Impedance sensing-based neurosurgery postoperative aspiration risk grading detection head band and system thereof

Abstract

The invention provides a neurosurgery postoperative aspiration risk grading detection head band and a neurosurgery postoperative aspiration risk grading detection system based on impedance sensing, and relates to the technical field of medical detection, the neurosurgery postoperative aspiration risk grading detection head band comprises a band body, a flexible electrode array, a signal processing unit and a wireless transmission chip, the flexible electrode array is integrated on the inner side of the band body and is used for collecting biological impedance signals of anterior cervical muscle groups; the signal processing unit is electrically connected with the flexible electrode array and is used for preprocessing and demodulating the impedance signal; the wireless transmission chip is connected with the signal processing unit and is used for uploading the processed signal to the cloud control platform; the flexible electrode array is integrated on the belt body, non-invasive and comfortable continuous monitoring is achieved, the electrodes are attached to the curved surface of the anterior cervical muscle group, stimulation of traditional invasive electrodes to the skin is avoided, the belt body is designed through elastic spandex, use is convenient, the device adapts to the head circumference of different patients, and the comfort of long-term wearing is ensured.

Description

Technical Field

[0001] This invention relates to the field of medical testing technology, and in particular to a headband and system for assessing the risk of postoperative aspiration in neurosurgery based on impedance sensing. Background Technology

[0002] Postoperative neurosurgical patients often have a significantly increased risk of aspiration due to swallowing dysfunction. Traditional detection methods, such as methylene blue staining, have low sensitivity and serious side effects and have been phased out. Although fiber optic endoscopic swallowing assessment and video fluoroscopic swallowing studies are the standard, they require specialized equipment and are invasive procedures, making it difficult to monitor continuously in the early postoperative period. Pepsin testing is costly and complicated, and cannot provide immediate feedback, limiting its clinical application. Current technologies can only identify aspiration after it has occurred and cannot assess the risk before or in the early postoperative period.

[0003] Bioimpedance sensing technology has the advantages of being non-invasive and real-time in muscle function assessment. For example, time-domain impedance plethysmography can characterize the microscopic equivalent circuit parameters of cervical neuromuscular tissue, and time-domain impedance imaging can reflect the macroscopic spatial distribution of swallowing organs. However, most existing impedance detection devices are fixed and not integrated into portable tools, making them inconvenient to use. In addition, existing technologies lack the function of combining impedance changes with aspiration risk grading algorithms, making it difficult to achieve personalized early warning. Therefore, this invention proposes a headband and system for detecting aspiration risk grading after neurosurgery based on impedance sensing to solve the problems existing in the prior art. Summary of the Invention

[0004] To address the aforementioned issues, this invention proposes a headband and system for assessing the risk of postoperative aspiration in neurosurgery based on impedance sensing. By integrating a flexible electrode array into the headband, non-invasive and comfortable continuous monitoring is achieved. The electrodes conform to the curved surface of the anterior cervical muscles, avoiding skin irritation caused by traditional invasive electrodes. The headband is made of elastic spandex, making it easy to use and adaptable to different patient head circumferences, ensuring long-term comfort.

[0005] To achieve the objectives of this invention, the invention is implemented through the following technical solution: a headband for detecting the risk grading of postoperative aspiration in neurosurgery based on impedance sensing, comprising a headband body, a flexible electrode array, a signal processing unit, and a wireless transmission chip. The flexible electrode array is integrated inside the headband body and is used to collect bioimpedance signals from the anterior neck muscles. The signal processing unit is electrically connected to the flexible electrode array and is used to preprocess and demodulate the impedance signals.

[0006] The wireless transmission chip is connected to the signal processing unit and is used to upload the processed signal to the cloud control platform.

[0007] Further improvements include: the flexible electrode array consists of a matrix of 15-30 graphene electrodes with an electrode spacing of 5-10 mm; the flexible electrode array covers key muscle groups; the contact impedance between the flexible electrode array and the skin is ≤5kΩ; it employs a conductive gel or dry electrode structure; and the surface of the dry electrode has a microneedle structure.

[0008] Further improvements include: the belt body is made of elastic spandex material, and the surface of the belt body is coated with a silver ion antibacterial layer; the breathability of the belt body is ≥2000g / (m²). 2 • 24h), tensile strength ≥ 50 N / cm.

[0009] A further improvement is that the signal processing unit preprocesses the impedance signal, including signal quality assessment, using the following formula:

[0010]

[0011] Wherein, SQ is the signal quality score (0-1, ≥0.6 is a valid signal), SNR is the impedance signal-to-noise ratio (dB, calculated by the ratio of signal peak value to noise mean), and BD is the baseline drift (Ω, the standard deviation of impedance fluctuation during resting period).

[0012] Further improvements include: the signal processing unit comprises a low-pass filter module, an analog-to-digital converter module, and an amplification module; the low-pass filter module is used to filter out 50Hz power line noise; the analog-to-digital converter module has a sampling frequency of 1000-2000Hz; and the amplification module is used to calculate the impedance amplitude and phase, with the formula for calculating the impedance amplitude being:

[0013]

[0014] Where |Z| is the impedance magnitude (Ω), Z real Z represents the real part of the impedance (Ω, reflecting the tissue's electrical resistance). imag This represents the imaginary part of the impedance (Ω, reflecting the capacitance characteristics of the tissue).

[0015] The neurosurgical postoperative aspiration risk grading detection system based on impedance sensing includes a cloud control platform and terminal devices. The cloud control platform is used to receive, store, and analyze impedance data uploaded by the wireless transmission chip. The cloud control platform has a built-in risk grading algorithm for real-time assessment of the risk level of the impedance data.

[0016] The terminal device communicates with the cloud control platform to display risk levels and early warning information.

[0017] Further improvements include: the cloud control platform also includes a continuous wavelet transform module and a neural network. The continuous wavelet transform module uses the db5 wavelet basis to generate a two-dimensional scale map based on impedance data. The neural network takes the scale map as input and outputs the swallowing event type based on the built-in CNN-BiLSTM deep learning model. The cloud control platform interfaces with the hospital information system to automatically synchronize patient clinical data and generate records.

[0018] A further improvement is made in that the risk grading algorithm includes the following formula:

[0019]

[0020] Where ΔZ is the impedance change (Ω), t is the duration of the swallowing action (seconds), Weight is the weighting coefficient (0.5-1.5), Offset is the offset (0.1-0.3), and the Risk value ranges from 0 to 1, with a higher value indicating a higher risk.

[0021] The formula for calculating the duration t of the swallowing action is:

[0022] t = tend - tstart

[0023] Where t is the duration of the swallowing action (seconds), tstart is the start time of the swallowing action (seconds, the moment when the impedance signal first exceeds twice the resting mean of the mean), and tend is the end time of the swallowing action (seconds, the moment when the impedance signal returns to the resting mean ± one standard deviation).

[0024] The risk grading algorithm also includes a dynamic threshold adjustment formula to update the warning thresholds T1 (high risk) and T2 (medium risk):

[0025] Tnew = Told × (1 + 0.1 × ErrRate)

[0026] Where Tnew is the updated threshold, Told is the initial threshold: T1 = 0.7, T2 = 0.4, and ErrRate is the historical risk assessment error rate: 0-0.5, calculated using clinical validation data.

[0027] A further improvement is made in the risk grading algorithm, where ΔZ is calculated using the following formula:

[0028] ΔZ = |Z swallowing period - Z resting period|

[0029] Among them, the Z-swallowing period is the impedance value when the swallowing action occurs, and the Z-resting period is the average impedance value in the 30 seconds before swallowing.

[0030] Further improvements include: the terminal device includes a mobile APP and a mobile tablet, which is used for risk level display, historical data query and warning threshold setting, and provides warnings and prompts based on the set warning threshold.

[0031] The beneficial effects of this invention are as follows:

[0032] 1. This invention integrates a flexible electrode array into the belt body to achieve non-invasive and comfortable continuous monitoring. The electrodes conform to the curved surface of the anterior cervical muscle group, avoiding the skin irritation caused by traditional invasive electrodes. The belt body is made of elastic spandex, which is convenient to use, adaptable to different patient head circumferences, and ensures long-term wearing comfort.

[0033] 2. This invention monitors the impedance changes of the anterior cervical muscle group in real time and uses a grading algorithm to realize aspiration risk warning. By detecting the impedance change rate and time parameters during swallowing, the risk level can be calculated and compared with a preset threshold to issue an early warning, which is more reliable.

[0034] 3. This invention integrates a cloud control platform and terminal devices to achieve multi-dimensional data analysis. The impedance signal collected by the headband is processed by a demodulation algorithm and then uploaded to the cloud control platform for storage and analysis. The cloud control platform has a built-in CNN-BiLSTM deep learning model for risk assessment. The terminal devices support remote viewing of real-time data and historical trends, which facilitates medical staff to dynamically adjust nursing strategies. Attached Figure Description

[0035] Figure 1 This is the front view of the present invention;

[0036] Figure 2 This is a schematic diagram of the system of the present invention. Detailed Implementation

[0037] To enhance understanding of the present invention, the present invention will be further described in detail below with reference to embodiments. These embodiments are only used to explain the present invention and do not constitute a limitation on the scope of protection of the present invention.

[0038] Example 1

[0039] according to Figure 1 , 2 As shown, this embodiment proposes a headband for detecting the risk of aspiration after neurosurgery based on impedance sensing, including a headband body, a flexible electrode array, a signal processing unit, and a wireless transmission chip. The flexible electrode array is integrated inside the headband body and is used to collect bioimpedance signals from the anterior neck muscles. The signal processing unit is electrically connected to the flexible electrode array and is used to preprocess and demodulate the impedance signals.

[0040] The wireless transmission chip is connected to the signal processing unit to upload the processed signal to the cloud control platform. When the headband is worn, the flexible electrode array conforms to the anterior neck muscles, and the signal processing unit collects impedance signals in real time. After low-pass filtering and analog-to-digital conversion, the amplitude and phase are calculated by the amplification module. The data is uploaded to the cloud control platform via the wireless transmission chip, where it undergoes continuous wavelet transform to generate a scaling map, which is then input into a neural network to identify the swallowing event type. The risk grading algorithm combines the impedance change rate (ΔZ), time parameter (t), and clinical data to calculate the Risk value and compare it with a threshold, issuing an alert through the terminal device.

[0041] The flexible electrode array consists of 15-30 graphene electrodes arranged in a matrix, with an electrode spacing of 5-10 mm, covering key muscle groups. The contact impedance between the flexible electrode array and the skin is ≤5kΩ. It employs a conductive gel or dry electrode structure, with the dry electrode surface featuring a microneedle structure. The band is made of elastic spandex and coated with a silver ion antibacterial layer. The breathability of the band is ≥2000g / (m²). 2 • 24h), tensile strength ≥ 50 N / cm. The signal processing unit includes a low-pass filter module, an analog-to-digital converter module, and an amplification module. The low-pass filter module is used to filter out 50Hz power line noise; the analog-to-digital converter module has a sampling frequency of 1000-2000Hz; and the amplification module is used to calculate impedance amplitude and phase.

[0042] The signal processing unit preprocesses the impedance signal, including signal quality assessment, using the following formula:

[0043]

[0044] Wherein, SQ is the signal quality score (0-1, ≥0.6 is a valid signal), SNR is the impedance signal-to-noise ratio (dB, calculated by the ratio of signal peak value to noise mean), and BD is the baseline drift (Ω, the standard deviation of impedance fluctuation during resting period).

[0045] The formula for calculating impedance magnitude is:

[0046]

[0047] Where |Z| is the impedance magnitude (Ω), Z real Z represents the real part of the impedance (Ω, reflecting the tissue's electrical resistance). imag This represents the imaginary part of the impedance (Ω, reflecting the capacitance characteristics of the tissue).

[0048] The neurosurgical postoperative aspiration risk grading detection system based on impedance sensing includes a cloud control platform and terminal devices. The cloud control platform is used to receive, store, and analyze impedance data uploaded by the wireless transmission chip. The cloud control platform has a built-in risk grading algorithm for real-time assessment of the risk level of the impedance data.

[0049] The terminal device communicates with the cloud control platform to display risk levels and early warning information. The impedance signal collected by the headband is processed by a demodulation algorithm and then uploaded to the cloud control platform for storage and analysis. The cloud control platform has a built-in CNN-BiLSTM deep learning model for risk assessment. The terminal device supports remote viewing of real-time data and historical trends, facilitating medical staff to dynamically adjust nursing strategies.

[0050] The cloud-based control platform also includes a continuous wavelet transform module and a neural network. The continuous wavelet transform module uses the db5 wavelet basis to generate a two-dimensional scale map based on impedance data. The neural network takes the scale map as input and outputs the swallowing event type based on the built-in CNN-BiLSTM deep learning model. The cloud-based control platform interfaces with the hospital information system to automatically synchronize patient clinical data and generate records.

[0051] The risk grading algorithm includes the following formula:

[0052]

[0053] Where ΔZ is the impedance change (Ω), t is the duration of the swallowing action (seconds), Weight is the weighting coefficient (0.5-1.5), Offset is the offset (0.1-0.3), and the Risk value ranges from 0 to 1, with a higher value indicating a higher risk.

[0054] The formula for calculating the duration t of the swallowing action is:

[0055] t = tend - tstart

[0056] Where t is the duration of the swallowing action (seconds), tstart is the start time of the swallowing action (seconds, the moment when the impedance signal first exceeds twice the resting mean of the mean), and tend is the end time of the swallowing action (seconds, the moment when the impedance signal returns to the resting mean ± one standard deviation).

[0057] The risk grading algorithm also includes a dynamic threshold adjustment formula to update the warning thresholds T1 (high risk) and T2 (medium risk):

[0058] Tnew = Told × (1 + 0.1 × ErrRate)

[0059] Where Tnew is the updated threshold, Told is the initial threshold: T1 = 0.7, T2 = 0.4, and ErrRate is the historical risk assessment error rate: 0-0.5, calculated using clinical validation data.

[0060] In the risk grading algorithm, ΔZ is calculated using the following formula:

[0061] ΔZ = |Z swallowing period - Z resting period|

[0062] Among them, the Z-swallowing period is the impedance value when the swallowing action occurs, and the Z-resting period is the average impedance value in the 30 seconds before swallowing.

[0063] The terminal devices include mobile apps and tablets. These devices are used for displaying risk levels, querying historical data, and setting warning thresholds. Warnings and alerts are issued based on the set warning thresholds.

[0064] Example 2

[0065] according to Figure 1 , 2 As shown in the figure, this embodiment proposes a headband and system for assessing the risk of aspiration after neurosurgery based on impedance sensing. For a 58-year-old female patient who underwent acoustic neuroma resection, the swallowing reflex was weakened on the first day after surgery. This invention was used for short-term monitoring for 48 hours to assess the risk of early feeding.

[0066] Headband configuration: It adopts an 8-electrode array (1×8 arrangement) to cover the key area of ​​the cricothyroid muscle, with an 8mm electrode spacing, dry electrode design (no conductive gel required), and a contact resistance of 4.5kΩ; the band width is 6cm, and the lightweight design is suitable for short-term wear.

[0067] Wearing procedure: The patient should wear the device while lying flat. After the Velcro is fastened, the impedance reference value will be automatically calibrated within 30 seconds.

[0068] Signal acquisition: The ADC sampling frequency is reduced to 1000Hz, and only the impedance amplitude change is monitored. The resting period reference value is updated every 5 seconds.

[0069] Risk calculation: ΔZ = |Z swallowing period - Z resting period|, where Z swallowing period is the peak impedance during eating;

[0070] The algorithm parameters are simplified to t = 2 seconds, Weight = 1.0, and Offset = 0.15.

[0071] On the second day after surgery, during lunchtime, while the patient was eating rice cereal, the system detected ΔZ = 15Ω, and calculated Risk = (15 / 2) × 1.0 + 0.15 = 7.65, triggering a medium-risk warning. On-site assessment by medical staff revealed delayed swallowing, and the patient was immediately switched to small, frequent feedings. Subsequent monitoring showed that the Risk value stabilized within the 0.3-0.5 range.

[0072] Results: 63 swallowing events were monitored within 48 hours, with 9 medium-risk warnings and no high-risk events. Patients had a VAS comfort score of 9.0 and no obvious foreign body sensation. Compared with traditional bedside observation, nursing manpower was reduced by 40%, and no aspiration-related complications occurred.

[0073] Example 3

[0074] according to Figure 1 , 2 As shown in the figure, this embodiment proposes a headband and system for assessing the risk of aspiration after neurosurgery based on impedance sensing. For a 72-year-old patient (postoperative hypertensive intracerebral hemorrhage), who was conscious but had difficulty swallowing on the first day after surgery, this system was used for 24-hour simple monitoring to quickly assess the risk of the first oral feeding.

[0075] Headband configuration: It adopts a 4-electrode array (1×4 arrangement), which only covers the core area of ​​the cricothyroid muscle. The electrode spacing is 10mm. The surface of the dry electrode is decorated with micro-convex texture to enhance contact. The contact resistance is ≤5kΩ. The width of the band is 4cm and the weight is ≤20g, which is suitable for the fragile skin of elderly patients.

[0076] Wearing procedure: Nursing staff assist the patient to sit up, wrap the headband around the neck and start the device with one button. The device will complete self-check and baseline value initialization within 10 seconds.

[0077] Signal acquisition: The ADC sampling frequency is reduced to 500Hz, and only the impedance amplitude core parameter is retained. The resting period reference value (Z resting period) is updated every 10 seconds.

[0078] Risk calculation: ΔZ = |Z swallowing period - Z resting period| (Z swallowing period is the maximum resistance value of swallowing action);

[0079] Simplified algorithm parameter settings: t = 2.5 seconds, Weight = 0.8 (basic weight for elderly patients), Offset = 0.1.

[0080] On the first day after surgery, when attempting to feed water at dinner, the system detected ΔZ = 12Ω, and calculated Risk = (12 / 2.5) × 0.8 + 0.1 = 3.94, triggering a medium-risk warning. Medical staff immediately stopped feeding water and instead tried feeding thick porridge. Monitoring showed that ΔZ dropped to 8Ω, and Risk = (8 / 2.5) × 0.8 + 0.1 = 2.62, changing the risk level to low risk.

[0081] Results: 32 swallowing events were monitored within 24 hours, with 3 medium-risk warnings, all of which were promptly addressed; the patient's comfort VAS score was 8.8, with no skin indentations or complaints of discomfort; the time required for the first feeding assessment was reduced by 75% compared to traditional methods, and no aspiration events occurred.

[0082] Example 4

[0083] according to Figure 1 , 2 As shown in the figure, this embodiment proposes a headband and system for detecting the risk of aspiration after neurosurgery based on impedance sensing. For a 50-year-old male patient (after craniotomy for traumatic brain injury), consciousness was restored on the 3rd day after surgery but swallowing function was to be evaluated. A rapid screening was required every 2 hours to determine whether oral feeding should be started.

[0084] Headband design: The headband is designed for single use and features four built-in flexible silver paste electrodes (arranged in a 1×4 pattern) with a 12mm electrode spacing. It is pre-coated with conductive adhesive, which can be easily applied by peeling it off. The contact resistance is ≤6kΩ. The headband weighs only 15g and has no complicated adjustment structure.

[0085] The signal processing unit and wireless transmission chip are integrated into the portable host (palm-sized), and pair automatically after powering on, requiring no parameter settings.

[0086] Signal acquisition: The sampling frequency is fixed at 800Hz, only the core data of impedance amplitude is retained, and the resting period reference value is automatically recorded every 30 seconds.

[0087] ΔZ = |Z peak value during swallowing - Z mean value during resting period|;

[0088] The simplified formula for calculating risk is:

[0089]

[0090] t is fixed at 3 seconds, Weight = 1.0, Offset = 0.15.

[0091] When the patient drank 5ml of warm water, the system detected ΔZ = 18Ω, and calculated Risk = (18 / 3) × 1.0 + 0.15 = 6.15, triggering a medium-risk warning. The medical staff then switched to feeding the patient lotus root powder (a thick liquid), and ΔZ dropped to 10Ω, Risk = 3.48 (low risk), indicating that small amounts of oral feeding could be started. The entire screening process took 1.5 hours.

[0092] Implementation value: Disposable headbands prevent cross-infection and reduce the cost per person by 60%; no professional training is required to operate, making them suitable for emergency or rapid assessment scenarios; screening accuracy is consistent with standard configurations, providing immediate evidence for clinical decision-making.

[0093] Example 5

[0094] according to Figure 1 , 2As shown in the figure, this embodiment proposes a headband and system for detecting the risk of aspiration after neurosurgery based on impedance sensing. For a 45-year-old patient who underwent surgery for a cerebellopontine angle tumor and was recovering at home after discharge, the patient had mild difficulty swallowing. The family members needed to monitor the risk of daily eating through a simple device and report the data to the medical team every week.

[0095] Headband design: It adopts a washable fabric headband with 8 built-in dry electrodes (arranged in 2×4) with an electrode spacing of 10mm. It can be quickly fixed with Velcro when worn, and the contact resistance is stable at 4-5kΩ; it is equipped with a rechargeable main unit (48 hours of battery life).

[0096] The cloud platform retains only the core algorithm, and the terminal APP interface is optimized to "high / medium / low" risk indicator lights, without complex data curves.

[0097] Routine monitoring: Patients wear the device while eating three meals a day. It automatically enters monitoring mode after being turned on and generates a risk snapshot every 5 minutes.

[0098] Algorithm adaptation: Considering interference from the home environment, the signal quality assessment formula is simplified to:

[0099]

[0100] (SNR≥25dB ​​is considered valid); the Risk threshold is relaxed to T1=0.8 (high risk) and T2=0.5 (medium risk) to reduce false alarms.

[0101] When the patient ate rice, the app displayed a yellow medium-risk light (Risk = 0.62). The family immediately reminded the patient to slow down and increase the number of chews. Subsequent monitoring showed that the Risk level dropped to 0.35 (green low-risk). After the weekly data was uploaded to the cloud, the medical team remotely adjusted dietary recommendations via the app, and no aspiration incidents occurred.

[0102] Implementation value: The operation process is simplified to 3 steps (power on - wear - eat), and the family members have a 100% understanding rate; the low power consumption design supports long-term monitoring, and a single charge can meet 5 days of use; remote data interaction reduces the frequency of follow-up visits and saves medical resources.

[0103] Example 6

[0104] according to Figure 1 , 2 As shown in the figure, this embodiment proposes a headband and system for detecting the risk of aspiration after neurosurgery based on impedance sensing. For an 8-year-old child patient (after brainstem tumor surgery) with impaired swallowing function, due to the child's low cooperation, a very simple device is required for short-term monitoring (≤30 minutes each time).

[0105] Headband design: The headband features a cartoon pattern and is 5cm wide. It has 6 built-in micro electrodes (arranged in 2×3), with an electrode diameter of 2mm and a spacing of 6mm, which fits the size of a child's neck. The edges of the electrodes are rounded to avoid skin scratches.

[0106] The main unit is reduced to the size of a lighter and weighs ≤10g. It is fixed to the outside of the headband with medical tape to reduce the feeling of a foreign object.

[0107] Rapid monitoring: Select a time when the child is awake and calm, and immediately give the child a small amount of juice to drink after wearing the device. The duration of each monitoring session should be controlled within 15 minutes.

[0108] Algorithm adjustment: Considering the characteristics of children's muscle development, the weight is set to 0.8 (reducing the weight), and t is fixed at 2 seconds;

[0109] The risk warning has been changed to vibration feedback (without sound alarm) to avoid startling the child.

[0110] Monitoring Results: During the trial feeding, the system detected ΔZ = 9Ω, and calculated Risk = (9 / 2) × 0.8 + 0.1 = 3.7, indicating a low risk. Small amounts of feeding can continue. The child showed no resistance during the monitoring process, and parents could assist with the operation. Data was synchronized to the pediatric medical terminal.

[0111] Implementation value: The child-specific design improves wearing compliance, increasing the monitoring success rate from 60% to 92%; the short-term monitoring mode is adapted to children's attention characteristics, and the amount of effective data collected in a single session meets the assessment needs; the miniaturized device reduces restrictions on children's activities and lowers the difficulty of care.

[0112] By simplifying hardware, adapting algorithms, and optimizing processes for different scenarios, the universality of this system in diverse clinical needs has been verified, providing a practical solution for aspiration risk monitoring in different scenarios.

[0113] This invention integrates a flexible electrode array into a headband, enabling non-invasive and comfortable continuous monitoring. The electrodes conform to the curved surface of the anterior cervical muscles, avoiding skin irritation caused by traditional invasive electrodes. The headband is made of elastic spandex, making it easy to use and adaptable to different patient head circumferences, ensuring long-term comfort. Furthermore, this invention monitors real-time impedance changes in the anterior cervical muscles and uses a grading algorithm to provide early warning of aspiration risk. By detecting the rate of impedance change and time parameters during swallowing, the risk level can be calculated and compared with a preset threshold, issuing an early warning for greater reliability. Simultaneously, this invention integrates a cloud control platform and terminal devices for multi-dimensional data analysis. The impedance signals collected by the headband are processed by a demodulation algorithm and uploaded to the cloud control platform for storage and analysis. The cloud control platform incorporates a CNN-BiLSTM deep learning model for risk assessment, and the terminal devices support remote viewing of real-time data and historical trends, facilitating dynamic adjustments to nursing strategies by medical staff.

[0114] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A headband for assessing the risk of postoperative aspiration in neurosurgery based on impedance sensing, comprising a band body, a flexible electrode array, a signal processing unit, and a wireless transmission chip, characterized in that: The flexible electrode array is integrated inside the belt body and is used to collect bioimpedance signals of the anterior neck muscles; the signal processing unit is electrically connected to the flexible electrode array and is used to preprocess and demodulate the impedance signals. The wireless transmission chip is connected to the signal processing unit and is used to upload the processed signal to the cloud control platform.

2. The headband for assessing the risk of postoperative aspiration in neurosurgery based on impedance sensing as described in claim 1, characterized in that: The flexible electrode array consists of 15-30 graphene electrodes arranged in a matrix, with an electrode spacing of 5-10 mm. The flexible electrode array covers key muscle groups, and the contact impedance between the flexible electrode array and the skin is ≤5kΩ. It adopts a conductive gel or dry electrode structure, and the surface of the dry electrode has a microneedle structure.

3. The headband for assessing the risk of postoperative aspiration in neurosurgery based on impedance sensing as described in claim 2, characterized in that: The belt body is made of elastic spandex material, and the surface of the belt body is coated with a silver ion antibacterial layer. The breathability of the belt body is ≥2000g / (m²). 2 • 24h), tensile strength ≥ 50 N / cm.

4. The headband for assessing the risk of postoperative aspiration in neurosurgery based on impedance sensing as described in claim 1, characterized in that: The signal processing unit preprocesses the impedance signal, including signal quality assessment, using the following formula: Wherein, SQ is the signal quality score (0-1, ≥0.6 is a valid signal), SNR is the impedance signal-to-noise ratio (dB, calculated by the ratio of signal peak value to noise mean), and BD is the baseline drift (Ω, the standard deviation of impedance fluctuation during resting period).

5. The headband for assessing the risk of postoperative aspiration in neurosurgery based on impedance sensing as described in claim 1, characterized in that: The signal processing unit includes a low-pass filter module, an analog-to-digital converter module, and an amplification module. The low-pass filter module is used to filter out 50Hz power line noise. The analog-to-digital converter module has a sampling frequency of 1000-2000Hz. The amplification module is used to calculate the impedance amplitude and phase. The formula for calculating the impedance amplitude is: Where |Z| is the impedance magnitude (Ω), Z real Z represents the real part of the impedance (Ω, reflecting the tissue's electrical resistance). imag This represents the imaginary part of the impedance (Ω, reflecting the capacitance characteristics of the tissue).

6. A neurosurgical postoperative aspiration risk grading detection system based on impedance sensing, applied to the headband for neurosurgical postoperative aspiration risk grading detection based on impedance sensing as described in any one of claims 1-5, characterized in that: It includes a cloud control platform and terminal equipment. The cloud control platform is used to receive, store, and analyze impedance data uploaded by the wireless transmission chip. The cloud control platform has a built-in risk classification algorithm for real-time assessment of the risk level of the impedance data. The terminal device communicates with the cloud control platform to display risk levels and early warning information.

7. The neurosurgical postoperative aspiration risk grading and detection system based on impedance sensing according to claim 6, characterized in that: The cloud-based control platform also includes a continuous wavelet transform module and a neural network. The continuous wavelet transform module uses the db5 wavelet basis to generate a two-dimensional scale map based on impedance data. The neural network takes the scale map as input and outputs the swallowing event type based on the built-in CNN-BiLSTM deep learning model. The cloud-based control platform interfaces with the hospital information system to automatically synchronize patient clinical data and generate records.

8. The neurosurgical postoperative aspiration risk grading and detection system based on impedance sensing according to claim 7, characterized in that: The risk grading algorithm includes the following formula: Where ΔZ is the impedance change (Ω), t is the duration of the swallowing action (seconds), Weight is the weighting coefficient (0.5-1.5), Offset is the offset (0.1-0.3), and the Risk value ranges from 0 to 1, with a higher value indicating a higher risk. The formula for calculating the duration t of the swallowing action is: t = tend - tstart Where t is the duration of the swallowing action (seconds), tstart is the start time of the swallowing action (seconds, the moment when the impedance signal first exceeds twice the resting mean of the mean), and tend is the end time of the swallowing action (seconds, the moment when the impedance signal returns to the resting mean ± one standard deviation). The risk grading algorithm also includes a dynamic threshold adjustment formula to update the warning thresholds T1 (high risk) and T2 (medium risk): Tnew = Told × (1 + 0.1 × ErrRate) Where Tnew is the updated threshold, Told is the initial threshold: T1 = 0.7, T2 = 0.4, and ErrRate is the historical risk assessment error rate: 0-0.5, calculated using clinical validation data.

9. The neurosurgical postoperative aspiration risk grading and detection system based on impedance sensing according to claim 8, characterized in that: In the risk grading algorithm, ΔZ is calculated using the following formula: ΔZ = |Z swallowing period - Z resting period| Among them, the Z-swallowing period is the impedance value when the swallowing action occurs, and the Z-resting period is the average impedance value in the 30 seconds before swallowing.

10. The neurosurgical postoperative aspiration risk grading and detection system based on impedance sensing according to claim 6, characterized in that: The terminal devices include mobile apps and tablets. These devices are used for displaying risk levels, querying historical data, and setting warning thresholds. Warnings and alerts are issued based on the set warning thresholds.

Images