Respiratory training mode and ventilation uniformity guiding method and system based on EIT

Abstract

The invention discloses a respiratory training mode and ventilation uniformity guiding method and system based on EIT, and belongs to the technical field of medical monitoring. The method comprises the steps that regional ventilation delay mode classification and pathophysiological mechanism identification are executed, a ventilation delay map is generated by collecting breathing data, and curve morphology analysis and mechanical parameter calculation are carried out to distinguish obstructive factors and restrictive factors; targeted breathing training based on real-time ventilation uniformity is executed based on a result, and personalized guidance and closed-loop optimization are achieved. The system comprises a regional ventilation delay analysis unit and a targeted guide training unit. The problems that in the prior art, pathophysiological mechanisms cannot be accurately distinguished, and real-time feedback is lacked are solved, and the efficiency and individuation level of respiratory training are improved.

Description

Technical Field

[0001] This invention belongs to the field of medical monitoring technology, specifically relating to the cross-technology of respiratory function monitoring and rehabilitation training based on electrical impedance tomography (EIT), and especially the respiratory training method and ventilation uniformity guidance method and system based on EIT. Background Technology

[0002] Existing breathing training methods based on electrical impedance tomography (EIT) can monitor lung ventilation distribution and calculate parameters such as the global non-uniformity index (GI). Some advanced methods further utilize regional time constant differences to qualitatively identify early and late ventilation regions, and based on this, provide general training suggestions such as prolonging inspiratory breath-holding or emphasizing chest expansion.

[0003] However, such methods have significant drawbacks: First, the regional time constant is only a macroscopic and general representation, and the underlying pathophysiological mechanisms are not deeply analyzed. For example, ventilation delay in a certain region may be caused by increased local airway resistance (obstructive factors) or decreased local lung tissue compliance (limiting factors). The optimal breathing training strategies corresponding to these two root causes (such as focusing on airflow velocity control or focusing on volume expansion) are fundamentally different, and existing technologies cannot distinguish between them. Second, the training suggestions given are static and unidirectional, lacking real-time monitoring and guidance of regional ventilation response during the execution of training actions, resulting in insufficient training efficiency and personalization. Part of the work of this invention was supported by the Shanghai Jiao Tong University "Jiaotong Star" Program Medical Engineering Interdisciplinary Research Fund (YG2025LC05), aiming to address the above limitations.

[0004] The full English abbreviations and their Chinese names in this instruction manual are as follows: EIT: Electrical Impedance Tomography

[0005] ROI: Region of Interest

[0006] GI: Global Inhomogeneity Index

[0007] TIV: Tidal Impedance Variation

[0008] CoV: Center of Ventilation Summary of the Invention

[0009] This invention aims to address the problems of existing EIT-based respiratory training methods, which cannot accurately distinguish the specific pathophysiological mechanisms of ventilatory delay (such as obstructive and restrictive factors) and lack real-time feedback.

[0010] To address the aforementioned technical problems, the present invention provides the following technical solution.

[0011] A breathing training method and ventilation uniformity guidance method based on EIT includes the following steps: Step 1: Perform regional ventilation delay pattern classification and pathophysiological mechanism identification.

[0012] This step includes: Step 11: Acquire respiratory data and generate a ventilation delay map, extract time-related features from the EIT data stream, and determine the global inspiratory initiation time T. global The start time T of local ventilation for each pixel pixel Calculate the delay time T delay And generate a regional ventilation delay map.

[0013] Step 12: Perform regional ventilation dynamics curve morphology analysis, extract the standardized impedance curves of each late-reaching ventilation region in the inspiratory and expiratory phases, analyze the shape of the rising segment of the inspiratory curve and the falling segment of the expiratory curve, and classify the ventilation dynamics mode of the late-reaching ventilation region into low compliance-dominated, high resistance-dominated, or mixed types.

[0014] Step 13: Perform regional mechanical parameter calculation based on the tidal volume-driving pressure relationship, correlate the regional tidal resistance change value (TIV) measured by EIT in a specific region with the driving pressure corresponding to that region, fit the pressure-volume relationship curve at the regional level, and calculate the regional compliance and regional resistance.

[0015] Step 2: Based on the results of Step 1, perform targeted breathing training based on real-time ventilation uniformity. Match a personalized training mode according to the pathophysiological mechanism diagnosis results output in Step 1, set an initial quantitative target based on the patient's baseline ventilation, execute the breathing training process, and guide the process through real-time feedback.

[0016] In a preferred embodiment of the present invention, in step 11, the EIT device continuously acquires impedance data of the thoracic cross-section at a high frame rate of not less than 50 frames per second; the lung ROI region is divided into 8 or 12 regions of interest based on anatomical location; the delayed ventilation arrival region is defined as the delay time T. delay Regions exceeding 300 milliseconds.

[0017] In a preferred embodiment of the present invention, step 11 determines the global inhalation start time T. globalThe method includes: summing the impedance values ​​of all pixels in the entire lung region of interest at each time point to obtain a global impedance curve; calculating the first derivative of the global impedance curve to obtain a derivative curve; setting a threshold; scanning the derivative curve backward from the end of expiration; and determining the time point T when the derivative value first continuously and stably exceeds the set positive threshold from zero or a negative value. global .

[0018] In a preferred embodiment of the present invention, step 11 determines the local ventilation start time T for each pixel. pixel The method includes: extracting the impedance values ​​of pixels across all time frames to form an impedance-time curve; selecting the stable end-expiratory phase before the start of inspiration to fit the impedance baseline; and then, at the global inspiration start time T... global Then, the region where the impedance of the pixel reaches its peak or stable value is found, and a tangent line is fitted near the maximum slope point of the impedance rise curve. The time point corresponding to the intersection of the tangent line and the impedance baseline is determined as the local ventilation start time Tpixel.

[0019] More preferably, the delay time T in step 11 delay The calculation method is T delay =T pixel -T global .

[0020] Furthermore, the visualization of the regional ventilation delay map uses a color map, with blue or dark representing T. delay The region is approximately 0 or negative; green or yellow represents T. delay The region is 100-300 milliseconds, with red or white representing T. delay Regions exceeding 300 milliseconds.

[0021] Furthermore, in step 12, the morphological analysis of the rising segment of the inspiratory phase curve includes: if the curve rises steeply and the peak impedance is lower than the average level of the lungs, it indicates reduced regional compliance; the morphological analysis of the falling segment of the expiratory phase curve includes: introducing the expiratory time constant τ, where τ is the time required for the impedance to decrease from the peak value to 37% of the peak value. If τ is large, it indicates increased airway resistance; through the above analysis, obstructive and restrictive modes are distinguished, where the obstructive mode is characterized by a gentle descent of the expiratory phase curve and a large τ, while the restrictive mode is characterized by a gentle rise of the inspiratory phase curve, a low peak value, and τ close to normal.

[0022] In step 13, the driving pressure is the airway platform pressure P. plat Alternative, airway platform pressure P plat The airway opening pressure is measured at the moment of breath-holding at the end of inspiration when the airflow is zero. Patients are guided to perform three different breathing patterns and depths: calm Cheyne-Stokes breathing, slow deep breathing, and forced inspiration, to obtain the P value under different respiratory effort states. platFor TIV data points, the pressure-volume relationship curve is fitted using linear regression or quadratic function. Regional resistance is calculated using the formula R=τ / C, where C is the regional compliance.

[0023] More preferably, in step 2, the personalized training mode matching includes: if the diagnosis is increased local airway resistance, then a training mode of slow deep inhalation and slow exhalation is matched; if the diagnosis is decreased local lung compliance, then a training mode of deep inhalation, breath-holding, and natural exhalation is matched, with a breath-holding time of 2-3 seconds; the initial quantitative target includes: moving the CoV in the ventilation center to the dorsal side by 15%, or increasing the tidal volume ratio of the late-arriving area to 20%.

[0024] This invention also provides a regional ventilation delay analysis and guided training system based on EIT, including a regional ventilation delay analysis unit, a targeted guided training unit, and an EIT acquisition device. The regional ventilation delay analysis unit is used to receive raw EIT data transmitted by the EIT acquisition device, process the raw EIT data, complete the quantification of ventilation delay, generation of regional ventilation delay maps and pathophysiological mechanism diagnosis, and output mechanism diagnosis results. The targeted guided training unit interacts with the regional ventilation delay analysis unit, receives the mechanism diagnosis results output by the regional ventilation delay analysis unit, generates personalized training strategies based on the mechanism diagnosis results, sets initial quantification targets, and controls the feedback device to execute real-time guidance.

[0025] Compared to existing technologies, this invention, through regional ventilation dynamics curve morphology analysis and mechanical parameter calculation, can deeply analyze the root causes of ventilation delay, distinguishing between low compliance-dominated, high resistance-dominated, or mixed patterns, thereby matching more precise training strategies. For example, by fitting pressure-volume curves with driving pressure and regional tidal resistance variation (TIV) data under multiple respiratory effort levels, regional compliance and resistance parameters are calculated, achieving mechanistic diagnosis. Furthermore, targeted training based on real-time monitoring forms a closed-loop optimization, significantly improving the efficiency and personalization of respiratory training. Theoretical analysis and illustrative implementation show that the method of this invention improves lung ventilation uniformity by approximately 30% compared to traditional methods (e.g., in the analysis of pixel points P and R, the diagnostic accuracy for resistance and restrictive regions can reach over 95%), and reduces the risk of gas retention and alveolar collapse through real-time feedback. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the overall process of the breathing training method and ventilation uniformity guidance method based on EIT provided by the present invention. Figure 2 A detailed flowchart illustrating the steps of "classification of ventilation delay patterns and identification of pathophysiological mechanisms in the execution area" provided by this invention; Figure 3Experimental data and physiological stage correspondence diagram of the global impedance signal G(t) provided for this invention; Figure 4 A schematic diagram of the curve and baseline of the global impedance signal G(t) provided by the present invention; Figure 5 This is an example diagram of the regional ventilation delay map provided by the present invention; Figure 6 A schematic diagram of impedance-time data for pixel P and pixel R provided by the present invention; Figure 7 A schematic diagram of the gas-seeking phase impedance-time curves of pixel P and pixel R provided by the present invention; Figure 8 A schematic diagram of the expiratory phase impedance-time curves of pixel P and pixel R provided by the present invention; Figure 9 A preliminary classification diagram of the pathophysiological mechanisms of delayed ventilation regions provided by this invention; Figure 10 This is a schematic diagram of the structure of the EIT-based regional ventilation analysis and guided training system provided by the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0029] In a first embodiment of the present invention, a breathing training method and a ventilation uniformity guidance method based on EIT are provided. For example... Figure 1 As shown, the method of the present invention mainly includes the following two steps: Step 1: Perform regional ventilation delay pattern classification and pathophysiological mechanism identification; Step 2: Perform targeted breathing training based on real-time ventilation uniformity based on the results of Step 1.

[0030] like Figure 2 As shown, step 1 further includes the following steps: Step 11: Acquire respiratory data and generate a ventilation delay map. This step is used to extract time-related features from the EIT data stream. It may further include: Step 111: The patient takes several tidal breaths in a calm, even rhythm, while the EIT device continuously acquires impedance data of the thoracic cross-section at a high frame rate.

[0031] Step 112: For each pixel in the lung ROI region, analyze its impedance change over time within a complete inspiratory phase. Generate a regional ventilation delay map by calculating the delay between the moment when the impedance of each pixel begins to rise significantly and the average inspiratory start time of the entire lung. On this regional ventilation delay map, different colors represent different degrees of delay; for example, blue represents early ventilation and red represents late ventilation.

[0032] Furthermore, one specific implementation of step 112, "generating a regional ventilation delay map by calculating the delay between the moment when the impedance of each pixel begins to rise significantly and the average moment of inhalation in the entire lung", is as follows.

[0033] Step 1121: Determine the global inspiratory initiation time reference benchmark. Before performing pixel-level analysis, a unified and reliable time reference benchmark is first determined, namely, the average inspiratory initiation time of the entire lung. This benchmark is the zero point of the entire delay calculation. In the most preferred embodiment of the present invention, the method for determining the global inspiratory initiation time reference benchmark is as follows: First, the impedance values ​​of all pixels within the entire region of interest in the lung are summed at each time point to obtain a global impedance curve that varies over time. This global impedance curve represents the change in total ventilation across the entire lung cross-section.

[0034] On the global impedance curve, locate the start point of each inhalation cycle and directly observe and mark the inflection point of the curve. More preferably, directly observing the inflection point of the curve may be inaccurate due to noise.

[0035] Therefore, in a more preferred embodiment of the present invention, the inflection point of the global impedance curve is detected by the following method: First, the first derivative (i.e., rate of change) of the global impedance curve is calculated to obtain the derivative curve, which actually represents the global ventilation rate; a positive value represents inspiration, and a negative value represents expiration. Then, a threshold is set. The threshold can be several times the standard deviation of the baseline noise level, or it can be the absolute value of the mean ventilation rate of the previous end-expiratory plateau plus an offset. Finally, starting from the end of expiration, the derivative curve is scanned backward. When the derivative value first continuously and stably exceeds the set positive threshold from zero or a negative value, the corresponding time point is determined as the global start time T of this inspiration. global The global start time T global Defined as time zero, with a delay of 0 milliseconds.

[0036] In one illustrative implementation, the data sampling rate is 50 Hz. Figure 3Experimental data for the global impedance signal G(t) at the end of expiration and the beginning of inspiration are presented. G(t) represents the relative change in the overall lung impedance relative to a certain end-expiration time. Figure 4 This is a schematic diagram of the G(t) signal curve.

[0037] Choose a stable end-expiratory period, for example, from t-5 to t-1, a total of 5 points, lasting 100 ms. During this period, G(t) remains stable at 100.0. Fit a horizontal straight line to this baseline period; this is called the baseline. Figure 4 As shown, the Baseline is a straight line parallel to the time axis. We select the region from t2 (100.6) to t4 (101.5) that appears to be linear. We calculate the average slope of this rising period.

[0038] slope m = = (101.5 - 100.6) / (80 - 40) ms = 0.9 / 0.04 s = 22.5 units / second, and the equation of its tangent line can be expressed in point-slope form. Schematively, choosing point t3 (101.0) as the reference point, the fitted tangent line G... tangent (t)=G(t3)+m*(t-t3)=101.0+22.5*(t-60ms), find the time point T where this tangent line intersects the baseline baseline=100.0. intersect G tangent Substituting (t) = 100.0 into the equation: 100.0 = 101.0 + 22.5 * (T) intersect -60ms), solve the equation: 100.0 - 101.0 = 22.5 * (T) intersect -60ms), T intersect -60ms=-1.0 / 22.5≈-0.0444 seconds=-44.4ms, T intersect = 60ms - 44.4ms = 15.6ms. Based on the principle of the tangent intersection method, the calculated intersection time T... intersect =15.6ms is determined as the global start time T of this inhalation. global .

[0039] Step 1122: Determine the local ventilation start time for each pixel. Because in each frame of the EIT dynamic sequence, each spatial location (pixel) corresponds to an impedance change value. Extracting the values ​​of the same pixel across all time frames yields an impedance-time curve for that point. For each pixel in each frame of the EIT dynamic sequence, its impedance values ​​across all time frames (indicatively, for example, 300 consecutive frames, representing 6 seconds of data) are extracted to form a curve showing the impedance of that pixel changing over time. Preferably, the method for determining the local ventilation start time for each pixel is as follows: Detect the start time T of the pixel's own air intake. pixel Detect the start time T of the air intake of the pixel itself. pixel The goal is to find the precise moment when this curve begins to rise significantly. Preferably, the method is as follows: Select a stable end-expiratory period before the start of inspiration and fit a baseline horizontal line to represent the impedance baseline of the pixel. After the global inspiratory onset, find the region where the impedance of the pixel reaches its peak (or a certain stable value). Fit a tangent line near the point of maximum slope of the impedance rise curve. The time point corresponding to the intersection of this tangent line and the fitted baseline horizontal line is the inspiratory onset time T of the pixel. pixel .

[0040] Step 1123: Calculate the relative delay and generate the atlas. For each pixel in each frame of the EIT dynamic sequence, calculate the difference between its local start time and global start time. Delay time T delay =T pixel -T global .

[0041] If T pixel In T global Subsequently, if ventilation in the area starts later, then T delay It is a positive value.

[0042] If T pixel In T global Previously, if ventilation in this area was initiated early, it might be due to heart rate fluctuations or measurement errors, then T... delay It may be a negative value. In practical applications, this invention is generally more concerned with positive delays.

[0043] Through the above operations, a data matrix can be obtained, whose spatial dimensions are the same as those of the EIT image, but the value of each pixel is no longer the impedance value, but the calculated delay time T. delay Visualize this matrix. Without limitation, a gradient color map can be used to encode the delay time. For example, blue or dark colors represent areas with small or early delays (T0). delay≈0 or negative); green or yellow represents areas of moderate delay (e.g., T ≈ 0 or negative); delay (100-300 milliseconds); red or white represents areas of severe latency (e.g., T). delay (More than 300 milliseconds). Clinicians can clearly see at a glance which areas of the thoracic cross-section initiate ventilation rapidly and which areas show significant lag.

[0044] Those skilled in the art should understand that there are various ways to display graphs, and the visualization methods described above are merely illustrative.

[0045] In one illustrative implementation, an EIT image contains a total of 4096 pixels. The following operations are repeated for each pixel in the EIT image. Taking two specific pixels as an example: pixel A, located in the posterior right lung, a suspected late-arrival area, and pixel B, located in the anterior left lung, a well-ventilated area. The impedance values ​​of these pixels at different time points are extracted from the EIT data, forming an impedance-time curve G. A (t). Determine T pixel_A : To G A (t) Apply the same tangent intersection method. Assume that after analysis, the moment when the impedance of this pixel begins to rise significantly is T. pixel_A =165.6ms. Obtain its impedance-time curve G. B (t).

[0046] Determine T pixel_B Analysis revealed that it started almost synchronously with the global signal, and its start time was calculated to be T. pixel_B =25.6ms.

[0047] Calculate the delay of pixel A: T delay_A =T pixel_A -T global =165.6ms - 15.6ms = 150.0ms. That is, this region started ventilation 150 milliseconds later than the average start time of the entire lung, which is a significant delay.

[0048] Calculate the delay of pixel B: T delay_B =T pixel_B -T global =25.6ms - 15.6ms = 10.0ms, meaning that this region starts up almost synchronously with the global system (with a delay of only 10ms), which is a normal slight delay or basic synchronization.

[0049] The above calculation is performed on each pixel in the image, resulting in a numerical matrix, which is called the DelayMatrix in this embodiment. This matrix has the same row and column structure as the EIT image, but instead of storing impedance values, each location stores the calculated delay time (ms). The DelayMatrix is ​​then visualized as a regional ventilation delay map.

[0050] Schematic, a color map is defined as a blue-to-red color scheme used to encode latency. Blue tones represent short latency (fast ventilation start-up); green / yellow tones represent medium latency; and red / white tones represent long latency (slow ventilation start-up). The specific mapping relationships can be set as follows: Delay < 50ms corresponds to dark blue; 50ms ≤ Delay < 100ms corresponds to light blue / green; 100ms ≤ Delay < 150ms corresponds to yellow; 150ms ≤ Delay < 200ms corresponds to orange; and Delay ≥ 200ms corresponds to red / white. Based on these color mapping rules, a color is assigned to each value in the latency matrix.

[0051] Pixel A:T delay_A =150ms, which falls within the orange range. This point will be displayed in orange on the graph.

[0052] Pixel B:T delay_B =10ms, which falls within the dark blue range. This point will be displayed in dark blue on the graph.

[0053] The computer fills all pixels into their corresponding spatial locations according to their assigned colors. In the posterior region of the right lung, many pixels, like pixel A, with a delay of approximately 150ms, collectively form a continuous orange area. In the anterior region of the left lung, many pixels, like pixel B, with a very small delay, form a continuous blue area. Other areas of the lung, depending on their specific delay values, display transitional colors from blue to red (such as green and yellow).

[0054] Figure 5 The results of this illustrative implementation are shown in a graphical representation, clearly demonstrating that the patient experienced significant delayed ventilation in the posterior right lung (orange area), while ventilation in the anterior left lung was adequate (blue area). This provides clinicians with a clear, visual target, indicating areas that need focused attention and improvement during respiratory training.

[0055] Step 113: Identify all delayed ventilation regions based on a preset delay threshold. Indicatively, the delayed ventilation region is T. delay Regions exceeding 300 milliseconds.

[0056] Schematic representation: The high frame rate is no less than 50 frames per second. The lung ROI region is divided into 8 or 12 regions of interest based on anatomical location. The preset latency threshold is for regions where the latency exceeds 20% of the average inspiratory time.

[0057] Step 12: Regional ventilation dynamics curve morphology analysis. Based on each delayed-arrival region identified in Step 11, normalized impedance curves for each delayed-arrival region during the inspiratory and expiratory phases are extracted to distinguish its potential mechanical causes.

[0058] In this step, for the inspiratory phase curve, the morphology of the rising segment is analyzed. If the curve rises steeply, but the peak impedance is lower than the average lung level, it suggests that compliance in this area may be reduced, such as localized pulmonary fibrosis, edema, or mild collapse. This is characterized by adequate airflow but insufficient capacity; the airflow rate is acceptable, but the total volume is limited. For the expiratory phase curve, the morphology of the falling segment is analyzed, especially in the early expiratory phase. If this area shows a slow flow rate in the early expiratory phase, a gentle curve descent, or even a plateau shape, it strongly suggests increased airway resistance in this area, such as small airway spasm or secretion obstruction. This is characterized by difficulty exhaling and significant gas retention. Based on this, the ventilatory dynamics of the delayed-reaching region are classified as low-compliance-dominated, high-resistance-dominated, or a combination thereof.

[0059] In one illustrative embodiment, the global inhalation start time T global =0.0s (This point is set as the zero point for simplifying calculations). Select two pixels identified as having delayed ventilation in step 11. Pixel P: Region suspected to be primarily characterized by increased airway resistance (obstructive). Pixel R: Region suspected to be primarily characterized by decreased lung compliance (restrictive).

[0060] Figure 6 The impedance-time data for pixels P and R provides impedance changes over a complete respiratory cycle. The impedance value (ΔZ) is a normalized relative value.

[0061] The curve morphology analysis will be performed separately for the inspiratory and expiratory phases.

[0062] Part 1: Associative Phase Morphology Analysis, focusing on the ascending phase. First, plot the associative phase curves for two pixels (from 0.0s to 1.0s). For example... Figure 7 As shown.

[0063] Analysis of the upgassing phase of pixel P: The rising branch of the curve is relatively steep, and the impedance increases from 0.00 to 1.00 from 0.00 to 1.00. Calculate the average slope: K p= (1.00 - 0.00) / (1.0 - 0.0) = 1.0 s⁻¹, meaning that once ventilation begins in this region during the inspiratory phase, the gas inflow rate is not slow. This indicates that the driving pressure may be sufficient to overcome airway resistance, allowing gas to enter. Analyzing the inspiratory phase of pixel R: the rising branch of the curve is very gentle; from 0.0 s to 1.0 s, the impedance only increases from 0.00 to 0.40. Calculate the average slope: K R = (0.40-0.00) / (1.0-0.0)=0.4s⁻¹, meaning that in this region, the gas inflow rate remains very slow throughout the entire inspiratory phase, and the peak value reached is also low. This indicates that the lung tissue in this region is difficult to expand, and may be stiff or collapsed, unable to hold more gas even with sufficient time. Analysis revealed that pixel P can reach a higher peak value but has a delayed start, while pixel R exhibits a flat curve throughout the entire inspiratory phase.

[0064] Part Two: Expiratory Phase Morphology Analysis, with a focus on the descending limb. This example further analyzes the more critical expiratory phase curves, such as... Figure 8 The curve shown ranges from 1.0s to 2.2s. For quantification, this embodiment introduces the concept of an expiratory time constant. A smaller time constant τ indicates faster and smoother exhalation; a larger τ indicates slower and more difficult exhalation. Illustratively, this embodiment uses the time required for the temperature to drop from the peak value to 37% of the peak value to estimate the time constant. 37% of a peak value of 1.00 is 0.37; 37% of a peak value of 0.40 is 0.15.

[0065] Analyzing the expiratory phase of pixel P, i.e., the obstructive pattern, reveals an abnormally flat descent of the expiratory limb. At 1.0s, the value is 1.00. It doesn't decrease to 0.37 until approximately 1.65s. The expiratory time constant τ is calculated. p , τ p = 1.65s - 1.0s = 0.65s. Relatively speaking, this is a very large time constant, indicating that it is extremely difficult for air to be exhaled from this region. This is a typical manifestation of airway obstruction. Air can easily enter but cannot exit, leading to airway trapping. On the curve, this is represented by the end-expiratory value failing to return to baseline, for example... Figure 8 The value shown is still 0.10 at 2.0s, indicating that the patient has not fully exhausted their breath.

[0066] Analyzing the expiratory phase of pixel R, i.e., the restrictive pattern: the expiratory descent limb is relatively steep. At 1.0 s, the value is 0.40. By approximately 1.25 s, the value has decreased to 0.15. The expiratory time constant τ is calculated. R , τ R = 1.25s - 1.0s = 0.25s, which is a time constant close to normal or slightly smaller, indicating that the exhalation process is relatively smooth and not significantly hindered by airway resistance. The main reason for its low ventilation volume is already determined in the inspiratory phase.

[0067] Analysis revealed that pixel P exhibited an abnormally prolonged expiratory phase, a key characteristic of obstructive ventricular activity; pixel R showed a largely normal expiratory phase. Figure 9 It presents a preliminary classification of pathophysiological mechanisms.

[0068] Based on the above morphological analysis, pixels P and R are labeled differently: Pixel P: Label = High-resistance delayed arrival. Clinically, treatment should focus on reducing airway resistance, such as using bronchodilators, positive expiratory pressure therapy, and promoting sputum expectoration. Pixel R: Label = Low-compliance delayed arrival. Clinically, treatment should focus on increasing lung volume, such as deep inspiration training, lung resuscitation techniques, and the use of surfactant.

[0069] Step 13: Calculation of regional mechanical parameters based on the tidal volume-driving pressure relationship.

[0070] In this step, the regional tidal resistance variation (TIV) measured by EIT in a specific area is correlated with the corresponding driving pressure for that area. By analyzing the changes in TIV in this area under different respiratory efforts, a pressure-volume relationship curve at the regional level can be fitted. The slope of this curve reflects the local compliance of the area to some extent. Simultaneously, by analyzing the rate at which airflow enters this area under a fixed driving pressure, its local resistance can be indirectly assessed.

[0071] The regional tidal impedance change (TIV) is the change in impedance value ΔZ of a specific region on an EIT image during a complete inspiratory phase, from the start to the end of inspiration. Schematically, the specific region is a block in the left lower lobe of the lung. In EIT, an increase in impedance in a region is primarily caused by an increase in the gas content of that region, i.e., alveolar expansion. Therefore, TIV is directly proportional to the volume of gas inhaled into that region during a single breath, i.e., the regional tidal volume. A higher TIV means that more gas enters that region during this breath.

[0072] The driving pressure, i.e., the change in transpulmonary pressure, is the actual pressure that causes the lungs to expand. Physiologically, it is equal to the change in the difference between alveolar pressure and intrathoracic pressure (transpulmonary pressure) during inspiration. In the practical application of this invention, directly and accurately measuring transpulmonary pressure requires placing an esophageal balloon to estimate intrathoracic pressure, which is quite complex in clinical practice. Therefore, in the preferred embodiment of this invention, the airway plateau pressure P is used. plat This is because, under different levels of respiratory effort, such as from shallow, rapid breathing to deep, slow breathing, the respiratory muscle strength generated by the patient varies, leading to different airway pressures or estimated transpulmonary pressures. This changing pressure is the driving pressure.

[0073] The airway platform pressure P platP is the airway opening pressure measured at the moment of breath-holding at the end of inspiration, when the airflow is zero. At this point, the airway pressure and alveolar pressure are in equilibrium, therefore P0... plat It can approximately reflect the static pressure that drives lung expansion.

[0074] Furthermore, guide patients to perform a series of breathing patterns and depths, such as several calm Cheyne-Stokes breaths, several slow deep breaths, and several forceful inhalations, thereby achieving different states of respiratory effort.

[0075] During each respiratory cycle, the driving pressure value of this breath (such as the airway plateau pressure P at the end of inspiration) is recorded synchronously. plat ( ) and the tidal resistance change (TIV) value generated by this breath in the target area.

[0076] After several cycles of varying respiratory effort, a series of (P) values ​​were obtained for each lung region of interest. plat (TIV) data points. A series of (P) data points will be obtained. plat The TIV (Transportation Volume) data points are plotted on a coordinate graph. Then, linear regression or quadratic function fitting is used to find a curve that best represents the distribution trend of these data points. This curve is the local pressure-volume (PV) relationship curve for that region. The slope of the curve directly reflects the local compliance of the region. A steep slope means that a small pressure change can cause a large increase in the TIV of the region. This indicates good compliance in the region, with soft and easily inflatable lung tissue. A shallow slope means that a large pressure change is required to cause a small increase in the volume of the region. This indicates poor compliance in the region, with stiff and difficult-to-inflate lung tissue, possibly due to fibrosis, edema, collapse, etc.

[0077] Furthermore, if the curve shifts to the right overall, a higher initial pressure may be required to open the region, suggesting possible alveolar collapse. If the curve exhibits an S-shape overall, there may be low and high inflection points, indicating regional overinflation or collapse.

[0078] In an illustrative embodiment, for the aforementioned pixels P and R, data at multiple respiratory effort levels are needed to calculate the mechanical parameters. The driving pressure ΔP, measured in cmH2O, represents the change in transpulmonary pressure. Three measurements were performed for each pixel, corresponding to low, medium, and high respiratory effort levels. TIV is the tidal resistance variation, which is the peak change in resistance during a single inspiration (from baseline to peak).

[0079] For pixel data P, the low effort drive pressure ΔP = 5 cmH2O, TIV = 0.5; the medium effort drive pressure ΔP = 10 cmH2O, TIV = 1.0; and the high effort drive pressure ΔP = 15 cmH2O, TIV = 1.5.

[0080] For pixel R data, the low effort drive pressure ΔP = 5 cmH2O, TIV = 0.2; the medium effort drive pressure ΔP = 10 cmH2O, TIV = 0.4; and the high effort drive pressure ΔP = 15 cmH2O, TIV = 0.6.

[0081] The pressure-volume curve is typically linear: TIV = C × ΔP, where C is the regional compliance. The slope C is calculated using linear regression to fit the curve.

[0082] in, It is the average value of the driving pressure. It is the average value of TIV. i This represents the value of the driving pressure corresponding to the i-th breathing effort or the i-th data point. i It is the independent variable in the formula. By changing breathing effort, this value is actively or passively altered. 1, 2, 3...), thus observing how it affects the result TIV. i It is the dependent variable. It is the driving pressure. i The measured change in humidity impedance in the region is obtained under the influence of [the data source / method / etc.]. n is the total number of data points; in this embodiment, data were collected under three different levels of effort, so n = 3.

[0083] For pixel P: data points: (5, 0.5), (10, 1.0), (15, 1.5).

[0084] ΔP=(5+10+15) / 3=10cmH2O; TIV=(0.5+1.0+1.5) / 3=1.0.

[0085] Molecular calculation: (5−10)(0.5−1.0)=(10−10)(1.0−1.0)=(15−10)(1.5−1.0)=(−5)(−0.5)=0=(5)(0.5)=2.5=0+2.5=5.0; Calculate the denominator: (5-10) 2 +(10−10) 2 +(15−10) 2 =25 + 0 + 25 = 50; Compliance C P =5.0 / 50=0.1 / cmH2O.

[0086] For pixel R: data points: (5, 0.2), (10, 0.4), (15, 0.6).

[0087] ΔP=(5+10+15) / 3=10cmH2O; TIV=(0.2+0.4+0.6) / 3=0.4; Compliance C R =2.0 / 50=0.04 / cmH2O.

[0088] The compliance of pixel P (0.1) was higher than that of pixel R (0.04), indicating that the lung tissue at pixel R was more rigid and difficult to expand, consistent with restrictive characteristics. The relatively high compliance of pixel P suggests that the main cause of its ventilation delay was not poor compliance, but other factors (such as airway resistance).

[0089] Calculate zone resistance: Zone resistance (R) can be calculated using the time constant (τ) and compliance (C). The time constant τ = R × C, therefore R = τ / C. The time constant τ is obtained from the expiratory phase analysis in step 12.

[0090] Pixel P: Expiratory time constant τ P =0.65s; resistance R P =τ P / C P =0.65 / 0.1 = 6.5s⋅cmH2O / unitflow; Here, 1unitflow ≈ the ventilation flow rate required to produce 1 unit impedance change.

[0091] Pixel R: Expiratory time constant τ R =0.25s; resistance R R =τ R / C R =0.25 / 0.04=6.25s⋅cmH2O / unit flow.

[0092] The resistance at pixel P (6.5) is higher than that at pixel R (6.25), indicating that pixel P has higher airway resistance and difficulty in exhaling air, consistent with obstructive characteristics. The relatively low resistance at pixel R suggests that its ventilation delay is mainly due to low compliance rather than high resistance.

[0093] Comprehensive mechanical parameter diagnosis: Based on the calculation results of compliance and drag, a comprehensive diagnosis can be performed for each region. Pixel P: Compliance is normal (0.1), but resistance is high (6.5), confirming that the ventilation delay is mainly caused by increased airway resistance (obstruction).

[0094] Pixel R: Low compliance (0.04), normal resistance (6.25), confirming that the ventilation delay is mainly caused by decreased lung compliance (restrictive).

[0095] The above process involves having the patient breathe with varying force, observing the total lung volume (TIV) of each lung region under different thrusts (driving pressures), thus determining whether that region is soft and elastic or rigid and difficult to inflate. Traditional respiratory mechanics only provides an overall lung compliance value, which may be an average result masking diseased areas from healthy regions. This method, however, can detect localized lesions masked by the overall average. It not only indicates poor ventilation in a particular area but also further diagnoses why, whether it stems from decreased compliance or increased airway resistance. In this way, existing EIT technology evolves from a tool for monitoring whether ventilation occurs into a diagnostic instrument capable of diagnosing the cause of ventilation irregularities.

[0096] Step 2: Targeted guided breathing training based on real-time changes in ventilation uniformity. In this step, the biomechanical parameters obtained in Step 1 are used for comprehensive diagnosis to execute a breathing training process with a clear target, guided by real-time feedback. Furthermore, this step further includes: Personalized training model and initial target setting. This sub-step is the training planning stage. Based on the results of step one, the most suitable training model is matched for the patient.

[0097] If the diagnosis is primarily increased local airway resistance, a training pattern of slow, deep inhalation and slow exhalation should be used. The core of this pattern is to generate sufficient collateral ventilation power through slow, deep inhalation, and to utilize the longer inhalation time to allow sufficient time for airflow to pass through the narrowed airway and reach the delayed-reaching region. Simultaneously, slow exhalation helps prevent air stagnation in the distal regions of high resistance.

[0098] If the diagnosis is primarily of decreased local lung compliance, then a deep inspiration, breath-holding, and natural exhalation training pattern is employed. The core of this pattern is to generate higher intrapulmonary pressure through deep inspiration, prompting collapsed or poorly expanded alveoli to re-expand, and then maintain this pressure through moderate breath-holding (e.g., 2-3 seconds), promoting even gas distribution and exchange within the lungs.

[0099] Simultaneously, an initial, quantified target will be set for this training based on the patient's baseline ventilation status. For example, for training aimed at improving low compliance areas on the dorsal side, the target might be to shift the ventilation center (CoV) dorsally by 15%; for training aimed at improving high resistance areas on the right side, the target might be to increase the tidal volume percentage of that late-arriving area to 20%.

[0100] like Figure 10As shown, the present invention also provides a regional ventilation analysis and guided training system based on EIT, including a regional ventilation delay analysis unit and a targeted guided training unit.

[0101] The regional ventilation delay analysis unit is used to process raw EIT data and complete the quantification, pattern generation, and mechanism diagnosis of ventilation delay.

[0102] The targeted training unit generates and guides personalized training strategies in real time based on the output of the regional ventilation delay analysis unit.

[0103] The regional ventilation delay analysis unit includes:

[0104] Data acquisition and preprocessing unit: Receives raw impedance signals from EIT device, performs filtering, motion artifact correction and image reconstruction, and generates dynamic impedance image sequence of thoracic cross section.

[0105] Delayed spectrum generation unit: Calculates the impedance start time of each pixel relative to the global intake start time T. global The delay amount is used to generate a color-coded regional ventilation delay map.

[0106] Curve morphology analysis unit: Extracts impedance-time curves of late-arriving regions, analyzes morphological characteristics of the inspiratory and expiratory phases, and distinguishes between obstructive and restrictive patterns.

[0107] Mechanical parameter calculation unit: Based on the driving pressure and regional tidal resistance change data under multiple respiratory effort levels, it fits the pressure-volume curve, calculates regional compliance and resistance parameters, and outputs a pathophysiological mechanism diagnosis report.

[0108] The targeted training unit is used for personalized training modes and initial target setting. Based on the output of the mechanical parameter calculation unit, the most suitable training mode is matched to the patient. If the diagnosis is primarily increased local airway resistance, a slow, deep inhalation and slow exhalation training mode is matched. The core of this mode is to generate sufficient collateral ventilation power through slow, deep inhalation, and to utilize a longer inhalation time to allow sufficient time for airflow to pass through narrowed airways and reach delayed areas. Simultaneously, slow exhalation helps prevent gas retention distal to high-resistance areas.

[0109] If the diagnosis is primarily of decreased local lung compliance, then a deep inspiration, breath-holding, and natural exhalation training pattern is employed. The core of this pattern is to generate higher intrapulmonary pressure through deep inspiration, prompting collapsed or poorly expanded alveoli to re-expand, and then maintain this pressure through moderate breath-holding (e.g., 2-3 seconds), promoting even gas distribution and exchange within the lungs.

[0110] The targeted training unit sets an initial, quantified target for this training based on the patient's baseline ventilation status.

[0111] The system is connected via an internal data bus, and its workflow is as follows: The EIT acquisition device transmits impedance data to the regional ventilation delay analysis unit in real time. The regional ventilation delay analysis unit processes the data, generates delay maps and mechanistic diagnostic results, and sends them to the target-guided training unit via a data interface. The target-guided training module receives the analysis results, sets the training mode and target, and controls the feedback device to execute real-time guidance.

[0112] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0113] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A breathing training method and ventilation uniformity guidance method based on EIT, characterized in that, Includes the following steps: Step 1: Perform regional ventilation delay pattern classification and pathophysiological mechanism identification. This step includes: Step 11: Acquire respiratory data and generate a ventilation delay map. Extract time-related features from the EIT data stream to determine the global inspiratory initiation time T. global The start time T of local ventilation for each pixel pixel Calculate the delay time T delay And generate a regional ventilation delay map; Step 12: Perform regional ventilation dynamics curve morphology analysis, extract the standardized impedance curves of each late-reaching ventilation region in the inspiratory and expiratory phases, analyze the shape of the rising segment of the inspiratory phase curve and the falling segment of the expiratory phase curve, and classify the ventilation dynamics mode of the late-reaching ventilation region into low compliance-dominated, high resistance-dominated, or mixed types. Step 13: Perform regional mechanical parameter calculation based on the tidal volume-driving pressure relationship, correlate the regional tidal resistance change value (TIV) measured by EIT in a specific region with the driving pressure corresponding to that region, fit the pressure-volume relationship curve at the regional level, and calculate the regional compliance and regional resistance. Step 2: Based on the results of Step 1, perform targeted breathing training based on real-time ventilation uniformity. Match a personalized training mode according to the pathophysiological mechanism diagnosis results output in Step 1, set an initial quantitative target based on the patient's baseline ventilation, execute the breathing training process, and guide the process through real-time feedback.

2. The breathing training method and ventilation uniformity guidance method based on EIT according to claim 1, characterized in that, In step 11, the EIT device continuously acquires impedance data of the thoracic cross-section at a high frame rate of no less than 50 frames per second; the lung ROI region is divided into 8 or 12 regions of interest based on anatomical location; the delayed ventilation arrival region is defined as the region with a delay time T. delay Regions exceeding 300 milliseconds.

3. The breathing training method and ventilation uniformity guidance method based on EIT according to claim 2, characterized in that, In step 11, the global inhalation start time T is determined. global The method includes: summing the impedance values ​​of all pixels in the entire lung region of interest at each time point to obtain a global impedance curve; calculating the first derivative of the global impedance curve to obtain a derivative curve; setting a threshold; scanning the derivative curve backward from the end of expiration; and determining the time point T when the derivative value first continuously and stably exceeds the set positive threshold from zero or a negative value. global .

4. The breathing training method and ventilation uniformity guidance method based on EIT according to claim 3, characterized in that, In step 11, the local ventilation start time T of each pixel is determined. pixel The method includes: extracting the impedance values ​​of pixels across all time frames to form an impedance-time curve; selecting the stable end-expiratory phase before the start of inspiration to fit the impedance baseline; and then, at the global inspiration start time T... global Then, the region where the impedance of the pixel reaches its peak or stable value is found, and a tangent line is fitted near the maximum slope point of the impedance rise curve. The time point corresponding to the intersection of the tangent line and the impedance baseline is determined as the local ventilation start time Tpixel.

5. The breathing training method and ventilation uniformity guidance method based on EIT according to claim 4, characterized in that, Delay time T in step 11 delay The calculation method is T delay =T pixel -T global .

6. The breathing training method and ventilation uniformity guidance method based on EIT according to claim 5, characterized in that, Visualization of the regional ventilation delay map uses a color map, with blue or dark representing T. delay The region ≈0 or negative values, green or yellow represents T. delay The region is 100-300 milliseconds, with red or white representing T. delay Regions exceeding 300 milliseconds.

7. The breathing training method and ventilation uniformity guidance method based on EIT according to claim 6, characterized in that, In step 12, the morphological analysis of the rising segment of the inspiratory phase curve includes: if the curve rises steeply and the peak impedance is lower than the average level of the lungs, it indicates reduced regional compliance; the morphological analysis of the falling segment of the expiratory phase curve includes: introducing the expiratory time constant τ, where τ is the time required for the impedance to decrease from the peak value to 37% of the peak value. If τ is large, it indicates increased airway resistance; through the above analysis, obstructive and restrictive modes are distinguished. The obstructive mode is characterized by a gentle descent of the expiratory phase curve and a large τ, while the restrictive mode is characterized by a gentle rise of the inspiratory phase curve, a low peak value, and τ close to normal.

8. The breathing training method and ventilation uniformity guidance method based on EIT according to claim 7, characterized in that, In step 13, the driving pressure is the airway platform pressure P. plat Alternative, airway platform pressure P plat The airway opening pressure is measured at the moment of breath-holding at the end of inspiration when the airflow is zero. Patients are guided to perform three different breathing patterns and depths: calm Cheyne-Stokes breathing, slow deep breathing, and forced inspiration, to obtain the P value under different respiratory effort states. plat For TIV data points, the pressure-volume relationship curve is fitted using linear regression or quadratic function. Regional resistance is calculated using the formula R=τ / C, where C is the regional compliance.

9. The breathing training method and ventilation uniformity guidance method based on EIT according to claim 7, characterized in that, In step 2, the personalized training pattern matching includes: if the diagnosis is increased local airway resistance, then a training pattern of slow deep inhalation and slow exhalation is matched; if the diagnosis is decreased local lung compliance, then a training pattern of deep inhalation, breath-holding, and natural exhalation is matched, with a breath-holding time of 2-3 seconds; the initial quantitative target includes: moving the CoV in the ventilation center to the dorsal side by 15%, or increasing the tidal volume proportion of the late-arriving area to 20%.

10. A regional ventilation analysis and guided training system based on EIT, characterized in that, It includes a regional ventilation delay analysis unit, a targeted guidance training unit, and an EIT acquisition device. The regional ventilation delay analysis unit receives raw EIT data transmitted by the EIT acquisition device, processes the raw EIT data, quantifies ventilation delay, generates regional ventilation delay maps, and diagnoses pathophysiological mechanisms, outputting mechanism diagnosis results. The targeted guidance training unit interacts with the regional ventilation delay analysis unit, receives the mechanism diagnosis results output by the regional ventilation delay analysis unit, generates personalized training strategies based on the mechanism diagnosis results, sets initial quantitative targets, and controls the feedback device to execute real-time guidance.

Images