A sleep intervention improvement system for alzheimer's patients

Abstract

The present application relates to sleep intervention, in particular to a kind of Alzheimer's disease patient sleep intervention improvement system, including control unit, control unit passes through electroencephalogram data acquisition module, electrocardiogram data acquisition module respectively the electroencephalogram data of patient, electrocardiogram data, and utilize current sleep stage identification module according to electroencephalogram data, identify the current sleep stage of patient, control unit passes through sleep staging model construction module and is based on the sleep staging model of patient's historical sleep data Construction, and utilize standard sleep stage determination module according to sleep staging model determines the standard sleep stage corresponding to the current sleep time of patient, control unit passes through sleep stage comparison module and compares current sleep stage and standard sleep stage, and generates sleep stage comparison result;The technical scheme provided by the present application can effectively overcome the defects that the sleep condition of Alzheimer's disease patient cannot be effectively improved in prior art.

Description

Technical Field

[0001] This invention relates to sleep intervention, specifically to a sleep intervention and improvement system for Alzheimer's disease patients. Background Technology

[0002] Alzheimer's disease typically begins in old age or pre-old age, with a slow onset and gradual progression, primarily manifesting as dementia. A family history of the disease is common. Alzheimer's disease is a persistent neurological disorder and the most prevalent cause of dementia. Symptoms include progressively worsening cognitive impairment (memory impairment, learning disabilities, attention deficits, spatial cognitive function, and problem-solving difficulties), leading to a gradual inability to adapt to society.

[0003] Improving sleep quality can effectively slow down the progression of Alzheimer's disease. Although current technologies can monitor people's sleep quality, they cannot effectively improve sleep for people with poor sleep quality, thus failing to effectively improve the sleep of Alzheimer's patients and creating a negative factor in curbing the progression of the disease. Summary of the Invention

[0004] (a) Technical problems to be solved

[0005] In view of the above-mentioned shortcomings of the existing technology, the present invention provides a sleep intervention and improvement system for Alzheimer's disease patients, which can effectively overcome the deficiency of the existing technology in that it cannot effectively improve the sleep of Alzheimer's disease patients.

[0006] (II) Technical Solution

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] A sleep intervention and improvement system for Alzheimer's disease patients includes a control unit. The control unit collects the patient's brainwave data and electrocardiogram (ECG) data through an EEG data acquisition module and an ECG data acquisition module, respectively. It identifies the patient's current sleep stage based on the EEG data using a current sleep stage identification module. The control unit constructs a sleep stage model based on the patient's historical sleep data through a sleep stage model construction module. It then determines the standard sleep stage corresponding to the patient's current sleep time using a standard sleep stage determination module. Finally, the control unit compares the current sleep stage with the standard sleep stage using a sleep stage comparison module and generates a sleep stage comparison result.

[0009] The control unit extracts features from the patient's EEG and ECG data using a feature extraction module, obtaining EEG and ECG features respectively. A feature fusion module then splices and fuses these features to determine the probability of the patient being in different sleep states. The control unit then uses a sleep state quantification module to obtain a current sleep state quantification index based on the probability of the patient being in different sleep states. Finally, a current sleep state recognition module identifies the patient's current sleep state based on this quantification index. Based on the sleep stage comparison results and the patient's current sleep state, the control unit uses an intervention sound wave playback module to play matching intervention sound waves to the patient.

[0010] Preferably, the current sleep stage recognition module identifies the patient's current sleep stage based on EEG data, including:

[0011] The patient's EEG data was filtered, and feature indicators were extracted from the filtered EEG data.

[0012] The extracted feature indicators are input into the machine learning algorithm to obtain the stage accuracy of the patient's sleep stage, and the patient's current sleep stage is identified based on the stage accuracy of the patient's sleep stage.

[0013] The staging accuracy rate is used to indicate the likelihood that a patient is in each sleep stage.

[0014] Preferably, the step of filtering the patient's electroencephalogram (EEG) data and extracting feature indicators from the filtered EEG data includes:

[0015] The patient's EEG data was filtered, and the filtered EEG data was iteratively filtered using a Gaussian convolution kernel to obtain multiple sub-components of the EEG data. Feature indicators were extracted from each sub-component.

[0016] Among them, the characteristic indicators include overall numerical distribution characteristics, instantaneous spectral characteristics, and interval spectral characteristics.

[0017] Preferably, the iterative filtering uses one-dimensional convolution with a window length of 1 second.

[0018] Preferably, the feature extraction module extracts features from the patient's electroencephalogram (EEG) data and electrocardiogram (ECG) data to obtain EEG features and ECG features, respectively, including:

[0019] The brainwave data feature extraction network is used to extract features from brainwave data within N seconds in sequence to obtain brainwave features;

[0020] Meanwhile, the ECG wave data feature extraction network is used to extract features from the ECG wave data within 10*N seconds in sequence to obtain ECG wave features.

[0021] Preferably, the network architecture of the EEG data feature extraction network includes a first input layer, a two-dimensional convolutional layer, a channel convolutional layer, a first average pooling layer, a 2D separable convolutional layer, a depthwise separable convolutional layer, a second average pooling layer, a first flattening layer, and a first fully connected layer.

[0022] Preferably, the network architecture of the ECG data feature extraction network includes a second input layer, a first convolutional layer, a third average pooling layer, a second convolutional layer, a fourth average pooling layer, a third convolutional layer, a fifth average pooling layer, a second flattening layer, and a second fully connected layer.

[0023] Preferably, the sleep state quantification module obtains a current sleep state quantification index based on the probability that the patient is in different sleep states, including:

[0024] Statistical analysis is performed on the quantitative indices of each sleep state obtained from the current time and the M consecutive times before it to determine the most frequently occurring sleep state.

[0025] Calculate the average of the sleep state quantification indices corresponding to the most frequent sleep states to obtain the current sleep state quantification index.

[0026] Preferably, the control unit plays matching intervention sound waves to the patient using the intervention sound wave playback module based on the sleep stage comparison results and the patient's current sleep state, including:

[0027] The control unit performs a comprehensive analysis of the sleep stage comparison results and the patient's current sleep state, and matches appropriate intervention sound waves based on the patient's preference for sound and preferred frequency, and plays the intervention sound waves to the patient using the intervention sound wave playback module.

[0028] (III) Beneficial Effects

[0029] Compared with existing technologies, the sleep intervention and improvement system for Alzheimer's disease patients provided by this invention has the following beneficial effects:

[0030] 1) The current sleep stage identification module identifies the patient's current sleep stage based on EEG data. The sleep staging model construction module constructs a sleep staging model based on the patient's historical sleep data. The standard sleep stage determination module determines the standard sleep stage corresponding to the patient's current sleep time based on the sleep staging model. The sleep stage comparison module compares the current sleep stage and the standard sleep stage and generates a sleep stage comparison result. By identifying the patient's current sleep stage through EEG data and determining the standard sleep stage corresponding to the patient's current sleep time through the construction of a sleep staging model, it is possible to effectively analyze the actual sleep stage that the patient is in at the current sleep time.

[0031] 2) The feature extraction module extracts features from the patient's EEG and ECG data to obtain EEG features and ECG features respectively. The feature fusion module splices and fuses the EEG and ECG features to obtain the probability of the patient being in different sleep states. The sleep state quantification module obtains the current sleep state quantification index based on the probability of the patient being in different sleep states. The current sleep state recognition module identifies the patient's current sleep state based on the current sleep state quantification index. By performing feature extraction and feature fusion on the patient's EEG and ECG data, the sleep quality can be effectively quantified, and thus the patient's current sleep state can be accurately identified.

[0032] 3) Based on the sleep stage comparison results and the patient's current sleep state, the control unit plays matching intervention sound waves to the patient using the intervention sound wave playback module. By comprehensively analyzing the sleep stage comparison results and the patient's current sleep state, and combining the patient's preference for sound and preferred frequency, the control unit matches appropriate intervention sound waves and plays them to the patient using the intervention sound wave playback module. Through active intervention, the sleep condition of Alzheimer's patients is effectively improved. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0034] Figure 1 This is a schematic diagram of the system of the present invention;

[0035] Figure 2 This is a schematic diagram of the process in this invention for generating sleep stage comparison results by comparing the current sleep stage with the standard sleep stage;

[0036] Figure 3 This is a schematic diagram of the process for identifying a patient's current sleep state based on a quantitative index of the current sleep state in this invention. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0038] A sleep intervention system to improve sleep in Alzheimer's patients, such as Figure 1 and Figure 2 As shown, the device includes a control unit. The control unit collects the patient's brainwave data and electrocardiogram (ECG) data through an EEG data acquisition module and an ECG data acquisition module, respectively. It then uses a current sleep stage identification module to identify the patient's current sleep stage based on the EEG data. The control unit constructs a sleep stage model based on the patient's historical sleep data through a sleep stage model construction module. It then uses a standard sleep stage determination module to determine the patient's standard sleep stage at the current sleep time based on the sleep stage model. Finally, the control unit compares the current sleep stage with the standard sleep stage through a sleep stage comparison module and generates a sleep stage comparison result.

[0039] The current sleep stage recognition module identifies the patient's current sleep stage based on EEG data, including:

[0040] The patient's EEG data was filtered, and feature indicators were extracted from the filtered EEG data.

[0041] The extracted feature indicators are input into the machine learning algorithm to obtain the stage accuracy of the patient's sleep stage, and the patient's current sleep stage is identified based on the stage accuracy of the patient's sleep stage.

[0042] The staging accuracy rate is used to indicate the likelihood that a patient is in each sleep stage.

[0043] Specifically, the patient's electroencephalogram (EEG) data is filtered, and feature indicators are extracted from the filtered EEG data, including:

[0044] The patient's EEG data was filtered, and the filtered EEG data was iteratively filtered using a Gaussian convolution kernel (the iterative filtering used one-dimensional convolution with a window length of 1s) to obtain multiple sub-components of the EEG data. Feature indicators were extracted from each sub-component.

[0045] Among them, the characteristic indicators include overall numerical distribution characteristics, instantaneous spectral characteristics, and interval spectral characteristics.

[0046] The above technical solution involves a current sleep stage identification module that identifies the patient's current sleep stage based on EEG data, a sleep staging model construction module that constructs a sleep staging model based on the patient's historical sleep data, a standard sleep stage determination module that determines the patient's standard sleep stage at the current sleep time based on the sleep staging model, and a sleep stage comparison module that compares the current sleep stage with the standard sleep stage and generates a sleep stage comparison result. By identifying the patient's current sleep stage through EEG data and determining the patient's standard sleep stage at the current sleep time through the construction of a sleep staging model, the actual sleep stage of the patient at the current sleep time can be effectively analyzed.

[0047] like Figure 1 and Figure 3 As shown, the control unit extracts features from the patient's EEG and ECG data through a feature extraction module, obtaining EEG features and ECG features respectively. Then, it uses a feature fusion module to splice and fuse the EEG and ECG features to obtain the probability that the patient is in different sleep states. The control unit uses a sleep state quantification module to obtain a current sleep state quantification index based on the probability that the patient is in different sleep states, and uses a current sleep state recognition module to identify the patient's current sleep state based on the current sleep state quantification index.

[0048] ① The feature extraction module extracts features from the patient's EEG and ECG data, obtaining EEG features and ECG features respectively, including:

[0049] The brainwave data feature extraction network is used to extract features from brainwave data within N seconds in sequence to obtain brainwave features;

[0050] Meanwhile, the ECG wave data feature extraction network is used to extract features from the ECG wave data within 10*N seconds in sequence to obtain ECG wave features.

[0051] Specifically, the network architecture of the EEG data feature extraction network includes a first input layer, a two-dimensional convolutional layer, a channel convolutional layer, a first average pooling layer, a 2D separable convolutional layer, a depthwise separable convolutional layer, a second average pooling layer, a first flattening layer, and a first fully connected layer.

[0052] Specifically, the network architecture of the ECG data feature extraction network includes a second input layer, a first convolutional layer, a third average pooling layer, a second convolutional layer, a fourth average pooling layer, a third convolutional layer, a fifth average pooling layer, a second flattening layer, and a second fully connected layer.

[0053] ② The sleep state quantification module obtains a current sleep state quantification index based on the probability of the patient being in different sleep states, including:

[0054] Statistical analysis is performed on the quantitative indices of each sleep state obtained from the current time and the M consecutive times before it to determine the most frequently occurring sleep state.

[0055] Calculate the average of the sleep state quantification indices corresponding to the most frequent sleep states to obtain the current sleep state quantification index.

[0056] The above technical solution involves a feature extraction module that extracts features from the patient's EEG and ECG data to obtain EEG features and ECG features respectively; a feature fusion module that splices and fuses the EEG and ECG features to obtain the probability of the patient being in different sleep states; a sleep state quantification module that obtains a current sleep state quantification index based on the probability of the patient being in different sleep states; and a current sleep state recognition module that identifies the patient's current sleep state based on the current sleep state quantification index. By performing feature extraction and feature fusion on the patient's EEG and ECG data, effective quantification of sleep quality can be achieved, thereby enabling accurate identification of the patient's current sleep state.

[0057] like Figure 1 As shown, the control unit plays matching intervention sound waves to the patient using the intervention sound wave playback module based on the sleep stage comparison results and the patient's current sleep state. Specifically, this includes:

[0058] The control unit performs a comprehensive analysis of the sleep stage comparison results and the patient's current sleep state, and matches appropriate intervention sound waves based on the patient's preference for sound and preferred frequency, and plays the intervention sound waves to the patient using the intervention sound wave playback module.

[0059] The above technical solution involves the control unit playing matching intervention sound waves to the patient using an intervention sound wave playback module based on the sleep stage comparison results and the patient's current sleep state. By comprehensively analyzing the sleep stage comparison results and the patient's current sleep state, and combining the patient's preference for sound and preferred frequency to match appropriate intervention sound waves, the solution achieves effective improvement in the sleep of Alzheimer's patients through active intervention.

[0060] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A sleep intervention and improvement system for Alzheimer's disease patients, characterized in that: The device includes a control unit. The control unit collects the patient's brainwave data and electrocardiogram (ECG) data through an EEG data acquisition module and an ECG data acquisition module, respectively. It also uses a current sleep stage identification module to identify the patient's current sleep stage based on the EEG data. The control unit constructs a sleep stage model based on the patient's historical sleep data through a sleep stage model construction module. It then uses a standard sleep stage determination module to determine the patient's standard sleep stage at the current sleep time based on the sleep stage model. Finally, the control unit compares the current sleep stage with the standard sleep stage through a sleep stage comparison module and generates a sleep stage comparison result. The control unit extracts features from the patient's EEG and ECG data using a feature extraction module, obtaining EEG features and ECG features respectively. A feature fusion module then splices and fuses these features to obtain the probability of the patient being in different sleep states. The control unit uses a sleep state quantification module to obtain a current sleep state quantification index based on the probability of the patient being in different sleep states. A current sleep state recognition module identifies the patient's current sleep state based on this quantification index. Finally, based on the sleep stage comparison results and the patient's current sleep state, the control unit uses an intervention sound wave playback module to play matching intervention sound waves to the patient. The current sleep stage recognition module identifies the patient's current sleep stage based on EEG data, including: The patient's EEG data was filtered, and feature indicators were extracted from the filtered EEG data. The extracted feature indicators are input into the machine learning algorithm to obtain the stage accuracy of the patient's sleep stage, and the patient's current sleep stage is identified based on the stage accuracy of the patient's sleep stage. Among them, the staging accuracy rate is used to indicate the likelihood that a patient is in each sleep stage; The filtering of the patient's electroencephalogram (EEG) data and the extraction of feature indicators from the filtered EEG data include: The patient's EEG data was filtered, and the filtered EEG data was iteratively filtered using a Gaussian convolution kernel to obtain multiple sub-components of the EEG data. Feature indicators were extracted from each sub-component. Among them, the characteristic indicators include overall numerical distribution characteristics, instantaneous spectrum characteristics, and interval spectrum characteristics; The iterative filtering uses one-dimensional convolution with a window length of 1 second.

2. The sleep intervention and improvement system for Alzheimer's disease patients according to claim 1, characterized in that: The feature extraction module extracts features from the patient's electroencephalogram (EEG) and electrocardiogram (ECG) data, obtaining EEG features and ECG features respectively, including: The brainwave data feature extraction network is used to extract features from brainwave data within N seconds in sequence to obtain brainwave features; Meanwhile, the ECG wave data feature extraction network is used to extract features from the ECG wave data within 10*N seconds in sequence to obtain ECG wave features.

3. The sleep intervention and improvement system for Alzheimer's disease patients according to claim 2, characterized in that: The network architecture of the EEG data feature extraction network includes a first input layer, a two-dimensional convolutional layer, a channel convolutional layer, a first average pooling layer, a 2D separable convolutional layer, a depthwise separable convolutional layer, a second average pooling layer, a first flattening layer, and a first fully connected layer.

4. The sleep intervention and improvement system for Alzheimer's disease patients according to claim 3, characterized in that: The network architecture of the ECG data feature extraction network includes a second input layer, a first convolutional layer, a third average pooling layer, a second convolutional layer, a fourth average pooling layer, a third convolutional layer, a fifth average pooling layer, a second flattening layer, and a second fully connected layer.

5. The sleep intervention and improvement system for Alzheimer's disease patients according to claim 3, characterized in that: The sleep state quantification module obtains a current sleep state quantification index based on the probability of the patient being in different sleep states, including: Statistical analysis is performed on the quantitative indices of each sleep state obtained from the current time and the M consecutive times before it to determine the most frequently occurring sleep state. Calculate the average of the sleep state quantification indices corresponding to the most frequent sleep states to obtain the current sleep state quantification index.

6. The sleep intervention and improvement system for Alzheimer's disease patients according to claim 1, characterized in that: The control unit, based on the sleep stage comparison results and the patient's current sleep state, uses the intervention sound wave playback module to play matching intervention sound waves to the patient, including: The control unit performs a comprehensive analysis of the sleep stage comparison results and the patient's current sleep state, and matches appropriate intervention sound waves based on the patient's preference for sound and preferred frequency, and plays the intervention sound waves to the patient using the intervention sound wave playback module.

Images