Method for detecting physiological condition of sleeping patient based on analysis of pulse waves

[00002] This application is based on and incorporates herein by reference Japanese Patent Applications No. 2000-271456 filed on Sep. 7, 2000 and NO. 2000-351713 filed on Nov. 17, 2000.

[00003] 1. Field of the Invention

[00004] The present invention relates to a method for detecting the physiological condition of a sleeping patient.

[00005] 2. Related Art

[00006] The sleep condition (depth of a sleep) of a patient is usually detected from a polysomnogram which shows signals which are simultaneously measured on the patient's body by a polysomnographer. The polysomnogram includes an electroencephalogram, an oclogram, an electromyogramn, an electrocardiogram and the like. The patient needs to be hospitalized for such measurement, because the polysomnographer is a large-scaled facility. Further, the patient feels uncomfortable during the measurement, because sensors are attached on his/her head and face. Accordingly, the patient's sleep is disturbed and, as a result, accurate data is not obtained.

[00007] JP-A-3-41926 discloses an alternative method for detecting a sleep condition of a patient. The method detects a respiration rate and a pulse rate of the patient, and determines the patient's sleep condition based on the detected respiration rate and pulse rate as follows. The Pulse period, which corresponds to the R-R period in the electrocardiogram, is calculated, and thereafter the sleep condition (REM sleep or non-REM sleep) is detected from the fluctuation of the pulse period. In order to obtain the pulse period, the peak (top or bottom) of every pulse wave should be accurately detected. When pulse waves fluctuate regularly, the peaks of the pulse waves can be detected accurately. However, pulse waves may fluctuate irregularly if the patient moves his/her body during sleep. In this case, the peaks of pulse waves may shift due to a factor other than the fluctuation of blood flow and consequently may be incorrectly detected. In order to obtain the pulse period based on the peaks of the pulse waves which may be provided as the result of the misdetection, complicated calculations are required. Accordingly, this method cannot readily detect the patient's sleep condition.

[00008] Further, the pulse period varies distinctly depending on whether the patient is in the REM sleep or non-REM sleep, only when the patient is healthy. When the patient is ill or old, the pulse period varies only slightly depending on whether the patient is in the REM sleep or non-REM sleep. Accordingly it is difficult to detect the sleep condition.

[00009] The diagnosis of sleep apnea syndrome is also made by examining all the signals in the polysomnogram. Therefore it is expensive and requires time and effort for a patient to have examination for sleep apnea syndrome, because the patient needs to be hospitalized. As a result, it is difficult to detect and treat the sleep apnea syndrome early.

[00010] The present invention has an object to provide a method for accurately detecting the sleep condition of a patient at home without executing complicated calculation and without being affected by noise due to the movement of the patient's body.

[00011] The present invention also has an object to provide a method for diagnosing sleep apnea syndrome (SAS) at home without imposing a burden on a patient.

[00012] A first method according to the present invention detects the sleep condition of a patient by analyzing the pulse waves of the patient. The first method measures pulse waves on the patient's body, and creates the envelope of the pulse waves by connecting the tops and bottoms of the pulse waves. The first method determines that the patient is in non-REM sleep, if the created envelope fluctuates regularly. It determines that the patient is in REM sleep, if the created envelope fluctuates irregularly.

[00013] The pulse height, which is a height of every pulse wave, or the baseline, which is a line connecting a middle point of every pulse wave, may be created instead of the envelope of the pulse waves. In this case, it is determined that the patient is in non-REM sleep, if the created pulse height or baseline fluctuates regularly. It is determined that the patient is in REM sleep, if the created pulse height or baseline fluctuates irregularly.

[00014] A second method according to the present invention diagnoses a patient as SAS by analyzing the pulse waves of the patient. The second method measures pulse waves on the patient's body, and analyzes the pulse wave data. Based on the result of the analysis, the second method determines whether the patient has SAS. Preferably, the type of the SAS is also determined based on the result of the analysis, if it is determined that the patient has SAS.

[heading-00030] (First Embodiment)

[00031] A method according to a first embodiment detects the sleep condition of a patient by a physiological condition detection apparatus shown in FIG. 1. The physiological condition detection apparatus includes a pulse wave sensor 1, a drive circuit 2, a data processing unit 3 and a display unit 4. The data processing unit 3 includes a detection circuit 3A, an A/D converter 3B, and a microcomputer 3C. The pulse wave sensor 1 is a well-known optical sensor, and worn on the patient's wrist 5 or finger. The sensor 1 includes a sensor housing 1b on which a window 1a is formed. The sensor 1 further includes a light emitting element 1c and a light receiving element 1d in the housing 1b. The drive circuit 2 drives the light emitting element 1c to emit light toward the patient's wrist 5. A portion of the emitted light penetrates the capillary arteriole 6 in the inside of the patient's wrist 5 and is absorbed by the hemoglobin in the blood. The rest of the emitted light is reflected and scattered by the capillary arteriole 6, and partly reaches the light receiving element 1d. As the amount of the hemoglobin in the blood fluctuates in waves due to the pulsation of the patient's blood, the amount of the light absorbed by the hemoglobin also fluctuates in waves. As a result, the amount of the light which is reflected by the capillary arteriole 6 and reaches the light receiving element 1d fluctuates in waves. This fluctuation in the amount of the light received by the light receiving element 1d is detected as pulse wave information. The pulse wave sensor 1 further converts the pulse wave information into electrical signals (i.e., voltage signals), and outputs the electrical signals to the data processing unit 3.

[00032] In the data processing unit 3, the detection circuit 3A receives and amplifies the electrical signals. The A/D converter 3B converts the amplified analog signals to digital signals at a sampling frequency of 100 Hz. The digital signals are outputted to the microcomputer 3C. The microcomputer 3C is programmed to detect the sleep condition of the patient by using the digital signals which represent pulse wave information detected from the sleeping patient as follows.

[00033] Referring to FIG. 2, pulse wave data is measured from the digital signals at step 10. The pulse wave data is filtered through a digital filter so that unwanted frequency component of the pulse waves are eliminated at step 20. The peak of every pulse wave is detected from the pulse wave data at step 30.

[00034] One of the envelopes of the pulse waves is created at step 40. FIG. 3 shows pulse waves detected from the patient. Two envelopes A, B can be created by connecting the peaks of the pulse waves as shown in FIG. 3. One envelope A is obtained by connecting the tops of the pulse waves, while the other envelope B is obtained by connecting the bottoms of the pulse waves. Either of the envelopes A, B may be created at step 40. FIGS. 4A and 4B show the envelopes of the pulse waves detected during the patient's REM sleep and non-REM sleep, respectively.

[00035] At step 50, the characteristics of the envelope A, B is detected at regular intervals as follows. A statistical index such as the average, maximum value, minimum value, variance, deviation or the like is employed for representing the characteristics of the envelope A, B over an interval. The employed index is calculated over and at regular intervals (i.e., 20 seconds) at step 50. At step 60, the sleep condition of the patient is detected based on the fluctuation of the calculated index. As shown in FIG. 5, the envelopes A, B fluctuate regularly while the pulse period fluctuates regularly. As shown in FIG. 6, the envelopes A, B fluctuate irregularly while the pulse period fluctuates irregularly. It is known that the patient is in non-REM sleep if the pulse period fluctuates regularly. It is also known that the patient is in REM sleep if the pulse period fluctuates irregularly.

[00036] Further, the envelopes A, B of the pulse waves actually fluctuate more irregularly during the patient's REM sleep than during the patient's non-REM sleep as shown in FIGS. 4A and 4B. In the case of non-REM sleep, the standard deviations calculated from the envelope data shown in FIG. 4B over intervals of 20 seconds are 0.15 (0-20 s), 0.16 (20-40 s) and 0.16 (40-60 s). That is, the standard deviations do not vary largely. In contrast to this, in the case of REM sleep, the standard deviations calculated from the envelope data shown in FIG. 4A over intervals of 20 seconds are 0.16 (0-20 s), 0.09 (20-40 s), 0.10 (40-60 s) and 0.15 (60-80 s) That is, the standard deviations vary largely Therefore, it may be determined that the patient is in non-REM steep, if the envelope A, B fluctuates regularly, that is, the statistical index varies within the predetermined range. Further, it may be determined that the patient is in REM sleep, if the envelope A, B fluctuates irregularly, that is, the statistical index varies exceeding the predetermined range.

[00037] Accordingly, it is determined at step 60 that the patient is in REM sleep, if the variation of the index is within 25%. Further it is determined at step 60 that the patient is in non-REM sleep, if the variation of the index is over 25%. The detected sleep condition is displayed by the display unit 4 at step 70.

[heading-00038] (Second Embodiment)

[00039] A method according to a second embodiment diagnoses a patient as sleep apnea syndrome (SAS) by a physiological condition detection apparatus (shown in FIG. 1) which has the similar configuration as the first embodiment. Further the method determines the type of SAS. The type of SAS is obstructive sleep apnea syndrome (OSAS), central sleep apnea syndrome (CSAS), or mixed sleep apnea syndrome (MSAS). In obstructive sleep apnoeaic (OSA) condition, the breathing movement of the patient's thoracic part and abdominal wall is maintained, but ventilation through the patient's mouth or nose is stopped due to partially obstructed upper airway. In central sleep apnoeaic (CSA) condition, movement of the patient's respiratory muscle is stopped due to a standstill of respiratory center or a disturbance of excitation-conduction. The patient is diagnosed as MSAS if he/she shifts from CSA condition to OSA condition.

[00040] In the physiological condition detection apparatus, similarly to the first embodiment, a pulse sensor 1 is driven by a drive circuit 2 and detects pulse wave information from the patient's body. The pulse sensor 1 outputs the pulse wave information as electrical signals to a data processing unit 3. In the data processing unit 3, a detection circuit 3A and an A/D converter 3B operates similarly to the first embodiment, and a microcomputer 3C receives digital signals from the A/D converter 3B. The microcomputer 3C is programmed to diagnose the patient as SAS by using the digital signals which represents the pulse wave information detected from the sleeping patient as follows.

[00041] Referring to FIG. 7, the received digital signals are processed at steps 10-30 and one of the envelopes A, B of the pulse waves is created at step 40 similarly to the first embodiment. However, both of the envelopes A, B may be created at step 40, if necessary at later steps.

[00042] FIG. 8 shows pulse waves detected from the patient in eupneic condition, and FIG. 9 shows pulse waves detected from the patient in OSA condition. The envelopes A, B shown in FIG. 9 fluctuate largely as compared with the envelopes A, B shown in FIG. 8. Specifically, the normalized amplitude of the envelope A, B (FIG. 9) in OSA condition is twice the normalized amplitude of the envelope A, B (FIG. 8) in eupneic condition or higher. Accordingly the amplitude of the envelope A, B is calculated at step 50A for detecting OSA condition. The calculated amplitude is normalized with respect to the amplitude of the pulse wave, because the amplitude of the envelope changes in proportion to the amplitude of the pulse wave. For example, the amplitude of the envelope is normalized by being divided by the amplitude of the pulse wave averaged over one period of the pulse waves.

[00043] FIG. 10 shows pulse waves detected from the patient in CSA condition. In CSA condition, movement of the patient's thoracic part is stopped. Therefore pulse wave data detected from the patient in CSA condition does not include signals which represent respiration. Accordingly, the envelope A, B fluctuates in a long period as shown in FIG. 10. Specifically, it fluctuates in a period of roughly 10 seconds (7-12 seconds). In contrast to this, the envelope A, B in eupneic condition or OSA condition does not fluctuate in a long period as shown in FIG. 8 or 9. Therefore the period of the fluctuation of the envelope A, B is calculated at step 50B for detecting CSA condition. The calculated period is also normalized.

[00044] At step 60A, it is determined that the patient is in OSA condition if the normalized amplitude of the envelope A, B is twice that in the eupnea condition or higher. At step 60B, it is determined that the patient is in CSA condition if the normalized period is equal to or longer than 7 seconds.

[00045] At step 70A, a diagnosis of SAS is made based on the results of steps 60A and 60B. Further the type of SAS is determined based on the results of steps 60A and 60B, if the patient is diagnosed as SAS. The patient is diagnosed as MSAS, if he/she shifts from CSA condition to OSA condition. The result of the diagnosis is displayed by the display unit 4.

[heading-00046] (Modifications)

[00047] In the first embodiment, the sleep condition of the patient may be detected based on the pulse height of the pulse waves as shown in FIG. 11. The pulse height is defined as the height of a pulse wave as shown in FIG. 3. The pulse height of every pulse wave is created instead of the envelope A, B at step 40. Further statistical index which represents the characteristics of pulse height is calculated over and at regular intervals at step 50. As shown in FIG. 4, the pulse height fluctuates regularly while the pulse period fluctuates regularly. As shown in FIG. 5, the pulse height fluctuates irregularly while the pulse period fluctuates irregularly. Accordingly, it may be determined at step 60 that the patient is in non-REM sleep, if the calculated index varies within a predetermined range. Further it may be determined at step 60 that the patient is in REM sleep, if the calculated index varies exceeding the predetermined range.

[00048] Further in the first embodiment, the sleep condition of the patient may be detected based on the baseline of the pulse waves as shown in FIG. 12. The baseline is a line which connects the middle point of the pulse height of every pulse wave as shown in FIG. 3. The baseline of the pulse waves is created instead of the envelope A, B at step 40. Further statistical index which represents the characteristics of baseline is calculated over and at regular intervals at step 50. As shown in FIG. 5, the baseline fluctuates regularly while the pulse period fluctuates regularly. As shown in FIG. 6, the baseline fluctuates irregularly while the pulse period fluctuates irregularly. Accordingly, it may be determined at step 60 that the patient is in non-REM sleep, if the calculated index varies within a predetermined range. Further it may be determined at step 60 that the patient is in REM sleep, if the calculated index varies exceeding the predetermined range.

[00049] In the second embodiment, another attribute of the envelope A, B of the pulse waves may be calculated instead of the amplitude of the envelope A, B. That is, another attribute of the envelope A, B may be employed for detecting OSA condition. When the patient is in OSA condition, a bottom B_A of the envelope A is very close to a top T_B of the envelope B as shown in FIG. 8. In contrast to this, a bottom B_A of envelope A and a top T_B of the envelope B is not close when the patient is in eupneic condition. Accordingly, the closeness of the bottom B_A of the envelope A and the top T_B of the envelope B may be calculated and normalized at step 50A. In this case, at step 60A, it is determined whether the patient is in OSA condition based on the normalized closeness of the bottom B_A of the envelope A and the top T_B of the envelope B.

[00050] Furthermore an attribute of the pulse waves may be employed for detecting OSA condition. When the patient is in OSA condition, the tops T_P and bottoms B_P of the pulse waves fluctuate largely as shown in FIG. 9. Therefore the degree of the fluctuation of the tops T_P and bottoms B_P may be calculated and normalized at step 50A, and it is determined at step 60A whether the patient is in OSA condition based on the normalized degree of the fluctuation of the tops T_P and bottoms B_P of the pulse waves. In this case, the envelope A, B is not used at step 50A, because the tops T_P and bottoms B_P of the pulse waves can be obtained without the envelope A, B of the pulse waves.

[00051] Moreover OSA condition may be detected based on the degree of the fluctuation of the pulse height of the pulse. FIGS. 13A-13C show pulse waves obtained by enlarging FIGS. 8, 9 and 10 with respect to the horizontal axis (time axis), respectively. The pulse height of the pulse waves detected from the patient in eupneic condition or CSA condition does not fluctuate largely. Specifically, the pulse height of the pulse waves fluctuate within a range of 20%. In contrast to this, when the patient is in OSA condition, the pulse height of the pulse waves fluctuates largely. Specifically, the minimum pulse height of the pulse waves is equal to or lower than 50% of the maximum pulse height of the pulse waves. Accordingly, the pulse height of every pulse wave may be calculated at step 50A. Further a ratio of the minimum pulse height to the maximum pulse height is calculated at step 50A. At step 60A, it is determined that the patient is in OSA condition if the ratio is equal to or lower than 50%.

[00052] Further the areas or lengths of the pulse waves may be employed for detecting OSA condition. FIGS. 14A-14C show pulse waves obtained by enlarging FIGS. 13A-13C with respect to the horizontal axis (time axis), respectively. The area of a pulse wave (pulse wave area) means the area formed by the pulse wave between two bottoms and the line which connects the bottoms or formed by the pulse wave between two tops and the line which connects the tops as shown in FIGS. 14A-14C. The length of a pulse wave (pulse wave length) means the length of the pulse wave between two bottoms or two tops as shown in FIGS. 14A-14C.

[00053] The pulse wave areas of the pulse waves detected from the patient in eupneic condition or in CSA condition do not vary largely as shown in FIG. 14A or 14C. Specifically, the pulse wave areas vary within a range of 20%. In contrast to this, the pulse wave areas of the pulse waves detected from the patient in OSA condition vary largely. Specifically, the minimum pulse wave area is equal to or less than 50% of the maximum pulse wave area. Accordingly, the pulse wave area of every pulse wave may be calculated at step 50A. Further the ratio of the minimum pulse wave area to the maximum pulse wave area is calculated at step 50A. It is determined at step 60A that the patient is in OSA condition, if the ratio is equal to or less than 50%.

[00054] On the other hand, the pulse wave lengths of the pulse waves detected from the patient in eupneic condition or in CSA condition do not vary largely as shown in FIG. 14A or 14C. Specifically, the pulse wave lengths vary within a range of 10%. In contrast to this, the pulse wave lengths of the pulse waves detected from the patient in OSA condition vary largely. Specifically, the minimum pulse wave length is equal to or less than 75% of the maximum pulse wave length. Accordingly, the pulse wave length of every pulse wave may be calculated at step 50A. Further the ratio of the minimum pulse wave length to the maximum pulse wave length is calculated at step 50A. It is determined at step 60A that the patient is in OSA condition, if the ratio is equal to or less than 75%.

[00055] Further OSA condition may be detected based on the difference between the fluctuation of the pulse waves and fluctuation of pulse waves detected from the patient in eupneic condition.

[00056] In the above embodiments and modifications, the pulse wave sensor 1 may be an ultrasonic sensor, a Doppler sensor or a pressure type sensor.

[00057] The effects of the above embodiments and modifications are as follows. In the first embodiment and its modifications, the methods detect the sleep condition of the patient without executing complicated calculations, because the regularities of the fluctuation of the envelope, the pulse height or the baseline is not significantly affected by the inaccuracy of the detected tops or bottoms of the pulse waves. In this way, the methods detect the sleep condition of the patient more simply.

[00058] Further in the first embodiment and its modifications, pulse wave information does not need to be obtained very accurately with respect to time. Therefore the methods can detect the sleep condition of the patient without being significantly affected by disturbance.

[00059] In the second embodiment and its modifications, the patient can readily have examination at home, because a diagnosis of SAS is made based on the pulse waves detected by the pulse sensor 1. As a result, SAS can be detected and treated at the earliest possible time.