Sleep cognition enhancement system and method based on deep brain stimulation

Abstract

The invention discloses a sleep cognition enhancement system and method based on deep brain stimulation. The sleep cognition enhancement system comprises a data acquisition unit, a sleep staging unit, a first detection unit, a second detection unit, a third detection unit and an electrical stimulation unit. Wherein the data acquisition unit is used for acquiring real-time SEEG signals and scalp electroencephalogram data of a user, so that real-time sleep staging is carried out through the sleep staging unit. Based on a real-time sleep staging result of a user, real-time detection of a slow wave rising stage, a spindle wave valley and a hippocampus sharp wave ripple is performed on the user through the three detection units, so that three detection results are output. Judging whether the three detection results meet a first preset condition or not; if yes, electrical stimulation is carried out on the three preset areas through an electrical stimulation unit; and if not, detecting the user again.

Description

Technical Field

[0001] The present invention relates to a sleep cognition enhancement system based on deep brain stimulation, and also to a corresponding sleep cognition enhancement method, belonging to the field of neural regulation technology. Background Art

[0002] Sleep is one of the most fundamental human physiological activities and the "golden hour" for cognitive map formation and memory consolidation. During this period, the brain spontaneously undergoes "neural replay," and hippocampal sharp wave ripples (SWRs) are the core biomarker driving this process. SWRs are the most synchronized neuronal cluster firing pattern in the mammalian brain. Each SWR is accompanied by the sequential replay of information learned during the day and the abstract processing of experience, thereby transforming short-term memories into long-term memories and ultimately building and optimizing cognitive maps.

[0003] Traditional memory consolidation theory further states that short-term memories are initially formed in the hippocampus and then gradually "transferred" to the neocortex during sleep to become long-term memories. During non-rapid eye movement (NREM) sleep, hippocampal SWRs, thalamic spindles, and cortical slow-wave oscillations form a "three-wave coupling" with specific phase relationships. This spatiotemporal coupling is considered a key mechanism for memory consolidation—knowledge acquired during the day is accurately replayed at night through "three-wave coupling," allowing the brain to organize and consolidate newly encoded memories.

[0004] Based on this theoretical framework, studies have confirmed that precisely applying sound stimulation during the up-state or down-state of slow waves can enhance or weaken memory for images paired with the sound, respectively. Similarly, transcranial electrical or magnetic stimulation during specific phases of slow waves or spindles can also modulate sleep structure and memory processing. These findings, for the first time, suggest the feasibility of improving memory consolidation during sleep through neuromodulation, and have become a cutting-edge topic in the fields of brain-computer interfaces, cognitive psychology, and clinical diagnosis and treatment.

[0005] However, existing research faces two major bottlenecks: First, real-time decoding accuracy is insufficient: it is not yet possible to simultaneously capture the precise phase of the spindle trough and hippocampal SWR on a millisecond scale, resulting in a blurred control window. Second, stimulation methods are limited: sound and transcranial electrical / magnetic stimulation are non-invasive techniques with low spatial resolution, making it difficult to precisely target deep structures such as the entorhinal cortex and hippocampus. Therefore, there is currently a lack of experimental evidence directly linking the "different states of slow waves, three-wave coupling, and memory consolidation" relationship. Furthermore, there is a lack of a sleep cognition enhancement system based on the "three-wave coupling" theory that can precisely stimulate deep brain regions. Summary of the Invention

[0006] The primary technical problem to be solved by the present invention is to provide a sleep cognition enhancement system based on deep brain stimulation.

[0007] Another technical problem to be solved by the present invention is to provide a method for enhancing sleep cognition based on deep brain stimulation.

[0008] In order to achieve the above technical objectives, the present invention adopts the following technical solutions:

[0009] According to a first aspect of an embodiment of the present invention, a sleep cognition enhancement system based on deep brain stimulation is provided, comprising:

[0010] Data acquisition unit, used to obtain the user's real-time SEEG signal and scalp EEG data;

[0011] a sleep staging unit connected to the data acquisition unit, configured to receive the user's real-time SEEG signals and scalp EEG data, perform real-time sleep staging, and output real-time sleep staging results;

[0012] a first detection unit connected to the sleep staging unit to detect the slow wave rising stage in real time based on the real-time sleep staging result and output a first detection result;

[0013] a second detection unit connected to the sleep staging unit to perform real-time detection of spindle wave troughs based on the real-time sleep staging result and output a second detection result;

[0014] a third detection unit connected to the sleep staging unit to perform real-time detection of hippocampal sharp wave ripples based on the real-time sleep staging result and output a third detection result;

[0015] The electrical stimulation unit is connected to the first detection unit, the second detection unit and the third detection unit, and is used to electrically stimulate the first preset area, the second preset area and the third preset area when the first detection result, the second detection result and the third detection result meet the first preset condition.

[0016] Preferably, the sleep staging unit includes:

[0017] a preprocessing module connected to the data acquisition unit to receive the SEEG signal and scalp EEG data of the user and perform data preprocessing;

[0018] a feature extraction module connected to the preprocessing module to extract corresponding signal features in real time based on the preprocessed data; wherein the signal features include at least: time domain signal standard deviation, interquartile range, skewness, kurtosis, and power of multiple frequency band signals;

[0019] The sleep recognition module is connected to the feature extraction module and is preset with a sleep staging model to receive the signal features and use the signal features of the previous period to predict the sleep stage at a predetermined time thereafter, thereby performing real-time sleep staging.

[0020] Preferably, the first detection unit includes:

[0021] a slow wave falling stage detection module, connected to the sleep staging unit, for detecting the slow wave falling stage in real time based on a preset first threshold after the sleep stage enters stage N3, and outputting a slow wave falling stage detection result;

[0022] a slow wave rising phase detection module, connected to the slow wave falling phase detection module, configured to perform real-time detection of the slow wave rising phase based on a preset second threshold when the slow wave falling phase detection result is lower than the first threshold, and output the slow wave rising phase detection result;

[0023] a slow wave length detection module, connected to the slow wave rising phase detection module, configured to detect the length of the slow wave when the detection result of the slow wave falling phase is lower than the first threshold and the detection result of the slow wave rising phase is higher than the second threshold, and output the slow wave length detection result;

[0024] Among them, when the detection result of the slow wave descending phase is lower than the first threshold, the detection result of the slow wave rising phase is higher than the second threshold, and the detection result of the slow wave length is consistent with the frequency of the slow wave, then the first detection result is that the slow wave is detected; otherwise, the first detection result is that the slow wave is not detected.

[0025] Preferably, the first detection unit further includes:

[0026] A threshold updating module is connected to the slow wave rising phase detection module and the slow wave falling phase detection module to update the first threshold and the second threshold based on all slow wave events in the N3 sleep period within a preset time period.

[0027] Preferably, the second detection unit includes:

[0028] an RMS power detection module, connected to the sleep staging unit, configured to perform RMS power detection based on a third threshold after the sleep stage enters N2 and N3 stages, and output an RMS power detection result; wherein, when the RMS power detection result is higher than the third threshold and lasts for more than a preset time, it indicates that the center of the spindle wave has been detected;

[0029] a power value detection module connected to the RMS power detection module to output a power value detection result; wherein, when the time derivative of the RMS power becomes zero and the RMS power begins to decrease, the maximum spindle power is detected;

[0030] a spindle end detection module connected to the RMS power detection module to output a spindle end detection result; wherein the spindle end is detected when the RMS power detection result changes from being higher than the third threshold to being lower than the third threshold;

[0031] a pause and resume detection module, connected to the RMS power detection module, the power value detection module, and the spindle end detection module, to pause or resume detection based on the RMS power detection result, the power value detection result, and the spindle end detection result;

[0032] The phase detection module performs real-time phase detection on the spindle waves based on a preset algorithm to output the phase detection results;

[0033] Among them, when the RMS power detection result, the power value detection result, the spindle wave end detection result and the phase detection result meet the second preset condition, the second detection result is that the spindle wave trough is detected; otherwise, the second detection result is that the spindle wave trough is not detected.

[0034] Preferably, the third detection unit includes:

[0035] a filtering module, connected to the sleep staging unit, for performing sequential bandpass filtering on the SEEG signal of the hippocampus after the sleep stage enters stage N3;

[0036] a pulse conversion module connected to the filtering module to define a baseline amplitude that a sharp wave ripple event must exceed based on a time window of a preset length, and convert the electrophysiological signal into an upward pulse event / downward pulse event according to the amplitude strength and direction;

[0037] a double-layer spiking neural network, connected to the spiking conversion module, to perform real-time detection of hippocampal sharp wave ripples based on the upward spiking event / downward spiking event, and output a third detection result;

[0038] Among them, when the third preset condition is met, the third detection result is that hippocampal sharp wave ripples are detected; otherwise, the third detection result is that hippocampal sharp wave ripples are not detected.

[0039] Preferably, the electrical stimulation unit comprises:

[0040] a first stimulation portion corresponding to a first preset area, for electrically stimulating the first preset area based on preset stimulation parameters;

[0041] a second stimulation portion corresponding to a second preset area, for electrically stimulating the second preset area based on preset stimulation parameters;

[0042] a third stimulation portion corresponding to a third preset area, for electrically stimulating the third preset area based on preset stimulation parameters;

[0043] Among them, the first preset area corresponds to the real-time detection of the rising phase of the slow wave, the second preset area corresponds to the real-time detection of the spindle wave trough, and the third preset area corresponds to the real-time detection of the hippocampal sharp wave ripples.

[0044] Preferably, the first preset area is the neocortex, the second preset area is the entorhinal cortex, and the third preset area is the white matter adjacent to the hippocampus CA1.

[0045] Preferably, the preset stimulation parameters are: 5 bidirectional pulse signals with a frequency of 100 Hz, a current intensity of 1.5 mA, a duration and interval time of two bidirectional pulses of 100 μs, and a duration of 50 ms for each stimulation.

[0046] According to a second aspect of an embodiment of the present invention, a method for enhancing sleep cognition based on deep brain stimulation is provided, comprising the following steps:

[0047] Obtain the user's real-time SEEG signals and scalp EEG data;

[0048] Based on the user's real-time SEEG signal and scalp EEG data, the user is subjected to real-time sleep staging, thereby outputting a real-time sleep staging result;

[0049] Based on the real-time sleep staging result, detecting the slow wave rising stage of the user in real time, and outputting a first detection result;

[0050] Based on the real-time sleep staging result, performing real-time detection of spindle wave troughs on the user, and outputting a second detection result;

[0051] Based on the real-time sleep staging result, performing real-time detection of hippocampal sharp wave ripples on the user, and outputting a third detection result;

[0052] If the first detection result, the second detection result, and the third detection result meet the first preset condition, then electrically stimulating the first preset area, the second preset area, and the third preset area through the electrical stimulation unit;

[0053] If any one of the first detection result, the second detection result and the third detection result does not meet the first preset condition, the user's sleep stage is re-classified based on the user's real-time SEEG signal and scalp EEG data, and the test is performed again until the first detection result, the second detection result and the third detection result meet the first preset condition.

[0054] Compared with the prior art, the present invention has the following technical effects:

[0055] (1) It can capture three key physiological events in parallel and in real time during non-rapid eye movement sleep: the rising phase of slow waves, the trough of spindle waves, and the hippocampal sharp wave ripples. With the help of millisecond-level sliding windows and specialized algorithms, it provides a precise time window for "targeted firing" for neural regulation, thereby improving the temporal resolution to an unprecedented level.

[0056] (2) The system's built-in electrical stimulation unit acts directly on the neocortex, entorhinal cortex, and hippocampal formation via implanted deep electrodes. Compared with non-invasive methods such as transcranial magnetic stimulation or transcranial electrical stimulation, this targeted stimulation has a spatial positioning error of less than 0.5 mm, significantly improving spatial resolution and ensuring effective intervention of deep circuits.

[0057] (3) By seamlessly combining the above-mentioned “high-precision time window” with the “high-precision spatial target”, the system not only provides a new closed-loop research tool for the formation of cognitive maps and memory processing mechanisms during sleep, but also brings potential new personalized treatment options for patients with cognitive and memory disorders such as Alzheimer’s disease. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 This is an overall schematic diagram of a sleep cognition enhancement system based on deep brain stimulation provided by the first embodiment of the present invention;

[0059] Figure 2 This is a diagram of an application scenario of the sleep cognition enhancement system based on deep brain stimulation provided by the first embodiment of the present invention;

[0060] Figure 3 This is an overall flow chart of the sleep cognition enhancement method based on deep brain stimulation provided by the second embodiment of the present invention;

[0061] Figure 4 This is a detailed flow chart of the sleep cognition enhancement method based on deep brain stimulation provided in the second embodiment of the present invention. DETAILED DESCRIPTION

[0062] The technical content of the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0063] An embodiment of the present invention provides a sleep cognition enhancement system based on deep brain stimulation. It is based on the mechanism of the slow waves, spindles and hippocampal sharp wave ripples in the non-rapid eye movement period of sleep in the formation of cognitive maps and the consolidation of memory processing. In combination with stereotactic electroencephalography (SEEG) / scalp electroencephalography (EEG) and neuroimaging (magnetic resonance imaging), real-time detection of the rising phase of slow waves, the trough of spindles and hippocampal sharp wave ripples is achieved through multiple pattern recognition and neural network methods. At the same time, deep brain stimulation technology is used to directly electrically stimulate the neocortex, entorhinal cortex and hippocampus, ultimately building a neural regulation brain-computer interface system based on the "three-wave coupling" of sleep.

[0064] First embodiment

[0065] like Figure 1 and Figure 2 As shown, the first embodiment of the present invention provides a deep brain stimulation-based sleep cognition enhancement system, comprising a data acquisition unit 1, a sleep staging unit 2, a first detection unit 3, a second detection unit 4, a third detection unit 5, and an electrical stimulation unit 6. The data acquisition unit 1 is used to acquire a user's real-time SEEG signals and scalp EEG data, which are then processed by the sleep staging unit 2 to perform real-time sleep staging on the user. After outputting the user's real-time sleep staging results, the first detection unit 3, the second detection unit 4, and the third detection unit 5 respectively detect the user's real-time slow wave rising phase, spindle wave trough, and hippocampal sharp wave ripples, thereby outputting three detection results. Finally, a determination is made as to whether the three detection results meet a first preset condition. If so, the electrical stimulation unit 6 electrically stimulates the three preset areas to improve the user's memory processing ability. If not, the user is retested until all three detection results meet the first preset condition.

[0066] The following describes in detail the structure and working process of each unit of the sleep cognition enhancement system:

[0067] (1) Data acquisition unit

[0068] In this embodiment, the data acquisition unit 1 is required to collect the user's SEEG signals and scalp EEG data. The SEEG signals are collected using a first method, while the scalp EEG data are collected using a second method. Combining these two data types improves the precision of subsequent data decoding, thereby increasing data detection accuracy.

[0069] It is understandable that the two data collection methods used in this embodiment are both conventional technical means in the field. In other embodiments, appropriate data collection methods can be selected as needed to achieve data collection, and no specific limitation is made here.

[0070] (2) Sleep staging unit

[0071] Reference Figure 1 As shown, in this embodiment, the sleep staging unit 2 includes a preprocessing module 21, a feature extraction module 22, and a sleep recognition module 23. Specifically, the preprocessing module 21 is connected to the data acquisition unit 1 to receive the user's SEEG signals and scalp EEG data and perform data preprocessing (for example, filtering, denoising, and standardization of the collected EEG signals) to improve signal quality and reliability. It is understood that these preprocessing steps help reduce interference and artifacts, making subsequent analysis more accurate. However, after the data preprocessing is completed, feature extraction needs to be performed by the feature extraction module 22.

[0072] The feature extraction module 22 is connected to the preprocessing module 21 to extract relevant signal features in real time based on the preprocessed data. In this embodiment, these signal features include at least: time domain signal standard deviation, interquartile range, skewness, kurtosis, and power of multiple frequency band signals. After feature extraction by the feature extraction module 22 is completed, the sleep recognition module 23 is used to perform real-time sleep stage classification.

[0073] The sleep recognition module 23 is connected to the feature extraction module 22 and pre-installed with a sleep staging model. The sleep staging model is trained using the user's EEG data (C3-M2 (C4-M1) and electrooculogram (EOG) from the previous night. The labels for this training data are derived from the currently popular Yasa sleep staging algorithm. The core of the Yasa sleep staging algorithm is the Light GBM classifier, a gradient boosting machine-based classifier that uses a decision tree model based on a learning algorithm. It reduces memory usage through sparse optimization and optimal segmentation of class features.

[0074] After the feature extraction module 22 completes feature extraction, the sleep recognition module 23 receives the extracted signal features and uses the signal features of the previous period (e.g., 30s or 60s, but not limited to this) to predict the sleep stage at a predetermined time thereafter (e.g., 3s or 5s, but not limited to this), thereby performing real-time sleep staging. In addition, in this embodiment, when performing sleep staging using the sleep staging model, a random forest algorithm is first used to perform real-time sleep staging on the scalp EEG, and Class A random undersampling is used to address the issue of the shorter duration of stages N3 and N1 compared to the awake period. It is understood that the random forest algorithm and Class A random undersampling are both common knowledge in the field and will not be elaborated on in detail here.

[0075] (3) First detection unit

[0076] like Figure 1 As shown, in this embodiment, the first detection unit 3 is connected to the sleep staging unit 2 to detect the slow wave rising phase in real time based on the real-time sleep staging results and output a first detection result. Specifically, the first detection unit 3 includes a slow wave falling phase detection module 31, a slow wave rising phase detection module 32, and a slow wave length detection module 33. The slow wave falling phase detection module 31 is used to perform real-time detection of the slow wave falling phase, the slow wave rising phase detection module 32 is used to perform real-time detection of the slow wave rising phase, and the slow wave length detection module 33 is used to detect the length of the slow wave.

[0077] Slow waves (0.16-1.25 Hz slow wave oscillations) are composed of periodic alternations of rising and falling slow wave phases. The rising phase manifests as an increase in EEG amplitude, while the falling phase manifests as a decrease in EEG amplitude below baseline. Each falling phase lasts approximately 100-300 milliseconds and alternates with rising phases, forming a slow wave oscillation.

[0078] In this embodiment, all slow waves in the sleep N3 stage are first extracted based on the user's neocortex and hippocampus sleep data from the previous night, and 75% of the slow wave descending stage is set as the first threshold, and 75% of the slow wave ascending stage is set as the second threshold, and the lengths of all slow waves are calculated at the same time.

[0079] Then, the SEEG signal is subjected to real-time serial bandpass filtering (0.16-1.25Hz) on the current night, and then data caches with lengths of 2s and 400s are defined respectively. Based on the real-time sleep staging results of the sleep staging unit 2, when the user's sleep enters the N3 stage, the first threshold is used to perform real-time detection of the slow wave falling stage. When the SEEG signal is lower than the first threshold of the slow wave falling stage, the 2s data cache is subjected to slow wave rising stage threshold detection. When the above-mentioned slow wave rising stage detection result is higher than the second threshold, the slow wave length detection module 33 is used to detect the length of the slow wave to ensure that the length of the slow wave meets the frequency of the slow wave.

[0080] Finally, when the detection result of the slow wave descending phase is lower than the first threshold, the detection result of the slow wave rising phase is higher than the second threshold, and the detection result of the slow wave length is consistent with the frequency of the slow wave, the first detection result output by the first detection unit 3 is that the slow wave is detected; otherwise, the first detection result output by the first detection unit 3 is that the slow wave is not detected.

[0081] In addition, in this embodiment, the first detection unit 3 further includes a threshold updating module 34. The threshold updating module is connected to the slow wave falling phase detection module 31 and the slow wave rising phase detection module 32 to update the first threshold and the second threshold based on all slow wave events in the N3 sleep data within a preset time period (i.e., the aforementioned 400 s data buffer).

[0082] (4) Second detection unit

[0083] like Figure 1 As shown, in this embodiment, the second detection unit 4 is connected to the sleep staging unit 2 to perform real-time detection of spindle troughs based on the real-time sleep staging results and output a second detection result. Specifically, the second detection unit 4 includes an RMS (Root Mean Square) power detection module 41, a power value detection module 42, a spindle end detection module 43, a pause and resume detection module 44, and a phase detection module 45.

[0084] The RMS power detection module 41 is connected to the sleep staging unit 2 and is configured to perform RMS power detection based on a third threshold after entering sleep stages N2 and N3, and output the RMS power detection result. When the RMS power detection result is higher than the third threshold (the mean EEG signal power value obtained from the user's previous night's data + 1.15 standard deviations) and lasts longer than a preset duration (250 milliseconds in this embodiment, corresponding to half the shortest spindle duration of 500 milliseconds), it indicates that the spindle center has been detected. Otherwise, it indicates that the spindle center has not been detected.

[0085] The power value detection module 42 is connected to the RMS power detection module 41 to output a power value detection result. When the time derivative of the RMS power becomes zero and the RMS power begins to decrease, the maximum spindle power is detected; otherwise, the maximum spindle power is not detected.

[0086] The spindle end detection module 43 is connected to the RMS power detection module 42 to output a spindle end detection result. When the RMS power detection result changes from being above a third threshold to being below the third threshold, the spindle end is detected; otherwise, the spindle end is not detected.

[0087] Pause and resume detection module 44 is connected to RMS power detection module 41, power value detection module 42, and spindle end detection module 43 to pause or resume detection based on the RMS power detection results, power value detection results, and spindle end detection results. Specifically, in this embodiment, if the RMS power signal remains above the third threshold for more than 2 seconds, detection is paused until at least three of the four signal characteristics again exceed their respective thresholds.

[0088] The phase detection module 45 performs real-time phase detection on the spindle waves based on a preset algorithm to output a phase detection result.

[0089] During the specific detection, all spindles in the N2 stage of sleep are first extracted based on the user's neocortex and hippocampus sleep data from the previous night, and 75% of the peak value of the spindle is set as the initial threshold, and the length of all spindles is calculated. Then, the SEEG signal is subjected to serial bandpass filtering (12-16Hz) on the current night, and data caches of lengths of 4s and 400s are defined respectively. Based on the real-time sleep staging results of the sleep staging module 2, when the user's sleep enters the N2 and N3 stages, when at least three of the following four signal characteristics meet their respective criteria and the duration of the spindle meets the requirements (that is, when the second preset condition is met), it can be determined that a spindle has been detected:

[0090] ①RMS power signal detection: After the SEEG RMS power signal exceeds the “entry threshold” (mean + 1.15 standard deviations), if the duration exceeds 250 milliseconds (corresponding to half of the shortest spindle duration of 500 milliseconds), it indicates that the spindle center has been detected.

[0091] ② Maximum power signal: When the time derivative of the RMS power of the SEEG becomes zero and the RMS power begins to decrease, the maximum spindle power is detected, which is assumed to be the center of the spindle.

[0092] ③ End of spindle wave: When the signal falls below the “entry threshold” again, the end of the spindle wave is detected.

[0093] ④ Pause and resume detection: If the RMS power signal remains above the "entry threshold" for more than 2 seconds, the detection will be suspended until at least three of the four signal characteristics exceed their respective thresholds again.

[0094] In addition to detecting spindle events, it is also necessary to estimate their oscillation phase in real time. In this embodiment, an open-source real-time phase estimation algorithm, phastimate, is used to achieve real-time detection and phase estimation of spindles, providing accurate and real-time spindle trough detection results.

[0095] (5) The third detection unit

[0096] like Figure 1 As shown, in this embodiment, the third detection unit 5 is connected to the sleep staging unit 2 to perform real-time detection of hippocampal sharp wave ripples based on the real-time sleep staging results and output a third detection result. Specifically, the third detection unit 5 includes a filtering module 51, a pulse conversion module 52, and a two-layer pulse neural network 53. Among them, the filtering module 51 is connected to the sleep staging unit 2 and is used to perform sequential bandpass filtering (70-180 Hz) on the hippocampal SEEG signal after the sleep stage enters stage N3, and then enter the signal pulse conversion stage.

[0097] The pulse conversion module 52 is connected to the filtering module 51 to define a baseline amplitude that a sharp wave ripple event must exceed based on a time window of a preset length, and convert the electrophysiological signal into an upward pulse event / downward pulse event based on the amplitude strength and direction. Specifically, in this embodiment, a baseline amplitude that a sharp wave ripple event must exceed is predefined within a time window of a specific length (the time window length in this embodiment is selected as 0.1s) (in this embodiment, the maximum signal amplitude of 0.05s of continuous non-overlapping signals within this time window is selected, and the average value at the upper and lower quartiles is used as the baseline amplitude). An asynchronous modulator is used to convert the electrophysiological signal into an upward pulse event / downward pulse event based on the amplitude strength and direction.

[0098] The two-layer spiking neural network 53 is connected to the pulse conversion module 52 to perform real-time detection of hippocampal sharp wave ripples based on upward / downward pulse events, and output a third detection result. Specifically, after the pulse event is obtained, it is input into the two-layer spiking neural network that integrates dynamic synaptic and neuronal discharges. Finally, by extracting the key features of the electrophysiological signal within this time period, the previously trained classifier with the best predictive performance is used to perform real-time detection of hippocampal sharp wave ripples.

[0099] It can be understood that, in this embodiment, when the preset condition is met, the third detection result is that hippocampal sharp wave ripples are detected; otherwise, the third detection result is that hippocampal sharp wave ripples are not detected.

[0100] (6) Electrical stimulation unit

[0101] like Figure 1 As shown, in this embodiment, the electrical stimulation unit 6 is connected to the first detection unit 3, the second detection unit 4, and the third detection unit 5, and is used to electrically stimulate the first preset area, the second preset area, and the third preset area when the first detection result, the second detection result, and the third detection result meet the first preset condition. The electrical stimulation unit 6 includes a first stimulation part 61, a second stimulation part 62, and a third stimulation part 63. The three stimulation parts correspond to the three preset areas respectively, so as to electrically stimulate the corresponding preset areas based on the preset stimulation parameters.

[0102] Specifically, in the present embodiment, the first preset area is the neocortex, corresponding to the real-time detection of the rising phase of the slow wave, and the first preset area is electrically stimulated by the first stimulation unit 61. The second preset area is the entorhinal cortex, corresponding to the real-time detection of the spindle wave trough, and the second preset area is electrically stimulated by the second stimulation unit 62. The third preset area is the white matter adjacent to the hippocampal CAI, corresponding to the real-time detection of the hippocampal sharp wave ripples, and the third preset area is electrically stimulated by the third stimulation unit 63. Among them, the first stimulation unit 61, the second stimulation unit 62 and the third stimulation unit 63 implement electrical stimulation based on the same preset stimulation parameters, which are: 5 bidirectional pulse signals with a frequency of 100 Hz, a current intensity of 1.5 mA, a duration and interval time of two bidirectional pulses of 100 μs, and a duration of 50 ms for each stimulation. In addition, in the present embodiment, the first preset condition is: the first detection result is the detection of slow waves, the second detection result is the detection of the spindle wave trough, and the third detection result is the detection of hippocampal sharp wave ripples. Therefore, based on the "three-wave coupling" of sleep, deep brain stimulation technology is used to directly electrically stimulate the neocortex, entorhinal cortex and hippocampus, and finally a sleep cognitive enhancement system based on deep brain stimulation is built.

[0103] Second embodiment

[0104] like Figure 3 and Figure 4 As shown, based on the above-mentioned first embodiment, the second embodiment of the present invention provides a method for enhancing sleep cognition based on deep brain stimulation, which specifically includes the following steps:

[0105] S1: Obtain the user's real-time SEEG signal and scalp EEG data.

[0106] Specifically, in this embodiment, the user's SEEG signal and scalp EEG data are collected in real time by the data collection unit 1. The SEEG signal is collected using a first method, and the scalp EEG data is collected using a second method.

[0107] S2: Based on the user's real-time SEEG signal and scalp EEG data, the user's sleep staging is performed in real time, thereby outputting real-time sleep staging results.

[0108] Specifically, after the user's SEEG signal and scalp EEG data are obtained based on step S1, first, the signal features are extracted after data preprocessing; then, based on a preset sleep staging model, the signal features of the previous period (for example, 30s or 60s, etc., but not limited to this) are used to predict the sleep stage at a predetermined time thereafter (for example, 3s or 5s, etc., but not limited to this), thereby performing real-time sleep staging.

[0109] S3: Perform sleep detection on the user based on the real-time sleep staging results.

[0110] Specifically, this step is divided into three aspects of detection, namely, real-time detection of the rising phase of the slow wave, real-time detection of the spindle wave trough, and real-time detection of the hippocampal sharp wave ripples.

[0111] In this embodiment, when the user is in sleep stage N3, the first detection unit 3 performs real-time detection of the user's slow wave rising phase and outputs a first detection result. The specific detection logic of the first detection unit 3 refers to the description of the first embodiment and is not repeated here. Furthermore, when the user is in sleep stage N3, the third detection unit 5 performs real-time detection of the user's hippocampal sharp wave ripples and outputs a third detection result. The specific detection logic of the third detection unit 5 refers to the description of the first embodiment and is not repeated here.

[0112] In addition, when the user is in sleep stage N2, the second detection unit 4 performs real-time detection of the user's spindle wave trough and outputs a second detection result. The specific detection logic of the second detection unit 4 refers to the description of the first embodiment and is not repeated here.

[0113] It should be noted that, in this embodiment, the above three aspects of detection are performed in parallel, thereby respectively realizing the detection of sharp wave ripples, slow waves and spindle waves, and the most core detection is the detection of sharp wave ripples.

[0114] S4: Implement electrical stimulation on the preset area.

[0115] After obtaining the first, second and third test results based on step S3, it is necessary to determine whether the three test results meet the first preset condition. If so, electrical stimulation is implemented. If not, sleep staging is performed again and tested again.

[0116] Specifically, if the first detection result indicates slow waves, the second detection result indicates spindles, and the third detection result indicates hippocampal sharp wave ripples, electrical stimulation is applied to the neocortex, entorhinal cortex, and white matter adjacent to the hippocampal CA1 using the first stimulation section 61, the second stimulation section 62, and the third stimulation section 63, respectively. Conversely, if any of the detection results indicate no detection, no electrical stimulation is applied.

[0117] In summary, the sleep cognition enhancement system and method based on deep brain stimulation provided by the embodiments of the present invention have the following beneficial effects:

[0118] First, in terms of time, the system uses a "three-stage parallel detection" mechanism: at the same moment, it performs real-time synchronous detection on the rising phase of the slow wave in stage N3, the trough of the spindle wave in stage N2, and the hippocampal sharp wave ripple in stage N3. With the help of a millisecond-level sliding time window, the system can respond very quickly after the event occurs. Among them, the spindle phase is completed by the real-time phase estimation algorithm phastimate, and the phase estimation error is only ±10ms; the hippocampal sharp wave ripple is processed by a two-layer pulse neural network to process high-frequency signals. Compared with the traditional FFT method, the time resolution is significantly improved. In addition, the slow wave detection threshold is not fixed, but is updated based on the rolling data of the previous 400s. It can adapt to individual differences and significantly reduce the false alarm rate.

[0119] Secondly, in the spatial dimension, the system delivers electrical stimulation with preset parameters directly to the neocortex, entorhinal cortex, and hippocampus through stereotactic deep electrodes. Compared with traditional non-invasive neuromodulation (transcranial magnetic or electrical stimulation positioning error > 10mm), the positioning error of stereotactic electrodes is < 0.5mm, which significantly improves the spatial resolution and greatly increases the success rate of stimulation of the entorhinal cortex. More importantly, the system simultaneously stimulates the three targets of "neocortex-entorhinal cortex-hippocampus" in a coordinated manner, forming a closed memory loop regulation, further enhancing the stimulation effect.

[0120] Finally, by organically combining the above-mentioned high temporal resolution with high spatial resolution, the system not only provides a new closed-loop research tool for complex scientific problems such as the formation of cognitive maps and memory processing mechanisms during sleep, but also brings practical solutions to improving cognitive functions in patients with cognitive and memory disorders such as Alzheimer's disease.

[0121] Therefore, the embodiments of the present invention solve the core problem of "imbalance in spatiotemporal resolution" in traditional neuromodulation technology (insufficient spatial resolution of non-invasive means and lack of temporal precision of deep stimulation), and provide a new closed-loop intervention paradigm for the treatment of cognitive impairment.

[0122] It should be noted that the above embodiments are merely examples, and the technical solutions of the various embodiments may be combined and are all within the scope of protection of the present invention.

[0123] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature identified as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0124] The above describes in detail the deep brain stimulation-based sleep cognitive enhancement system and method provided by the present invention. For those skilled in the art, any obvious modification without departing from the essence of the present invention would constitute an infringement of the present invention's patent rights and would result in corresponding legal liability.

Claims

1. A sleep cognitive enhancement system based on deep brain stimulation, characterized in that include: Data acquisition unit, used to obtain the user's real-time SEEG signal and scalp EEG data; a sleep staging unit connected to the data acquisition unit, configured to receive the user's real-time SEEG signals and scalp EEG data, perform real-time sleep staging, and output real-time sleep staging results; a first detection unit connected to the sleep staging unit to detect the slow wave rising stage in real time based on the real-time sleep staging result and output a first detection result; a second detection unit connected to the sleep staging unit to perform real-time detection of spindle wave troughs based on the real-time sleep staging result and output a second detection result; a third detection unit connected to the sleep staging unit to perform real-time detection of hippocampal sharp wave ripples based on the real-time sleep staging result and output a third detection result; The electrical stimulation unit is connected to the first detection unit, the second detection unit and the third detection unit, and is used to electrically stimulate the first preset area, the second preset area and the third preset area when the first detection result, the second detection result and the third detection result meet the first preset condition.

2. The sleep cognition enhancement system according to claim 1, characterized in that The sleep staging unit includes: a preprocessing module connected to the data acquisition unit to receive the SEEG signal and scalp EEG data of the user and perform data preprocessing; a feature extraction module connected to the preprocessing module to extract corresponding signal features in real time based on the preprocessed data; wherein the signal features include at least: time domain signal standard deviation, interquartile range, skewness, kurtosis, and power of multiple frequency band signals; The sleep recognition module is connected to the feature extraction module and is preset with a sleep staging model to receive the signal features and use the signal features of the previous period to predict the sleep stage at a predetermined time thereafter, thereby performing real-time sleep staging.

3. The sleep cognition enhancement system according to claim 1, characterized in that The first detection unit includes: a slow wave falling stage detection module, connected to the sleep staging unit, for detecting the slow wave falling stage in real time based on a preset first threshold after the sleep stage enters stage N3, and outputting a slow wave falling stage detection result; a slow wave rising phase detection module, connected to the slow wave falling phase detection module, configured to perform real-time detection of the slow wave rising phase based on a preset second threshold when the slow wave falling phase detection result is lower than the first threshold, and output the slow wave rising phase detection result; a slow wave length detection module, connected to the slow wave rising phase detection module, configured to detect the length of the slow wave when the detection result of the slow wave falling phase is lower than the first threshold and the detection result of the slow wave rising phase is higher than the second threshold, and output the slow wave length detection result; Among them, when the detection result of the slow wave descending phase is lower than the first threshold, the detection result of the slow wave rising phase is higher than the second threshold, and the detection result of the slow wave length is consistent with the frequency of the slow wave, then the first detection result is that the slow wave is detected; otherwise, the first detection result is that the slow wave is not detected.

4. The sleep cognition enhancement system according to claim 3, characterized in that The first detection unit further includes: A threshold updating module is connected to the slow wave rising phase detection module and the slow wave falling phase detection module to update the first threshold and the second threshold based on all slow wave events in the N3 sleep period within a preset time period.

5. The sleep cognition enhancement system according to claim 1, characterized in that The second detection unit includes: an RMS power detection module, connected to the sleep staging unit, configured to perform RMS power detection based on a third threshold after the sleep stage enters N2 and N3 stages, and output an RMS power detection result; wherein, when the RMS power detection result is higher than the third threshold and lasts for more than a preset time, it indicates that the center of the spindle wave has been detected; a power value detection module connected to the RMS power detection module to output a power value detection result; wherein, when the time derivative of the RMS power becomes zero and the RMS power begins to decrease, the maximum spindle power is detected; a spindle end detection module connected to the RMS power detection module to output a spindle end detection result; wherein the spindle end is detected when the RMS power detection result changes from being higher than the third threshold to being lower than the third threshold; a pause and resume detection module, connected to the RMS power detection module, the power value detection module, and the spindle end detection module, to pause or resume detection based on the RMS power detection result, the power value detection result, and the spindle end detection result; The phase detection module performs real-time phase detection on the spindle waves based on a preset algorithm to output the phase detection results; Among them, when the RMS power detection result, the power value detection result, the spindle wave end detection result and the phase detection result meet the second preset condition, the second detection result is that the spindle wave trough is detected; otherwise, the second detection result is that the spindle wave trough is not detected.

6. The sleep cognition enhancement system according to claim 1, characterized in that The third detection unit includes: a filtering module, connected to the sleep staging unit, for performing sequential bandpass filtering on the SEEG signal of the hippocampus after the sleep stage enters stage N3; a pulse conversion module connected to the filtering module to define a baseline amplitude that a sharp wave ripple event must exceed based on a time window of a preset length, and convert the electrophysiological signal into an upward pulse event / downward pulse event according to the amplitude strength and direction; a double-layer spiking neural network, connected to the spiking conversion module, to perform real-time detection of hippocampal sharp wave ripples based on the upward spiking event / downward spiking event, and output a third detection result; Among them, when the third preset condition is met, the third detection result is that hippocampal sharp wave ripples are detected; otherwise, the third detection result is that hippocampal sharp wave ripples are not detected.

7. The sleep cognition enhancement system according to claim 1, characterized in that The electrical stimulation unit comprises: a first stimulation portion corresponding to a first preset area, for electrically stimulating the first preset area based on preset stimulation parameters; a second stimulation portion corresponding to a second preset area, for electrically stimulating the second preset area based on preset stimulation parameters; a third stimulation portion corresponding to a third preset area, for electrically stimulating the third preset area based on preset stimulation parameters; Among them, the first preset area corresponds to the real-time detection of the rising phase of the slow wave, the second preset area corresponds to the real-time detection of the spindle wave trough, and the third preset area corresponds to the real-time detection of the hippocampal sharp wave ripples.

8. The sleep cognition enhancement system according to claim 7, wherein: The first preset area is the neocortex, the second preset area is the entorhinal cortex, and the third preset area is the white matter adjacent to the hippocampus CAI.

9. The sleep cognition enhancement system according to claim 7, wherein: The preset stimulation parameters are: 5 bidirectional pulse signals with a frequency of 100 Hz, a current intensity of 1.5 mA, a duration and interval time of two bidirectional pulses of 100 μs, and a duration of 50 ms for each stimulation.

10. A method for enhancing sleep cognition based on deep brain stimulation, characterized in that The steps include: Obtain the user's real-time SEEG signals and scalp EEG data; Based on the user's real-time SEEG signal and scalp EEG data, the user is subjected to real-time sleep staging, thereby outputting a real-time sleep staging result; Based on the real-time sleep staging result, detecting the slow wave rising stage of the user in real time, and outputting a first detection result; Based on the real-time sleep staging result, performing real-time detection of spindle wave troughs on the user, and outputting a second detection result; Based on the real-time sleep staging result, performing real-time detection of hippocampal sharp wave ripples on the user, and outputting a third detection result; If the first detection result, the second detection result, and the third detection result meet the first preset condition, then electrically stimulating the first preset area, the second preset area, and the third preset area through the electrical stimulation unit; If any one of the first detection result, the second detection result and the third detection result does not meet the first preset condition, the user's sleep stage is re-classified based on the user's real-time SEEG signal and scalp EEG data, and the test is performed again until the first detection result, the second detection result and the third detection result meet the first preset condition.

Images