System and method for improving sleep breathing disorder based on nerve stimulation
By activating the hypoglossal and phrenic nerves, combined with fixed-flow positive pressure support and closed-loop control using an LSTM model, the problem of synergistic regulation of airway expansion and respiratory drive, which is currently unattainable with existing devices, has been solved, thus achieving highly efficient treatment for sleep apnea.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHANDONG UNIV
- Filing Date
- 2026-03-04
- Publication Date
- 2026-04-21
AI Technical Summary
Existing equipment cannot achieve dual coordinated regulation of airway dilation and respiratory drive. The stimulation intensity lacks dynamic adaptation to respiratory cycle, sleep stage and body position changes. A closed-loop control network for multi-source biological signals has not been established, resulting in a mismatch between treatment intervention and physiological needs.
By activating the hypoglossal and phrenic nerves through a neural stimulation module, and combining this with a fixed-flow positive pressure support module, a closed-loop control network is established using an LSTM model to achieve a dual closed-loop control strategy. Parameters are dynamically adjusted to reduce timing control errors.
It achieves synergistic regulation of upper airway dilation and respiratory drive, reduces irritative microarousing, and improves the precision and effectiveness of treatment.
Smart Images
Figure CN121891713A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of medical technology, and in particular relates to a system and method for improving sleep-disordered breathing based on nerve stimulation. Background Technology
[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.
[0003] Sleep-disordered breathing (SDB) is a group of diseases characterized by abnormal breathing rhythms or ventilation during sleep. Obstructive sleep apnea-hypopnea syndrome (OSAHS) and central sleep apnea syndrome (CSAS) are the most common types. OSAHS is mainly caused by upper airway collapse and obstruction during sleep, leading to apnea and insufficient ventilation. Its incidence is higher in adults, increasing with age and obesity rates. CSAS, on the other hand, is caused by abnormal respiratory center regulation, resulting in impaired respiratory drive. These sleep-disordered breathing diseases not only severely affect patients' sleep quality, causing symptoms such as daytime sleepiness, fatigue, and poor concentration, but also lead to various complications in the long term, posing a significant threat to patients' physical health and quality of life.
[0004] Mechanical ventilation therapy using continuous positive airway pressure (CPAP) based on a face mask often results in complications such as nasal ulcers due to mask pressure, gastrointestinal bloating caused by airflow impact, and psychological resistance. Furthermore, CPAP has an improvement rate of less than 35% for central sleep apnea, and in mixed respiratory events, it may exacerbate bicarbonate buildup by interfering with spontaneous breathing rhythms due to positive pressure.
[0005] Implantable hypoglossal nerve stimulators (such as Inspire) are surgically implanted into the chest wall to generate a pulse generator. While this improves the upper airway opening rate, it can lead to other symptoms such as pneumothorax and surgical infection. More importantly, this technology only targets a single point of the hypoglossal nerve and cannot address the complex pathophysiological abnormalities present in most patients, such as delayed diaphragmatic contraction (mean lag time > 0.5 seconds) and paradoxical breathing. As a result, the residual index of central-obstructive mixed apnea remains as high as 18.7 ± 6.2 breaths / hour.
[0006] The single-channel stimulation mode of percutaneous hypoglossal nerve stimulation units often leads to a 0.3-0.8 second timing misalignment between pharyngeal expansion and thoracic respiratory movements, resulting in decreased effectiveness during dynamic regulation across sleep cycles. While transcervical pharyngeal nerve stimulation devices can increase tidal volume, they completely fail to prevent pharyngeal collapse (blood oxygen saturation still drops below 82%). Furthermore, existing devices all employ open-loop control strategies, resulting in a synchronization error of 23.5±7.4% between stimulation intensity and respiratory phase (inspiratory / expiratory phase), causing most patients to experience stimulative micro-arousals, which in turn disrupts sleep continuity.
[0007] In summary, the existing technology has the following main drawbacks: (1) Existing equipment either focuses on upper airway mechanical dilation (CPAP) or only stimulates a single nerve target (hypoglossal / phrenic nerve), which cannot achieve dual synergistic regulation of "airway dilation-breathing drive"; (2) The lack of millisecond-level dynamic adaptation between stimulation intensity and respiratory cycle, sleep stage and body position changes leads to a mismatch between treatment intervention and physiological needs; (3) Positive pressure ventilation and electrical stimulation operate independently, and a closed-loop control network based on multi-source biological signals (snoring, blood oxygen, chest and abdominal movements, etc.) has not been established. Summary of the Invention
[0008] To overcome the shortcomings of the prior art, this invention provides a system and method for improving sleep-disordered breathing based on nerve stimulation. The system activates the hypoglossal nerve and phrenic nerve through a nerve stimulation module, thereby enhancing tongue muscle tone and diaphragmatic contraction force. A fixed-flow positive pressure support module provides continuous positive airway pressure ventilation. The control unit module establishes a closed-loop control network based on multi-source biological signals, identifies respiratory event types through an LSTM model, and dynamically adjusts parameters using a dual closed-loop control strategy to reduce timing control errors and achieve multi-signal synchronization.
[0009] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions: The first aspect of this invention provides a system for improving sleep-disordered breathing based on nerve stimulation; A system for improving sleep-disordered breathing based on neural stimulation includes: A face mask is used to collect nasopharyngeal airflow, snoring, and vibration signals; The nerve stimulation module includes: a hypoglossal nerve stimulation unit, which is composed of a flexible biocompatible electrode array; and a phrenic nerve stimulation unit, which is composed of a ring bipolar electrode. A fixed-flow positive pressure support module, comprising a pressure regulating unit and a positive pressure valve, for providing continuous positive airway ventilation; A multi-source data acquisition module, connected to a control unit module, includes a chest and abdominal motion sensor and a blood oxygen sensor, used to acquire respiratory motion amplitude data and blood oxygen saturation data; The control unit module includes a data processing core and a dual closed-loop controller. The data processing core identifies respiratory event types based on an LSTM model and establishes a stimulus-ventilation response matrix. The dual closed-loop controller is used to adjust the stimulation intensity of the neural stimulation module in real time.
[0010] As a further technical solution, the electrode array of the hypoglossal nerve stimulation unit includes 16 contacts with a thickness of ≤0.3mm. The stimulation parameters are frequency 20-50Hz, pulse width 200-500μs, and current intensity 5-30mA. Automatic targeting and positioning are achieved based on impedance feedback.
[0011] As a further technical solution, the annular bipolar electrode of the phrenic nerve stimulation unit has an outer diameter of 25mm, a contact resistance of <1kΩ, a stimulation trigger delay of <30ms, and is matched with the respiratory phase in real time.
[0012] As a further technical solution, the pressure regulating unit is a fixed high-flow gas with a continuously adjustable flow rate of 4-20 cm H2O. It supports small-range dynamic compensation with a flow rate accuracy of ±0.5 cm H2O, automatic gradient compensation based on 0.2 cm H2O / 100 mL tidal volume, and dynamic adjustment of the proportional valve opening based on real-time airway resistance to ensure that the target pressure deviation is less than 0.3 cm H2O.
[0013] As a further technical solution, the positive pressure valve is set to 80-100% of the set pressure to ensure that the expiratory resistance is ≥2 cm H2O / L / s. When a signal of diaphragmatic relaxation at the end of expiration is detected, the opening range of the positive pressure valve increases to 120%. When the pressure exceeds 25 cm H2O, it is forcibly opened to release pressure and prevent the risk of overpressure.
[0014] As a further technical solution, the LSTM model identifies respiratory event types through a three-level judgment mechanism, including primary screening, intermediate classification, and advanced verification.
[0015] As a further technical solution, the dual closed-loop controller is used to adjust the stimulation intensity of the nerve stimulation module in real time, including inner loop control and outer loop control, wherein the inner loop control takes precedence over the outer loop control, positive pressure ventilation is applied first, and after the airway is opened, the hypoglossal nerve stimulation is delayed.
[0016] A second aspect of the present invention provides a method for improving sleep-disordered breathing based on nerve stimulation.
[0017] A method for improving sleep-disordered breathing based on neural stimulation includes: Data on chest and abdominal movement amplitude, nasopharyngeal airflow, and blood oxygen saturation were collected and preprocessed. The preprocessed signal is input into the LSTM classification model to extract time-domain and frequency-domain features. The three-level judgment mechanism outputs the judgment result and confidence level of obstructive or central sleep apnea. Based on the judgment results and confidence level, a dual closed-loop controller is used to adjust the stimulation intensity and positive pressure ventilation parameters in real time to achieve coordinated regulation of airway expansion and respiratory drive.
[0018] As a further technical solution, the three-level judgment mechanism includes primary screening, intermediate classification, and advanced verification.
[0019] As a further technical solution, the dual closed-loop controller is used to adjust the stimulation intensity of the nerve stimulation module in real time, including inner loop control and outer loop control, wherein the inner loop control takes precedence over the outer loop control, positive pressure ventilation is applied first, and after the airway is opened, the hypoglossal nerve stimulation is delayed.
[0020] The above one or more technical solutions have the following beneficial effects: (1) This invention achieves coordinated regulation of upper airway expansion and respiratory drive by simultaneously stimulating the hypoglossal nerve and the phrenic nerve through a nerve stimulation module. Hypoglossal nerve stimulation enhances tongue muscle tone, prevents the tongue from falling back, and maintains upper airway patency; phrenic nerve stimulation enhances diaphragmatic muscle contraction, improving respiratory depth and efficiency. This dual stimulation mechanism effectively overcomes the limitations of single nerve stimulation.
[0021] (2) The present invention ensures precise matching of nerve stimulation and respiratory phase through millisecond-level synchronous control, avoids mismatch between stimulation and physiological needs, and reduces the occurrence of stimuli-induced micro-arousals.
[0022] (3) This invention establishes a closed-loop control network based on multi-source biological signals (such as snoring, blood oxygenation, chest and abdominal movements, etc.) to monitor and adjust parameters in real time. The LSTM model is used to identify respiratory event types, and the stimulation intensity and positive pressure ventilation parameters are dynamically adjusted according to the event type to ensure the accuracy and effectiveness of the intervention.
[0023] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description
[0024] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.
[0025] Figure 1 This is a schematic diagram of a system structure for improving sleep-disordered breathing based on neural stimulation in the first embodiment.
[0026] Figure 2 This is a flowchart of the method in the second embodiment.
[0027] In the diagram, 1 is the face mask, 2 is the hypoglossal nerve stimulation unit, 3 is the phrenic nerve stimulation unit, 4 is the fixed flow positive pressure support module, 5 is the multi-source data acquisition module, and 6 is the control unit module. Detailed Implementation
[0028] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.
[0029] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations of the present invention.
[0030] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.
[0031] Example 1 This embodiment discloses a system for improving sleep-disordered breathing based on nerve stimulation; like Figure 1 As shown, a system for improving sleep-disordered breathing based on neural stimulation includes: Mask 1, which is used to collect nasopharyngeal airflow, snoring and vibration signals; The nerve stimulation module includes a hypoglossal nerve stimulation unit 2 and a phrenic nerve stimulation unit 3, which are used to stimulate the hypoglossal nerve and the phrenic nerve with electrical pulses.
[0032] The hypoglossal nerve stimulation unit 2 consists of a flexible biocompatible electrode array containing 16 contacts with a thickness ≤0.3 mm. The stimulation parameters are a frequency of 20-50 Hz, a pulse width of 200-500 μs, and a current intensity of 5-30 mA. Based on impedance feedback automatic targeting technology, the positioning error is <1.5 mm.
[0033] The phrenic nerve stimulation unit 3 is composed of a ring bipolar electrode with an outer diameter of 25 mm, a contact resistance of <1kΩ, a stimulation trigger delay of <30ms, and real-time matching with the respiratory phase.
[0034] The fixed flow positive pressure support module 4 includes a pressure regulating unit and a positive pressure valve, which are used to provide continuous positive airway ventilation. The pressure regulating unit is for a fixed high flow rate gas, continuously adjustable from 4-20 cm H2O, supports dynamic compensation of the flow rate within a small range (±0.5 cm H2O accuracy), supports automatic gradient compensation based on tidal volume (0.2 cm H2O / 100mL), and dynamically adjusts the opening of the proportional valve based on real-time airway resistance (feedback from a differential pressure sensor) to ensure that the target pressure deviation is <0.3 cm H2O.
[0035] The positive pressure valve is set to 80-100% of the set pressure to ensure that the expiratory resistance is ≥2 cm H2O / L / s. When a diaphragm relaxation signal is detected at the end of expiration, the opening of the positive pressure valve increases to 120%. When the pressure exceeds 25 cm H2O, it is forcibly opened to release pressure (response time <50ms) to prevent the risk of system overpressure.
[0036] Multi-source data acquisition module 5, which is connected to control unit module 6, includes a chest and abdominal motion sensor and a blood oxygen sensor, used to acquire respiratory motion amplitude data and blood oxygen saturation data; Control unit module 6 includes a data processing core and a dual closed-loop controller; the data processing core identifies respiratory event types based on an LSTM model and establishes a stimulus-ventilation response relationship matrix. The LSTM model first acquires chest and abdominal movement amplitude, nasopharyngeal airflow, and blood oxygen saturation based on MEMS sensors, differential pressure sensors, and SpO2 sensors, and then performs preprocessing. The preprocessed signal is input into the LSTM model to extract time-domain and frequency-domain features. The model outputs the determination result and confidence level of obstructive or central sleep apnea through a three-level judgment mechanism, which includes primary screening, intermediate classification and advanced validation.
[0037] Primary screening: breathing interval > 10 seconds and airflow < 15% of baseline.
[0038] Intermediate classification: Distinguish between obstructive and central types using an LSTM model (confidence > 85%).
[0039] Advanced validation: combining esophageal pressure fluctuations with diaphragmatic electromyographic signals.
[0040] Based on the judgment results and confidence levels, a dual closed-loop controller is used to adjust the stimulation intensity and positive pressure ventilation parameters in real time to achieve coordinated regulation of airway expansion and respiratory drive. The dual closed-loop controller is used to adjust the stimulation intensity of the neurostimulation module in real time, including inner-loop control and outer-loop control. The inner-loop control is used for constant pressure and flow, controlling the flow rate, which remains constant with respiration, with a slight pressure increase of 2-6 cm H2O. The outer-loop control increases the stimulation intensity based on the SpO2 decrease rate (ΔSpO2 / Δt > 3% / min). Furthermore, the inner-loop control takes precedence over the outer-loop control, applying positive pressure ventilation first, and then delaying the initiation of hypoglossal nerve stimulation after airway opening.
[0041] The specific working process includes: after the patient wears the device, the system automatically performs electrode impedance detection (takes 20 seconds) to confirm that the impedance of the hypoglossal nerve stimulation area is less than 5kΩ and the impedance of the phrenic nerve area is less than 1.2kΩ.
[0042] The turbine system was started and pre-pressurized, and the air leakage from the mask was controlled within 5 L / min.
[0043] During the obstructive event detection process, when nasopharyngeal airflow drops to 12% of the baseline value and there is paradoxical breathing (phase difference greater than 90°) with chest and abdominal movements, the LSTM model outputs an obstructive probability of 92%, triggering a synergistic treatment strategy.
[0044] When the nasopharyngeal airflow is interrupted for 12 seconds, the amplitude of chest and abdominal movement is less than 5% of the baseline value, the electromyographic signal of the diaphragm disappears, the LSTM model outputs a central probability of 88%, and the central intervention mode is activated.
[0045] At this point, dynamic parameter adjustments are made, increasing the nerve stimulation intensity of the hypoglossal nerve stimulation unit to 20mA, raising the positive pressure ventilation to 12cm H2O, and triggering the nerve stimulation of the phrenic nerve stimulation unit 50 ms before the inspiratory phase. The upper airway opens within 0.8 seconds after stimulation (snoring disappears as detected by a vibration sensor).
[0046] Hypoglossal nerve stimulation was maintained at baseline intensity (10mA), positive pressure ventilation was increased to 10cm H2O, and phrenic nerve stimulation was triggered 100 ms at the end of expiration (preemptively activating the next inspiratory phase). As SpO2 recovered from 82% to 94%, the stimulation intensity was gradually reduced to baseline level to avoid overstimulation leading to microarousal. Simultaneously, based on body position sensor data (supine position → lateral position), the hypoglossal nerve stimulation frequency was automatically adjusted from 35 Hz to 25 Hz (reducing excessive tongue muscle contraction during lateral positioning).
[0047] Example 2 This embodiment discloses a method for improving sleep-disordered breathing based on nerve stimulation; like Figure 2 As shown, a method for improving sleep-disordered breathing based on neural stimulation includes: Step S1: Collect data on chest and abdominal movement amplitude, nasopharyngeal airflow, and blood oxygen saturation, and perform preprocessing. Step S2: Input the preprocessed signal into the LSTM classification model, extract time domain features and frequency domain features, and output the judgment result and confidence level of obstructive or central sleep apnea through a three-level judgment mechanism. In step S2, the preprocessed polysomnography signal is input into an LSTM classification model. This model employs a bidirectional network architecture, extracting the temporal features (including amplitude variance, zero-crossing rate, peak-to-trough interval, etc.) and frequency domain features (0.1-0.3Hz fundamental respiratory frequency components and their harmonic distribution) of the respiratory waveform signal in parallel through temporal convolutional layers. A three-level judgment mechanism is executed sequentially: Primary screening: respiratory interval > 10 seconds and airflow < 15% of baseline. Intermediate classification: LSTM model distinguishes between obstructive and central respiratory symptoms (confidence > 85%). Advanced validation: Combining esophageal pressure fluctuations with diaphragmatic electromyography signals.
[0048] Step S3: Based on the judgment result and confidence level, a dual closed-loop controller is used to adjust the stimulation intensity and positive pressure ventilation parameters in real time to achieve coordinated regulation of airway expansion and respiratory drive.
[0049] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.
Claims
1. A system for improving sleep-disordered breathing based on neural stimulation, characterized in that, include: A face mask is used to collect nasopharyngeal airflow, snoring, and vibration signals; The nerve stimulation module includes: a hypoglossal nerve stimulation unit, which is composed of a flexible biocompatible electrode array; and a phrenic nerve stimulation unit, which is composed of a ring bipolar electrode. A fixed-flow positive pressure support module, comprising a pressure regulating unit and a positive pressure valve, for providing continuous positive airway ventilation; A multi-source data acquisition module, connected to a control unit module, includes a chest and abdominal motion sensor and a blood oxygen sensor, used to acquire respiratory motion amplitude data and blood oxygen saturation data; The control unit module includes a data processing core and a dual closed-loop controller. The data processing core identifies respiratory event types based on an LSTM model and establishes a stimulus-ventilation response matrix. The dual closed-loop controller is used to adjust the stimulation intensity of the neural stimulation module in real time.
2. The system for improving sleep-disordered breathing based on nerve stimulation as described in claim 1, characterized in that, The electrode array of the hypoglossal nerve stimulation unit contains 16 contacts with a thickness of ≤0.3mm. The stimulation parameters are frequency 20-50Hz, pulse width 200-500μs, and current intensity 5-30mA. Automatic targeting and positioning are achieved based on impedance feedback, with a positioning error of <1.5mm.
3. The system for improving sleep-disordered breathing based on nerve stimulation as described in claim 1, characterized in that, The annular bipolar electrode of the phrenic nerve stimulation unit has an outer diameter of 25 mm, a contact resistance of <1 kΩ, a stimulation trigger delay of <30 ms, and is matched with the respiratory phase in real time.
4. The system for improving sleep-disordered breathing based on nerve stimulation as described in claim 1, characterized in that, The pressure regulating unit is for a fixed high-flow-rate gas, continuously adjustable from 4-20 cm H2O, supports small-range dynamic compensation with a flow rate accuracy of ±0.5 cm H2O, supports automatic gradient compensation based on 0.2 cm H2O / 100 mL tidal volume, and dynamically adjusts the proportional valve opening based on real-time airway resistance to ensure that the target pressure deviation is less than 0.3 cm H2O.
5. The system for improving sleep-disordered breathing based on neural stimulation as described in claim 1, characterized in that, The positive pressure valve is set to 80-100% of the set pressure to ensure that the expiratory resistance is ≥2 cm H2O / L / s. When a signal of diaphragmatic relaxation at the end of expiration is detected, the opening range of the positive pressure valve increases to 120%. When the pressure exceeds 25 cm H2O, it is forcibly opened to release pressure and prevent the risk of overpressure.
6. The system for improving sleep-disordered breathing based on neural stimulation as described in claim 1, characterized in that, The LSTM model identifies respiratory event types through a three-level judgment mechanism, including primary screening, intermediate classification, and advanced validation.
7. The system for improving sleep-disordered breathing based on neural stimulation as described in claim 1, characterized in that, The dual closed-loop controller is used to adjust the stimulation intensity of the nerve stimulation module in real time, including inner loop control and outer loop control. The inner loop control takes precedence over the outer loop control. Positive pressure ventilation is applied first, and after the airway is opened, hypoglossal nerve stimulation is started with a delay.
8. A method for improving sleep-disordered breathing based on neural stimulation, employing the system for improving sleep-disordered breathing based on neural stimulation as described in any one of claims 1-7, characterized in that, include: Data on chest and abdominal movement amplitude, nasopharyngeal airflow, and blood oxygen saturation were collected and preprocessed. The preprocessed signal is input into the LSTM classification model to extract time-domain and frequency-domain features. The three-level judgment mechanism outputs the judgment result and confidence level of obstructive or central sleep apnea. Based on the judgment results and confidence level, a dual closed-loop controller is used to adjust the stimulation intensity and positive pressure ventilation parameters in real time to achieve coordinated regulation of airway expansion and respiratory drive.
9. A method for improving sleep-disordered breathing based on nerve stimulation as described in claim 8, characterized in that, The three-level judgment mechanism includes primary screening, intermediate classification, and advanced verification.
10. A method for improving sleep-disordered breathing based on nerve stimulation as described in claim 8, characterized in that, The dual closed-loop controller is used to adjust the stimulation intensity of the nerve stimulation module in real time, including inner loop control and outer loop control. The inner loop control takes precedence over the outer loop control. Positive pressure ventilation is applied first, and after the airway is opened, hypoglossal nerve stimulation is started with a delay.