Method for noninvasive clinical monitoring of intracranial edema

Abstract

The invention discloses a method for noninvasive clinical monitoring of intracranial edema, and relates to the technical field of medical monitoring, and the method comprises the following steps: S001, arranging an annular tension acquisition structure at the periphery of a forehead attached optical sensor, obtaining a skin tension change curve, and enabling the skin tension change curve to correspond to an optical reflection signal point by point according to time, a tension coupling observation band is formed; and S002, the tension rising speed is extracted based on the tension coupling observation band, a shake doubt area is formed in the time slice when the tension rising speed changes abruptly, and a shake fingerprint list is generated in the shake doubt area according to the tension amplitude. Through synchronous observation of tension and optical signals, pseudo peak light energy isolation reconstruction and light source adaptive scanning, the accuracy and stability of intracranial edema monitoring are improved, the risk of false alarm and excessive intervention is reduced, and high precision and clinical practicability are achieved.

Description

Technical Field

[0001] This invention relates to the field of medical monitoring technology, specifically to a non-invasive clinical monitoring method for intracranial edema. Background Technology

[0002] Non-invasive clinical monitoring of intracranial edema refers to acquiring peripheral physiological signals related to the intracranial state through the skin surface via an intelligent sensing system, without puncture, implantation, or contact with the intracranial cavity. These signals include brain optical reflexes, brain electrical rhythms, cerebral blood flow pulsations, skull micro-vibrations, fundus optic nerve changes, and scalp physiological fluctuations. The intelligent sensing system continuously monitors these signals, dynamically calculates them, and performs medical pattern recognition to assess in real-time the increased pressure, tissue expansion, and blood flow changes caused by increased intracranial tissue fluid. This allows for 24 / 7, bedside clinical monitoring of the development, trends, and risk of acute deterioration of intracranial edema. This monitoring method avoids the trauma, infection, and surgical risks associated with traditional intracranial pressure probes, enabling physicians to observe the dynamics of intracranial lesions continuously, promptly, and quantitatively. It provides a safer, more convenient, and real-time means of managing the condition of patients with brain injury, stroke, and neurological intensive care.

[0003] The existing technology has the following shortcomings:

[0004] Current non-invasive optical monitoring solutions largely rely on attached sensors to form a stable optical path on the forehead. However, in non-fixed states, sudden coughing, convulsions, or involuntary tension can cause a rapid change in the tension of the forehead skin, leading to abrupt changes in the optical refractive index of superficial tissues over microsecond timescales. This abrupt change disrupts the propagation balance of the original optical path, creating abnormally strong reflected echoes and resulting in bright false peaks in the monitoring signal. These false peaks are misidentified by current monitoring systems as spike signals indicating a rapid expansion of brain tissue water content within a short period. This causes edema prediction models to enter extreme warning states, triggering mandatory intervention procedures inconsistent with the actual condition, including excessive dehydration, excessive compression therapy, and sudden adjustments to ventilation parameters. Such erroneous interventions can cause adverse damage to the patient's brain perfusion, neural metabolism, and circulatory stability in a short period, and may even exacerbate existing intracranial damage.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide a non-invasive clinical monitoring method for intracranial edema, in order to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for non-invasive clinical monitoring of intracranial edema, comprising the following steps:

[0008] S001, A ring-shaped tension acquisition structure is set around the optical sensor attached to the forehead to obtain the skin tension change curve, and the skin tension change curve is mapped to the optical reflection signal point by point according to time to form a tension coupling observation zone;

[0009] S002, based on the tension coupling observation zone, the tension rise rate is extracted, and a swaying suspicion zone is formed in the time segment where the tension rise rate changes sharply. A swaying fingerprint list is generated in the swaying suspicion zone according to the tension amplitude.

[0010] S003, map the shaking fingerprint list to the optical reflection signal trajectory, lock the bright peak in the optical reflection signal according to the time index of the shaking fingerprint list, form a pseudo peak candidate set, and generate a pseudo peak masking band according to the tension amplitude of the shaking fingerprint list.

[0011] S004, a multi-ring polarization splitter structure is introduced in the optical path acquisition link around the pseudo-peak masking band, and the reflected light energy of the time point marked by the pseudo-peak masking band is introduced into the side buffer channel. In the side buffer channel, a smooth reference trajectory is generated according to the label of the pseudo-peak masking band, and the smooth reference trajectory is used to perform peak reduction and valley filling on the main monitoring light curve to form a sway suppression edema observation curve.

[0012] S005 constructs an extracranial optical flow self-compliant traction array based on the sway-inhibited edema observation curve. It uses the tension rhythm in the sway-inhibited edema observation curve to drive the incident angle of the light source and drive the spot to perform a swaying scan on the forehead, so that the optical flow trajectory forms a continuous cross-irradiation path. In the cross-irradiation path, the intracranial edema signal is accumulated and the instantaneous spurious peaks generated by sway are canceled, thereby completing the stable dynamic monitoring of intracranial edema.

[0013] Preferably, step S001 includes:

[0014] An optical sensor is attached to the central area of ​​the forehead, and tension sensing units are arranged around the optical sensor. A medical-grade flexible adhesive material is used to maintain a tight fit, and an initial zero stress reference value is set for the tension sensing units to establish a mechanical measurement reference.

[0015] The strain change values ​​of the tension sensing unit during different facial activities are collected, and the collected data are mapped to the forehead reference plane through the elliptical geometric center model to form a set of tension change curves that are consistent with the time axis and correspond to the spatial points.

[0016] Based on the tension change curve group, a snapshot sequence of tension spatial distribution is constructed at each moment, and the snapshot sequence is linked with the synchronously acquired optical reflection signal to form an index binding structure in time point by point;

[0017] The tension snapshot sequence after index binding is combined with the optical reflection signal sequence to form a tension coupling observation band organized in time sequence, which is used to locate the corresponding tension distribution state when the optical signal is abnormal.

[0018] Preferably, in the step of constructing the tension spatial distribution snapshot sequence, the collected tension change curves are adjusted by mean and normalized by amplitude to ensure that the data of all tension sensing units have a unified response range, thereby enhancing the comparability and synchronicity of tension changes.

[0019] Preferably, step S002 includes:

[0020] The tension increment value is extracted point by point over time from the tension change curve in the tension coupling observation zone to form a tension rise velocity spectrum, and the suspected swaying area is delineated based on the abrupt change segment of the rise velocity.

[0021] Extract the maximum tension amplitude value of the tension sensing unit within the swaying suspicion area, establish the correspondence between tension channels and tension amplitude, and form a data entry consisting of time index, tension channel number and tension peak value;

[0022] All data entries are arranged in chronological order to form a shaking fingerprint list, and the corresponding time period and tension point are marked on the tension change curve to realize the mapping between time period and tension amplitude.

[0023] Dynamic sliding matching is performed on the tension coupling observation band based on the shaking fingerprint list to identify repeated shaking patterns and assign stable labels to form a high-confidence shaking fingerprint.

[0024] Preferably, step S003 includes:

[0025] The list of shaken fingerprints is analyzed one by one, and synchronously aligned in the optical reflection signal trajectory based on the time index. High-brightness peaks in the optical signal are extracted to form a set of pseudo-peak candidates.

[0026] The tension amplitude information in the shaking fingerprint list is bound to the pseudo-peak candidate set to form a composite interference label with tension intensity label;

[0027] Arrange the candidate set of pseudo-peaks in chronological order, expand the time interval before and after each peak to form a masking segment, and transfer the tension intensity label to the corresponding masking area;

[0028] All masked areas are embedded in the optical reflection signal trajectory to form a coupled observation composite curve, which is used to identify the interference intensity and provide positioning coordinates for subsequent signal processing.

[0029] Preferably, when forming the pseudo-peak masking band, the masking segments extending before and after each pseudo-peak candidate point are uniformly marked with grayscale, and the masking segments are divided into different intensity levels according to the tension intensity label, so that the optical reflection signal can obtain intensity grade expression within the masking segment, which is used to improve the accuracy and stability of subsequent peak clipping processing.

[0030] Preferably, step S004 includes:

[0031] A multi-ring polarization splitter structure is introduced into the optical acquisition path to direct the reflected light energy of the pseudo-peak mask at the marked time point into the side buffer channel;

[0032] In the side buffer channel, the optical signal is standardized according to the tension label of the pseudo-peak masking strip and spliced ​​to form a time-aligned smooth reference trajectory.

[0033] The signal in the main monitoring light curve during the corresponding masking time period is replaced point by point with the smooth reference trajectory to achieve peak reduction and valley filling while preserving the continuity of the curve.

[0034] The processed master monitoring light curve was checked as a whole, and a shaking inhibition edema observation curve was generated for subsequent optical flow traction and pathological identification.

[0035] Preferably, the generation of the smooth reference trajectory is further defined as follows: the reflected light energy is amplitude compressed in the side buffer channel with tension labels as weights, and the sampling density is kept consistent with the main monitoring light curve on the time axis, so that the replaced curve forms a continuous transition in the masking section and maintains the stability of the light reflection trend.

[0036] Preferably, step S005 includes:

[0037] Using the tension rhythm extracted from the edema suppression observation curve as a reference driving signal, the incident angle of the light source is driven and the spot landing point on the forehead forms a dynamic displacement.

[0038] The driving spot performs a fine-amplitude oscillating scan on the forehead, forming a cross optical flow path with a cross density distribution;

[0039] By jointly analyzing the optical reflection signals collected in the cross optical flow path with the stable segment of the edema observation curve suppressed by shaking, the segment of the actual dynamic change of edema is extracted.

[0040] The extracted stable segment was fused with the edema observation curve of shaking inhibition to construct a dynamic monitoring curve of intracranial edema for stable dynamic monitoring.

[0041] Preferably, the oscillating scanning of the light spot landing point sets a time delay between each light source, so that the cross optical flow paths form a staggered coverage in the spatial distribution, and the consistency of the reflected signals in the optical flow path determines whether the stable signal segment is outside the tension interference shield.

[0042] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0043] This invention, based on the simultaneous observation of skin tension and light reflection signals, effectively establishes a direct correspondence between biomechanical changes and optical anomalies. This prevents spurious peaks caused by shaking from being mistaken for edema signals, significantly reducing the risk of medical over-intervention due to false alarms. By introducing a multi-ring polarization shunt structure and a side buffer channel, optical-level isolation and reconstruction of abnormal reflection energy are achieved, preserving the integrity and continuity of the signal waveform. Finally, by using shaking rhythm to drive light source scanning, adaptive adjustment and cross-illumination of the optical flow path under tension are achieved, effectively enhancing the redundancy and consistency of the optical signal and significantly improving the accuracy and dynamic stability of monitoring the true state of intracranial edema. This method combines accuracy, stability, and clinical adaptability, providing a novel and highly reliable solution for non-invasive brain monitoring. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0045] Figure 1 This is a flowchart of a non-invasive clinical monitoring method for intracranial edema according to the present invention. Detailed Implementation

[0046] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0047] This invention provides, for example Figure 1 The method shown includes the following steps for non-invasive clinical monitoring of intracranial edema:

[0048] S001, A ring-shaped tension acquisition structure is set around the optical sensor attached to the forehead to obtain the skin tension change curve, and the skin tension change curve is mapped to the optical reflection signal point by point according to time to form a tension coupling observation zone;

[0049] This step discloses a key component of a method for non-invasive clinical monitoring of intracranial edema: constructing a tension sensing path around an optical sensor attached to the forehead, and mapping the acquired tension change curves point-by-point with the optical reflection signals to achieve a high-precision tension coupling observation zone construction process. The specific implementation steps are as follows:

[0050] An optical sensor is attached to the central area of ​​the patient's forehead, ensuring a close fit to the skin surface using a medical-grade flexible adhesive material. Eight to twelve tension sensing units are equidistantly arranged around the sensor, forming a closed ring around the circumference. Each tension sensing unit consists of a deformable flexible strain layer and an embedded fiber Bragg sensor core, with a highly breathable medical coating on the outer layer to enhance wearing comfort. By adjusting the number and spacing of the tension sensing units, for example, placing one unit every 30 degrees, comprehensive coverage of tension changes across the entire forehead skin can be achieved. In this step, by setting an initial zero-stress reference value for each sensing unit, it is ensured that the numerical response of subsequent tension change curves is based on the actual stretching amplitude, thus establishing a unified mechanical measurement benchmark.

[0051] Multi-channel skin tension response data is generated by real-time acquisition of strain changes from all tension sensing units under natural conditions and different stimulus conditions. The data acquisition frequency is set to 1,000 times per second, covering common facial muscle activities such as rest, conversation, coughing, head tilting, and frowning. To ensure spatial consistency of the data, the data collected by each group of tension sensing units are mapped to coordinates, and the position of each data point is uniformly normalized to the forehead reference plane using an elliptical geometric center model. Furthermore, the acquired raw tension value curves are uniformly resampled into equally spaced curves of 1,000 points per second, forming a group of tension change curves with consistent time axis and corresponding spatial points. By performing mean adjustment and amplitude normalization on each group of tension change curves, all data curves have a unified amplitude response range in subsequent steps, enhancing the contrast and synchronicity of tension changes.

[0052] Based on the obtained tension change curve set, the tension values ​​at each moment are horizontally aggregated according to the time axis sequence. That is, at any time point, the corresponding values ​​of all tension units arranged in a ring are taken to construct a tension spatial distribution snapshot sequence. Each snapshot represents the instantaneous distribution state of forehead skin tension in two-dimensional space at that moment. At the same time, another synchronously acquired optical reflection signal curve is also recorded at the same time sampling frequency, ensuring a point-by-point correspondence with the tension change curve on the time axis. In this process, the tension snapshot sequence and the optical reflection signal curve are indexed and bound, that is, each optical reflection data point corresponds to a complete tension distribution state. The coupled data structure constructed in this way can accurately locate the corresponding tension background state at any time point when the optical reflection signal changes. For example, when the optical reflection curve jumps abnormally at 3.562 seconds, the system can immediately call the 3562nd sample in the tension snapshot sequence to obtain the complete distribution map of skin tension at that moment, thereby analyzing whether there is any non-optical interference.

[0053] The tension snapshot sequence, after index binding, is combined with the optical reflection signal sequence to construct a tension-coupled observation band. The tension-coupled observation band is a multidimensional dataset organized chronologically, with each band containing complete temporal and spatial tension distributions and corresponding optical reflection responses. In practice, the visualization system displays the content of the tension-coupled observation band as a heatmap, enabling dynamic playback of tension and optical signals. In bedside clinical use, doctors can intuitively determine whether abnormal optical signal states in the coupled observation band are caused by the patient's voluntary movements by observing the tension distribution state at those moments. For example, if the tension at the upper and lower edges of the forehead increases rapidly simultaneously within a certain time period, and the optical reflection signal also shows synchronous pulse-like spikes, the system will mark this segment as a possible swaying interference. By constructing the tension-coupled observation band, not only is precise time-point matching of skin tension and optical reflection achieved, but a data foundation is also provided for subsequent interference identification and spurious peak suppression, significantly improving the reliability and stability of non-invasive monitoring.

[0054] S002, based on the tension coupling observation zone, the tension rise rate is extracted, and a swaying suspicion zone is formed in the time segment where the tension rise rate changes sharply. A swaying fingerprint list is generated in the swaying suspicion zone according to the tension amplitude.

[0055] For time-series data in tension-coupled observation bands, a method is proposed to identify periods of swaying interference and extract their features. This method uses the rate of increase in skin tension as the dynamic identification criterion, and combines this with the tension amplitude to generate a traceable list of swaying feature data, providing a basis for subsequent spurious peak removal. The specific steps are as follows:

[0056] The tension change curves in the constructed tension coupling observation zone were analyzed point-by-point, and the tension increment values ​​between adjacent sampling points were extracted second by second along the time axis. With a sampling frequency of 1,000 times per second, each second of the tension change curve contained 1,000 data points, and the increment value of each point reflected the rate of change of local skin tension at that time point. Taking a cough as an example, the increment value of the tension sensing units on both sides of the forehead jumped sharply from 0.3 units to 3.1 units between 0.820 seconds and 0.827 seconds, with an average increase rate exceeding 0.4 units per millisecond. This rate of increase significantly exceeded the upper limit of normal static fluctuations. After mapping the increase rates of all tension channels to the time axis, a tension increase rate spectrum was formed, in which the time periods of steep increase slopes could be visually observed. By setting medically empirical thresholds, such as considering an increase rate exceeding 2.5 units per second as a signal of involuntary rapid muscle activity, multiple time segments of high-speed tension changes could be delineated in the spectrum. Each delineated time segment constitutes a preliminary suspected area of ​​swaying.

[0057] Within each identified suspected swaying zone, the maximum tension amplitude value of all tension-sensing units within that zone is further extracted, and a direct correspondence between tension channels and amplitude values ​​is established. For example, in a suspected swaying zone from 0.820 seconds to 0.860 seconds, the maximum amplitudes of tension-sensing units numbered 4, 6, and 10 are 4.7, 5.3, and 4.2 units, respectively. These values ​​reflect the differences and intensities of force on the skin in different directions during this time period. All tension channel numbers, corresponding tension amplitudes, time index start points, and durations are integrated to form data entries, which are then arranged in chronological order to constitute a continuous data list of swaying features. To improve descriptive clarity, each entry is labeled with the trigger time, maximum tension channel number, and corresponding tension peak value. For example, the first entry is: start time 0.820 seconds, channel number 4, maximum tension 4.7 units, duration 40 milliseconds. This list is called the swaying fingerprint list, representing a sequence of swaying feature labels constructed in the temporal and spatial dimensions.

[0058] After the vibration fingerprint list is constructed, it is synchronously labeled with the tension change curves in the original tension coupling observation zone. Specifically, the time period corresponding to each entry in the list is marked with a uniform color on the linear tension change curve, and the maximum tension point and channel number of the entry are overlaid on the tension curve graph. This visualization not only provides a location basis for subsequent steps but also facilitates medical personnel in quickly determining the distribution range and intensity of vibration interference during actual observation. This binding method creates a clear mapping between time periods and characteristic amplitudes, allowing for direct access to the time index corresponding to the vibration fingerprint list for precise interference removal when analyzing optical reflection signals.

[0059] By using a pre-constructed list of sway fingerprints, dynamic sliding matching is performed on continuous time periods throughout the tension coupling observation band to verify whether similar sway patterns recur in subsequent time periods. For example, if the same channel combination is found to experience rapid tension rise events of similar duration under similar tension amplitudes in two different time periods, this pattern is considered to have stable sway behavior characteristics. In this case, the sway pattern can be assigned a stability label, marked as a high-confidence sway fingerprint, providing credible interference evidence for subsequent interference identification steps of optical reflection signals. This approach not only achieves systematic encoding of the time, intensity, and channel distribution of tension disturbance events but also makes the sway characteristics traceable and identifiable. The final generated list of sway fingerprints becomes the core basis for fusion analysis with optical signals, supporting the entire process of subsequent pseudo-peak locking and processing.

[0060] S003, map the shaking fingerprint list to the optical reflection signal trajectory, lock the bright peak in the optical reflection signal according to the time index of the shaking fingerprint list, form a pseudo peak candidate set, and generate a pseudo peak masking band according to the tension amplitude of the shaking fingerprint list.

[0061] By utilizing the time index and tension amplitude information in the shaking fingerprint list, high-brightness abnormal peaks in the optical reflection signal trajectory are screened, a candidate set of pseudo-peaks is further extracted, and a pseudo-peak masking band is constructed for subsequent interference suppression, thereby achieving clean processing and signal-to-noise enhancement of optical data. The specific implementation steps are as follows:

[0062] The generated list of swaying fingerprints is analyzed one by one, and synchronized with the optical reflection signal trajectory based on the time index. In the full-time curve of the optical reflection signal, each second contains one thousand sampling points. The start time and duration of each record in the swaying fingerprint list precisely correspond to a specific segment of the optical signal within that time period. For example, if a fingerprint record occurs between 0.860 and 0.910 seconds, the corresponding segment in the optical signal curve is points 860 to 910. By iterating through the light reflection intensity within this interval, it is possible to observe whether there are any instantaneous peak values ​​more than three times the normal baseline amplitude. If a sudden increase in intensity occurs, and the time index of this peak point falls exactly within the corresponding time period in the swaying fingerprint list, then this peak is identified as an initial candidate point for a false peak. Following this method, all swaying fingerprint entries are traversed, and all suspected strong reflection signal points are located and marked one by one on the optical reflection signal curve, constructing an initial candidate set of false peaks composed of three elements: time point, peak intensity, and signal amplitude.

[0063] The initial set of candidate pseudo-peaks is analyzed for effectiveness, and features are superimposed using the tension amplitude information of each entry in the shake fingerprint list. Each shake record contains the maximum tension channel number and the corresponding tension value, which reflects the physiological intensity of the corresponding shake action. For example, at a candidate pseudo-peak time point of 0.892 seconds, the tension of channel 6 in its corresponding shake record is 5.1 units, significantly higher than the baseline static level. Based on this, the tension value is projected onto the optical pseudo-peak as a tension intensity label. This label describes the degree of interference that the peak may be subject to at the tension level and also serves as a basis for reduction in subsequent processing. In this way, each pseudo-peak not only possesses intensity information in its optical signal but also carries the underlying tension-driven characteristics, forming a composite interference label with both physical and physiological factors.

[0064] After binding all pseudo-peaks to tension labels, the pseudo-peak masking band construction stage begins. Each point in the pseudo-peak candidate set is arranged sequentially along the time axis, and dynamic protection intervals are set before and after each peak time point, for example, extending forward and backward by fifteen milliseconds to form a total of thirty milliseconds of local masking segments. These masking segments are considered high-interference sensitive time windows and are marked as pseudo-peak masking areas on the optical reflection signal trajectory using grayscale processing. Simultaneously, the tension intensity label bound to the central peak point of each masking segment is transferred to the entire masking area, so that each masking band not only has a clear time boundary but also a tension level label representing the degree of physiological impact. In this way, interference events of different intensities can be treated differently on the optical curve; for example, masking bands with tension intensities exceeding 5 units can be marked with dark gray, those between 3 and 5 units with light gray, and those below 3 units with pale gray. This multi-level masking method facilitates different processing strategies for pseudo-peaks of different intensities in subsequent implementation stages, improving the flexibility and accuracy of overall optical signal processing.

[0065] After the pseudo-peak masking band is constructed, it is fully embedded into the original optical reflection signal trajectory to form a coupled observation composite curve. This composite curve not only preserves the time-series characteristics of the original optical reflection but also visually reflects the distribution and intensity of the shaking interference periods through the masking band. In actual clinical monitoring, this composite curve can serve as an auxiliary reference for physicians. When confirming the presence of a pathological edema trend, the marked pseudo-peak masking band area should be avoided to prevent external interference from being mistaken for internal lesion signals. Furthermore, this composite curve provides basic positioning coordinates for subsequent optical curve shaping, enabling the spurious peak position, interference amplitude, and duration to be quantified. Thus, the entire process from shaking fingerprint list to optical signal pseudo-peak locking and masking construction is completed, establishing a stable data support system for achieving high-precision non-invasive intracranial edema monitoring.

[0066] S004, a multi-ring polarization splitter structure is introduced in the optical path acquisition link around the pseudo-peak masking band, and the reflected light energy of the time point marked by the pseudo-peak masking band is introduced into the side buffer channel. In the side buffer channel, a smooth reference trajectory is generated according to the label of the pseudo-peak masking band, and the smooth reference trajectory is used to perform peak reduction and valley filling on the main monitoring light curve to form a sway suppression edema observation curve.

[0067] Based on the calibrated pseudo-peak masking band, a multi-ring polarization shunt structure is introduced into the acquisition link by adjusting the optical path structure and guiding the energy channel. This effectively isolates the abnormal reflection energy caused by swaying interference and reconstructs a smooth reference trajectory. Finally, the original monitoring light curve is adjusted to generate an edema observation curve with anti-swaying capability. The specific implementation steps are as follows:

[0068] A multi-layered optical path structure is constructed within the light acquisition path attached to the forehead skin surface. This structure includes a set of main incident light sources, a main reflection acquisition path, and two side buffer channels parallel to the main path. A multi-ring polarization shunt structure is implanted at the boundary point of the main reflection acquisition path. This structure consists of continuously arranged polarization guiding layers, refraction angle adjustable guiding prisms, and energy separation films. Each shunt structure is layered and identified according to the polarization direction of the incident light, and guides a portion of the reflected light into the side channels at a set angle. For example, when a specific reflection signal triggers a pseudo-peak masking band marker at a time point of 0.875 seconds, the reflected light acquired at this time point is preferentially shunt into the left buffer channel, temporarily removed from the main acquisition path. This structural arrangement ensures that during the time period marked by the pseudo-peak masking band, all abnormal reflection energy will not enter the main curve trajectory, but will instead be diverted to the buffer space used to construct the reference curve. The multi-ring polarization shunt structure can identify and respond to the incident direction, wavelength disturbance and polarization state of the optical signal in real time, thereby ensuring that the shunt action has sufficient response accuracy and energy stability in time, and is suitable for microsecond-level signal processing requirements.

[0069] A continuous light energy acquisition sequence is established in the side buffer channel, and each segment of the incoming reflected light signal is standardized according to the interference tags provided by the pseudo-peak masking band. During each segment of the diverted light signal, energy weight normalization is performed according to the amplitude level of the tension tag. For example, if the tension tag in the masking band is 5.4 units, the light energy value within that segment is compressed proportionally, converting it into a smooth signal point. These diverted signals are then stitched together in a continuous processing manner to form a complete time-series reference light curve with continuous amplitude. Each signal point in the reference light curve is reconstructed into a smooth transition curve with the masking band's time index as the horizontal axis and the reflected value after interference amplitude adjustment as the vertical axis, at a density of 1,000 points per second. This curve has two key characteristics: first, its composition is entirely derived from the abnormal signal within the pseudo-peak time segment but has been converted to a standard amplitude; second, its time structure strictly follows the time axis arrangement of the main light curve, thus providing a strictly aligned compensation trajectory for subsequent signal adjustments.

[0070] The constructed smooth reference trajectory serves as the basis for peak clipping, applied to the corresponding masking time periods in the main beam curve to replace outliers and maintain waveform continuity. During processing, all masked time period signal points in the main beam curve are replaced point-by-point by the equal-time signal values ​​in the smooth reference trajectory, thereby reducing abnormal reflection intensity and filling curve gaps caused by interference. For example, between 0.875 seconds and 0.910 seconds, there are two sudden signal peaks in the main beam curve, at 6.2 and 7.8 units respectively, corresponding to tension masking bands marked as 5.4 and 5.7 units. The main beam curve for this time period is then completely replaced with smoothed values ​​from the reference trajectory for the same time period, such as 4.5 and 4.8 units, preserving curve continuity while effectively eliminating the interference of sudden peaks on the overall trend. This operation is performed point-by-point within each masking band, with a ten-millisecond smooth transition boundary segment reserved before and after each segment to avoid abrupt changes at signal splicing points.

[0071] After replacing all masking time periods in the master light curve, the processing results were compared with the original curve to generate an edema observation curve with sway suppression capability. This curve maintains the same time length and sampling density as the original monitoring light curve in its overall structure, but it no longer contains abnormally high-value signals caused by sway interference in the region of spurious peaks, while also preserving the optical reflection changes caused by normal physiological responses. The sway-suppressed edema observation curve can serve as the basic data input for subsequent medical judgment and dynamic trend identification, and also provides a stable, accurate, and interference-removing optical reflection sequence for subsequent active optical flow traction. The construction of this observation curve not only improves the signal reliability of non-invasive brain monitoring but also significantly reduces the risk of excessive clinical intervention due to misjudgment, laying a solid foundation for real-time monitoring of cerebral edema.

[0072] S005, based on the sway-inhibition edema observation curve, constructs an extracranial optical flow self-compliant traction array. It uses the tension rhythm in the sway-inhibition edema observation curve to drive the incident angle of the light source and drive the spot to perform sway scanning on the forehead, so that the optical flow trajectory forms a continuous cross-irradiation path. In the cross-irradiation path, the intracranial edema signal is accumulated and the instantaneous spurious peaks generated by sway are canceled, thereby completing the stable dynamic monitoring of intracranial edema.

[0073] Based on the obtained sway-inhibited edema observation curves, an active traction mechanism was further constructed to enhance the stability of optical acquisition. This mechanism guides the optical flow path through tension rhythm to form continuous cross-irradiation in the forehead, achieving stable acquisition and interference cancellation of intracranial edema signals. The specific steps are as follows:

[0074] Using the tension rhythm extracted from the sway-induced edema observation curve as a reference driving signal, a multi-point light source incident array is established on the outer surface of the forehead. This array is symmetrically arranged around the original main light spot illumination center, covering the entire forehead extending to the temples. Each group of incident lights has a controllable optical axis adjustment structure. By observing the tension change cycle and amplitude distribution in the sway-induced edema observation curve, the tension guiding direction represented by each time segment within each complete tension rhythm cycle is calculated. Taking a tension rhythm with a two-second cycle as an example, if the tension continuously increases and the high-frequency changes are concentrated in the left side of the forehead within 0.5 to 1.2 seconds, the incident angle adjustment mechanism of the right light source will be activated during the corresponding time period, slightly shifting its illumination direction to the upper left, forming a reverse pull on the light flow in the tension area. Simultaneously, the illumination position of the central light spot is shifted by 0.8 mm along the tension guiding direction, forming a dynamic light spot landing point displacement. This dynamic adjustment is performed in time segments, ensuring that each small segment of the illumination path within the entire tension cycle precisely corresponds to the dominant direction of the current tension state.

[0075] After initial orientation adjustment, each group of light spots is driven to perform a fine-amplitude oscillating scan on the forehead, forming a planar cross-optical flow structure. Each light spot oscillates at a micro-scale angle around a reference point within a local area, with the oscillation amplitude controlled within ±1.2 mm. The oscillation period strictly references the sub-period of the tension rhythm; for example, when the main tension period is two seconds, the oscillation period is one-tenth of the main period, meaning one oscillation cycle is completed every 0.2 seconds. A micro-time delay is set between each light spot to form mutually staggered cross-scanning trajectories. Throughout the oscillating scan, the light spots are ensured to maintain an adjacent, non-overlapping distribution, and the number of cross-scans is accumulated within the scan coverage area, ultimately forming a spatial distribution network with an optical flow trajectory cross-density of no less than thirty times per square centimeter across the entire forehead. This network structure allows any facial area to receive repeated irradiation from multiple directions and different time points per unit time, significantly improving the consistency and redundancy of light signal propagation within the tissue.

[0076] Optical reflection signals collected from cross-path irradiation are jointly analyzed with stable segments in the edema suppression observation curve to extract relatively stable segments of intracranial edema signals by comparing irradiation paths. In cross-paths, optical flows from different paths penetrate the same area at multiple angles. If a signal waveform shows a similar trend in the irradiation results of all paths and occurs outside the tension interference shield, it is determined to be a genuine intracranial tissue response. For example, within 0.900 to 0.960 seconds, the midline region of the forehead receives irradiation from five different directions, and the obtained reflected wave peak values ​​are 3.2, 3.4, 3.3, 3.3, and 3.5 units, respectively, with a difference of no more than 0.3 units. The reflected signal during this time period can be considered highly stable and unaffected by spurious peaks caused by spurious irradiation. Such signal segments verified by multi-path sampling consistency are extracted as segments of real dynamic edema changes. The same screening operation is performed in all cross-paths, and finally, a high-confidence optical reflection curve encompassing the true response information is spliced ​​together.

[0077] Using the selected stable light reflection segment as the core data, this segment is fused with the entire sway-inhibition edema observation curve to construct the final dynamic monitoring curve for intracranial edema. This curve retains the temporal continuity structure of the sway-inhibition curve while using the stable signal segment confirmed by cross-path sampling as the main characteristic peak and fluctuation benchmark, replacing and compensating for segments with low signal-to-noise ratios in the original curve. The fused curve possesses the completeness of signal spurious peak removal processing and also achieves physical cancellation of asymmetric interference and abnormal local muscle responses at the spatial scanning level, thereby significantly improving the sensitivity to the actual intracranial pathological state. This monitoring curve can be continuously output and updated in real time, suitable for bedside monitoring scenarios for stroke, traumatic brain injury, and neurocritical care patients, ensuring a highly reliable basis for intracranial edema assessment in dynamic environments. Thus, the dynamic steady-state optical flow construction method using active spot traction and path cross-structure has achieved the goal of stable dynamic monitoring of intracranial edema changes.

[0078] This invention, based on the simultaneous observation of skin tension and light reflection signals, effectively establishes a direct correspondence between biomechanical changes and optical anomalies. This prevents spurious peaks caused by shaking from being mistaken for edema signals, significantly reducing the risk of medical over-intervention due to false alarms. By introducing a multi-ring polarization shunt structure and a side buffer channel, optical-level isolation and reconstruction of abnormal reflection energy are achieved, preserving the integrity and continuity of the signal waveform. Finally, by using shaking rhythm to drive light source scanning, adaptive adjustment and cross-illumination of the optical flow path under tension are achieved, effectively enhancing the redundancy and consistency of the optical signal and significantly improving the accuracy and dynamic stability of monitoring the true state of intracranial edema. This method combines accuracy, stability, and clinical adaptability, providing a novel and highly reliable solution for non-invasive brain monitoring.

[0079] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A method for non-invasive clinical monitoring of intracranial edema, characterized in that, Includes the following steps: S001, A ring-shaped tension acquisition structure is set around the optical sensor attached to the forehead to obtain the skin tension change curve, and the skin tension change curve is mapped to the optical reflection signal point by point according to time to form a tension coupling observation zone; S002, based on the tension coupling observation zone, the tension rise rate is extracted, and a swaying suspicion zone is formed in the time segment where the tension rise rate changes sharply. A swaying fingerprint list is generated in the swaying suspicion zone according to the tension amplitude. S003, map the shaking fingerprint list to the optical reflection signal trajectory, lock the bright peak in the optical reflection signal according to the time index of the shaking fingerprint list, form a pseudo peak candidate set, and generate a pseudo peak masking band according to the tension amplitude of the shaking fingerprint list. S004, a multi-ring polarization splitter structure is introduced in the optical path acquisition link around the pseudo-peak masking band, and the reflected light energy of the time point marked by the pseudo-peak masking band is introduced into the side buffer channel. In the side buffer channel, a smooth reference trajectory is generated according to the label of the pseudo-peak masking band, and the smooth reference trajectory is used to perform peak reduction and valley filling on the main monitoring light curve to form a sway suppression edema observation curve. S005 constructs an extracranial optical flow self-compliant traction array based on the sway-inhibited edema observation curve. It uses the tension rhythm in the sway-inhibited edema observation curve to drive the incident angle of the light source and drive the spot to perform a swaying scan on the forehead, so that the optical flow trajectory forms a continuous cross-irradiation path. In the cross-irradiation path, the intracranial edema signal is accumulated and the instantaneous spurious peaks generated by sway are canceled, thereby completing the stable dynamic monitoring of intracranial edema.

2. The method for non-invasive clinical monitoring of intracranial edema according to claim 1, characterized in that, Step S001 includes: An optical sensor is attached to the central area of ​​the forehead, and tension sensing units are arranged around the optical sensor. A medical-grade flexible adhesive material is used to maintain a tight fit, and an initial zero stress reference value is set for the tension sensing units to establish a mechanical measurement reference. The strain change values ​​of the tension sensing unit during different facial activities are collected, and the collected data are mapped to the forehead reference plane through the elliptical geometric center model to form a set of tension change curves that are consistent with the time axis and correspond to the spatial points. Based on the tension change curve group, a snapshot sequence of tension spatial distribution is constructed at each moment, and the snapshot sequence is linked with the synchronously acquired optical reflection signal to form an index binding structure in time point by point; The tension snapshot sequence after index binding is combined with the optical reflection signal sequence to form a tension coupling observation band organized in time sequence, which is used to locate the corresponding tension distribution state when the optical signal is abnormal.

3. The method for non-invasive clinical monitoring of intracranial edema according to claim 2, characterized in that, In the step of constructing the tension spatial distribution snapshot sequence, the collected tension change curves are adjusted by mean and normalized by amplitude to ensure that the data of all tension sensing units have a unified response range, thereby enhancing the comparability and synchronicity of tension changes.

4. The method for non-invasive clinical monitoring of intracranial edema according to claim 2, characterized in that, Step S002 includes: The tension increment value is extracted point by point over time from the tension change curve in the tension coupling observation zone to form a tension rise velocity spectrum, and the suspected swaying area is delineated based on the abrupt change segment of the rise velocity. Extract the maximum tension amplitude value of the tension sensing unit within the shaking suspicion area, establish the correspondence between tension channels and tension amplitude, and form a data entry consisting of time index, tension channel number and tension peak value; All data entries are arranged in chronological order to form a shaking fingerprint list, and the corresponding time period and tension point are marked on the tension change curve to realize the mapping between time period and tension amplitude. Dynamic sliding matching is performed on the tension coupling observation band based on the shaking fingerprint list to identify repeated shaking patterns and assign stable labels to form a high-confidence shaking fingerprint.

5. A method for non-invasive clinical monitoring of intracranial edema according to claim 4, characterized in that, Step S003 includes: The list of shaken fingerprints is analyzed one by one, and synchronously aligned in the optical reflection signal trajectory based on the time index. High-brightness peaks in the optical signal are extracted to form a set of pseudo-peak candidates. The tension amplitude information in the shaking fingerprint list is bound to the pseudo-peak candidate set to form a composite interference label with tension intensity label; Arrange the candidate set of pseudo-peaks in chronological order, expand the time interval before and after each peak to form a masking segment, and transfer the tension intensity label to the corresponding masking area; All masked areas are embedded in the optical reflection signal trajectory to form a coupled observation composite curve, which is used to identify the interference intensity and provide positioning coordinates for subsequent signal processing.

6. A method for non-invasive clinical monitoring of intracranial edema according to claim 5, characterized in that, When forming the pseudo-peak masking band, the masking segments extending before and after each pseudo-peak candidate point are uniformly marked with grayscale, and the masking segments are divided into different intensity levels according to the tension intensity label, so that the optical reflection signal can obtain intensity grade expression within the masking segment, which is used to improve the accuracy and stability of subsequent peak clipping processing.

7. A method for non-invasive clinical monitoring of intracranial edema according to claim 5, characterized in that, Step S004 includes: A multi-ring polarization splitter structure is introduced into the optical acquisition path to direct the reflected light energy from the pseudo-peak mask at the marked time point into the side buffer channel. In the side buffer channel, the optical signal is standardized according to the tension label of the pseudo-peak masking strip and spliced ​​to form a time-aligned smooth reference trajectory. The signal in the main monitoring light curve during the corresponding masking time period is replaced point by point with the smooth reference trajectory to achieve peak reduction and valley filling while preserving the continuity of the curve. The processed master monitoring light curve was checked as a whole, and a shaking inhibition edema observation curve was generated for subsequent optical flow traction and pathological identification.

8. A method for non-invasive clinical monitoring of intracranial edema according to claim 7, characterized in that, The generation of the smooth reference trajectory is further defined as follows: the reflected light energy is amplitude compressed in the side buffer channel with tension labels as weights, and the sampling density is kept consistent with the main monitoring light curve on the time axis, so that the replaced curve forms a continuous transition in the masking section and maintains the stability of the light reflection trend.

9. A method for non-invasive clinical monitoring of intracranial edema according to claim 7, characterized in that, Step S005 includes: Using the tension rhythm extracted from the edema suppression observation curve as a reference driving signal, the incident angle of the light source is driven and the spot point on the forehead is dynamically displaced. The driving spot performs a fine-amplitude swing scan on the forehead, forming a cross optical flow path with a cross density distribution; By jointly analyzing the optical reflection signals collected in the cross optical flow path with the stable segment of the edema observation curve suppressed by shaking, the true dynamic change segment of edema is extracted. The extracted stable segment was fused with the edema observation curve of shaking inhibition to construct a dynamic monitoring curve of intracranial edema for stable dynamic monitoring.

10. A method for non-invasive clinical monitoring of intracranial edema according to claim 9, characterized in that, The oscillating scanning of the light spot landing point sets a time delay between each light source, so that the cross optical flow paths form a staggered coverage in the spatial distribution, and the consistency of the reflected signals in the optical flow path determines whether the stable signal segment is outside the tension interference shield.

Images