A device and method for assessing collateral circulation based on multimodal monitoring

Abstract

This invention discloses a collateral circulation assessment device and method based on multimodal monitoring. The device includes a physiological induction module, a functional near-infrared spectroscopy (fNIRS) monitoring module, an electroencephalogram (EEG) monitoring module, a control unit, and a data processing unit. It utilizes physiological stimulation to induce changes in cerebral blood flow and simultaneously acquires cerebral blood oxygenation signals and EEG signals. The data processing unit analyzes and fuses these multimodal signals to calculate a collateral circulation assessment index, objectively assessing the blood supply status of cerebral collateral circulation. This invention can detect cerebral blood flow regulation capacity and neurological functional responses in real time under non-invasive conditions. Compared with traditional DSA and CTA imaging assessment methods, it has advantages in safety, speed, and repeatability, significantly improving the accuracy and convenience of cerebral collateral circulation assessment. It is suitable for bedside monitoring and prognostic assessment of collateral circulation function in patients with cerebrovascular diseases such as stroke.

Description

Technical Field

[0001] This invention belongs to the field of imaging examination, and specifically relates to a collateral circulation assessment device and method based on multimodal monitoring. Background Technology

[0002] Cerebral collateral circulation refers to a compensatory mechanism in which blood is supplied to the ischemic area from other vascular pathways when cerebral arteries are narrowed or blocked. Good collateral circulation can reduce infarct volume, improve patient prognosis, and lower the risk of stroke recurrence. Therefore, accurately assessing the degree of cerebral collateral circulation is crucial for treatment decisions in diseases such as ischemic stroke.

[0003] Currently, clinical assessment of collateral circulation primarily relies on imaging examinations. Digital subtraction angiography (DSA) is widely recognized as the gold standard for assessing cerebral collateral circulation. DSA can dynamically and intuitively display cerebral blood flow at various stages, accurately reflecting the source, direction, and extent of collateral blood flow. However, DSA examination has disadvantages such as invasive operation, risk of complications, X-ray exposure, and iodine contrast agent nephrotoxicity. It is also expensive and requires professional personnel to operate, limiting its clinical application. As an alternative, non-invasive vascular imaging techniques such as CT angiography (CTA) and magnetic resonance angiography (MRA) are widely used for collateral circulation assessment, with CTA becoming a commonly used method. Plain and reconstructed CTA images can show the occlusion of intracranial main vessels and the distribution of leptomeningeal collateral vessels. Multiphase CTA, by adding delayed scans, more fully displays the temporal filling characteristics of collateral blood supply, improving the reliability of collateral circulation imaging. However, single-phase CTA may underestimate the degree of collateral circulation due to limitations in acquisition timing and is limited by image resolution, failing to adequately display fine and slow collateral blood flow. MRA is effective in visualizing collateral vessels of the main trunk, such as the Circle of Willis, but it has low sensitivity to small peripheral vessels. Contrast-enhanced MRA or high-resolution sequences can improve this, but still have limitations. In addition, CT perfusion imaging (CTP) and MR perfusion (PWI / ASL, etc.) can quantitatively assess the perfusion status and collateral compensation function of brain tissue, but these imaging techniques all require expensive large-scale equipment, resulting in high examination costs and making them unsuitable for continuous bedside monitoring.

[0004] Besides imaging examinations, ultrasound technology can also be used for the indirect evaluation of collateral circulation function. Transcranial Doppler ultrasound (TCD) can preliminarily assess the patency of collateral circulation by monitoring changes in blood flow velocity, making it a useful tool for assessing collateral compensation in patients with internal carotid artery stenosis / occlusion. In particular, introducing physiological stimulation to detect cerebral blood flow retention (CVR) function is one method of TCD assessment of collateral circulation: by inducing cerebral vasodilation stimulation through CO2 inhalation, intravenous injection of acetazolamide, or brief breath-holding, changes in cerebral blood flow velocity before and after stimulation can provide information on cerebral vascular autoregulation and collateral circulation status. However, TCD test results are highly dependent on the operator's experience and the patient's skull acoustic window conditions, resulting in relatively poor repeatability and reliability. In addition, the above-mentioned induction tests mostly use a single parameter (such as blood flow velocity) to reflect changes in blood supply, making it difficult to comprehensively characterize the functional state of brain tissue.

[0005] In recent years, some artificial intelligence algorithms have begun to be applied to the analysis of stroke images, such as automatic collateral circulation scoring and prognostic prediction based on brain images. However, these AI image assessment methods essentially rely on high-quality image data, and their results are still limited by the temporal and spatial resolution of the imaging modality, and cannot overcome the problems of invasive examinations and dependence on large equipment. In summary, existing technologies for assessing cerebral collateral circulation are either invasive and risky, or lack a reflection of real-time function, making it difficult to guide clinical decision-making in a timely and comprehensive manner. Therefore, there is an urgent need for a safe, non-invasive, bedside-accessible, multi-parameter comprehensive assessment technology to address the above shortcomings. This invention addresses this need by providing a collateral circulation assessment device and method based on a combination of physiological induction and multimodal monitoring to achieve an objective quantitative assessment of cerebral collateral circulation blood supply capacity. Summary of the Invention

[0006] The purpose of this invention is to provide a collateral circulation assessment device and method based on multimodal monitoring. This invention solves the problems of existing technologies in cerebral collateral circulation assessment, which are either invasive and risky or lack real-time functional reflection, making it difficult to guide clinical decision-making in a timely and comprehensive manner.

[0007] To achieve the above objectives, the present invention provides the following technical solution, comprising: a collateral circulation assessment device and method based on multimodal monitoring, comprising: a physiological evoked module, a functional near-infrared spectroscopy (fNIRS) monitoring module, an electroencephalogram (EEG) monitoring module, a control unit, and a data processing unit; wherein, the physiological evoked module is used to apply physiological stimulation to the subject to induce changes in cerebral blood flow; the fNIRS monitoring module is used to collect the concentration signals of oxyhemoglobin and deoxyhemoglobin in the subject's brain; the EEG monitoring module is used to collect the subject's electroencephalogram (EEG) activity signals; the control unit is connected to the physiological evoked module, the fNIRS monitoring module, and the EEG monitoring module respectively, and is used to control the triggering sequence of the physiological evoked stimulation and synchronize the data acquisition of the above monitoring modules; the data processing unit is connected to the fNIRS monitoring module and the EEG monitoring module, and is used to analyze and process the acquired multimodal signals to assess the state of cerebral collateral circulation, wherein the data processing unit assesses collateral circulation through the following steps and formulas: (i) (ii) Calculate the brain oxygenation change parameter ΔO based on the fNIRS signal, where ΔO is the difference between the average oxygenated hemoglobin concentration before and after physiological induction; (iii) Calculate the brain electrical change parameter ΔE based on the EEG signal, where ΔE is the change in the ratio of slow wave power to fast wave power in EEG before and after physiological induction; (iv) Combine ΔO and ΔE to obtain the collateral circulation assessment index CCI, where CCI = α·f(ΔO) + β·g(ΔE), where α and β are preset weighting factors, and f(·) and g(·) are preset functions, thereby characterizing the patency of blood supply to the brain collateral circulation.

[0008] Furthermore, the physiological induction module includes a respiratory control device for instructing or controlling the subject to hold their breath, change their respiratory rate rhythm, or inhale a predetermined concentration of carbon dioxide mixture to induce a physiological vasodilation response in the brain.

[0009] Furthermore, the functional near-infrared spectroscopy monitoring module includes multiple near-infrared light sources and multiple photodetectors. These light sources and detectors are arranged on the surface of the subject's head to form multiple monitoring channels for recording blood oxygen dynamics signals in different brain regions. The fNIRS monitoring module uses at least two different wavelengths of near-infrared light for detection, thereby distinguishing and acquiring changes in the concentrations of oxyhemoglobin and deoxyhemoglobin.

[0010] Furthermore, the EEG monitoring module includes multiple EEG electrodes, which are arranged on the subject's scalp at preset positions to simultaneously acquire multi-channel EEG signals; preferably, the EEG monitoring module and the fNIRS monitoring module are integrated on the same head-mounted device to achieve synchronous acquisition and spatial correspondence of EEG signals and near-infrared spectral signals.

[0011] Furthermore, the control unit is used to record the trigger time of the physiological evoked stimulus and provide the data processing unit with the timestamp corresponding to the time point; the data processing unit aligns and synchronizes the physiological evoked event with the data collected by the fNIRS monitoring module and the EEG monitoring module according to the timestamp, so as to use it for subsequent evoked response analysis.

[0012] Furthermore, the data processing unit performs filtering and noise reduction and baseline correction on the fNIRS signal, extracts the average value of oxyhemoglobin before and after physiological induction and the signal change amplitude, and obtains the brain oxygenation change parameter ΔO; the data processing unit performs filtering and noise reduction and spectrum analysis on the EEG signal, calculates the change in the slow wave to fast wave power ratio in the EEG signal before and after physiological induction, and obtains the EEG change parameter ΔE.

[0013] Furthermore, the data processing unit is used to assess the neurovascular coupling status based on the time correlation between the fNIRS signal and the EEG signal: when ΔO and ΔE are positively correlated or basically synchronous, it is determined that the collateral circulation function is good; if the correlation between ΔO and ΔE is weak or asynchronous, it indicates that the collateral circulation blood supply may be insufficient.

[0014] Furthermore, the data processing unit compares the collateral circulation assessment index (CCI) with a pre-established threshold or benchmark model to classify or alarm the collateral circulation status; when the CCI is lower than the threshold, it outputs a warning signal indicating poor collateral circulation, and when the CCI is higher than the threshold, it outputs a warning signal indicating good collateral circulation.

[0015] Furthermore, a method for assessing collateral circulation based on multimodal monitoring includes the following steps: (1) Apply a predetermined physiological evoked stimulus to the subject to induce hemodynamic changes in the brain; (2) The concentration signals of oxyhemoglobin and deoxyhemoglobin in the corresponding regions of the subject's brain were collected using the functional near-infrared spectroscopy monitoring module, and the electroencephalogram (EEG) signals were collected simultaneously using the EEG monitoring module. (3) The acquired near-infrared spectral signals and EEG signals are sent to the data processing unit for analysis, including calculating the change in oxygenated hemoglobin concentration ΔO and the change in EEG slow wave / fast wave power ratio ΔE before and after physiological induction, and fusing ΔO and ΔE to obtain the collateral circulation assessment index CCI; (4) The collateral circulation blood supply status of the subject’s brain is assessed based on the CCI and the assessment results are output.

[0016] Furthermore, the physiological evoked stimulation in step (1) is achieved by having the subject hold their breath for 20–30 seconds or inhale a predetermined concentration of carbon dioxide gas to induce a cerebral vasodilation response; the collateral circulation assessment index (CCI) in step (3) is calculated according to the formula CCI = α·f(ΔO) + β·g(ΔE), where α and β are preset weighting factors.

[0017] Compared with the prior art, the present invention has the following beneficial effects: Compared with the prior art, the present invention has the following beneficial effects: Safe, non-invasive, and repeatable: This invention employs in vitro physiological stimulation and non-invasive monitoring methods, eliminating the need for intubation and angiography, radiation exposure, and contrast agent toxicity, thus avoiding the invasive risks of DSA examination. The entire assessment process is safe and patient-friendly, and dynamic observation can be repeated as needed.

[0018] Dynamic functional assessment: Unlike static methods such as CTA / MRA that only provide vascular anatomy images, this invention observes reactive changes in cerebral blood flow through physiologically induced stimuli, enabling quantitative assessment of the functional reserve and regulatory capacity of cerebral collateral circulation. Compared to single-phase imaging assessment, this approach achieves real-time dynamic monitoring of collateral circulation blood supply status.

[0019] High Accuracy Through Multimodal Fusion: This invention integrates dual-modal information from fNIRS blood oxygenation signals and EEG neural electrical signals for a comprehensive assessment of cerebral perfusion and neurological function. Studies have shown that integrating fNIRS-EEG can extract neurovascular coupling features that cannot be obtained from a single modality, improving the accuracy and reliability of the assessment results. Through multimodal cross-validation, this approach reduces the impact of artifacts or fluctuations in single signals on the results, thereby more accurately reflecting the effectiveness of collateral circulation. Attached Figure Description

[0020] Figure 1 This is a block diagram of the overall structure of the device of the present invention. Detailed Implementation

[0021] This invention provides a collateral circulation assessment device based on physiological induction and multimodal monitoring. Its overall working principle involves inducing changes in cerebral blood flow by controlling subjects to undergo physiological stimuli such as CO2 inhalation or breath-holding, while simultaneously recording brain responses using multimodal monitoring methods such as functional near-infrared spectroscopy (fNIRS) and electroencephalography (EEG). Subsequently, relevant parameters are calculated to assess collateral blood supply function. The following detailed description of the specific implementation method, including the structure of each module, connection relationships, control flow, time synchronization mechanism, and algorithmic parameter calculation method, is provided.

[0022] (1) Module Composition and Functional Structure of the Device: This device mainly includes a physiological induction module, an fNIRS monitoring module, an EEG monitoring module, an optional transcranial Doppler (TCD) monitoring module, a central control module, and a data processing module. The physical structure and function of each module are as follows: Physiological evoked module: This module applies controlled respiratory stimulation to subjects to induce changes in cerebral hemodynamics. For example, it may include a medical gas control unit that supplies a specific concentration of CO2 mixture to the subject through a mask to temporarily increase blood carbon dioxide partial pressure; or it may include a prompting device (auditory or visual signal) to instruct the subject to hold their breath (pause breathing) for several seconds. Through CO2 inhalation or breath-holding, mild hypercapnia is induced in the subject, leading to cerebral vasodilation and changes in cerebral blood flow. This physiological evoked method is safe and controllable, designed to test the responsiveness of cerebral blood vessels and the compensatory level of collateral circulation.

[0023] The fNIRS monitoring module is used to monitor changes in blood oxygenation in the cerebral cortex. This module contains multiple near-infrared light sources and detector pairs, installed in specific areas of the head, such as the frontotemporal region, to cover the frontal and temporal lobes. Near-infrared light at a specific wavelength penetrates the scalp and skull to illuminate brain tissue, and the detectors receive the light signals scattered by the tissue. Changes in the concentrations of oxyhemoglobin (HbO) and deoxyhemoglobin (HbR) can be calculated using Beer-Lambert's law. The fNIRS monitoring module provides a curve of local cortical blood oxygen saturation over time, reflecting the local blood flow and oxygen supply to the brain. Positioning the probe in the frontotemporal region allows for focused observation of blood oxygenation changes in the area supplied by the middle cerebral artery, which is closely related to collateral circulation function.

[0024] EEG monitoring module: Used to record changes in brain electrical activity. This module consists of multi-lead EEG electrodes and amplifiers. The electrodes are placed on the scalp surface according to the international 10-20 system. In collateral circulation assessment, preferred locations are the central and parietal lobe regions (such as electrode sites C3, C4, Cz, and Pz) to monitor neural electrical activity throughout the brain, especially in the motor and sensory cortex. The EEG signal is amplified with high input impedance and acquired through analog-to-digital conversion, with a typical sampling rate of 500 Hz to 1000 Hz, ensuring millisecond-level time resolution. The EEG monitoring module can provide information on changes in neural activity during physiological stimulation, such as changes in the power of slow or fast waves in the EEG spectrum. These changes reflect alterations in cerebral blood and oxygen supply from a neurophysiological perspective.

[0025] TCD monitoring module: Used to monitor blood flow velocity in large cerebral vessels, mainly consisting of a transcranial Doppler ultrasound probe and a signal processing unit. The probe is typically placed at the temporal window, aimed at feeding arteries such as the middle cerebral artery (MCA) to record blood flow velocity waveforms. The TCD can measure baseline status and the average blood flow velocity of cerebral arteries after stimulation. This module can supplement local blood oxygenation changes in fNIRS, providing direct macroscopic hemodynamic indicators for calculating parameters such as the breath-holding index. Since collateral circulation directly affects changes in cerebral arterial blood flow velocity under hypercapnia stimulation, TCD monitoring helps quantify collateral blood supply capacity.

[0026] Central Control Module: Used for unified control and coordination of the above modules. The central control module is generally composed of an embedded microcontroller or industrial computer, with built-in timing and triggering circuits. Its functions include: ① Timing control of the experimental process, such as starting baseline acquisition, issuing stimulus start commands, controlling stimulus duration, and terminating acquisition; ② Distribution of synchronous trigger signals to ensure that all monitoring modules start recording with a unified time reference (see the time synchronization mechanism below); ③ Interaction with the physiological induction module, outputting control signals such as opening the CO2 supply valve or issuing a breath-holding prompt to precisely trigger physiological stimulation at a preset time point; ④ Real-time monitoring of feedback signals, such as CO2 concentration from the gas concentration sensor, motion signals from the subject's chest and abdominal breathing sensor, or TCD waveform morphology, to determine whether the subject correctly performed breath-holding as instructed or inhaled a sufficient concentration of CO2; ⑤ Judging the effectiveness of the induction based on feedback (e.g., premature cessation of breathing by the subject will result in ineffective stimulation). If an abnormality is detected, the current experiment can be terminated and restarted after the subject recovers to ensure the validity and safety of the data.

[0027] Data Processing Module: This module is used for the unified storage, processing, and analysis of acquired multimodal physiological signals. It can be a software function within the central control unit or a separate computing unit. The module receives HbO / HbR concentration change data from the fNIRS module, multichannel EEG data from the EEG module, and blood flow velocity data from the TCD module (if a TCD is configured). It performs preprocessing (e.g., filtering and denoising, baseline correction) and feature extraction on each data stream, and aligns and fuses data from different modalities according to the synchronization time axis. The module is also responsible for calculating various assessment indicators, including the rate of change of blood oxygen saturation ΔHbO, breath-holding index BHI, reaction time delay Tpeak, EEG band power change rate ΔP, and the comprehensive multimodal fusion index M (these algorithms will be described in detail below). After calculation, the module compares the obtained indicators with pre-set thresholds or models, providing an assessment result of collateral circulation function, such as a numerical score or qualitative grade (good, average, poor, etc.), for clinical reference.

[0028] (2) Module connection relationship and signal flow: The modules of this device are connected into a whole through wired or wireless means. Signals and data flow in a predetermined direction to complete the entire process of side-branch cycle evaluation. The central control module, as the core scheduling unit, establishes control and data communication connections with other modules.

[0029] Regarding the control signal path, the central control module connects to the physiological induction module and each monitoring module via a control bus or digital interface. When data acquisition needs to be initiated or stimulation triggered, the control module sends a start-acquisition command signal (e.g., trigger level or command message) to the fNIRS, EEG, and TCD monitoring modules, and simultaneously sends a stimulus execution command to the physiological induction module (e.g., an electrical signal to open the CO2 supply valve, or a prompt to the subject to begin breath-holding via a buzzer / display). Each monitoring module enters its working state upon receiving the start command, acquiring its own signal data in real time. Throughout the process, the central control module continuously outputs necessary control signals according to a preset time sequence: for example, when the breath-holding duration reaches the set time, a prompt device is used to instruct the subject to resume normal breathing, or the gas supply solenoid valve is closed after the predetermined CO2 inhalation time.

[0030] Regarding data flow, raw data collected by each monitoring module is sent to the storage unit of the data processing module or central control module via data transmission lines. The EEG module typically transmits multi-channel EEG digital signals to the control / processing unit via high-speed data lines (such as USB or Ethernet); the fNIRS module transmits HbO / HbR concentration data at a lower sampling rate (e.g., 10Hz or higher) via a serial communication interface or network interface; the TCD module outputs instantaneous blood flow velocity signals via analog voltage output (after A / D conversion) or digital communication. All data streams converge at the central control / processing module, are then appended with a unified time stamp, and stored in a buffer in preparation for subsequent synchronous analysis. During this process, the control module also receives various feedback signals: such as the status of the respiratory sensor, CO2 concentration sensor readings, or monitoring module self-test status signals. If an abnormality is detected (e.g., the subject prematurely ends breath-holding, resulting in CO2 not reaching the target level), the control module terminates the current trial via a control path and records it as invalid, thereby preventing erroneous data from entering the analysis.

[0031] (3) Side loop evaluation process and control scheduling logic: When using this device to perform a side loop function evaluation, the control module coordinates each module according to the preset process. The typical steps are as follows: Baseline Acquisition Phase: Before initiating physiological stimulation, the central control module first activates each monitoring module to enter baseline data acquisition mode. At this time, the subject is at rest and breathing normally. The fNIRS module records baseline brain tissue oxygenation status, the EEG module records baseline EEG activity characteristics, and the TCD module records baseline cerebral artery blood flow velocity. Baseline acquisition lasts for a predetermined time (e.g., 30 seconds) to obtain stable initial state data. The data processing module calculates the average values ​​of each indicator during the baseline period as a reference benchmark for subsequent analysis (e.g., baseline HbO concentration, baseline power of each EEG band, baseline blood flow velocity, etc.).

[0032] Physiological evoked phase (stimulation initiation and maintenance): After the baseline, the central control module simultaneously triggers the physiological evoked module and data recording to enter the stimulation phase. In the CO2 inhalation protocol, the control module opens the CO2 supply control valve, allowing premixed gas to be delivered to the subject through a mask for inhalation for a set period (e.g., 30 seconds), thereby inducing an increase in the subject's blood CO2 partial pressure. In the breath-holding protocol, the control module issues instructions (such as sound or screen prompts) to the subject to begin holding their breath, also for approximately 30 seconds or the duration specified by the physician. During this evoked phase, each monitoring module continues to operate: fNIRS monitors the dynamic changes in brain HbO / HbR in real time, EEG records EEG signals to capture immediate responses in neural activity, and TCD tracks changes in blood flow velocity. The control module precisely times the stimulation duration and can monitor the subject's condition during the process: for example, if it detects that the subject cannot tolerate the full duration (premature exhalation, CO2 concentration not reaching the target, etc.), it can terminate the stimulation early according to a preset strategy and record the trial failure or adjust the stimulation protocol and retry. Under normal circumstances, when the predetermined stimulus duration ends, the control module immediately proceeds to the next step.

[0033] Recovery Phase (After Stimulation): The control module issues a command to stop physiological stimulation: shut off the CO2 gas supply or instruct the subject to resume normal breathing. Afterward, each monitoring module continues to collect data for a period of time (e.g., another 60 seconds) to record the recovery process of cerebral blood flow and neural activity after stimulation stops. During this phase, the fNIRS signal often peaks a few seconds after stimulation ends and then gradually recovers; the EEG signal may show a gradual return to baseline spectral changes; and the blood flow velocity measured by TCD will also return to baseline levels. The control module ensures continuous data recording during the recovery phase and sends a stop-collection command to all modules after the predetermined monitoring time has been completed.

[0034] Data Validation and Feedback: After a complete procedure, the central control module can perform preliminary verification of the completeness and validity of the collected data. For example, it checks whether the breath-holding duration meets the required parameters, whether the CO2 concentration curve matches expectations, and whether there are severe motion artifacts in the fNIRS / EEG. If data is missing or has quality issues, the control module can decide whether the subject needs to rest and repeat the test. If the data is valid, it proceeds to the data processing and analysis phase.

[0035] Data Analysis and Result Output: The data processing module performs synchronous analysis on the complete multimodal data (the specific analysis algorithm is described in Part (4) below). The processing module first aligns the EEG, fNIRS, and TCD data along the time axis based on the synchronization signal, and then calculates the parameters of each indicator. Based on the calculation results, it compares them with the built-in discrimination criteria and finally gives the evaluation result of collateral circulation function. The central control module displays or stores the result through the user interface. If necessary, the result may include specific indicator values ​​and corresponding explanations, such as prompts like "Collateral circulation function in the left middle cerebral artery region is normal (comprehensive index M=85, normal range ≥80)". After the entire process is completed, the device enters standby mode, ready for the next measurement or shutdown.

[0036] (4) Multimodal Data Time Synchronization Mechanism: To ensure that signals from different sources such as EEG, fNIRS, and TCD can accurately correspond to the same time axis, this invention adopts a time synchronization mechanism combining hardware and software. At the hardware level, the central control module is equipped with a high-precision clock and a synchronization trigger circuit: at key time points such as the start of baseline acquisition and the start of stimulation, the control module sends a unified trigger pulse signal to each monitoring module through the synchronization output port. For example, a TTL synchronization signal is sent simultaneously to the EEG amplifier and the fNIRS host, triggering them to record the event marker at that moment; if the TCD module can receive external triggers, it also receives the marker signal synchronously. In this way, each data stream will include a common time zero-point reference in its respective record. In addition, during the entire monitoring period, the control module can periodically send synchronization pulses (such as every few seconds) or use a shared clock signal to calibrate the time base drift of each device. EEG and fNIRS can also share a unified sampling clock in the integrated design to achieve millisecond-level synchronization accuracy.

[0037] At the software level, the data processing module adds timestamps to the received data and executes an alignment algorithm: using the control module's clock as the global standard, the time axes of EEG, fNIRS, and TCD data are offset and corrected according to the trigger event markers. For example, if EEG data is sampled in milliseconds and fNIRS data is sampled at 100ms intervals, the processing module maps the fNIRS data points to the EEG timeline using a time interpolation method, or resamples each data point with a common time step. This unified time resampling and interpolation ensures that events (such as changes in blood flow and neural signals) at the same point in time in different modal signals can be directly compared. Through the combination of hardware synchronization triggering and software time alignment, multimodal physiological signal data synchronization is achieved, providing a reliable foundation for subsequent fusion analysis.

[0038] (5) Algorithm Indicator Calculation and Fusion Decision: In the data processing module, based on the synchronized multimodal data, a series of key indicator parameters for evaluating the function of the collateral circulation are extracted and calculated. These mainly include: a. Blood oxygenation rate change ΔHbO and breath-holding index BHI: These indicators reflect the magnitude of cerebral hemodynamic changes under induced stimulation. The blood oxygenation rate change ΔHbO is calculated based on changes in oxyhemoglobin concentration measured by fNIRS, while the breath-holding index BHI is calculated based on changes in blood flow velocity measured by TCD.

[0039] For the calculation of ΔHbO, the data processing module first determines the subject's baseline HbO concentration before stimulation (e.g., the average HbO concentration over several seconds at the end of the baseline phase), and the peak HbO concentration after stimulation (usually reaching its maximum within seconds after stimulation ends). Let baseline HbO represent the baseline oxyhemoglobin concentration, peak HbO represent the maximum oxyhemoglobin concentration reached after stimulation, and T represent the stimulation duration (e.g., the number of seconds of breath-holding). Then, the rate of change in blood oxygenation ΔHbO can be defined as the percentage change in the relative concentration of oxyhemoglobin per unit time, for example, according to the formula:

[0040] The ΔHbO calculated by the above formula is expressed in "% per second," representing the percentage increase in oxygenated hemoglobin concentration relative to baseline during the induction process per second. The larger the ΔHbO value, the more significant the increase in local brain oxygen supply caused by the induction stimulus, indicating a more adequate collateral circulation response; conversely, if ΔHbO is close to zero or even negative, it may indicate that blood oxygenation in that area has not increased significantly or has decreased, suggesting insufficient collateral circulation compensation.

[0041] For the calculation of the Breath Holding Index (BHI), the data processing module extracts relevant parameters from the TCD signal. First, it obtains the baseline mean cerebral artery blood flow velocity (Vbaseline, e.g., the baseline value of the mean flow velocity of the MCA) before stimulation, and the peak flow velocity (Vpeak, e.g., the instantaneous maximum flow velocity at the end of breath-holding, or the highest mean flow velocity within several seconds after stimulation) that occurs during or after breath-holding or CO2 inhalation induction. The Breath Holding Index is defined as the percentage change in cerebral blood flow velocity per unit time, and its calculation formula can be expressed as:

[0042] Where T is the duration of breath-holding (seconds), Vpeak and Vbaseline are the peak and baseline flow velocities at the end of breath-holding, respectively. The calculated BHI is expressed as a percentage per second (%). For example, if the average MCA flow velocity increases from the baseline of 40 cm / s to 48 cm / s during a 30-second breath-hold, then Vbaseline = 40, Vpeak = 48, the percentage increase is 20%, and dividing by 30 seconds gives BHI ≈ 0.67 % / s. The normality of collateral circulation response is usually evaluated by whether BHI exceeds a certain threshold. For example, BHI > 0.69 % / s is generally considered a normal response, while a value below this indicates reduced cerebrovascular reactivity. A higher BHI indicates that cerebral blood vessels can dilate sufficiently and blood flow increases significantly under high CO2 / hypoxia stimulation, reflecting good collateral circulation capacity; conversely, a low BHI indicates limited vascular dilation and poor collateral compensation.

[0043] It should be noted that for implementations without a TCD module, the amplitude of blood flow response can still be approximated using indicators such as ΔHbO from fNIRS; however, with a TCD, BHI, as a classic indicator, will provide a more objective evaluation. The data processing module can simultaneously calculate the BHI of both middle cerebral arteries (if both sides are monitored) to compare the differences in the function of the left and right collateral circulations.

[0044] b. Response Delay Time (Tpeak) and Regional Delay Difference: This indicator is used to assess the temporal agility of cerebral blood flow and oxygenation responses, as well as the synchronicity between different regions. The data processing module identifies the response curve after evoked stimulation in each signal and calculates the time required to reach the peak value, i.e., Tpeak. Specifically, for fNIRS signals, timing can begin at the start of the stimulus, and the point at which the HbO concentration rises to its maximum value can be found. The time difference between this point and the start of the stimulus is the HbO response delay time Tpeak (HbO) for that region. Similarly, for blood flow velocity curves measured by TCD, the point at which the flow velocity peak occurs is found, yielding the blood flow velocity response Tpeak (CBF). If certain features in the EEG signal (such as frequency band power variations) also have peaks or maximum deviations, Tpeak (EEG) can also be calculated accordingly. However, EEG changes are usually instantaneous or slightly different from the blood flow peak and can only be used as a reference.

[0045] Under normal circumstances, the brain's blood flow response to high CO2 or breath-holding stimulation occurs within a short time. A high response speed indicates unobstructed collateral circulation pathways and sensitive regulation. If the Tpeak is significantly prolonged, it indicates a sluggish vascular response to stimulation, possibly due to insufficient collateral circulation, requiring a longer time to deliver blood to the target area. The data processing module can compare Tpeaks from different brain regions or sides, calculating regional delay differences. For example, the difference between the Tpeak of the fNIRS signal in the left frontotemporal region and the corresponding Tpeak in the right side is calculated as ΔTleft-right; or the difference in response delay between the anterior circulation (frontal region) and the posterior circulation (occipital region) can be compared. If the Tpeak in a particular side or region is consistently much longer than in other regions, this regional delay difference suggests that the collateral blood supply pathway in that region may not be fast enough. The processing module can set a threshold; when the regional Tpeak difference exceeds a predetermined value (e.g., more than 3 seconds), it is marked as an abnormal delay, thus alerting clinicians to potential problems with blood flow supply to the corresponding region.

[0046] Furthermore, when calculating Tpeak, the data processing module can employ smoothing filtering and thresholding methods to improve robustness. For example, the peak value can be defined as the time point at which a stable maximum value is reached or a certain percentage of the relative increase is achieved, in order to reduce misjudgments caused by signal noise. Ultimately, the Tpeak metric provides time-dimensional information for the evaluation of side circulation, complementing the amplitude information of ΔHbO and BHI.

[0047] c. EEG Band Power Change Rate ΔP: This indicator assesses changes in brain function before and after stimulation from a neurophysiological perspective. The data processing module performs spectral analysis on the EEG signal, extracting the power spectral density from the EEG data through Fast Fourier Transform (FFT) or wavelet transform, and dividing the power according to classical frequency bands, such as delta waves (0.5–4 Hz), theta waves (4–8 Hz), alpha waves (8–13 Hz), beta waves (13–30 Hz), etc. First, the average power values ​​Pbaseline(δ), Pbaseline(θ), Pbaseline(α)… for each frequency band during the baseline phase are calculated. Then, the average power values ​​Pstimulus(δ), Pstimulus(θ), Pstimulus(α)… for the frequency bands after stimulation (e.g., a time window after the end of stimulation or a stable period in the later stage of stimulation) are calculated.

[0048] The rate of change of power in an EEG band, ΔP, is defined as the percentage change in power of the stimulus state relative to the baseline state. For each band i, it can be calculated as follows:

[0049] For example, in the alpha band, if the baseline alpha power is 50 μV 2 After stimulation, the α power decreased to 40 μV. 2 If ΔPα = ((40-50) / 50)×100% = -20%, it indicates that the α power decreased by 20%. Typically, in hypercapnia or ischemic states, brain EEG shows a trend of increased low-frequency power and decreased high-frequency power; that is, δ and θ power may increase while α and β power decrease. This is because insufficient cerebral blood supply slows neuronal activity and increases synchronicity, reflected in the EEG as increased slow waves and decreased fast waves. Therefore, this device pays particular attention to the degree of low-frequency enhancement and high-frequency attenuation. The data processing module can select ΔP values ​​from one or more representative frequency bands as indicators, such as changes in ΔPα or the δ / α power ratio. If a subject shows a significant increase in slow-wave power after stimulation (e.g., ΔPδ is positive and large) accompanied by a significant decrease in fast waves (ΔPα is negative and large), this may confirm that their brain tissue is significantly affected by hypercapnia / hypoxia, suggesting possible insufficient collateral circulation (because normal individuals should be able to maintain adequate cerebral oxygen supply, thus keeping the EEG relatively stable). Conversely, if the EEG spectrum shows almost no change before and after stimulation, it indicates that brain tissue function is minimally affected, and collateral blood supply may be adequate. It should be noted that EEG signals are sensitive to individual differences and noise; therefore, ΔP analysis is usually combined with other indicators for comprehensive judgment.

[0050] d. Multimodal Fusion Index M and Decision Logic: To provide a comprehensive assessment of collateral circulation function, this invention designs a multimodal fusion index M. This index combines the aforementioned parameters and fuses information from different modalities through a weighted algorithm, improving the accuracy and robustness of the assessment. The data processing module first normalizes each basic index, converting indices with different dimensions into comparable dimensionless values. For example, based on the normal population benchmark, ΔHbO, BHI, Tpeak, ΔP, etc., can be linearly scaled or standardized to the 0-1 range using empirical formulas (larger values ​​indicate better function or stronger response; the delay time Tpeak can be taken as its reciprocal or negative to form a "response speed" index to maintain consistency). Let the processed normalized indices be H (representing the amplitude of blood flow / blood oxygenation response, such as the fusion of ΔHbO or BHI), R (representing reaction speed, such as calculated based on Tpeak), and E (representing EEG changes, such as calculated based on ΔP). The multimodal fusion index M can be calculated using a weighted summation:

[0051] The weighting coefficients wH, wR, and wE are pre-set based on the importance of each modality to the assessment or determined through training with a large number of samples. For example, in some embodiments, a higher weight can be given to the amplitude of blood flow response because BHI / ΔHbO directly reflects the collateral circulation blood supply capacity; at the same time, appropriate weights can be given to response velocity and EEG changes to comprehensively reflect different aspects of information. More complex nonlinear fusion models (such as classifiers based on logistic regression or neural networks) can also be introduced to combine the various indicators, but for the sake of clarity, linear weighting is used as an example here.

[0052] The calculated M value will serve as a comprehensive score for collateral circulation function. The device's built-in decision logic compares M with predetermined thresholds to arrive at an evaluation conclusion. For example, assuming that calibration determines the M value range to be between 0 and 100, the following settings can be configured: M ≥ 80 indicates good collateral circulation function with sufficient compensation; 50 ≤ M < 80 indicates moderate or partial compensation function; and M < 50 indicates poor collateral circulation blood supply function. In some cases, only a single threshold can be set, for example, M below 60 is considered abnormal, and above 60 is considered normal. Specific thresholds and grading standards can be adjusted and optimized based on clinical trial data. The decision logic relies not only on the absolute value of M but also on the combination patterns of individual indicators. For example, if M is close to the critical value but one of the individual indicators is extremely abnormal, further investigation may be suggested.

[0053] Finally, the data processing module outputs the evaluation results based on the decision-making logic, including the M-index value and the collateral circulation function level. The central control module can store this result and present it to the operator on the display screen. For example, the display could be: "Comprehensive index M=85, collateral circulation blood supply function is good," or "Comprehensive index M=45, collateral circulation function is weakened, further imaging examination is recommended." By using the multimodal fusion index M, this device can provide an objective and quantitative evaluation of the collateral circulation status, which is more comprehensive and reliable than a single index.

[0054] In summary, the specific embodiments of this invention achieve quantitative assessment of cerebral collateral circulation function by organically combining physiological induction methods and multimodal brain monitoring technology. Each module has a clearly defined function and operates collaboratively: the central control module precisely schedules the induction and acquisition process, the time synchronization mechanism ensures high alignment of EEG, fNIRS, and TCD data, and the data processing module extracts and fuses various feature indicators for judgment. This device can provide multi-dimensional information on the amplitude and speed of cerebrovascular response, as well as changes in neurological function, thereby comprehensively assessing collateral blood supply capacity. In clinical applications, this device is expected to serve as a non-invasive and real-time detection tool, assisting physicians in assessing the collateral circulation status of patients with ischemic cerebrovascular disease, evaluating stroke risk or treatment effectiveness, and providing a basis for relevant decision-making. The above embodiments are only used to illustrate the technical solution of this invention; all equivalent substitutions or modifications made within the spirit and principle of this invention should be included within the scope of protection of this invention.

Claims

1. A device and method for assessing lateral circulation based on multimodal monitoring, characterized in that, include: The system comprises a physiological evoked stimulus module, a functional near-infrared spectroscopy (fNIRS) monitoring module, an electroencephalogram (EEG) monitoring module, a control unit, and a data processing unit. The physiological evoked stimulus module is used to apply physiological stimulation to the subject to induce changes in cerebral blood flow. The fNIRS monitoring module is used to acquire local oxyhemoglobin and deoxyhemoglobin concentration signals in the subject's brain. The EEG monitoring module is used to acquire the subject's electroencephalogram (EEG) activity signals. The control unit is connected to the physiological evoked stimulus module, the fNIRS monitoring module, and the EEG monitoring module, respectively, and is used to control the triggering sequence of the physiological evoked stimulus and synchronize the data acquisition of the aforementioned monitoring modules. The data processing unit is connected to the fNIRS monitoring module and the EEG monitoring module, and is used to analyze and process the acquired multi-mode signals to assess the state of cerebral collateral circulation. The data processing unit assesses collateral circulation through the following steps and formulas: (i) calculating the cerebral oxygen change parameter ΔO based on the fNIRS signal, where ΔO is the difference between the average oxyhemoglobin concentrations before and after physiological evoked stimulus; (ii) The EEG signal is used to calculate the EEG change parameter ΔE, where ΔE is the change in the ratio of slow wave power to fast wave power of EEG before and after physiological induction; (iii) the ΔO and ΔE are fused to obtain the collateral circulation assessment index CCI, where CCI = α·f(ΔO) + β·g(ΔE), α and β are preset weighting factors, and f(·) and g(·) are preset functions, which characterize the patency of blood supply to the brain collateral circulation.

2. The apparatus according to claim 1, characterized in that, The physiological induction module includes a respiratory control device for instructing or controlling the subject to hold their breath, change their respiratory rate rhythm, or inhale a predetermined concentration of carbon dioxide mixture to induce a physiological vasodilation response in the brain.

3. The apparatus according to claim 1, characterized in that, The functional near-infrared spectroscopy monitoring module includes multiple near-infrared light sources and multiple photodetectors. These light sources and detectors are arranged on the surface of the subject's head to form multiple monitoring channels for recording blood oxygen dynamics signals in different brain regions. The fNIRS monitoring module uses at least two different wavelengths of near-infrared light for detection, thereby distinguishing and acquiring changes in the concentrations of oxyhemoglobin and deoxyhemoglobin.

4. The apparatus according to claim 1, characterized in that, The EEG monitoring module includes multiple EEG electrodes, which are arranged on the subject's scalp at preset positions to simultaneously acquire multi-channel EEG signals. Preferably, the EEG monitoring module and the fNIRS monitoring module are integrated on the same head-mounted device to achieve synchronous acquisition and spatial correspondence of EEG signals and near-infrared spectral signals.

5. The apparatus according to claim 1, characterized in that, The control unit is used to record the trigger time of the physiological evoked stimulus and provide the data processing unit with the timestamp corresponding to the time point; the data processing unit aligns and synchronizes the physiological evoked event with the data collected by the fNIRS monitoring module and the EEG monitoring module according to the timestamp for subsequent evoked response analysis.

6. The apparatus according to claim 1, characterized in that, The data processing unit performs filtering and noise reduction and baseline correction on the fNIRS signal, extracts the average value of oxyhemoglobin before and after physiological induction and the signal change amplitude, and obtains the brain oxygenation change parameter ΔO; the data processing unit performs filtering and noise reduction and spectrum analysis on the EEG signal, calculates the change in the slow wave to fast wave power ratio in the EEG signal before and after physiological induction, and obtains the EEG change parameter ΔE.

7. The apparatus according to claim 1, characterized in that, The data processing unit is used to assess the neurovascular coupling status based on the time correlation between the fNIRS signal and the EEG signal: when ΔO and ΔE are positively correlated or basically synchronous, it is determined that the collateral circulation function is good; if the correlation between ΔO and ΔE is weak or asynchronous, it indicates that the collateral circulation blood supply may be insufficient.

8. The apparatus according to claim 1, characterized in that, The data processing unit compares the collateral circulation assessment index (CCI) with a pre-established threshold or benchmark model to classify or alert the collateral circulation status. When the CCI is below the threshold, a warning signal indicating poor side circulation is output; when the CCI is above the threshold, a warning signal indicating good side circulation is output.

9. A method for assessing collateral circulation based on multimodal monitoring, characterized in that, Includes the following steps: (1) Apply a predetermined physiological evoked stimulus to the subject to induce hemodynamic changes in the brain; (2) The concentration signals of oxyhemoglobin and deoxyhemoglobin in the corresponding regions of the subject's brain were collected using the functional near-infrared spectroscopy monitoring module, and the electroencephalogram (EEG) signals were collected simultaneously using the EEG monitoring module. (3) The acquired near-infrared spectral signals and EEG signals are sent to the data processing unit for analysis, including calculating the change in oxygenated hemoglobin concentration ΔO and the change in EEG slow wave / fast wave power ratio ΔE before and after physiological induction, and fusing ΔO and ΔE to obtain the collateral circulation assessment index CCI; (4) The collateral circulation blood supply status of the subject’s brain is assessed based on the CCI and the assessment results are output.

10. The method according to claim 9, characterized in that: The physiological evoked stimulation in step (1) is achieved by having the subject hold their breath for 20–30 seconds or inhale a predetermined concentration of carbon dioxide gas to induce a cerebral vasodilation response; the collateral circulation assessment index (CCI) in step (3) is calculated according to the formula CCI = α·f(ΔO) + β·g(ΔE), where α and β are preset weighting factors.

Images