Detailed Description
It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
The embodiment of the application provides a portable blood oxygen signal OSA intelligent detection method, which comprises the following steps:
Step a, continuously collecting photoelectric volume pulse wave signals, and detecting pulse wave peaks of the photoelectric capacitance pulse wave signals in real time to obtain pulse wave peak value sequences;
step b, generating an amplitude differential sequence reflecting the amplitude change of the pulse wave and a period differential sequence reflecting the period interval change of the pulse in parallel based on the pulse wave crest value sequence;
c, analyzing the rhythm synchronicity of the amplitude differential sequence and the periodic differential sequence in a preset time window, wherein the rhythm synchronicity is obtained by calculating the symbol consistency ratio of the elements of the amplitude differential sequence and the corresponding elements of the periodic differential sequence, and the symbol consistency ratio represents the symbol consistency degree of the two sequences;
Step d, comparing the symbol consistency ratio obtained in real time with a reference symbol consistency ratio, wherein the reference symbol consistency ratio is determined in a stable breathing state of a user in an initial monitoring stage;
And e, when the real-time symbol consistency ratio is continuously lower than a first preset threshold value of the reference symbol consistency ratio, determining that the current signal segment is a rhythm desynchronization state signal segment.
Preferably, in the step c, the symbol consistency ratio is calculated by counting the number of times that the product symbol of the element of the amplitude differential sequence and the corresponding element of the periodic differential sequence is positive in a preset time window, and taking the ratio of the number of times to the total heartbeat number in the time window as the symbol consistency ratio.
Preferably, the reference symbol uniformity ratio is adaptively established by calculating a moving average of the symbol uniformity ratio in a steady breathing state during an initial period of user monitoring.
Preferably, the method further comprises the step of outputting a high risk level alert when the cumulative duration or frequency of occurrence of the determined dyssynchrony status signal segment reaches or exceeds a preset reference threshold.
Preferably, after the step e of judging that the rhythm desynchronization state signal segment occurs, the method further comprises the steps of calculating a second order differential value of the amplitude differential sequence and a second order differential value of the periodic differential sequence at the occurrence time point of the rhythm desynchronization state signal segment, and if any one of the absolute value of the second order differential value of the amplitude differential sequence or the absolute value of the second order differential value of the periodic differential sequence is larger than a preset morphology discrimination threshold, marking the event as a high-frequency interference segment and reducing or eliminating the weight of the event when the risk index calculation is performed.
Preferably, the preset morphology discrimination threshold is set based on historical physiological data for distinguishing relatively smooth changes induced by respiratory effort from transient drastic changes induced by high frequency signal disturbances of the respiratory source.
Preferably, the method further comprises the steps of continuously monitoring the body movement of the user through the acceleration sensor, and suspending the judgment in the step e when the body movement amplitude is detected to be larger than a preset body movement threshold value so as to avoid interference of body movement artifact on judgment of the rhythm desynchronization state signal segment.
Preferably, the method further comprises the steps of continuously calculating the modulation energy of the amplitude differential sequence and the modulation energy of the periodic differential sequence during the period when the occurrence of the dyssynchrony state signal segment is not determined in the step e, the modulation energy being obtained by calculating the variance of the differential sequence during the period, and when the modulation energy of the amplitude differential sequenceAnd the modulated energy of the periodic differential sequenceAnd identifying a specific signal segment representing the energy silence of the signal modulation when the energy duration is continuously lower than the lower limit threshold of the respective energy base line within the preset duration.
Preferably, the lower threshold of the energy base line is set to be the lower proportion value of the corresponding modulation energy in the normal and steady sleep state of the user, and the range of the lower proportion value is 0.1 to 0.3.
Preferably, the modulation energy base line in the normal and steady sleep state is adaptively established by moving average of the variances of the amplitude differential sequence and the period differential sequence in the steady respiration state of the user monitoring initial stage, and meanwhile, in the invention, the setting of the reference symbol consistency ratio is adaptively established based on the steady respiration state of the user in the initial monitoring stage and by calculating the moving average value of the symbol consistency ratio at the moment, the method can dynamically adjust the reference value to adapt to individual physiological differences. In specific application, a user calculates a preliminary reference value through stable respiration state data in an initial state of monitoring, then the system carries out self-adaptive updating according to data monitored by the user for a long time, the updating period of the reference value is closely related to the stability of the monitored data, when the system detects that the respiration state is obviously changed, the reference value is corrected according to new data so as to ensure that the respiration characteristics of each user are fully considered, for preset thresholds, the adaptability to individual differences of different users is firstly considered, the setting of the thresholds is not only based on the percentage change of the consistency ratio of the reference symbol, but also the individual physiological characteristics of the user is considered, the selection range of the preset thresholds is usually between 20% and 30% of the reference value, the specific range is optimized according to the actual test and the physiological response of the user, and the threshold range is an engineering trade-off between normal fluctuation based on the respiration rhythm change and the specific mode change of possible signals so as to ensure that the system can effectively identify the respiration abnormality of a patient with light and medium OSA, and meanwhile, the misjudging normal physiological fluctuation is avoided. The reasonable setting of the threshold value is to achieve the optimal balance between the sensitivity and the specificity, so as to ensure the accuracy of recognition; in terms of event recognition, the present invention judges an ineffective respiratory effort event and a rhythm dyssynchrony state signal segment by analyzing the rhythm synchronicity of an amplitude differential sequence and a period differential sequence, wherein the ineffective respiratory effort event generally refers to an improvement of airflow caused by failure of respiratory effort of a patient when an airway is blocked, such event is characterized by occurrence of a rhythm decoupling phenomenon, which is reflected by a significant decrease in a symbol coincidence rate, and can be detected in an early stage before blood oxygen saturation decreases, and becomes an important basis for early screening of light and medium OSA, while the rhythm dyssynchrony state signal segment refers to a respiratory effort event caused by complete or partial obstruction of an airway, and is characterized in that the symbol coincidence rate is continuously decreased and is accompanied by an increase in physiological interference when monitored by an oxygen signal, and therefore, by comparing the rhythm synchronicity rate with a reference value, the system can effectively distinguish the two events, and in order to ensure accuracy of the process, the system also combines a second order differential calculation and a morphology mechanism, recognizes different types of artifacts, for example, when a preset differential value of a signal is eliminated, and the morphology is beyond a discrimination threshold, and a final discrimination is judged that the morphology is beyond a final discrimination threshold. The method is characterized in that in order to prevent the interference of body movement artifact on the judgment of a rhythm desynchronization state signal segment, a system monitors the body movement of a user in real time through an acceleration sensor and decides whether to suspend event judgment according to a preset body movement amplitude threshold. When the body movement amplitude exceeds a preset body movement threshold, the system pauses the judgment of the respiratory event of the current data segment, and the selection standard of the body movement threshold is set according to the typical body movement amplitude variation range of the wearing position of the equipment. Specifically, in general, the body movement amplitude threshold is set at about 0.5G to ensure that the body movement is not misjudged as an artifact in the normal body movement range, and the body movement monitoring is set to take the adaptability under different use environments into consideration, for example, the body movement amplitude varies in different sleep stages, sleep postures and living environments, so that the system adjusts the body movement threshold according to the long-term monitoring data of the user to ensure that the influence of the body movement artifact is minimized under the changeable environment, which is an extension implementation known to those of ordinary skill in the art.
The embodiment 1 provides a portable blood oxygen signal OSA intelligent detection method, which is based on a photoelectric volume pulse wave signal acquired by a single blood oxygen sensor, and further calculates the synchronization degree of the two differential sequences on a rhythm layer by extracting a pulse wave crest value sequence and constructing an amplitude differential sequence and a period differential sequence corresponding to the pulse wave crest value sequence, and accordingly identifies a rhythm synchronization state signal section, performs artifact removal by combining morphological characteristics on the basis, and judges a specific signal section representing signal modulation energy silence when a specific accumulation condition is met; in particular, in the signal acquisition stage, a conventional photoelectric volume sensor is adopted to realize continuous sampling of pulse wave signals of a user, pulse waveforms reflecting heart beat activities are obtained usually through photoelectric detection carried out at the positions of fingertips or earlobes, the system carries out peak extraction on the pulse waveforms in real time to construct a pulse wave crest value sequence, the peak value sequence provides basic data for subsequent differential processing, the stability of the peak value sequence directly influences the accuracy of the whole judgment flow, a multiscale smoothing and pseudo-peak correction strategy can be adopted in actual implementation to improve the peak extraction robustness under the condition of low signal-to-noise ratio, in addition, in the construction process of the pulse wave peak value sequence, in order to ensure the stability and accuracy of peak extraction, the amplitude threshold setting is obtained based on the average peak-valley value of input signals in a static section, the peak-valley value is typically set to be fifty percent of the average value, the peak value is used for eliminating pseudo-peak responses caused by instantaneous disturbance or noise, and a pseudo-peak correction mechanism is introduced by the threshold strategy, and the system specifically comprises firstly, after the peak candidate points are initially detected, a time window is set according to the duration of typical pulse cycle of adjacent heart beat cycles, judging whether the candidate points form local maximum centers, if a plurality of adjacent candidate points exist, keeping the maximum amplitude in a local time window, and removing the rest, and secondly, triggering adjacent peak re-comparison and replacement processes by a system under the condition that the distance between two adjacent peaks is obviously lower than the lower limit of an average heart beat period so as to correct a false overcritical peak structure induced by high-frequency interference and ensure that a pulse wave peak sequence finally formed has good time alignment characteristic and amplitude representativeness. Based on the peak value sequences, the system respectively constructs two differential sequences, namely an amplitude differential sequence used for representing amplitude variation between the peaks of continuous pulse waves, wherein elements of the amplitude differential sequence are composed of differences of adjacent peaks, a period differential sequence used for reflecting time interval variation between the peaks, elements of the period differential sequence are composed of variation quantity of continuous heart beat intervals, the two differential sequences respectively reveal local fluctuation of signals from two dimensions of an amplitude domain and a time domain and are key characteristic dimensions for identifying specific mode respiratory effort states of the signals, the system sets a sliding time window used for calculating rhythm synchronization indexes between the two differential sequences in real-time monitoring, namely a symbol consistency ratio, wherein the ratio is defined as the number of times that the product symbol of the corresponding element of the amplitude differential sequence and the period differential sequence is positive, and the number of times is compared with the total heart beat number in the time window, so that the obtained symbol consistency ratio can be used for quantifying the consistency degree of the two differential sequences in the symbol direction, and accordingly reflecting the rhythm coupling relation of respiratory modulation signals in pulse waves.
In the real-time detection, if the system finds that the symbol consistency ratio in the current period is continuously lower than a first preset threshold value and the descending state continuously exceeds a set minimum rhythm, the system considers that the synchronization state is possible to happen, wherein the preset threshold value is set to be in a state of being less than a preset minimum rhythm, the system sets up the synchronization state, the error rate is not equal to the preset threshold value, the error rate is reduced to be in a normal state, the error rate is not equal to the preset threshold value, and the error rate is not equal to the normal value, the system can be in a normal state, and the physiological error rate is not equal to the normal value.
In order to further improve the judgment accuracy, after the preliminary judgment event occurs, the system also calculates a second order difference value for the amplitude difference sequence and the periodic difference sequence, judges whether the absolute value of the second order difference value exceeds a morphological judgment threshold value, if the absolute value of the second order difference value exceeds the threshold value range, the secondary event can be judged as an artifact event caused by a heart factor so as to be eliminated, the morphological judgment threshold value is set according to historical physiological data and is usually used for distinguishing a relative slowly varying phenomenon caused by respiratory effort from severe fluctuation caused by sudden events such as high-frequency signal disturbance of a non-respiratory source, and the like, so that the recognition capability of the system to the non-respiratory factor disturbance is enhanced. Further, during the period when no rhythm desynchronization state signal segment is detected, the system calculates the modulation energy of the amplitude differential sequence and the period differential sequence continuously, the modulation energy is obtained by calculating the variance of the differential sequence in the period, so as to quantify the dynamic activity degree of the pulse wave signal, when the two modulation energy values simultaneously continuously fall below the respective energy base line lower limit threshold values in the set period, a specific signal segment representing the silence of the modulation energy is identified, and during the period when no rhythm desynchronization state signal segment is detected, the system has clear engineering boundary for ensuring the identification of the specific signal segment representing the silence of the modulation energy of the amplitude differential sequence and the period differential sequence, the duration can be set to be not less than the period of a complete respiratory rhythm, and generally corresponds to a sliding analysis window of thirty seconds to sixty seconds, the selection of the period is based on the specific signal segment representing the silence of the modulation energy is physiologically significant continuously characterized, and during the period when the silence mechanism of the central air flow is driven to be lost, The respiratory motion and blood flow modulation tend to silence, and thus the dual significant attenuation of the modulation energy in the pulse wave form is caused, if the modulation energy values of the amplitude differential sequence and the period differential sequence are continuously kept below the lower limit threshold of the energy base line in the time window, and the state continues to the window ending time point without interruption, the system judges that the signal segment is the occurrence interval of a specific signal segment representing the silence of the modulation energy of the signal, the judging logic is based on the characteristic that the central apnea leads to the complete interruption of the respiratory drive, and further shows that the silence of the signal is reduced, so that the low modulation energy becomes an important indicator of the type of event, the lower limit threshold of the modulation energy is set according to the corresponding modulation energy level of a user in a normal and steady sleep state, and the lower limit threshold of the modulation energy is usually ten percent to thirty percent of the level as the lower limit of the proportion, meanwhile, the energy base line value can also be adaptively generated in a moving average mode in the initial monitoring stage, so as to ensure the judging capability matched with the individual characteristics of the user.
Meanwhile, in the process, the elements of each sequence are marked by time sequence numbers to ensure the uniqueness of the one-to-one correspondence, so that the misjudgment risk caused by sequence misplacement is avoided, the setting logic of the ratio is not only based on mathematical operation, but also based on physiological coupling characteristics of pulse wave signals in a respiratory drive state, wherein in a normal state, the pulse wave amplitude and the periodic modulation have a tendency of synchronously rising or falling, and in a specific mode state of the signals or respiratory rhythm disorder, the decoupling behavior with inconsistent directivity is usually generated, the symbol consistency ratio is obviously reduced, and meanwhile, the system introduces a dynamic balance engineering mechanism based on the fundamental technical quantity of the threshold, namely the sensitivity of the rhythm decoupling state and the system can be balanced and balanced. If the threshold is set too high, the system may miss invalid respiratory effort signals in a particular mode phase of the mild signal, whereas if the threshold is too low, it may be too sensitive to short-time noise or body movement disturbances, resulting in an increase in false alarm rate. Therefore, in actual deployment, the selection of the threshold is based on the consistency ratio of the reference symbol established in the stable respiratory state in the monitoring initial stage, the relative judgment threshold is set through the proportion reduction quantity, the typical range is twenty percent to thirty percent of the reference value, the algorithm is ensured to have enough recognition sensitivity and judge stability, and the signal acquisition module introduces a multi-scale smoothing mechanism and a pseudo-peak correction strategy for further improving the stability of the system under the condition of low signal-to-noise ratio. In the pulse wave peak value extraction process, local extremum judgment is adopted to combine with amplitude threshold control, and the threshold value is taken from fifty percent of average peak-valley difference in the rest section so as to filter high-frequency false peaks. Meanwhile, a peak validity screening mechanism is established by carrying out joint judgment on the time interval and the amplitude gradient of adjacent peaks, so that good time alignment and representativeness of pulse wave crest value sequences are ensured at an input end, a set of robust acquisition links which are based on signal stability and are constrained by feature continuity is established by introducing the strategy, a reliable support is provided for subsequent differential calculation and rhythm analysis, key parameters such as symbol consistency ratio, second order differential values of the amplitude differential sequences and the periodic differential sequences, energy base line lower limit threshold and the like are not set by adopting absolute constants by a system, stable state data obtained in an individual initial monitoring stage are combined, self-adaptive generation is carried out through a sliding average and proportion setting mechanism, the suitability defect caused by static setting is avoided by the method for establishing parameter reference based on individual dynamic features, and simultaneously the system is allowed to be iterated and optimized according to actual monitoring data in long-term operation, so that a high robustness judging frame with endogenous feedback regulation capability is established.
The embodiment 2 discloses a specific implementation flow of a portable blood oxygen signal OSA intelligent detection method, which is based on photoelectric volume pulse wave signals acquired by a single blood oxygen sensor and aims to realize reliable detection of specific signal segments of obstructive respiratory effort and characterization signal modulation energy silence under the condition of not depending on additional sensors; the method comprises collecting pulse wave signals with photoelectric volume sensor, smoothing the signals with multiple scales to improve stability, extracting pulse wave peak value based on neighborhood extremum comparison and amplitude threshold constraint to form peak value sequence, which is used as basis of subsequent differential analysis to directly influence accuracy of integral judgment, generating two differential sequences in parallel based on peak value sequence, wherein one differential sequence is amplitude differential sequence for reflecting amplitude variation between adjacent pulse wave peak values, the other differential sequence is periodic differential sequence for representing time interval fluctuation between continuous heart beats, the differential sequences are respectively obtained from amplitude domain and time domain of signals to reveal local dynamic characteristics of pulse wave in sliding time window, and the system calculates positive number of times of corresponding elements of the differential sequences and compares the positive number of times with total heart times in the time period, the symbol consistency ratio is obtained as a quantization index of the rhythm synchronization.
In order to establish an individuation judgment standard, the system automatically selects a section of stable breathing period data in an initial resting state of a user wearing equipment, symbol consistency ratio in the period is obtained in a moving average mode, so that a standard value of the user is constructed, if the current symbol consistency ratio is continuously lower than a relative threshold value of the standard value minus a preset proportion (such as 20%) and continuously exceeds a set shortest time length in a subsequent real-time monitoring process, the system judges that a signal section of a rhythm desynchronization state possibly occurs, the judgment mechanism adopts a relative ratio mode to set the threshold value, and is favorable for adapting to physiological difference among individuals, and the universality and practicability of the system are improved; in order to further eliminate abnormal waveform interference caused by non-respiratory factors, after a suspected event is identified, the system respectively carries out second-order differential calculation on an amplitude differential sequence and a periodic differential sequence, if the absolute value of any two-order differential sequence exceeds a morphological discrimination threshold, the event is considered to be possibly caused by a cardiac factor, the event is judged to be an artifact and is removed, the morphological discrimination threshold is set based on a physiological fluctuation rule disclosed in the prior art, is commonly used for distinguishing a slowly varying signal caused by respiratory effort and a sudden change signal caused by abnormal heart rhythm, the system further integrates an acceleration sensor to monitor a body movement state in consideration of signal disturbance possibly caused by body position transformation of a user during sleeping at night, if the detected body movement amplitude exceeds a preset threshold, the threshold is set according to the body movement amplitude variation range of a typical wearing part, the system temporarily suspends the judging process of a rhythm out-of-synchronization state signal section, and resumes the judging operation after the body movement tends to be stable, the processing mechanism realizes cooperative control between signal layer rhythm analysis and physical layer interference suppression, and improves the anti-interference capability of the whole detection system in a daily environment.
In the time period when the signal section of the rhythm dyssynchrony state is not judged, the system also continuously evaluates the modulation activity of the differential signal, specifically, the system calculates the variance of the amplitude differential sequence and the periodic differential sequence in a set time window, so as to measure the fluctuation degree of pulse fluctuation, when the modulation energy of the two is simultaneously lower than the preset lower limit (such as 10% to 30%) of the respective baseline value in a continuous period, the system judges the state as a specific signal section representing the silence of the signal modulation energy, the baseline value can be adaptively determined through the fluctuation level of a user in the initial resting period so as to adapt to individual difference and improve judgment precision, and through orderly matching of the stages, the method can realize high-robustness identification of obstructive and central respiratory events on the basis of a single sensor, avoids the dependence on the traditional blood oxygen saturation hysteresis index, and has stronger instantaneity and identification sensitivity.
The embodiment 3 constructs a set of portable blood oxygen signal acquisition and analysis platform, which comprises a PPG signal acquisition module, a sensor module and a signal analysis module, wherein the PPG signal acquisition module adopts a photoelectric volume sensor, is worn on the fingertip of a subject through finger-clip design, continuously acquires an original photoelectric volume pulse wave signal, and the working frequency of the sensor can be adjusted according to actual requirements so as to ensure the stability of signal acquisition; the data preprocessing and feature extraction module processes the original photoelectric volume pulse wave signal to detect the pulse wave peak value of the photoelectric volume pulse wave signal in real time so as to acquire a pulse wave crest value sequence, the processing process can comprise a multi-scale smoothing and pseudo-peak correction strategy aiming at improving the robustness of the pulse wave crest value extraction under the condition of low signal-to-noise ratio, the differential sequence generation and rhythm synchronization analysis module generates an amplitude differential sequence reflecting the amplitude variation of the pulse wave and a period differential sequence reflecting the interval variation of the pulse period in parallel based on the pulse wave crest value sequence, the module analyzes the rhythm synchronization of the amplitude differential sequence and the period differential sequence in a preset time window, the rhythm synchronization is obtained by calculating the symbol consistency ratio of elements of the amplitude differential sequence and corresponding elements of the period differential sequence, the symbol consistency ratio represents the symbol consistency degree of the two sequences, in the experiment, the preset time window can be set to comprise the duration time of a typical respiration modulation period in a calculation mode of effectively capturing the rhythm coupling relation in waves, the symbol consistency ratio can count the number of times that the product of the elements of the amplitude differential sequence and the period differential sequence is the corresponding to the positive value of the period differential sequence, the ratio of the times to the total heart beat times in the time window is used as the symbol consistency ratio; the reference value self-adaptive establishment and event judgment module compares a symbol consistency ratio obtained in real time with a reference symbol consistency ratio, wherein the reference symbol consistency ratio can be determined in a stable breathing state of a user monitoring initial stage, specifically, the reference symbol consistency ratio can be adaptively established by calculating a moving average value of the symbol consistency ratio in the stable breathing state of the user monitoring initial stage, when the real-time symbol consistency ratio is continuously lower than a first preset threshold value of the reference symbol consistency ratio, the current signal section can be determined to be a rhythm synchronization state signal section, the first preset threshold value can be set as a proportional fall of a reference value to enhance adaptability to different individuals, for example, the proportion can be selected between twenty percent and thirty percent, the setting logic is that the error judgment probability of normal physiological fluctuation is reduced while the specific mode of a sensitive identification signal is abnormal, the artifact authentication and body motion monitoring module continuously monitors the user body motion through an acceleration sensor, when the body motion amplitude is detected to be larger than the preset body motion threshold value, the judgment of the rhythm synchronization state signal section can be suspended, in addition, the differential phase of the differential phase is prevented from being an absolute value, the differential phase is prevented from being an interference signal, and the differential phase is prevented from being an interference signal section, and the differential phase is prevented from being a differential phase, the differential phase is detected, and the differential phase is abnormal when the differential phase is abnormal, and the differential phase is detected, the central apnea judging module can continuously calculate the modulation energy of the amplitude differential sequence and the modulation energy of the periodic differential sequence in a period when the signal section of the rhythm dyssynchrony state does not judge to occur, the modulation energy is obtained by calculating the variance of the differential sequence in the period, when the modulation energy of the amplitude differential sequence and the modulation energy of the periodic differential sequence are simultaneously lower than the lower limit threshold of the respective energy base line in the preset duration, a specific signal section representing the silence of the signal modulation energy can be identified, the lower limit threshold of the energy base line can be respectively set to be a lower limit proportion value of the corresponding modulation energy in the normal and steady sleep state of the user, the range of the lower limit proportion value is 0.1 to 0.3, and the modulation energy base line in the normal and sleep state can be self-adaptively established by the variance of the amplitude differential sequence and the periodic differential sequence in the steady respiratory state in the initial monitoring stage of the user.
The test aims at verifying the recognition capability of the portable blood oxygen signal OSA intelligent detection method on invalid respiratory effort events, Firstly, continuously collecting photoelectric volume pulse wave signals through a photoelectric volume sensor, performing multi-scale smoothing on the collected original signals to improve signal stability, and then detecting pulse wave peak values of the photoelectric capacitance pulse wave signals in real time through a strategy based on neighborhood extremum comparison and amplitude threshold constraint to obtain pulse wave peak value sequences, wherein the stability of the pulse wave peak value sequences directly influences the accuracy of subsequent judgment; generating an amplitude differential sequence and a period differential sequence in parallel based on a pulse wave crest value sequence, wherein the amplitude differential sequence is used for reflecting amplitude variation between adjacent pulse wave peaks, the period differential sequence is used for representing time interval fluctuation between continuous heart beats, the number of times that the product symbol of an element of the amplitude differential sequence and a corresponding element of the period differential sequence is positive is calculated in a preset time window, the ratio of the number of times to the total number of heart beats in the time window is taken as a symbol consistency ratio to obtain rhythm synchronization, the ratio quantifies the consistency degree of the two differential sequences in the symbol fluctuation direction, a reference value is generated by moving average of the symbol consistency ratio calculated in the stable time period in a stable respiration state of a user monitoring initial stage, in real-time detection, if the symbol consistency ratio of the current time period is continuously lower than a first preset threshold value of the reference symbol consistency ratio, the threshold value can be set to be a descending amount of the ratio of twenty percent to thirty percent of the reference value, and the descending state continuously exceeds a set minimum duration, the current signal section is determined to be a rhythm synchronization state signal section, the preset threshold value is aimed at enhancing adaptability to different individuals, the method comprises the steps of determining a specific mode of a respiratory effort-induced relatively slowly varying phenomenon and a sudden fluctuation caused by a high-frequency signal disturbance emergency of a non-respiratory source, continuously monitoring user body movement through an acceleration sensor, suspending the determination of the rhythm desynchronizing state signal section when the detected body movement amplitude is larger than a preset body movement threshold value, avoiding the interference of the body movement artifact on the detection result, continuously calculating the modulation energy of the amplitude difference sequence and the period difference sequence in a time period when the absolute value of any two level difference values is not determined to be larger than a preset morphology discrimination threshold value, setting the morphology discrimination threshold value according to historical physiological data, simultaneously, continuously monitoring the user body movement through the acceleration sensor, simultaneously determining the threshold value when the two modulation energy values are simultaneously lower than the corresponding threshold value in a set time period, and automatically adapting to the threshold value by the self-adaptive threshold value, and generating a self-adapting mode of the self-adapting to the threshold value of the threshold value when the two modulation energy values are simultaneously lower than the set in a set time period, and the modulation energy of the amplitude difference sequence can be continuously calculated in a time period when the amplitude difference signal section is not determined to be larger than the preset morphology discrimination threshold value.
The test focuses on the detailed analysis of the performance of the method in the aspects of identifying ineffective respiratory effort, identifying artifacts and distinguishing different apnea types, and the test observes that when an ineffective respiratory effort event occurs to a user, the symbol consistency ratio shows a remarkable and continuous descending trend, when the user is in a normal respiratory state, relatively stable rhythm synchronism exists between an amplitude difference and a period difference, the symbol consistency ratio is usually kept at a higher level, when a rhythm desynchronization state signal section occurs, abnormal fluctuation of negative pressure in the lung is caused by airway obstruction, a rhythm decoupling phenomenon occurs between circulation dynamics and a nerve regulation mechanism, and further the descending trend of the ratio is caused, and when the ineffective respiratory effort event occurs, the trough of the pulse is usually earlier than the remarkable descending of blood oxygen saturation, which indicates that a monitoring focus can be advanced from a traditional airway obstruction result (such as blood oxygen descending) to a specific mode process of signals by capturing a pulse wave time sequence-amplitude decoupling phenomenon, so that when a patient has a remarkable abnormal respiratory index can not be identified through the abnormal physiological index decoupling phenomenon; effective discrimination of cardiac artifact and motion artifact, during the course of the test, for the scenario in which artifact may be introduced, the method exhibits its discrimination capability, if the absolute value of the second order differential value exceeds a morphological discrimination threshold, the system may identify an artifact event caused by a cardiac factor, the morphological discrimination threshold being intended to distinguish a relatively slowly varying phenomenon caused by respiratory effort from a severe fluctuation caused by a high frequency signal disturbance incident of a non-respiratory source, in addition, when the body movement amplitude exceeds a preset body movement threshold value, the system can suspend the judging process of the rhythm desynchronization state signal section, thereby effectively avoiding the interference of the body movement artifact on the detection result, constructing a cooperative barrier between the rhythm analysis of the signal layer and the interference monitoring of the physical behavior layer, remarkably enhancing the actual adaptability of the method in daily life, distinguishing the obstructive from the central apnea, distinguishing the specific signal section representing the signal modulation energy silence by monitoring the abnormal attenuation of the differential sequence modulation energy, and in the time section of the signal section without judging the occurrence of the rhythm desynchronization state, when the modulation energy of the amplitude differential sequence and the periodic differential sequence is continuously lower than the lower limit threshold value of the respective energy base line in the preset duration, identifying the specific signal section representing the signal modulation energy silence, wherein the judging logic is based on the characteristic that the central apnea leads to complete interruption of the respiratory drive, and further represents the silence of the signal, therefore, the low modulation energy becomes an important indicator of the type event, and the single blood oxygen sensor has the accurate decision-making capability to distinguish the central apnea according to the accurate treatment capability.
Embodiment 4. This embodiment describes an intelligent detection method for OSA using portable blood oxygen signals with reference to fig. 1 to 3. As shown in figure 1, three-axis acceleration data are output to a main processor in real time by an accelerometer to extract and analyze body movement amplitude values, when the body movement amplitude meets the amplitude <0.5G condition, the main processor enters an alt state and triggers and enables breath analysis, during the period, a current data segment is marked as an effective signal mark and is submitted to an event analysis module to be further processed, if the amplitude is detected to be more than or equal to 0.5G, the system enters a high body movement state, the system carries out pause event judgment operation and synchronously caches original data to be used for subsequent recovery, after the body movement event is ended, the system enters a recovery monitoring stage, if a user is detected to keep still for 2 seconds and return to a stable signal state, the main processor carries out restart analysis and carries out recovery mark, analysis judgment on a respiratory event is recovered, and a body movement threshold value of 0.5G and a pause delay of 200ms are also definitely set in the diagram and are used for triggering and controlling boundary conditions of each processing logic, and the flow effectively avoids influence of spurious interference on the accuracy and robustness of the synchronization state judgment of the rhythm signal segment by carrying out intelligent pause and breath analysis under the body movement interference conditions.
As shown in fig. 2, fig. 2 (a) shows the change of the rhythm synchronization ratio with the lapse of the monitoring time (hh: mm) in the form of a line graph, wherein different event types are marked with the state a, the state B, and the state C, showing that the rhythm synchronization ratio gradually decreases from the initial approach of 1.0 in the period of 00:00 to 03:00, a significant decrease occurs between 01:30 to 02:30 until 03:00 returns to about 0.9, the fluctuation feature of the synchronization between the amplitude differential sequence and the periodic differential sequence in the pulse wave is reflected, the trend of the change of the amplitude differential sequence in the pulse wave is highlighted, the reference line shown in the graph is used for judging the trigger threshold of the decrease of the synchronization, the gray background emphasizes the change range of the index, the histogram of fig. 2 (B) shows the distribution of the amplitude modulation activity value corresponding to the sleep stage states in the same monitoring time, the marks include the awake stage, the N1 stage, the N2 stage, the N3 stage and the REM stage, the overall trend from 00:00 to 03:00, the decrease in particular, and the silence stage (REM 02) in the low-stage and the silence state).
As shown in fig. 3, firstly, amplitude differential sequence and periodic differential sequence are continuously acquired and peak value extracted through photo capacitance product pulse wave signals, basic characteristic data of subsequent analysis are built, a reference symbol consistency ratio is built under a stable respiratory state by the system, an individuation baseline is formed through a moving average reference mode, then the differential sequence is calculated to calculate a symbol consistency ratio (amplitude and periodic differential symbol matching degree), and is compared with a reference value, if the real-time symbol consistency ratio is lower than a reference threshold value, a rhythm desynchronization state signal section is judged, an abnormal analysis stage is entered, in order to improve judgment accuracy and reduce misjudgment, the system introduces cardiac artifact elimination, specifically comprises two-step differential fractal analysis and body movement monitoring pause logic, thereby effectively identifying interference caused by high-frequency signal disturbance and body movement artifact of a non-respiratory source, on the premise that a blocking event is not triggered, the system also monitors modulation activity, if modulation energy is continuously lower than the baseline, a specific signal section representing signal modulation energy silence is identified, thereby the recognition scheme based on rhythm synchronization form change and energy triple characteristic judgment is built, and the high respiratory event recognition scheme is in a technical scheme of environment is ensured, and the technical scheme is suitable for the environment. As shown in fig. 4, the trend of the change in the symbol uniformity ratio during 24-hour monitoring is shown. The solid line indicates the real-time symbol consistency ratio, the dash-dot line indicates the reference value, and the broken line indicates the determination threshold value (70% of the reference value). When the real-time ratio is continuously below the threshold, the system determines a segment of the dyssynchrony state signal, as shown in fig. 5, illustrating the change in modulation energy of the amplitude differential sequence and the periodic differential sequence at different respiratory states. The solid line represents amplitude differential sequence modulation energy, the long dashed line represents periodic differential sequence modulation energy, the dotted line represents the energy baseline lower threshold, and the relationship between body movement amplitude and breath analysis state is shown in fig. 6. The upper part shows the body movement amplitude variation (solid line), and the lower part shows the system analysis status (square mark indicates normal analysis, triangle mark indicates suspension analysis).
In the embodiment 5, in the construction process of the differential sequence, a system firstly adopts a pulse peak extraction method combining neighborhood local extremum judgment and amplitude threshold control based on continuously acquired photoelectric volume pulse wave signals, a pulse peak value sequence is generated by matching with a pseudo peak suppression strategy, the stability of heart beat alignment is enhanced, a sampling frequency standardization module is arranged in the peak extraction process of the system, a numbering marking mechanism is introduced to ensure the sequence consistency in different heart rate intervals, on the basis, the system respectively calculates amplitude differences and period differences between adjacent peaks to form an amplitude differential sequence and period differential sequence for reflecting the change trend of blood oxygen signals, and ensures that the two differential sequences have a one-to-one element relation at each time node in a numbering association mode, in the aspect of generating a rhythm synchronization index, a sliding analysis window with fixed duration is generally set to be thirty to forty seconds, the setting comprehensively considers the stability of the time sequence characteristics of the respiration rhythms and algorithm responses, and can cover a representative signal change mode, the amplitude differential sequence and the period interference effectively, namely the differential window has a one-to-one element correspondence with the pulse window, and the pulse rate is a positive-to-ten-second ratio in the generation aspect of the pulse synchronization index, and the window has a corresponding element-to the pulse rate coefficient corresponding relation in the pulse window is compared with the pulse sequence in the pulse sequence.
In order to adapt to individual differences, the system selects a signal segment of a user in a resting state to construct a reference value in an initial stage of wearing equipment, in a specific process, the system adopts multi-window smoothing calculation to obtain a symbol consistency ratio in the initial stage, and establishes a statistical expectation by eliminating abnormal values deviating from an average level, as an individual reference symbol consistency ratio judged by a subsequent event, the mode can realize baseline self-learning on the premise of not depending on additional physiological parameters, is suitable for diversity of basic states among users, the system continuously monitors the change condition of the current symbol consistency ratio in actual operation, if the ratio in a continuous time period is found to drop in amplitude to exceed a set proportion of the reference value, and the duration reaches a system response threshold, the system judges whether a respiratory rhythm decoupling phenomenon exists in the time period, and further judges whether a rhythm synchronization state signal segment is formed by combining other signal characteristics, the proportion change threshold is generally set within a range of twenty five percent to thirty percent, the selection comprehensively considers the requirement of distinguishing between the rhythm fluctuation amplitude and the abnormal state in a normal sleep process, the system also introduces a second-order difference calculation mechanism for differential state, carries out differential calculation for differential sequence, and carries out differential processing on the differential sequence in respect to the actual operation, and the differential sequence, and the differential amplitude is respectively carries out differential processing to detect the difference value in the state to be higher than the critical state, if the differential amplitude is higher than the threshold, the absolute fluctuation state is higher than the threshold, the absolute state is possible, and the respiratory state is judged is different from the normal state, the respiratory state has a difference state, and the respiratory state has a severe state has been compared with the difference, the system judges the event as a non-respiratory artifact event and excludes the event from the final event record, the setting principle of the morphological threshold is based on the change characteristics of typical heart beat waveform mutation, the influence rule of common noise types on blood oxygen signals is considered, and the dynamic adaptability of the event is maintained through a smooth updating mechanism.
In the detection process of central apnea, the system carries out fluctuation intensity assessment on the amplitude differential sequence and the periodic differential sequence, and adopts sliding windows to count respective signal variances, the calculated variance value is used for reflecting the blood oxygen signal modulation activity degree in the time period, if the variance value is lower than the preset energy base line lower limit in a plurality of continuous time windows, the system is regarded as a modulation energy silence state, and the system recognizes that a specific signal segment representing signal modulation energy silence possibly exists according to the fluctuation intensity assessment; in the aspect of body movement interference judgment, the system collects three-axis acceleration signals of a wearing part in real time, calculates the current body movement amplitude in a vector synthesis mode, builds a dynamic interference identification model in combination with the judging time length, when the body movement amplitude value continuously exceeds a zero five-gravity acceleration unit and the duration exceeds two hundred milliseconds, the system enters a high interference state and temporarily pauses an respiratory event judging process, all signal data enter a buffer zone, and then, if the system detects that the acceleration signals continuously recover to the static state and keep exceeding two seconds, the system re-enables the event analyzing process, and brings the buffer zone data into subsequent processing, the judging condition adopts an amplitude threshold and time threshold dual judging mechanism to ensure good anti-interference capability on non-respiratory related factors such as night body position adjustment, slight shake and the like, the system realizes the cooperative detection of obstructive and central sleep respiratory events under the single blood oxygen signal input condition through the time sequence coordination, the data intercommunication and the logic linkage among the modules, and ensures the robustness and the practicability of the system under different use environments and user states.
Example 6 this example is intended to illustrate in detail the specific calculation of the core index of the present invention, the symbol uniformity ratio, the overall calculation being broken down into the following steps:
step 1, acquiring a pulse wave peak value sequence, continuously acquiring pulse wave signals through a photo volume (PPG) sensor, and detecting peaks of the pulse wave signals in real time, thereby acquiring a pulse wave peak value sequence, wherein the pulse wave peak value sequence can be formally expressed as:, wherein,Represents the firstInformation of peak points of pulse waves.First of allPeak amplitude of each pulse wave.First of allThe time at which the peak of the pulse wave occurs.
Step 2, generating an amplitude differential sequence and a period differential sequence based on the pulse wave peak value sequenceThe system generates a differential sequence of the two cores in parallel. Amplitude difference sequence) The sequence reflects the variation of peak amplitude between successive pulse waves. Its first stageIndividual elementsThe definition is as follows: Periodic differential sequence [ ]) The sequence reflects the variation of the interval of successive heart cycles. First, it is necessary to calculate the pulse cycle sequence from the peak time sequenceIts first oneIndividual elementsDefined as the time interval between two adjacent peaks: Subsequently, a periodic differential sequenceIs the first of (2)Individual elementsDefined as the difference between two adjacent pulse periods: By the definition above, elements of the amplitude differential sequenceElements of a sequence differing from cyclesBy common heartbeat index numberA one-to-one correspondence is formed, and a foundation is laid for subsequent synchronicity analysis.
Step 3, calculating symbol consistency ratio) Within a predetermined time window (e.g., 30-40 seconds, symbol uniformity ratioStatistics of the elements of the amplitude differential sequence within the windowCorresponding elements to periodic differential sequencesThe number of times the product sign is positive, and the ratio of the number of times to the total number of heartbeats in the window is taken as a final result, and the calculation formula can be expressed as follows: , wherein,Is the total number of heartbeats within the preset time window; is an index of a differential sequence from 1 to;Is a counting function, when the condition in brackets is @, the counting function is that) And when true, the value is 1, and when the condition is false, the value is 0. The symbol consistency ratio proposed by the inventionThe concept is based on the deep insight of specific mode physiological process of Obstructive Sleep Apnea (OSA) signals, and aims to solve the fundamental limitation that the prior art can only passively rely on hysteretic obstruction results such as blood oxygen saturation reduction and the like to judge, and the key principle is that during normal and steady breathing, the respiratory motion of a human body generates synchronous modulation on the amplitude and period of a photoplethysmogram pulse wave (PPG) signal through intrathoracic pressure change and autonomic nerve regulation, and the synchronous modulation appears as physiological coupling of the two change trends, and when an OSA patient generates airway obstruction and performs ineffective respiratory effort, severe and irregular intrathoracic negative pressure fluctuation can damage the regulation path, so that the amplitude and period change of pulse waves generate obvious decoupling phenomenon, and the formula is used for quantitatively capturing the conversion process from the coupling to the decoupling by calculating an amplitude differential sequence) And periodic differential sequence) Judging whether the change directions of the two are consistent so as to identify the heartbeat in the coupling state, and finallyThe ratio, namely the ratio of the coupled heart rate to the total heart rate, constructs a brand new OSA detection dimension independent of blood oxygen saturation, and the formula can directly sense the specific mode process of the core signal in the early stage of occlusion, so that the effective screening of the patients with light and medium degree is realized.
The physical meaning of the formula is to quantify the degree of synchronization of the amplitude variation trend with the period variation trend,Higher values of (a) indicate better synchronicity and smoother breathing, and conversely, indicate that a rhythmic decoupling phenomenon has occurred, possibly indicating the occurrence of ineffective breathing efforts.
Step 4. In order to make the calculation process of the present invention clearer, a simplified numerical example is now provided. Assuming that the system acquires 6 continuous pulse wave peaks in a time window, the total heart beat times. See table 1, raw data (corresponding to step 1).
Table 1. Pulse wave crest value sequence raw data example table.
Intermediate sequence calculation (corresponding to step 2) for calculating pulse period:
s;
s;
s;
s;
s;
Calculating a differential sequenceAndTable 2 of the following Table shows the slave indexTo the point ofBecause of (a) the calculation result of (a)Is required to use)。
Table 2. Differential sequence calculation results example table.
Table 3 symbol consistency determination procedure example table.
The final ratio is calculated by taking the sum of times (3 times) with positive product as the numerator and the total number of heartbeats in the window as the denominator(6 Times),The calculated real-time symbol uniformity ratio (0.5 in this example) is fed to the event decision module and compared to a reference symbol uniformity ratio established by moving averages. If the value continues to fall below the baseline threshold, the system determines that a rhythm dyssynchrony status signal segment has occurred.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof.
Finally, it should be noted that the above-mentioned embodiments are merely for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications and equivalents may be made to the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention.