CN117598685A - Respiration detection device, method and system acting on nerve regulation and control - Google Patents

Abstract

The invention relates to a respiration detection device, method and system for neuromodulation, wherein the device comprises a front end part and a rear end part; the front end part includes: a respiration signal collector for collecting raw signal data related to respiration; the signal extraction unit is used for preprocessing the original signal data related to respiration to obtain preprocessed data; a signal transmission unit for transmitting the preprocessing data to the back-end component; the rear end piece includes: the algorithm processing unit is used for carrying out threshold processing on the preprocessing data to obtain respiratory signal monitoring data; and the abnormality judgment unit is used for carrying out abnormality judgment on the respiratory signal monitoring data to obtain an abnormality judgment result. The invention can accurately monitor and collect the breathing activity data of the individual in real time.

Description

A breathing detection device, method and system for neural regulation

Technical field

The present invention relates to the technical field of medical devices, and in particular to a respiratory detection device, method and system for neural regulation.

Background technique

Using Internet of Things technology, various wearable or portable physiological signal collection devices can be used to collect respiratory information of special groups in real time and long-term, and artificial intelligence technology can be used to analyze the health status of patients, thereby achieving uninterrupted health monitoring and achieving corresponding goals. In recent years, the pressure and fast pace of life have led to the emergence of a large number of mentally sub-healthy people. At the same time, as the economic level has improved significantly, people have become more concerned about physical health and have higher requirements for life comfort. For some diseases such as rare muscles, Clonic dystonia syndrome, stress gastric dysfunction, etc. Neuromodulation can adjust or improve body functions by affecting the activities of the central nervous system without causing damage to brain tissue or postoperative sequelae. Therefore, neuromodulation Medical treatments are gradually attracting attention, and breathing is an important physiological process that is closely related to the regulation and health of the body. Breath detection systems are designed to monitor and evaluate an individual's respiratory activity to provide real-time, accurate measurements of parameters such as respiratory status, respiratory rate and depth.

Existing medical technologies often have problems such as drug side effects or untimely treatment. For example, sleeping pills to aid sleep have problems of dependence and addiction, and can also cause damage to liver and kidney function. For special groups who lack human care, when the disease breaks out, Failure to provide timely rescue is life-threatening.

Although there are currently some IoT monitoring solutions for medical purposes, most of these solutions are complex in design, high in cost, not comfortable to wear, and cannot provide timely and effective treatment to patients.

Contents of the invention

The present invention provides a respiratory detection device, method and system for neural regulation, which can monitor the user's physical condition in real time and deal with it in a timely manner through neural regulation treatment means with good safety performance in case of abnormality, so as to satisfy the user's corresponding physiological condition. need.

The technical solution adopted by the present invention to solve the technical problem is to provide a respiratory detection device for neural regulation, including a front-end component and a back-end component;

The front-end components include:

Respiration signal collector, used to collect original signal data related to respiration;

A signal extraction unit, used to preprocess the original signal data related to breathing to obtain preprocessed data;

A signal transmission unit for transmitting the preprocessed data to the back-end component;

The backend components include:

An algorithm processing unit, used to perform threshold processing on the preprocessed data to obtain respiratory signal monitoring data;

An abnormality judgment unit is used to judge the abnormality of the respiratory signal monitoring data and obtain an abnormality judgment result.

The signal extraction unit includes:

An analog-to-digital converter, used for analog-to-digital conversion of the original signal data related to respiration to obtain a digital signal related to respiration;

A filter, used to perform filtering and noise reduction processing on the respiration-related digital signal to obtain a filtered signal;

An amplifier is used to amplify the filtered signal to obtain preprocessed data.

The filter adopts a Butterworth low-pass filter with a cutoff frequency of 10 Hz, which is used to extract respiratory-related digital signals less than or equal to 8 Hz as filtered signals.

The algorithm processing unit includes:

Set subunits for setting valley threshold and peak threshold;

a recording subunit, configured to record data below the valley threshold in the preprocessed data as valley values and data above the peak threshold as peak values;

Select the calculation subunit for selecting the latest N peaks or valleys, and counting the interval time between adjacent peaks or valleys among the N peaks or valleys, and calculating the average of the interval times, and dividing the The average value is used as respiratory signal monitoring data.

The setting unit is also used to set the minimum interval time; when recording the valley value and peak value, the recording subunit compares the interval time between the current valley value or peak value and the previous valley value or peak value with the minimum interval time. If the current valley value or peak value is If the interval between the valley value or peak value and the previous valley value or peak value is less than the minimum interval time, the current valley value or peak value will not be recorded.

The abnormality judgment unit includes:

Determining subunit, used to determine the relationship between the respiratory signal monitoring data and time;

The classification subunit is used to input the relationship between the respiratory signal monitoring data and time into an abnormality determination model to obtain an abnormality determination result; wherein, the abnormality determination model is obtained through machine learning using a training set, wherein, The training set covers normal and disease respiratory signal monitoring data. It uses time series analysis methods to extract features from the respiratory signal monitoring data and annotate them based on existing medical knowledge.

The respiratory signal collector is one or more of a humidity sensor, a temperature sensor or a stress sensor.

The technical solution adopted by the present invention to solve the technical problem is to provide a breathing detection method for neural regulation, which includes the following steps:

Collect raw signal data related to breathing;

Preprocess the original signal data related to respiration to obtain preprocessed data;

Perform threshold processing on the preprocessed data to obtain respiratory signal monitoring data;

Abnormality judgment is performed on the respiratory signal monitoring data to obtain an abnormality judgment result.

Preprocessing the original signal data related to respiration to obtain preprocessed data specifically includes:

Perform analog-to-digital conversion on the original signal data related to respiration to obtain a digital signal related to respiration;

Perform filtering and noise reduction processing on the respiration-related digital signals to obtain filtered signals;

The filtered signal is amplified to obtain preprocessed data.

Performing threshold processing on the preprocessed data to obtain respiratory signal monitoring data specifically includes:

Set valley threshold and peak threshold;

Record the data below the valley threshold in the preprocessed data as valley values and the data above the peak threshold as peak values;

Select the most recent N peaks or valleys, count the intervals between adjacent peaks or valleys among the N peaks or valleys, calculate the average of the intervals, and use the average as respiratory signal monitoring data .

The technical solution adopted by the present invention to solve the technical problem is to provide a neural regulation system, including a feedback component, which is characterized in that it also includes the above-mentioned respiratory detection device for neural regulation, and the feedback component determines based on the abnormality The result is user-selective neuromodulation.

beneficial effects

Due to the adoption of the above technical solution, the present invention has the following advantages and positive effects compared with the existing technology: the present invention can monitor and collect individual respiratory activity data in real time and accurately, and continuously monitor respiratory status and patterns. , providing immediate neurological treatment, which realizes closed-loop health management of real-time monitoring and feedback treatment, and can effectively protect the life safety of users in a timely manner. Among them, the respiratory signal collector for monitoring can use non-invasive sensors, which eliminates the need for surgery or invasive operations and is convenient for users to wear and use.

Description of drawings

Figure 1 is a schematic diagram of a respiratory detection device for neural regulation according to the first embodiment of the present invention;

Figure 2 is a flow chart of a breathing detection method for neural regulation according to the second embodiment of the present invention;

Figure 3 is a schematic diagram of a neuromodulation system according to a third embodiment of the present invention.

Detailed ways

The present invention will be further described below in conjunction with specific embodiments. It should be understood that these examples are only used to illustrate the invention and are not intended to limit the scope of the invention. In addition, it should be understood that after reading the teachings of the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms also fall within the scope defined by the appended claims of this application.

The first embodiment of the present invention relates to a respiratory detection device for neural regulation. As shown in Figure 1, it includes a front-end component and a rear-end component.

Among them, the front-end components include: a respiratory signal collector, used to collect original signal data related to breathing; a signal extraction unit, used to preprocess the original signal data related to breathing to obtain preprocessed data; a signal transmission unit , used to transmit the preprocessed data to the backend component.

The back-end component includes: an algorithm processing unit for performing threshold processing on the preprocessed data to obtain respiratory signal monitoring data; and an abnormality judgment unit for performing abnormal judgment on the respiratory signal monitoring data to obtain abnormality judgment results.

The respiratory signal collector in this embodiment may be one or more of a humidity sensor, a temperature sensor or a stress sensor. When the respiratory signal collector is a humidity sensor, it is placed near the nose of the monitored person to monitor the respiratory frequency of the monitored person through the humidity change during the breathing process; when the respiratory signal collector is a temperature sensor, it can also be placed Monitor the respiratory rate of the monitored person through temperature changes during the breathing process near the nose of the monitored person; when the respiratory signal collector is a stress sensor, the stress sensor can be an electronic skin, so that it can be placed on the chest surface of the monitored person The monitored person's respiratory rate is monitored through changes in chest muscle expansion during breathing.

The signal extraction unit in this embodiment includes: an analog-to-digital converter, used to perform analog-to-digital conversion on the original signal data related to respiration to obtain a digital signal related to respiration; a filter, used to convert the respiration-related original signal data to digital. The relevant digital signals are filtered and noise-reduced to obtain a filtered signal; the amplifier is used to amplify the filtered signal to obtain preprocessed data.

After obtaining the original signal data related to breathing collected by the respiratory signal collector, it is necessary to perform analog-to-digital conversion on these original signal data, convert the changes in external physical quantities obtained at the front end of the sensor into weak electrical change signals, and then analyze the weak electrical changes. The signal is filtered and noise-reduced, and then a first-level low-noise operational amplifier is differentially input for signal amplification. The on-chip ADC of the Arduino minimum system board is used to collect the amplified analog quantity and store the data to obtain the respiratory signal preprocessing data. Among them, the Arduino minimum system board has both WIFI and ADC. The ADC included can read the analog voltage from 0 to 3.3V and convert the input analog voltage signal into a digital signal in the range of 0-1023, which is 3.3V voltage. Divided into 1024 parts, the minimum voltage change of 3.2mV can be recognized, and the built-in analog-to-digital converter can be used to achieve analog-to-digital conversion of the original signal data. The WIFI on the Arduino minimum system board can be used as the signal transmission unit of the front-end component in this embodiment, which can transmit each collected signal to the back-end component. This WIFI can transmit data at high speed without limiting the sampling frequency of the ADC. When performing the above filtering process, considering that the human breathing frequency is within 8Hz, a Butterworth low-pass filter with a cutoff frequency of 10Hz and a stopband attenuation of 60dB can be used. In situations where high-speed communication is required, FIFO devices can also be used as the communication means between the respiratory signal extraction device and the PC. FIFO is an advanced first-in-first-out memory, that is, the data read in first is read out first. The access time of the FIFO memory itself is generally tens of ns. Using FIFO, the data can be sent to the FIFO first. Once the FIFO is full, an interrupt is applied to the device. This can save the time of waiting and querying, and also reduce the number of interrupts. , filtering the data collected in the last 10 seconds in real time.

The algorithm processing unit in this embodiment includes: a setting subunit, used to set the valley threshold, peak threshold and minimum interval time; a recording subunit, used to record the data below the valley threshold in the preprocessed data as Valley values and data higher than the peak threshold are recorded as peak values; the calculation subunit is selected to select the nearest N peaks or valleys, and count the intervals between adjacent peaks or valleys among the N peaks or valleys. time, and calculate the average value of the interval time, and use the average value as the respiratory signal monitoring data.

In this embodiment, respiratory signal monitoring data is obtained by performing threshold processing on the preprocessed data, where the valley threshold is set to 20% of the minimum value of the signal data, and the peak threshold is set to 120% of the maximum value of the signal data. Among them, the minimum value is the lowest voltage displayed by the signal collected during normal breathing, and the maximum value is the highest voltage displayed by the signal collected during normal breathing.

In this embodiment, when recording preprocessed data, the data below the valley threshold is recorded as valley value and the data above the peak threshold is recorded as peak value. When recording the valley value, it is necessary to compare the current valley value with the previous valley value. The interval time between values is compared with the set minimum interval time. If the interval time between the current valley value and the previous valley value is less than the set minimum interval time, the current valley value will not be recorded. Similarly, when recording the peak value, the current valley value needs to be recorded. The interval between the peak value and the previous peak is compared with the set minimum interval. If the interval between the current peak and the previous peak is less than the set minimum interval, the current peak value will not be recorded.

Specifically, according to the actual data collected by the sensor, when the monitored person is breathing normally, the highest voltage collected is Vmax and the lowest voltage is Vmin. Therefore, the valley threshold is set to 20%*Vmin and the peak threshold is set to 120%*. Vmax, record the data below the valley threshold as 20%*Vmin, and record the data above the peak threshold as 120%*Vmax. At the same time, set the shortest interval to 0.1 seconds, that is, the latest trough (or peak) and If the interval between the previous wave peaks (or troughs) is less than 0.1 seconds, it will not be recorded. The influence of false peaks is further eliminated by setting the valley threshold, peak threshold and minimum interval time. There is almost no error in the respiratory peak data obtained at this time. The data is also processed first in first out to select the last ten respiratory peaks or troughs. The average of the last nine respiratory intervals is used as the interval of each breath. This obtains the respiratory rate.

The abnormality judgment unit of this embodiment includes: a determination subunit, used to determine the relationship between the respiratory signal monitoring data and time; a classification subunit, used to input the relationship between the respiratory signal monitoring data and time into the abnormality judgment model, Obtain an abnormality determination result; wherein, the abnormality determination model is obtained by using a training set through machine learning, wherein the training set covers normal and disease respiratory signal monitoring data, and the respiratory signal monitoring data is analyzed through a time series analysis method. Feature extraction is performed and annotated based on existing medical knowledge.

In this embodiment, after the respiratory frequency is obtained, a graph of respiratory frequency and time can be drawn by determining the subunit, and the graph of respiratory frequency and time is input into the abnormality determination model to determine whether there is an abnormality in breathing. Among them, the abnormality determination model collects measurement curve data covering normal and disease respiratory frequencies and performs preprocessing, such as image enhancement and image standardization into tensor format, and uses time series analysis methods to extract features of the preprocessed data. The data is annotated based on existing medical knowledge as a training set, and the model is trained using the training set based on appropriate optimizer, loss function, learning rate and other hyperparameters, that is, the image is passed into the model, the optimizer gradient is cleared, and the calculation Loss, backpropagation of gradients, updating parameters, processing output, and finally calculating training accuracy and performing cross-validation, parameter adjustment, evaluation and optimization to ensure the generalization ability of the model.

The second embodiment of the present invention relates to a respiration detection method for neural regulation. As shown in Figure 2, it includes the following steps: collecting original signal data related to respiration; preprocessing the original signal data related to respiration. Process to obtain preprocessed data; perform threshold processing on the preprocessed data to obtain respiratory signal monitoring data; perform abnormality judgment on the respiratory signal monitoring data to obtain abnormality judgment results.

This embodiment corresponds to the devices of the first embodiment, and the corresponding specific contents can be referred to each other and will not be described again here.

The third embodiment of the present invention relates to a neural regulation system. As shown in Figure 3, it includes a feedback component and a breathing detection device for neural regulation of the first embodiment. The feedback component determines the user's response to the abnormality determination result. Selective neuromodulation. Among them, the feedback component can drive related equipment according to the abnormal respiratory rate determination results. Related equipment includes non-invasive neuromodulation and invasive neuromodulation. Non-invasive neuromodulation can be vagus nerve stimulation, which can help the monitored person enter deep sleep and improve sleep quality; invasive neuromodulation can be transcranial direct current stimulation, which can alleviate the symptoms of schizophrenia, or it can be slow cortical potential neurofeedback , to treat epilepsy seizure symptoms.

The workflow of this neuromodulation system is as follows:

The respiratory signal collector is worn by the person being monitored to obtain the original signal data related to the person's breathing; the respiratory signal extraction unit performs front-end conditioning on the original signal data related to breathing to obtain respiratory signal preprocessing data; the respiratory signal preprocessing data It is sent to the back-end component through the signal transmission unit; the back-end component performs threshold processing on the preprocessed data to obtain the respiratory signal monitoring data (i.e., respiratory frequency), and displays it on the interface through the display module; the respiratory signal monitoring data is input into the abnormality judgment unit , the abnormality judgment unit predicts whether there is an abnormality in the respiratory signal monitoring data through a preset model, and obtains the abnormality judgment result; the feedback component selectively performs neuromodulation treatment on the monitored person based on the respiratory signal abnormality judgment result and the specific situation of the monitored person.

In this embodiment, for users with different needs, machine learning and data analysis technology can be used to obtain target state respiratory frequency characteristic information, so as to provide timely treatment when an abnormal state of the user is detected.

In one embodiment, patients with epilepsy can wear a flexible nasal patch integrated with a respiratory signal collector. When epilepsy attacks, significantly changing physical quantities of temperature and humidity are converted into changing characteristic electrical quantities, which are processed by the signal extraction unit It is transmitted to the mobile terminal through the signal transmission unit for interface display and signal analysis in the respiratory signal judgment device. At this time, the user's breathing frequency has significant changing characteristics, such as repeated apnea during sleep, that is, stopping breathing for a period of time; respiratory frequency There may be an unstable rhythm pattern, with alternating fast and slow breathing rates. The respiratory signal judgment device generates corresponding control flow based on this characteristic to drive the feedback system to initiate neurological treatment measures, such as setting appropriate stimulation parameters and providing high-frequency stimulation intracranial electrodes to prevent epileptic seizure activity. If no epileptic seizure activity is detected within the specified time range, Normal signals will trigger relevant early warnings and send help signals to the outside world.

In another embodiment, during deep sleep, the breathing rate is relatively slow and steady. The depth of breathing is relatively large, the movement of the chest and abdomen is relatively large, the inhalation and exhalation times of breathing are relatively long, and the breathing changes are relatively gentle, usually characterized by low frequency and high amplitude. In contrast, breathing during light sleep is characterized by brisk and regular breathing. The respiratory rate is relatively high, but usually exhibits smaller fluctuations and is more stable in frequency. During light sleep, the breathing time is shorter, the depth of breathing and the range of chest movement are relatively small. The user applies the electronic skin integrated with the respiratory signal collector at night, and the tensile stress physical quantity is converted into a changing characteristic electrical quantity. After being processed by the signal extraction unit, it is transmitted to the mobile terminal through the signal transmission unit for interface display and is displayed in the respiratory signal judgment device. Signal analysis, if the respiratory signal judgment device does not analyze the target signal type within the specified time range, a control signal will be generated to act on the feedback system to initiate neurological treatment measures to help the user achieve the expected sleep quality.

It is not difficult to find that the present invention can monitor and collect individual respiratory activity data in real time and accurately, continuously monitor respiratory status and patterns, and provide instant neurological treatment. It realizes closed-loop health management of real-time monitoring and feedback treatment, and can Protect users’ life safety promptly and effectively. Among them, the respiratory signal collector for monitoring can use non-invasive sensors, which eliminates the need for surgery or invasive operations and is convenient for users to wear and use.

Claims (10)

1. A respiration detection device acting on neuromodulation, comprising a front end part and a rear end part;

the front end piece includes:

a respiration signal collector for collecting raw signal data related to respiration;

the signal extraction unit is used for preprocessing the original signal data related to respiration to obtain preprocessed data;

a signal transmission unit for transmitting the preprocessing data to the back-end component;

the back end piece includes:

the algorithm processing unit is used for carrying out threshold processing on the preprocessing data to obtain respiratory signal monitoring data;

and the abnormality judgment unit is used for carrying out abnormality judgment on the respiratory signal monitoring data to obtain an abnormality judgment result.

2. The respiration detection device acting on neuromodulation according to claim 1, wherein the signal extraction unit comprises:

the analog-to-digital converter is used for carrying out analog-to-digital conversion on the original signal data related to respiration to obtain a digital signal related to respiration;

the filter is used for carrying out filtering noise reduction treatment on the digital signals related to respiration to obtain filtering signals;

and the amplifier is used for amplifying the filtered signal to obtain preprocessed data.

3. The respiration detection device acting on neuromodulation according to claim 2, wherein the filter employs a butterworth low-pass filter having a cut-off frequency of 10Hz for extracting a respiration-related digital signal of less than or equal to 8Hz as the filtered signal.

4. The respiration detection device for neuromodulation according to claim 1, wherein the algorithm processing unit comprises:

a setting subunit, configured to set a valley threshold and a peak threshold;

a recording subunit configured to record, as a valley value, data lower than a valley threshold value and record, as a peak value, data higher than a peak threshold value in the preprocessed data;

and the selecting and calculating subunit is used for selecting the nearest N peaks or valleys, counting the interval time of each adjacent peak or valley in the N peaks or valleys, calculating the average value of the interval time, and taking the average value as respiratory signal monitoring data.

5. The respiration detection device acting on neuromodulation as in claim 4, wherein the setting unit is further for setting a minimum interval time; and when the record subunit records the valley value and the peak value, comparing the interval time between the current valley value or the peak value and the last valley value or the peak value with the lowest interval time, and if the interval time between the current valley value or the peak value and the last valley value or the peak value is smaller than the lowest interval time, not recording the current valley value or the peak value.

6. The respiration detection device acting on neuromodulation according to claim 1, wherein the abnormality determination unit includes:

a determining subunit, configured to determine a relationship between the respiratory signal monitoring data and time;

the classification subunit is used for inputting the relation between the respiratory signal monitoring data and time into an abnormality judgment model to obtain an abnormality judgment result; the abnormal judgment model is obtained by training a training set in a machine learning mode, wherein the training set covers normal and disease respiratory signal monitoring data, and the respiratory signal monitoring data is subjected to feature extraction by a time sequence analysis method and marked according to the existing medical knowledge.

7. The neuromodulation breath detection device of claim 1, wherein the breath signal collector is one or more of a humidity sensor, a temperature sensor, or a stress sensor.

8. A respiration detection method acting on neuromodulation, comprising the steps of:

acquiring raw signal data related to respiration;

preprocessing the original signal data related to respiration to obtain preprocessed data;

performing threshold processing on the preprocessed data to obtain respiratory signal monitoring data;

and carrying out abnormality judgment on the respiratory signal monitoring data to obtain an abnormality judgment result.

9. The method for detecting respiration acting on neuromodulation according to claim 8, wherein the thresholding of the preprocessed data to obtain respiration signal monitoring data comprises:

setting a valley threshold and a peak threshold;

recording data below a valley threshold value and data above a peak threshold value in the preprocessed data as peaks;

selecting the nearest N peaks or valleys, counting the interval time of each adjacent peak or valley in the N peaks or valleys, calculating the average value of the interval time, and taking the average value as respiratory signal monitoring data.

10. A neuromodulation system comprising a feedback component, further comprising a respiratory detection apparatus for neuromodulation as in any of claims 1-7, the feedback component selectively neuromodulating a user based on the abnormality determination.