结果发送至所述服务器和/或所述用户终端；其中，所述预处理模块包括：带通滤波器，用于滤除带外噪声信号，得到仅保留了带内信息的所述生理信号组合信号；中值滤波器，用于对去噪后的所述生理信号组合信号进行基线偏移消除； 平滑滤波器，用于对进行基线偏移消除后的所述生理信号组合信号进行去噪处理，以消除信号带内的噪声，得到真实的原始生理信号组合信号；差分滤波器，用于对消除信号带内噪声的所述原始生理信号组合信号进行滤波，得到待分析的所述时频生理信号组合信号。
有益效果：本发明中通过将采集的原始生理信号依次通过带通滤波器、中值滤波器、平衡滤波器、差分滤波器进行预处理得到真实的时频生理信号，并输入预先训练好的神经网络模型进行识别，相较于仅通过带通滤波和中值滤波后进行特征提取，然后根据提取的特征进行识别的方式，由于输入神经网络模型的是真实的生理信号，不会因为采用特征提取而遗漏或丢失数据，因此，其检测准确率更高。本发明中通过采集多种生理信号，且每种信号组合中都包括最直接反映被监测用户癫痫发作时的放电模式的脑电信号，相较于无脑电信号的多信号检测方式，大大提高了精确率；另一方面，通过设置在被监测用户脑部固定位置的四个微电极采集脑电波，使得便于携带，无需被监测用户到医院或采用大型设备进行检测，使得用户可以随时监测。
附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作一简单地介绍。在所有附图中，类似的元件或部分一般由类似的附图标记标识。附图中，各元件或部分并不一定按照实际的比例绘制。显而易见地，下面描述中的附图是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其它的附图。
图1为本发明一示例性实施例的癫痫检测系统的功能模块图；
图2a和图2b分别为本发明一示例性实施例的癫痫检测系统中四个脑电信号微电极粘贴于头部固定位置的示意图；
图3a为采集到的被监测用户的原始脑电波信号；
图3b为图3a所示的原始脑电波信号经过带通滤波后的脑电波信号；
图3c为图3b所示的脑电波信号经过中值滤波处理后的脑电波信号；
图3d为图3c所示的脑电波信号经过平滑滤波处理后的脑电波时频信号；
图4a至图4g为分别采用两种癫痫检测手环对30例被监测用户进行检测的识别结果。
具体实施方式
为使本发明实施例的目的、技术方案和优点更加清楚，下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述。显然，所描述的实施例是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。本文中，使用用于表示元件的诸如“模块”、“部件”或“单元”的后缀仅为了有利于本发明的说明，其本身没有特定的意义。因此，“模块”、“部件”或“单元”可以混合地使用。本文中，术语“上”、“下”、“内”、“外”“前”、“后”、“一端”、“另一端”等指示的方位或位置关系为基于附图所示的方位或位置关系，仅是为了便于描述本发明和简化描述，而不是指示或暗示所指的装置或元件必须具有特定的方位、以特定的方位构造和操作，因此不能理解为对本发明的限制。此外，术语“第一”、“第二”仅用于描述目的，而不能理解为指示或暗示相对重要性。本文中，除非另有明确的规定和限定，术语“安装”、“设置有”、“连接”等，应做广义理解，例如“连接”，可以是固定连接，也可以是可拆卸连接，或一体地连接；可以是机械连接，可以是直接相连，也可以通过中间媒介间接相连，可以是两个元件内部的连通。对于本领域的普通技术人员而言，可以具体情况理解上述术语在本发明中的具体含义。本文中“和/或”包括任何和所有一个或多个列出的相关项的组合。本文中“多个”意指两个或两个以上，即其包含两个、三个、四个、五个等。
实施例1：参见图1，为本发明的癫痫检测系统的一示例性实施例的功能模块图，具体地，该癫痫检测系统包括：信号采集模块，用于采集被监测用户的多种生理信号；其中，该生理信号包括：脑电波信号和心电信号；预处理模块，用于对信号采集模块所采集的生理信号进行预处理；具体地，该预处理模块包括：带通滤波器，用于滤除带外噪声信号，得到仅保留了带内信息的所述生理信号；中值滤波器，用于对去噪后的所述生理信号进行基线偏移消除；平滑滤波器，用于对进行基线偏移消除后的所述生理信号进行去噪处理，以消除信号带内的噪声，得到真实的原始生理信号；差分滤波器，用于对消除信号带内噪声的所述生理信号进行滤波，得到待分析的时频生理信号；数据分析模块，用于通过神经网络模型对经过预处理的信号进行数据分析，以识别被监测用户是否发生癫痫；具体地，预先收集大量的癫痫患者(包括不同类型的癫痫患者)癫痫发作时的生理信号，并利用上述的预处理模块将所收集的生理信号进行预处理后，作为训练神经网络模块的训练样本；具体地，可采用CNN神经网络模型；预警模块，用于当上述数据分析模块识别出被监测用户癫痫发作时，进行预 警。
本实施例中，首先通过带通滤波器滤除带外噪声信号，只保留生理信号带内信息对于带内信号，如果存在基线偏移，信号特征检测会受基线漂移影响，存在大概率的误判，所以必须进行基线偏移消除；消除基线偏移后的信号带内还存在带内噪声，这时通过自适应平滑滤波器对带内噪声进行滤除，更真实的还原原始信号；为了使生理信号的信号变化特征更加明显，用差分滤波器对其进行滤波，滤波后的信号幅度变化更明显，送入神经网络CNN后检测准确率更高。具体地，如图3a所示，为采集到被监测用户的脑电波信号，如图3b所示，该脑电波信号经过带通滤波器处理后的脑电波信号，相较于原始信号，消除了带外干扰噪声，如图3c所示，该脑电波信号经过中值滤波器处理后的脑电波信号，消除了基线漂移，如图3d所示，该脑电波信号经过平滑滤波器处理后的脑电波信号，消除了信号中的毛刺。最后通过差分滤波器将信号进行放大。也即通过上述四个滤波器提高了信号质量，从而使得将高质量的生理信号输入数据分析模块进行机器学习，或者自动识别，大大提高了识别的精确率。
上述四个滤波器的顺序不能改变，如果先进行差分再进行基线消除和平滑滤波，则带内的高幅度变化噪声毛刺等信号也会因为幅度变化大在后续的平滑滤波器时误判为生理信号。
相较于现有技术中，为了提取特征，因此，仅对所采集到的信号进行带通滤波和中值滤波；然后进行特征提取，并将提取后的时域或者频域特征型号送入CNN神经网络的方式，而本实施例中，由于采用直接将所采集的生理信号进行预处理后得到的时频信号输入预先训练好的CNN神经网络进行癫痫识别，大大提高了检测精确率。
为了佐证本发明的该癫痫检测系统的精确率，分别采用特征提取的方式和本实施例的方式对30例被监测用户进行癫痫检测，具体地，该30例患者分别左右手穿戴两种癫痫检测系统，检测结果如图4a-图4g。针对该30例被监测用户，本实施例的该癫痫检测系统的识别出30例被监测用户均为癫痫发作，与医院诊断结果一致，其精确率为100％，而采用特征提取方式进行癫痫识别的装置，识别30例被监测用户中28例为癫痫发作，有2例与医院诊断结果不一致，其精确率为93.33％。
在一些实施例中，该信号采集模块具体包括：至少四个可贴合于被监测用户脑部，用于采集被监测用户的脑电信号的脑电信号采集微电极；与脑电信号采集微电极集成在一起 的，用于将脑电信号采集微电极所采集的脑电信号发送至预处理模块的无线通信单元；与预处理模块、数据分析模块、预警模块集成在一起，用于采集被监测用户的心电信号的心率传感器。
在一些实施例中，采集脑电信号的四个微电极作为可分别粘贴于被监测用户脑部上四个固定位置，参见图2a和图2b。通过设置该四个微电极，并将其与无线通信单元集成在一起，但独立于上述数据分析模块、预处理模块和预警模块，也即该四个微电极和无线通信单元作为该检测系统的便携式附件，从而使得无需被监测用户到医院等医疗机构用专用设备采集脑电信号，并且也便于携带，且隐蔽性较好，也便于被监测用户更好地随时进行监测。当然，进一步还可在两者之间设置一个模数转换器。
在一些实施例中，上述信号采集模块还包括：与上述预处理模块、上述数据分析模块、上述预警模块集成在一起，且与上述数据分析模块相连的，分别用于采集肌电信号的肌电信号采集电极、用于采集加速度的加速度传感器，以及用于采集角速度的陀螺仪传感器。其中，上述心率传感器、上述预处理模块、上述数据分析模块、上述预警模块集成于一可穿戴装置上，上述可穿戴装置包括手环。
在一些实施例中，上述预警模块包括：语音单元，与上述数据分析模块相连，用于当数据分析模块识别出被监测用户癫痫发作时，自动播报预存的求救语音信息；和/或，预警通知单元，与上述数据分析模块相连，用于当数据分析模块识别出被监测用户癫痫发作时，通过物联网向预先关联的监护人的移动终端发送预警消息。
当被监测用户外出，尤其是旅游或出差过程中，若癫痫发作，周围人员通常并不清楚知晓其具体情况，自然也就无法立即给予相应的救助，因此，通过设置一个语音单元来播放预先存储的求救语音信息，例如，“患者癫痫发作，请拨打XXXXX”、“患者癫痫发作，请将患者随身携带的XXX包里的XXX放置到患者嘴部”···，不仅使得周围的人群能够明确知晓被监测用户的具体情况，还能够即时根据语音信息做出相应的救助。另一方面，还可通过预警通知单元即时通知监护人被监测用户癫痫发作，从而使得监护人可以迅速做出相应的措施，例如，当监护人离被监测用户较近时，可迅速来到被监测用户身边以进行救助。
在一些实施例中，上述预警模块还包括：定位单元，与所述数据分析模块相连，用于当所述数据识别出所述用户癫痫发作时，向所述数据分析模块反馈所述被监测用户当前的定位信息，并获取最近的医疗机构的医疗机构信息；自动求助单元，与所述数据分析模块 相连，用于当所述数据分析模块识别出所述用户癫痫发作时，将所述定位信息和所述生理信号发送至所述医疗机构。当被监测用户外出，尤其是旅游或出差过程中，若癫痫发作，通过该定位单元(例如，GPS等)能够准确定位到被监测用户当前的位置信息，从而向监护人发送预警通知的同时，也可向其发送定位信息，使得即使监护人离被监测用户较远也能够掌握其动态；同时，还可自动根据被监测用户当前的定位信息搜索到最近的医疗机构，从而向其(例如，该医疗机构的急救门诊)发送相应的求救信息(包括定位信息、生理信号和被监测用户基本信息等)。
在一些实施例中，该癫痫检测系统还包括：无线通信模块，与所述数据分析模块相连，用于与预先绑定的所述被监测用户或监护人的移动终端进行数据通信。
在一些实施例中，上述数据分析模块包括：第一数据分析单元，用于当已指定所述被监测用户的发作类型时，根据所述发作类型匹配到相应的信号组合方式，并根据所匹配到的信号组合方式触发所述信号采集模块采集相应的生理信号；或者，当未指定所述被监测用户的发作类型时，获取预设的默认信号组合方式，并根据所述默认信号组合方式触发所述信号采集模块采集相应的生理信号；第二数据分析单元，用于利用预先训练好的神经网络模型对经过所述预处理模块预处理后的所述生理信号进行数据分析，以识别所述被监测用户是否发生癫痫。
在一些实施例中，上述信号组合方式包括：第一组合方式：脑电信号、心电信号+加速度；第二组合方式：脑电信号+加速度；第三组合方式：脑电信号+心电信号+肌电信号+加速度；第四组合方式：脑电信号+肌电信号；第五组合方式：脑电信号+肌电信号+加速度。
下面结合工作原理，对本实施例的该癫痫检测系统进行说明。
当被监测用户首次启动该癫痫检测系统时，可通过该癫痫检测系统的输入模块(例如，触摸屏等)选择自身的癫痫发作类型；数据分析模块根据该被监测用户所选定的癫痫发作类型，触发信号采集模块采集相应的生理信号；其中，该数据分析模块中预先存储有每种癫痫发作类型对应的生理信号组合方式，因此，一旦选定一种癫痫发作类型，就会自动根据该癫痫发作类型匹配到相应的生理信号组合，然后触发信号采集模块中相应的传感器采集相应的生理信号；信号采集模块在数据分析模块的触发作用下，采用相应的生理信号，并发送至预处理模块进行预处理，然后再将预处理后的生理次信号发送至数据分析模块进行数据分析，以识别该被监测用户是否发生癫痫；若是，则预警模块自动语音播报报警信号，从而使得被 监测用户癫痫发作时，可以吸引附近人并告知具体情况，进而使得附近的人可以提供帮助，并且还通过无线通信模块将预警信息发送至预先关联的监护人的移动终端，使得监护人随时掌握被监测用户的状态。当然，进一步地，当数据分析模块识别出被监测用户发生癫痫时，该预警模块根据被监测用户当前的定位信息自动搜索最近的医疗机构，并拨打相应的求救电话，同时发送定位信息至该医疗机构，从而提高救助的效率。
当然，若用户未选定发作类型时，默认采集所有的生理信号(即脑电信号、心电信号、加速度、角速度、肌电信号)，并按照上述所有的信号组合进行数据分析，以进行识别，然后根据被监测用户或其监护人的反馈动态确定一个信号组合，针对该被监测用户，后期都以将只采集该信号组合中的各生理信号即可，不再考虑其他信号组合，以降低功耗。
实施例2：基于上述的癫痫检测系统，本发明还提供了一种癫痫预警系统，其上述的癫痫检测系统，服务器和用户终端，其中，该癫痫检测系统和用户终端通过无线网络与服务器进行数据通信；具体地，该癫痫检测系统，用于采集被监测用户的生理信号，并利用内置的预处理模块对生理信号进行预处理后再进行数据分析，然后通过无线网络将识别结果和/或所采集的生理信号和/或数据分析结果发送至所述服务器和/或所述用户终端。在一些实施例中，该用户终端包括该被监测用户的移动终端，和/或，监护人的移动终端，从而使得被监测用户和/或监护人可以随时了解监测情况。
实施例3：基于上述实施例1的癫痫检测系统，本发明还提供了一种癫痫检测方法，具体地，该癫痫检测方法包括步骤：通过数据分析模块获取被监测用户的癫痫发作类型，并根据癫痫发作类型在数据库中匹配得到相应的信号组合方式；通常，被监测用户首次使用该癫痫检测系统，例如，首次佩戴手环形式的该癫痫检测系统时，会在癫痫检测系统中设定自己的癫痫发作类型，从而实现盖癫痫检测系统的初始化，即自动根据所设定的癫痫发作类型匹配相应的信号组合方式，从而使得后续检测过程中，按照所匹配到的信号组合方式来触发信号采集模块中相应的信号采集单元采集相应的生理信号；具体地，该信号组合方式包括：第一组合方式：脑电信号、心电信号+加速度；第二组合方式：脑电信号+加速度；第三组合方式：脑电信号+心电信号+肌电信号+加速度；第四组合方式：脑电信号+肌电信号；第五组合方式：脑电信号+肌电信号+加速度；当然，在另一些实施例中，若首次使用该系统的被监测用户未指定癫痫发作类型时，该数据分析模块触发该信号采集模块按照默认的信号组合方式来采集相应的生理信号；优选地，默认采集所有的生理信号，或者默认匹配到上述第三种信号组合方式；通过所述数据分析模块触发所述信号采集模块按照所匹配到的信号组合方式 采集相应的生理信号；例如，当数据分析模块匹配到第一组合方式时，触发信号采集模块仅采集脑电信号、心电信号和加速度即可；通过上述预处理模块对信号采集模块所采集到的生理信号进行预处理，得到时频生理信号；通过所述数据分析模块对经过所述预处理模块进行预处理的所述视频生理信号进行数据分析，以识别所述被监测用户是否发生癫痫；若是，通过数据分析模块触发预警模块进行预警。
在一些实施例中，触发该预警模块进行预警的方式包括以下任一种或多种：1)向预先关联的监护人的移动终端发送预警信息；2)按照预存的语音信息自动进行语音播报；3)获取被监测用户当前的定位信息，以及附近的医疗机构，并将该被监测用户的定位信息和基本信息、所采集到的生理信号发送至该医疗机构，以使得该医疗机构能够提前做好救助准备工作，并准确获取被监测用户的位置。具体地的预警方式可根据被监测用户的发作类型和严重程度预先设定。
需要说明的是，在本文中，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者装置不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者装置所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括该要素的过程、方法、物品或者装置中还存在另外的相同要素。上面结合附图对本发明的实施例进行了描述，但是本发明并不局限于上述的具体实施方式，上述的具体实施方式仅仅是示意性的，而不是限制性的，本领域的普通技术人员在本发明的启示下，在不脱离本发明宗旨和权利要求所保护的范围情况下，还可做出很多形式，这些均属于本发明的保护之内。

Claims (7)

  1. 一种癫痫检测系统，其特征在于，包括：
    信号采集模块，用于采集被监测用户的多种生理信号；所述多种生理信号包括：脑电波信号和心电信号、加速度、肌电信号；
    预处理模块，与所述信号采集模块相连，用于对所述信号采集模块所采集的多种生理信号的信号组合信号进行预处理；
    数据分析模块，与所述预处理模块相连，用于通过神经网络模型对经过预处理得到的时频生理信号组合信号进行数据分析，以识别所述被监测用户是否发生癫痫；
    预警模块，与所述数据分析模块相连，用于当所述数据分析模块识别出所述被监测用户癫痫发作时，进行预警；
    其中，所述预处理模块包括：带通滤波器，用于滤除带外噪声信号，得到仅保留了带内信息的生理信号组合信号；中值滤波器，用于对去噪后的所述生理信号组合信号进行基线偏移消除；平滑滤波器，用于对进行基线偏移消除后的所述生理信号组合信号进行去噪处理，以消除信号带内的噪声，得到真实的原始生理信号组合信号；差分滤波器，用于对消除信号带内噪声的所述原始生理信号组合信号进行滤波，得到待分析的所述时频生理信号组合信号；
    其中，所述神经网络模型的训练样本是预先收集到的大量不同类型癫痫患者癫痫发作时的生理信号组合信号，并依次经过所述预处理模块中的带通滤波器、中值滤波器、平滑滤波器、差分滤波器进行预处理；而不同类型癫痫对应于相应的信号组合方式，所述信号组合方式包括：第一组合方式：脑电信号、心电信号+加速度；第二组合方式：脑电信号+加速度；第三组合方式：脑电信号+心电信号+肌电信号+加速度；第四组合方式：脑电信号+肌电信号；第五组合方式：脑电信号+肌电信号+加速度；
    其中，所述数据分析模块包括：
    第一数据分析单元，用于当已指定所述被监测用户的发作类型时，根据所述发作类型匹配到相应的信号组合方式，并根据所匹配到的信号组合方式触发所述信号采集模块采集相应的生理信号；或者，当未指定所述被监测用户的发作类型时，获取预设的默认信号组合方式，并根据所述默认信号组合方式触发所述信号采集模块采集相应的生理信号；
    第二数据分析单元，用于对经过所述预处理模块预处理后的所述生理信号组合信号进行数据分析，以识别所述被监测用户是否发生癫痫。
  2. 根据权利要求1所述的一种癫痫检测系统，其特征在于，所述信号采集模块包括：至少四个可贴合于所述被监测用户脑部，用于采集所述被监测用户的脑电信号的脑电信号采集微电极；与所述脑电信号采集微电极集成在一起的，用于将所述脑电信号采集微电极所采集的脑 电信号发送至所述预处理模块的无线通信单元；与所述预处理模块、所述数据分析模块、所述预警模块集成在一起，用于采集所述被监测用户的心电信号的心率传感器。
  3. 根据权利要求2所述的一种癫痫检测系统，其特征在于，所述信号采集模块还包括：与所述预处理模块、所述数据分析模块、所述预警模块集成在一起，且与所述数据分析模块相连的，分别用于采集肌电信号的肌电信号采集电极、用于采集加速度的加速度传感器，以及用于采集角速度的陀螺仪传感器。
  4. 根据权利要求3所的一种癫痫检测系统，其特征在于，所述心率传感器、所述预处理模块、所述数据分析模块、所述预警模块集成于一可穿戴装置上，所述可穿戴装置包括手环。
  5. 根据权利要求4所述的一种癫痫检测系统，其特征在于，所述预警模块包括：
    语音单元，与所述数据分析模块相连，用于当所述数据分析模块识别出所述用户癫痫发作时，自动播报预存的求救语音信息；和/或，预警通知单元，与所述数据分析模块相连，用于当所述数据分析模块识别出所述用户癫痫发作时，通过物联网向预先关联的监护人的移动终端发送预警消息；和/或，所述癫痫检测系统还包括：无线通信模块，与所述数据分析模块相连，用于与预先绑定的所述被监测用户或监护人的移动终端进行数据通信。
  6. 根据权利要求5所述的一种癫痫检测系统，其特征在于，所述预警模块还包括：
    定位单元，与所述数据分析模块相连，用于当所述数据识别出所述用户癫痫发作时，向所述数据分析模块反馈所述被监测用户当前的定位信息，并获取最近的医疗机构的医疗机构信息；自动求助单元，与所述数据分析模块相连，用于当所述数据分析模块识别出所述用户癫痫发作时，将所述定位信息和所述生理信号发送至所述医疗机构。
  7. 一种癫痫预警系统，其特征在于，包括根据权利要求4至6中任一所述的癫痫检测系统，服务器和用户终端，所述癫痫检测系统和所述用户终端通过无线网络与所述服务器进行数据通信；其中，癫痫检测系统，用于采集被监测用户的生理信号组合信号，并通过内置的预处理模块对所述生理信号组合信号进行预处理后再进行数据分析，然后通过所述无线网络将所采集的所述生理信号组合信号和/或数据分析结果发送至所述服务器和/或所述用户终端；
    其中，所述预处理模块包括：带通滤波器，用于滤除带外噪声信号，得到仅保留了带内信息的所述生理信号组合信号；中值滤波器，用于对去噪后的所述生理信号组合信号进行基线偏移消除；平滑滤波器，用于对进行基线偏移消除后的所述生理信号组合信号进行去噪处理，以消除信号带内的噪声，得到真实的原始生理信号组合信号；差分滤波器，用于对消除信号带内噪声的所述原始生理信号组合信号进行滤波，得到待分析的所述时频生理信号组合信号。
PCT/CN2024/073897 2022-12-02 2024-01-24 癫痫检测方法及系统 WO2024114834A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211540428.6A CN115770050B (zh) 2022-12-02 2022-12-02 癫痫检测方法及系统
CN202211540428.6 2022-12-02

Publications (1)

Publication Number Publication Date
WO2024114834A1 true WO2024114834A1 (zh) 2024-06-06

Family

ID=85391301

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2024/073897 WO2024114834A1 (zh) 2022-12-02 2024-01-24 癫痫检测方法及系统

Country Status (2)

Country Link
CN (1) CN115770050B (zh)
WO (1) WO2024114834A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115770050B (zh) * 2022-12-02 2023-09-08 重庆医科大学附属第二医院 癫痫检测方法及系统

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050197590A1 (en) * 1997-01-06 2005-09-08 Ivan Osorio System for the prediction, rapid detection, warning, prevention, or control of changes in activity states in the brain of a subject
US20090124870A1 (en) * 2006-06-07 2009-05-14 Hobo Heeze B.V. Patient monitoring system for the real-time detection of epileptic seizures
US20090171168A1 (en) * 2007-12-28 2009-07-02 Leyde Kent W Systems and Method for Recording Clinical Manifestations of a Seizure
CN103476326A (zh) * 2011-01-28 2013-12-25 纽罗斯凯公司 用于病发作检测和监控的干式传感器eeg/emg和运动传感系统
CN111616682A (zh) * 2020-05-31 2020-09-04 天津大学 基于便携式脑电采集设备的癫痫发作预警系统及应用
CN112353368A (zh) * 2020-05-08 2021-02-12 北京理工大学 一种基于反馈调节的多输入信号癫痫发作检测系统
CN114099174A (zh) * 2020-06-09 2022-03-01 首都医科大学宣武医院 一种癫痫患儿用监测系统及方法
CN115770050A (zh) * 2022-12-02 2023-03-10 重庆医科大学附属第二医院 癫痫检测方法及系统

Family Cites Families (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1820444B1 (en) * 2006-02-17 2011-09-21 General Electric Company Detection of epileptiform activity
US20100228103A1 (en) * 2009-03-05 2010-09-09 Pacesetter, Inc. Multifaceted implantable syncope monitor - mism
MX370873B (es) * 2010-10-15 2020-01-08 Brain Sentinel Inc Metodo y aparato para detectar convulsiones.
US9681836B2 (en) * 2012-04-23 2017-06-20 Cyberonics, Inc. Methods, systems and apparatuses for detecting seizure and non-seizure states
US20200138318A1 (en) * 2012-06-12 2020-05-07 Brain Sentinel, Inc. Methods and systems for seizure analysis
US11033220B2 (en) * 2013-12-05 2021-06-15 Livanova Usa, Inc. Systems and methods of limb-based accelerometer assessments of neurological disorders
US20150282755A1 (en) * 2014-04-02 2015-10-08 King Fahd University Of Petroleum And Minerals System and method for detecting seizure activity
US10980469B2 (en) * 2015-10-08 2021-04-20 Brain Sentinel, Inc. Method and apparatus for detecting and classifying seizure activity
CN108714020B (zh) * 2018-05-18 2020-07-07 中国人民解放军陆军军医大学第三附属医院（野战外科研究所） 自适应心理调节仪、实时眼动延时求取方法及存储介质
CN109620248B (zh) * 2019-02-14 2021-12-14 北京瑞尔唯康科技有限公司 癫痫发作监测系统
US10743809B1 (en) * 2019-09-20 2020-08-18 CeriBell, Inc. Systems and methods for seizure prediction and detection
CN111671419B (zh) * 2020-06-12 2022-02-11 山东大学 一种基于脑电信号的癫痫早期检测与识别方法及系统
CN112043261B (zh) * 2020-09-29 2023-09-22 黑龙江吉远健康科技有限公司 一种基于肢端动作信号的癫痫发作状态识别方法
CN114631780A (zh) * 2021-12-02 2022-06-17 北京理工大学 一种基于多级分类的低功耗可穿戴癫痫发作检测系统
CN114176598B (zh) * 2021-12-22 2023-09-29 昆明市儿童医院 基于多模态的癫痫发病预警方法及智能手环装置
CN114916940A (zh) * 2022-05-17 2022-08-19 上海交通大学 基于专家知识的心律失常信号检测方法、系统及终端

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050197590A1 (en) * 1997-01-06 2005-09-08 Ivan Osorio System for the prediction, rapid detection, warning, prevention, or control of changes in activity states in the brain of a subject
US20090124870A1 (en) * 2006-06-07 2009-05-14 Hobo Heeze B.V. Patient monitoring system for the real-time detection of epileptic seizures
US20090171168A1 (en) * 2007-12-28 2009-07-02 Leyde Kent W Systems and Method for Recording Clinical Manifestations of a Seizure
US20150164403A1 (en) * 2007-12-28 2015-06-18 Cyberonics, Inc. Method for detecting neurological and clinical manifestations of a seizure
CN103476326A (zh) * 2011-01-28 2013-12-25 纽罗斯凯公司 用于病发作检测和监控的干式传感器eeg/emg和运动传感系统
CN112353368A (zh) * 2020-05-08 2021-02-12 北京理工大学 一种基于反馈调节的多输入信号癫痫发作检测系统
CN111616682A (zh) * 2020-05-31 2020-09-04 天津大学 基于便携式脑电采集设备的癫痫发作预警系统及应用
CN114099174A (zh) * 2020-06-09 2022-03-01 首都医科大学宣武医院 一种癫痫患儿用监测系统及方法
CN115770050A (zh) * 2022-12-02 2023-03-10 重庆医科大学附属第二医院 癫痫检测方法及系统

Also Published As

Publication number Publication date
CN115770050B (zh) 2023-09-08
CN115770050A (zh) 2023-03-10

Similar Documents

Publication Publication Date Title
US11553869B1 (en) Multi-channel brain or cortical activity monitoring and method
US10485471B2 (en) System and method for identifying ictal states in a patient
WO2024114834A1 (zh) 癫痫检测方法及系统
JP2018149355A (ja) 監視のための装置および方法
EP2911578B1 (en) Systems for detecting brain-based bio-signals
US10213152B2 (en) System and method for real-time measurement of sleep quality
WO2017075856A1 (zh) 基于小波分析远程心电监护与预警系统及方法
CN101637384B (zh) Ecg监视系统中信号质量指示和假警报减少的系统和方法
Pham et al. WAKE: a behind-the-ear wearable system for microsleep detection
US10694946B2 (en) Dual EEG non-contact monitor with personal EEG monitor for concurrent brain monitoring and communication
EP3150121A1 (en) Device for prediction and warning of hypoglycaemic attack
JPH07508185A (ja) 生理学的監視のためのセンサ及びシステム
DE102007020995A1 (de) System zur Kontrolle physiologischer Funktionen
CN204950942U (zh) 用于心电监测的手机壳
TW201020145A (en) System for health management used in vehicle and method thereof
CN112037916A (zh) 共享多功能防猝死生理信息检测系统及其方法
Charrad et al. ECG patch monitor: a telemedicine system for remote monitoring and assisting patients during a heart attack
CN113907756A (zh) 一种基于多种模态的生理数据的可穿戴系统
KR102475035B1 (ko) 생체 신호를 이용한 사용자 상태 모니터링 방법 및 이를 실행하는 시스템
CN114795149A (zh) 一种可穿戴心脑血管疾病监测装置
CN107485387B (zh) 穴位自动跟踪系统以及方法
Vijayprasath et al. Assessment of relationship between heart rate variability and drowsiness of post operative patients in driving conditions
WO2017147977A1 (zh) 一种贴片式健康监测设备
DE202019103671U1 (de) Tragbares Gehirnfunktionsüberwachungsgerät
US20190159691A1 (en) Method and system for cardiac health monitoring