Sleep state identification apparatus and implantable closed-loop stimulation system for use in sleep state

Abstract

Provided in the present invention are a sleep state identification apparatus and an implantable closed-loop stimulation system for use in a sleep state, which are applied to the technical field of medical instruments. The apparatus comprises a processor, wherein the processor is used for preliminarily determining, on the basis of user state information and / or a clock reading, whether a user is in a sleep state, when the user is in a suspected sleep state, acquiring a local field potential (LFP) collected by an electrode and an acceleration collected by a triaxial accelerometer, on the basis of the LFP and the acceleration, re-determining whether the user is in a sleep state, and when the user is in the sleep state, determining a sleep stage on the basis of the LFP. The system comprises the sleep state identification apparatus and a controller, wherein the controller is used for performing proportional regulation and stimulation signal output by using a stimulation amplitude corresponding to the sleep stage. In the present invention, a sleep state of a user is accurately identified, stimulation parameters are intelligently regulated and stimulation signals are output, so as to realize automatic and deep brain stimulation for the user, thereby achieving personalized treatment and improving therapeutic effects.

Description

Sleep state recognition device and implantable closed-loop stimulation system suitable for sleep state Technical Field

[0001] This invention relates to the field of medical device technology, specifically to a sleep state recognition device and an implantable closed-loop stimulation system suitable for sleep states. Background Technology

[0002] Implantable neurostimulation systems, including deep brain stimulation (DBS), have shown significant efficacy in treating various refractory neurological diseases. A pulse generator and electrodes are implanted in the body. The pulse generator delivers electrical pulses to specific areas of the brain via the electrodes to control disease symptoms. External devices can communicate with the pulse generator, adjusting its stimulation parameters to achieve different stimulation effects. For example, the stimulation location can be changed by adjusting the polarity of the contact points, and the range of influence can be altered by modifying the amplitude, pulse width, and frequency. Physicians program these parameters based on experience and the patient's response, establishing a fixed set for continuous stimulation.

[0003] However, patients' physiological states fluctuate. Taking Parkinson's patients as an example, their physiological state is related to various factors such as medication and sleep, and their need for and intensity of stimulation will vary. If continuous stimulation consumes a lot of power, especially for non-rechargeable products, continuous high-frequency stimulation can affect normal neural circuits to some extent, potentially impacting the patient's cognition, behavior, and emotions. Summary of the Invention

[0004] In view of this, the present invention provides a sleep state recognition device, including a processor;

[0005] The processor is used to initially determine whether the user is in a sleep state based on user status information and / or clock readings, wherein the user status information includes at least one of acceleration, circadian rhythm, electrocardiogram, and electroencephalogram; when the user is in a suspected sleep state, it acquires the local field potential (LFP) collected by the electrodes of the implantable closed-loop stimulation system and the acceleration collected by the triaxial accelerometer of the implantable closed-loop stimulation system; it further determines whether the user is in a sleep state based on the local field potential (LFP) and the acceleration; when the user is in a sleep state, it determines the sleep stage based on the local field potential (LFP).

[0006] Optionally, a preliminary determination of whether the user is asleep may be made, including:

[0007] Determine whether the clock reading is within a preset time period;

[0008] When the clock reading is within a preset time period, it is determined that the user is in a suspected sleep state;

[0009] If the clock reading is not within a preset time period, it is determined that the user is not in a sleep state.

[0010] Optionally, a preliminary determination of whether the user is asleep may be made, including:

[0011] At preset intervals, the system initially determines whether the user is asleep based on acceleration.

[0012] Temporal features of triaxial acceleration are extracted, including the mean and standard deviation of acceleration.

[0013] The mean acceleration and standard deviation of acceleration are respectively input into two pre-trained sleep state prediction models to predict the probability of sleep state, and the two prediction results are multiplied to obtain the sleep probability value.

[0014] Based on the sleep probability value, a preliminary judgment can be made as to whether the user is in a sleep state;

[0015] When the sleep probability value is greater than or equal to the preset probability value, the user is determined to be in a suspected sleep state;

[0016] When the sleep probability value is less than the preset probability value, it is determined that the user is not in a sleep state.

[0017] Optionally, determine again whether the user is asleep, including:

[0018] Determine whether the local field potential LFP has a frequency division artifact;

[0019] When the local field potential LFP does not have frequency division artifacts, time-domain features are extracted from the triaxial acceleration to obtain the mean acceleration and the standard deviation of acceleration.

[0020] The mean acceleration and standard deviation of acceleration are respectively input into two pre-trained sleep state prediction models to predict the probability of sleep state, and the two prediction results are multiplied to obtain the sleep probability value.

[0021] Based on the sleep probability value, determine again whether the user is in a sleep state;

[0022] When the sleep probability value is greater than or equal to the preset probability value, it is determined that the user is in a sleep state;

[0023] When the sleep probability value is less than the preset probability value, it is determined that the user is not in a sleep state.

[0024] Optionally, sleep stages are determined based on the local field potential (LFP), including:

[0025] Power spectrum estimation and denoising are performed on the local field potential LFP;

[0026] The denoised power spectrum is merged into multiple frequency band energies and then normalized.

[0027] The normalized energy across multiple frequency bands is reduced in dimensionality to obtain multiple dimensionality-reduced features, including the first dimensionality-reduced feature. Second dimensionality reduction feature Third dimensionality reduction feature Fourth dimensionality reduction feature ;

[0028] The multiple dimensionality-reduced features are input into a pre-built ARIMA model to obtain a set of predicted values. ;

[0029] According to the predicted value group Sleep stages are determined, including wakefulness, REM sleep, and non-REM sleep.

[0030] Optionally, the predicted numerical group Including the first dimensionality reduction feature prediction value Second dimensionality reduction feature prediction value Third dimensionality reduction feature prediction value and the fourth dimensionality reduction feature prediction value First dimensionality reduction feature prediction value Second dimensionality reduction feature prediction value Third dimensionality reduction feature prediction value and the fourth dimensionality reduction feature prediction value Each contains multiple predicted values;

[0031] According to the predicted value group Determining sleep stages includes:

[0032] Calculate the mean of the predicted values ​​for each dimensionality reduction feature;

[0033] The preset first optimal classification threshold and the first dimensionality reduction feature prediction value are used. The mean values ​​were compared to determine whether the REM or wakefulness phase, or the non-REM phase, were used;

[0034] When the condition is determined to be either REM or wakefulness, a second optimal classification threshold and the second dimensionality reduction feature prediction value will be preset. The mean values ​​were compared to distinguish between REM and wakefulness periods.

[0035] The present invention also provides an implantable closed-loop stimulation system suitable for sleep states, including the aforementioned sleep state recognition device and a controller;

[0036] The controller is used to proportionally adjust the output stimulation signal using stimulation amplitudes corresponding to different sleep stages, wherein the stimulation amplitudes are different for different sleep stages.

[0037] Optionally, when it is initially determined that the user is not asleep, the clock reading is used to determine whether it is nighttime. If it is nighttime, a stimulation signal is output using a preset stimulation amplitude S0.

[0038] When the user is determined to be awake again, a stimulation signal is output using a preset stimulation amplitude S0.

[0039] Optionally, when it is determined that the local field potential LFP has a frequency division artifact, a preset stimulation amplitude S0 is used to output the stimulation signal.

[0040] Optionally, for the waking phase in sleep stages, a preset stimulation amplitude S0 is used to output the stimulation signal;

[0041] For the non-rapid eye movement (NREM) sleep stage, the corresponding stimulus amplitude S1 of the output stimulus signal is:

[0042] ,

[0043] in, N1 represents the upper limit of non-rapid eye movement (NREM) stimulation, and N3 represents the lower limit of NREM stimulation.

[0044] For the REM sleep stage, the corresponding stimulus amplitude S2 of the output stimulus signal is:

[0045] ,

[0046] in, denoted as the sleep state coefficient during REM sleep, where P is the upper limit of REM stimulation and T is the lower limit of REM stimulation.

[0047] The sleep state recognition device provided in this application combines user state information (such as acceleration, circadian rhythm, electrocardiogram, and electroencephalogram) with clock readings for preliminary judgment, enabling rapid determination of whether the user is likely in a sleep state. A second judgment is made by acquiring local field potentials (LFP) and accelerations from electrodes of the implanted closed-loop stimulation system, further improving the accuracy and reliability of the judgment. This dual-judgment mechanism helps reduce false positives and false negatives, ensuring the accuracy of subsequent sleep staging.

[0048] The implantable closed-loop stimulation system for sleep states provided in this application, after the processor determines that the user is asleep and performs sleep staging, the controller determines the corresponding stimulation mode for the user based on the sleep staging results and outputs corresponding stimulation signals according to the stimulation mode, thereby improving the accuracy of stimulation parameters. By accurately identifying the user's sleep state and stage using a sleep state recognition device, the controller can intelligently adjust stimulation parameters and output stimulation signals to achieve automatic stimulation of the deep brain regions of the user, optimizing the treatment effect. Attached Figure Description

[0049] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0050] Figure 1 is a schematic diagram of the process of the processor recognizing the sleep state in the sleep state recognition device in an embodiment of the present invention;

[0051] Figure 2 is a flowchart illustrating a method for the processor in the sleep state recognition device of the present invention to initially determine whether the user is in a sleep state.

[0052] Figure 3 is a schematic flowchart of another method for the processor in the sleep state recognition device of the present invention to initially determine whether the user is in a sleep state.

[0053] Figure 4 is a schematic diagram of the process by which the processor in the sleep state recognition device in an embodiment of the present invention determines whether the user is in a sleep state again.

[0054] Figure 5 is a schematic diagram of the process by which the processor determines sleep stages in the sleep state recognition device according to an embodiment of the present invention. Detailed Implementation

[0055] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0056] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0057] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can also refer to the internal connection of two components; and they can refer to a wireless connection or a wired connection. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0058] Furthermore, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0059] This invention provides an implantable closed-loop stimulation system suitable for sleep states. The system includes a sleep state recognition device, electrodes, and a controller. The sleep state recognition device includes an implantable pulse generator, which includes a processor and a triaxial accelerometer. The processor can recognize the human sleep state, and the triaxial accelerometer is used to collect acceleration information. The electrodes are used to collect local field potentials (LFP). The controller is used to proportionally adjust the output stimulation signal using stimulation amplitudes corresponding to different sleep stages. Different sleep stages correspond to different stimulation amplitudes. The controller is communicatively connected to the processor and can configure parameters related to stimulation signal adjustment based on the processor's recognition results.

[0060] In other embodiments, the sleep state recognition device can be integrated into an implantable pulse generator or an external programming device. That is, sleep state recognition can be performed by the implantable pulse generator, or by the external programming device which then sends the recognition result to the implantable pulse generator, or by both implantable and external devices. For example, acceleration information can be collected by the external device and sent to the implantable pulse generator. Therefore, in other embodiments, the implantable closed-loop stimulation system also includes an external programming device.

[0061] This invention provides a sleep state recognition device, including a processor, as shown in FIG1. ​​The processor is used to perform operations including the following steps:

[0062] S1, based on user status information and / or clock readings, a preliminary determination is made as to whether the user is in a sleep state, wherein the user status information includes at least one of acceleration, circadian rhythm, electrocardiogram, and electroencephalogram; when it is preliminarily determined that the user is in a suspected sleep state, the processor executes step S2.

[0063] S2, acquire the local field potential LFP collected by the electrodes of the implantable closed-loop stimulation system and the acceleration collected by the triaxial accelerometer of the implantable closed-loop stimulation system;

[0064] S3, determine again whether the user is in a sleep state based on the local field potential LFP and acceleration; when it is determined again that the user is in a sleep state, the processor executes step S4;

[0065] S4, sleep stages are determined based on local field potential (LFP).

[0066] As shown in Figure 2, the processor's initial determination of whether the user is in a sleep state in step S1 specifically includes:

[0067] S11a, determine whether the clock reading is within a preset time period. If the clock reading is within the preset time period, proceed to step S12a; if the clock reading is not within the preset time period, proceed to step S13a.

[0068] S12a, determines that the user is in a suspected sleep state;

[0069] S13a, determined that the user is not in a sleep state.

[0070] The clock reading of the implanted pulse generator is determined, and the preset time period is the general time period of the user's sleep period, such as 21:00-7:00. Other times are considered to be the awake period during the day. The preset time period can be configured in the implanted pulse generator through the controller.

[0071] If the clock reading of the implantable pulse generator is between 21:00 and 7:00, the user is determined to be in a suspected sleep state, and the processor executes step S2; if the clock reading of the implantable pulse generator is not between 21:00 and 7:00, the user is determined not to be in a sleep state, that is, the user is in the daytime awake period, and the processor can output the result that the user is not in a sleep state and is in the daytime awake period.

[0072] And / or, as shown in Figure 3, at preset intervals, a preliminary judgment is made as to whether the user is in a sleep state based on acceleration, wherein,

[0073] S11b extracts time-domain features of triaxial acceleration, including the mean acceleration and the standard deviation of acceleration;

[0074] S12b: Input the mean acceleration and standard deviation of acceleration into two pre-trained sleep state prediction models to predict the probability of sleep state, and multiply the two prediction results to obtain the sleep probability value;

[0075] S13b: Based on the sleep probability value, initially determine whether the user is in a sleep state, and determine whether the sleep probability value is greater than or equal to the preset probability value; when the sleep probability value is greater than or equal to the preset probability value, the processor executes step S14b; when the sleep probability value is less than the preset probability value, the processor executes step S15b.

[0076] S14b, determines that the user is in a suspected sleep state;

[0077] S15b determines that the user is not in a sleep state.

[0078] The preset probability value can be pre-set in the implanted pulse generator via the controller. For example, the processor can initially determine whether the user is asleep based on acceleration. This sleep determination can be performed every 30 minutes. Specifically, when the user is initially determined to be asleep, a time window of 7 seconds can be set based on the triaxial acceleration data (X-axis, Y-axis, and Z-axis) from the triaxial accelerometer. Multiple triaxial accelerations are collected within each window. For the triaxial accelerations within each 7-second window, the mean and standard deviation of the triaxial accelerations are calculated. The mean and the maximum standard deviation of the triaxial accelerations within the three windows are used as acceleration features and input into two LDA models respectively. The posterior probability of the current sleep state is output. The two posterior probabilities are multiplied to obtain the final sleep probability predicted by acceleration. When the sleep probability is greater than or equal to 0.5, the processor determines that the user is in a suspected sleep state and executes step S2. When the sleep probability is less than 0.5, the processor determines that the user is not asleep. The two LDA models mentioned above were pre-trained using the mean and maximum standard deviation calculated from historical acceleration data of the user during sleep and wakefulness, respectively. After determining the sleep state, a timer can be established, and sleep staging can be performed after a certain period.

[0079] In another embodiment, the processor in step S1, when initially determining whether the user is in a sleep state, may also:

[0080] At preset time intervals, the system determines whether the user is asleep based on user status information (physiological information) and clock readings. User status information (physiological information) includes acceleration, circadian rhythm, electrocardiogram (ECG), and electroencephalogram (EEG). The steps are the same as those described above for determining whether the user is not asleep using only clock readings, and will not be repeated here.

[0081] In this embodiment, the processor continuously analyzes the user's state information at certain intervals to accurately determine the user's state, so that the controller can accurately adjust the stimulation parameters.

[0082] As shown in Figure 4, in step S3, the processor determines again whether the user is in a sleep state based on the local field potential LFP and acceleration, including:

[0083] S31, determine whether there is a frequency division artifact in the local field potential LFP. Preferably, determine whether there is a frequency division artifact in the local field potential LFP below 38Hz. When the processor determines that there is no frequency division artifact in the local field potential LFP, then execute step S32;

[0084] The determination of whether a frequency-division artifact exists is as follows: The energy ratio of the oscillation signal of the current local field potential (LFP) at the 1 / 4 frequency division, obtained by the processor through electrodes, to the fractal signal in the same frequency band is used as the artifact detection feature. The local field potential (LFP) without frequency-division artifacts is collected based on daytime open-loop stimulation (stimulation output using a preset stimulation amplitude S0 during the day), and the 95th percentile of the daytime data feature is used as the artifact detection threshold. The daytime data feature refers to the local field potential (LFP) without frequency-division artifacts during the daytime. During the determination process, the artifact detection feature is compared with the artifact detection threshold. If the threshold is exceeded, a 1 / 4 frequency-division artifact is determined to exist; if the threshold is not exceeded, a 1 / 4 frequency-division artifact is determined to not exist.

[0085] In some other embodiments, the LFP collected during the daytime with a preset stimulus amplitude S0 can be transmitted to the controller, which calculates the artifact judgment threshold and transmits it to the processor.

[0086] In other embodiments of the present invention, the frequency division artifact determination may also be carried out by: obtaining the oscillation signal energy of the current local field potential LFP at the 1 / 4 frequency division, calculating the average energy of all signals of the local field potential LFP in a specific frequency band collected by daytime open-loop stimulation (stimulation output when using a preset stimulation amplitude S0 during the day), expanding the average energy by 10 times, and then taking the median of the average energy of all frequency bands expanded by 10 times. The oscillation signal energy of the current local field potential LFP at the 1 / 4 frequency division is compared with the median. If it exceeds the median, it is determined that a 1 / 4 frequency division artifact exists. If it does not exceed the median, it is determined that a 1 / 4 frequency division artifact does not exist.

[0087] S32, perform time-domain feature extraction on triaxial acceleration to obtain the mean acceleration and standard deviation of acceleration;

[0088] S33, input the mean acceleration and standard deviation of acceleration into two pre-trained sleep state prediction models to predict the probability of sleep state, and multiply the two prediction results to obtain the sleep probability value;

[0089] S34, determine again whether the user is in a sleep state based on the sleep probability value, and determine whether the sleep probability value is greater than or equal to the preset probability value; when the sleep probability value is greater than or equal to the preset probability value, the processor executes step S35; when the sleep probability value is less than the preset probability value, the processor executes step S36.

[0090] S35 determines that the user is in a sleep state;

[0091] S36, determined that the user is not in a sleep state.

[0092] The processor performs frequency division artifact detection on the local field potential (LFP) collected from the electrodes to ensure the accuracy of the LFP, thereby determining the user's sleep stage.

[0093] As shown in Figure 5, in step S4, the processor determines the sleep stage based on the local field potential (LFP), specifically including:

[0094] S41 performs power spectrum estimation and denoising on the local field potential LFP.

[0095] Specifically, the current local field potential (LFP) is subjected to a Fast Fourier Transform (NFFT) using the sampling frequency up to the nearest power of 2 data points. Then, a Pwelch power spectrum estimation is performed with a 1-second window and a 50% overlap (50% overlap between adjacent signal segments) to obtain the 0-fs / 2Hz PSD. Next, the IRASA method is used to estimate the fractal signal PSD of the current LFP. The portion of the 0-fs / 2Hz PSD above 38Hz is replaced with the fractal signal PSD portion to achieve local field potential LFP denoising and improve its accuracy.

[0096] For processing non-uniformly sampled signals with 2^N data points, the efficient algorithm of Fast Fourier Transform (FFT) can be utilized. Furthermore, 2^N data points provide sufficient frequency resolution for signal analysis in the frequency domain. "0-fs / 2Hz" indicates a frequency range from zero to half the sampling frequency (fs / 2), because according to the sampling theorem, a discrete signal cannot contain frequency components exceeding half its sampling frequency.

[0097] S42 merges the denoised power spectrum into multiple frequency band energies and normalizes them. The default frequency band range is 2-3Hz, 4-7Hz, 8-14Hz, 15-25Hz, 26-38Hz, and 39-50Hz. The denoised PSD is merged into the energies of the above six frequency bands and normalized.

[0098] S43, dimensionality reduction is performed on the normalized energy of multiple frequency bands to obtain multiple dimensionality reduction features, including the first dimensionality reduction feature. Second dimensionality reduction feature Third dimensionality reduction feature Fourth dimensionality reduction feature .

[0099] Specifically, the LDA model is used to perform feature dimensionality reduction on the local field potential (LFP) based on normalized energy across multiple frequency bands, applicable to different sleep stages and sleep depth states, and the first dimensionality-reduced feature is output. Second dimensionality reduction feature Third dimensionality reduction feature Fourth dimensionality reduction feature .

[0100] S44. Input multiple dimensionality-reduced features into the pre-built ARIMA model to obtain the predicted numerical group. .

[0101] The ARIMA model is a prediction model with an input step size of 10 and a prediction step size of 10, constructed based on historical dimensionality reduction features. This model assumes that the information in the time series is correlated at the input step size, but uncorrelated between input step sizes. The model segments the input dimensionality reduction features according to the input step size (10), obtaining 10 subsequences. After obtaining the predicted value of a certain subsequence, this predicted value is used as the input for the next subsequence, and so on, until the entire sequence is predicted. In the prediction of each subsequence, the ARIMA model makes predictions based on the autoregressive term, differencing term, and moving average term of the sequence. The autoregressive term reflects the past information of the sequence, the differencing term reflects the non-stationarity of the sequence, and the moving average term reflects the noise component of the sequence.

[0102] S45, based on the predicted value group Sleep stages are determined, including wakefulness, REM sleep, and non-REM sleep.

[0103] Among them, the predicted numerical group Including the first dimensionality reduction feature prediction value Second dimensionality reduction feature prediction value Third dimensionality reduction feature prediction value and the fourth dimensionality reduction feature prediction value The ARIMA model has a prediction step size of 10, and each dimensionality reduction feature prediction contains 10 prediction values. For example, the first dimensionality reduction feature... Inputting the data into the ARIMA model will generate 10 predicted values ​​for the first dimensionality reduction features. ;

[0104] Therefore, the processor bases its decisions on the predicted numerical groups. Determining sleep stages specifically includes:

[0105] Calculate the mean of the predicted values ​​of each dimensionality reduction feature; calculate the mean of the 10 predicted values ​​of each dimensionality reduction feature.

[0106] The preset first optimal classification threshold and the first dimensionality reduction feature prediction value are used. The mean values ​​were compared to determine the REM or wakefulness phase and the non-REM phase.

[0107] First, a classification model is trained using the training set data. Then, this model is used to predict the values ​​of samples in the training set. Next, based on the prediction results and the actual sleep stage category labels of the samples, the model's performance metrics, such as AUC, are calculated. By trying different classification thresholds, different AUC values ​​can be obtained. Finally, the classification threshold corresponding to the highest AUC is selected as the optimal classification threshold. The mean features and threshold traversal intervals for the two categories of data (REM / Wake and NREM) are determined from the training set data. Based on the mean features and threshold traversal intervals, the corresponding classification thresholds are selected. With the maximum AUC as the target, a preset first optimal classification threshold can be determined. The preset first optimal classification threshold is then compared with the predicted values ​​of the first dimensionality reduction features. Compare with the mean, if If the value is less than the preset first optimal classification threshold, it is determined to be a non-rapid eye movement (NREM) period; otherwise, it is a rapid eye movement (REM) period or a wakeful period (REM / Wake).

[0108] When the condition is determined to be either REM or wakefulness, a second optimal classification threshold and the second dimensionality reduction feature prediction value will be preset. The mean values ​​were compared to distinguish between REM and wakefulness periods.

[0109] Since it can only determine whether a period is REM or wake, but cannot specifically distinguish between REM and wake, it is necessary to further differentiate between them once the period is identified as REM or wake. Specifically, this involves determining the mean features and threshold ranges for both REM and wake categories from the training data. Based on these mean features and threshold ranges, a corresponding classification threshold is selected, and the maximum AUC is used as the objective to determine a preset second optimal classification threshold. This preset second optimal classification threshold is then compared with the predicted values ​​of the second dimensionality reduction features. Compare with the mean, if If the value is less than the preset second optimal classification threshold, it is determined to be the wakefulness stage; otherwise, it is determined to be the REM stage. It should be noted that the preset optimal classification threshold can also be determined using other methods. The above mainly describes the determination of sleep stages.

[0110] After the processor in the sleep state recognition device provided by this invention recognizes the user's sleep state, the controller in the implantable closed-loop stimulation system of this invention will output a stimulation signal according to the sleep state. Specifically:

[0111] In the first scenario, when the processor initially determines whether the user is asleep based on the sleep probability value in step S12b, if the sleep probability is less than a preset probability value, it determines that the user is not asleep and outputs the result that the user is not asleep. At this time, the controller in the implanted closed-loop stimulation system will determine whether the user is at night based on the clock reading. When it is determined that the user is at night, the controller outputs a stimulation signal using a preset stimulation amplitude S0. The preset stimulation amplitude S0 is an open-loop stimulation parameter and is a fixed value. Since the processor needs to re-determine the human body state according to a preset interval when the controller executes a certain stimulation mode, so that the controller can adjust the stimulation mode in time after the state changes, when outputting a stimulation signal using a stimulation amplitude S0, the processor executes S1 again after a preset interval to adapt to individual needs and state changes.

[0112] In the second scenario, if the processor determines in step S31 that there is a frequency division artifact in the local field potential LFP, then the processor outputs the result that there is a frequency division artifact in the local field potential LFP. At this time, the controller in the implanted closed-loop stimulation system will output the stimulation signal using the preset stimulation amplitude S0.

[0113] The third method involves the processor, in step S45, determining the predicted value group... After determining the sleep stage, the specific amplitude of the stimulation signal output by the controller is as follows:

[0114] During the waking period, the controller outputs a stimulation signal using a preset stimulation amplitude S0;

[0115] During the non-rapid eye movement (NREM) phase, the corresponding stimulus amplitude S1 of the controller output stimulus signal is:

[0116] ,

[0117] in, N1 represents the sleep state coefficient during non-rapid eye movement (NREM) sleep, N2 represents the upper limit of NREM stimulation, and N3 represents the lower limit of NREM stimulation. N1 and N3 are pre-set stimulation parameters.

[0118] During REM (Rapid Eye Movement) phase, the corresponding stimulus amplitude S2 of the stimulus signal output by the controller is:

[0119] ,

[0120] in, P represents the sleep state coefficient during REM sleep, T represents the upper limit of REM stimulation, and P and T represent the lower limit of REM stimulation. These are pre-set stimulation parameters.

[0121] This invention utilizes two model parameters, miuN1 and miuN3, pre-trained from the ARIMA model. miuN1 and miuN3 represent the mean values ​​of NREM stage 1 and NREM stage 3 sleep depth, respectively. Training data, including physiological data such as LFP, can be obtained offline from the training data.

[0122] Where, miuN1 < miuN3;

[0123] like If ≤ miuN1, then the sleep state coefficient during non-rapid eye movement (NREM) sleep is... =0;

[0124] like If ≥miuN3, then the sleep state coefficient during non-rapid eye movement (NREM) sleep is... =1;

[0125] miuN1< <miuN3, then the sleep state coefficient during non-rapid eye movement (NREM) sleep. .

[0126] This invention utilizes two pre-trained ARIMA model parameters, miuPhasic and miuTonic, which are the mean values ​​representing the depth states of phasic and tonic REM sleep, respectively. Training data, including physiological data such as LFP, can be obtained from the training data. miuPhasic and miuTonic are then acquired offline from the training data.

[0127] Among them, miuPhasic > miuTonic;

[0128] like If the value is ≥miuPhasic, then the sleep state coefficient during REM sleep is... =1;

[0129] like If ≤miuTonic, then the sleep state coefficient during REM sleep is... =0;

[0130] miuTonic < <miuPhasic, then the sleep state coefficient during REM sleep. .

[0131] In this embodiment, the controller integrates the discrimination and proportional adjustment based on NREM and REM sleep depth states, which helps to achieve precise sleep control while reducing the total amount of cumulative stimulation and improving the therapeutic effect on users.

[0132] In summary, this invention provides an implantable closed-loop stimulation system suitable for sleep states. It determines the appropriate stimulation mode for the user based on sleep stage results and outputs corresponding stimulation signals according to the stimulation mode, improving the accuracy of stimulation parameters. By accurately identifying the user's sleep state and stage using a sleep state recognition device, the controller can intelligently adjust stimulation parameters and output stimulation signals to achieve automatic stimulation of the deep brain, enabling personalized treatment and improving therapeutic efficacy.

[0133] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0134] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more blocks of the flowchart illustrations and / or one or more blocks of the block diagrams.

[0135] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more block diagrams.

[0136] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, such that the instructions, which execute on the computer or other programmable apparatus, provide steps for implementing the functions specified in one or more flowcharts and / or one or more block diagrams.

[0137] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A sleep state recognition device, characterized in that, Including the processor; The processor is used to initially determine whether the user is in a sleep state based on user status information and / or clock readings, wherein the user status information includes at least one of acceleration, circadian rhythm, electrocardiogram, and electroencephalogram; when the user is in a suspected sleep state, it acquires the local field potential (LFP) collected by the electrodes of the implantable closed-loop stimulation system and the acceleration collected by the triaxial accelerometer of the implantable closed-loop stimulation system; and further determines whether the user is in a sleep state based on the local field potential (LFP) and the acceleration. When the user is asleep, sleep stages are determined based on the local field potential (LFP).

2. The apparatus according to claim 1, characterized in that, A preliminary assessment of whether a user is asleep includes: Determine whether the clock reading is within a preset time period; When the clock reading is within a preset time period, it is determined that the user is in a suspected sleep state; If the clock reading is not within a preset time period, it is determined that the user is not in a sleep state.

3. The apparatus according to claim 1, characterized in that, A preliminary assessment of whether a user is asleep includes: At preset intervals, the system initially determines whether the user is asleep based on acceleration. Temporal features of triaxial acceleration are extracted, including the mean and standard deviation of acceleration. The mean acceleration and standard deviation of acceleration are respectively input into two pre-trained sleep state prediction models to predict the probability of sleep state, and the two prediction results are multiplied to obtain the sleep probability value. Based on the sleep probability value, a preliminary judgment can be made as to whether the user is in a sleep state; When the sleep probability value is greater than or equal to the preset probability value, the user is determined to be in a suspected sleep state; When the sleep probability value is less than the preset probability value, it is determined that the user is not in a sleep state.

4. The apparatus according to claim 1, characterized in that, To determine if the user is asleep again, the following steps are taken: Determine whether the local field potential LFP has a frequency division artifact; When the local field potential LFP does not have frequency division artifacts, time-domain features are extracted from the triaxial acceleration to obtain the mean acceleration and the standard deviation of acceleration. The mean acceleration and standard deviation of acceleration are respectively input into two pre-trained sleep state prediction models to predict the probability of sleep state, and the two prediction results are multiplied to obtain the sleep probability value. Based on the sleep probability value, determine again whether the user is in a sleep state; When the sleep probability value is greater than or equal to the preset probability value, it is determined that the user is in a sleep state; When the sleep probability value is less than the preset probability value, it is determined that the user is not in a sleep state.

5. The apparatus according to claim 1, characterized in that, Determining sleep stages based on the local field potential (LFP) includes: Power spectrum estimation and denoising are performed on the local field potential LFP; The denoised power spectrum is merged into multiple frequency band energies and then normalized. The normalized energy across multiple frequency bands is reduced in dimensionality to obtain multiple dimensionality-reduced features, including the first dimensionality-reduced feature. Second dimensionality reduction feature Third dimensionality reduction feature Fourth dimensionality reduction feature ； The multiple dimensionality-reduced features are input into a pre-built ARIMA model to obtain a set of predicted values. ； According to the predicted value group Sleep stages are determined, including wakefulness, REM sleep, and non-REM sleep.

6. The apparatus according to claim 5, characterized in that, The predicted value group Including the first dimensionality reduction feature prediction value Second dimensionality reduction feature prediction value Third dimensionality reduction feature prediction value and the fourth dimensionality reduction feature prediction value First dimensionality reduction feature prediction value Second dimensionality reduction feature prediction value Third dimensionality reduction feature prediction value and the fourth dimensionality reduction feature prediction value Each contains multiple predicted values; According to the predicted value group Determining sleep stages includes: Calculate the mean of the predicted values ​​for each dimensionality reduction feature; The preset first optimal classification threshold and the first dimensionality reduction feature prediction value are used. The mean values ​​were compared to determine whether the REM or wakefulness phase, or the non-REM phase, were used; When the condition is determined to be either REM or wakefulness, a second optimal classification threshold and the second dimensionality reduction feature prediction value will be preset. The mean values ​​were compared to distinguish between REM and wakefulness periods.

7. An implantable closed-loop stimulation system suitable for sleep states, characterized in that, The device includes the sleep state recognition device according to any one of claims 1-6, and a controller; The controller is used to proportionally adjust the output stimulation signal using stimulation amplitudes corresponding to different sleep stages, wherein the stimulation amplitudes are different for different sleep stages.

8. The system according to claim 7, characterized in that, When it is initially determined that the user is not asleep, the clock reading is used to determine whether it is nighttime. If it is nighttime, the preset stimulation amplitude S0 is used to output the stimulation signal. When the user is determined to be awake again, a stimulation signal is output using a preset stimulation amplitude S0.

9. The system according to claim 7, characterized in that, When it is determined that the local field potential LFP has a frequency division artifact, a stimulation signal is output using a preset stimulation amplitude S0.

10. The system according to claim 7, characterized in that, For the waking phase in sleep stages, a preset stimulation amplitude S0 is used to output the stimulation signal; For the non-rapid eye movement (NREM) sleep stage, the corresponding stimulus amplitude S1 of the output stimulus signal is: , in, N1 represents the upper limit of non-rapid eye movement (NREM) stimulation, and N3 represents the lower limit of NREM stimulation. For the REM sleep stage, the corresponding stimulus amplitude S2 of the output stimulus signal is: , in, denoted as the sleep state coefficient during REM sleep, where P is the upper limit of REM stimulation and T is the lower limit of REM stimulation.

Images