568768 玖、發明說明 【發明所屬之技術領域】 本發明係關於-種分析方法，特別指一種將任意活動 狀態下血壓計量測所得之脈動訊號進行處理以作為心臟機 能判斷指標之脈動訊號與心臟機能相關性分析方法。 5【先前技術】 心臟病習稱「隱形頭號殺手」，然_般求診者於醫院 檢查出患有心臟病時，往往已近末期而治療不易。事實上 ’就目前全球醫療技術而言，心臟病之初期及中期往往難 以查覺，此-尚待突破之醫學盲點導致相當比例之心臟病 10 突發致死案例。 相較於心臟病之預防而言，近年來歸功於血壓計價格 不斷下降而更趨低廉，得以深入—般家庭成為居家保健基 本設備之一’使高血壓之預防警示人人皆可為之。血壓計 中又以操作簡便之電子式血壓計尤受家庭個人所歡迎，電 15子式血壓計除量測收縮壓及舒張壓外，並同時能量測脈搏 數等脈搏訊號供量測者參考，然由於該脈搏訊號僅係電子 式血壓計所提供之一附屬功能，故通常未經進一步利用分 析而不受重視。 【發明内容】 20 ®此，本發明之首一目的，即在提供-種利用血壓計 量測之脈動訊號作為心臟機能判斷指標之脈動訊號與心臟 機能相關性分析方法。 本發明之次一目的，在於提供一種適用於受試者任何 活動狀態之脈動訊號與心臟機能相關性分析方法。 7 於是，本發明脈動訊號與心臟機能相關性分析方法， 係用以將任意活動狀態下由一血壓計量測所得之一受絮者 於一定時間間距内之一脈動訊號進行處理，以作為該受古式 者、臟機能之參考指標，該方法包括：（丨）將該脈動訊號 進行能量頻譜轉換；（2)將該轉換獲得之一能量頻譜正規 化；及（3)由該正規化之能量頻譜計算一預先定義之心臟 指數，而以該心臟指數供判斷該受試者心臟機能之參考指 標。 少曰 本發明並揭示一種利用脈動訊號判定心臟機能之方法 ，包括：（1)將一受試者之脈動訊號進行能量頻譜轉換； (2)將該轉換獲得之一能量頻譜正規化；及由該正規化 之能量頻譜計算一預先定義之心臟指數；及（4)依據該心 臟指數判定該受試者之心臟機能狀態。 【實施方式】 本發明之前述以及其他技術内容、特點與優點，於以 下配合參考圖式之一較佳實施例詳細說明中，將可清楚明 白。 首先如第一圖所示，本發明脈動訊號與心臟機能相關 性分析方法之較佳實施例，主要包括：一脈動量測步驟 11、一刖段讯號去除步驟12、一頻域轉換步驟13、一能 量頻譜轉換步驟14、一正規化步驟15、一心臟指數定義 步驟16及一心臟指數計算步驟ι7、 本實施例中脈動量測步驟丨丨係採用一每秒可偵測16 點脈動訊號之電子血壓計就受試者進行脈動量測，其讀取 568768 之70整原始訊號31圖形如第二圖所例示。於該圖中， 虛線代表壓力，一開始血壓計之脈壓帶因充氣泵之充氣而 · 使壓力升南，達高點後泵之變頻馬達停止運轉而逐漸洩氣 減壓，實線則代表量測所得之脈動訊號，前段脈動訊號 5 311因參雜有泵馬達之機械運轉干擾而含較多雜訊，待泵 馬達停止而無干擾後，後段脈動訊號312即下降且呈明顯 · 規律性而利於後續分析。 自原始時域訊號31轉換至頻域中，所呈現之脈動訊 號頻譜分布狀況，將有助於進一步了解、判讀信號特徵及 鲁 10 潛藏意義。本實施例中經前段訊號去除步驟12去除前段 脈動汛號311後（理由於下詳述），頻域轉換步驟13係採 用快速傅立葉轉換（fast Fourier transform，FFT)演 异法，其運算式如以下[式一]所示，然其他可將時域訊號 31轉換至頻域之習知方法亦可適用。 χ(η) = Σχ〇^)Κ [式一] 其中，％=3 ® Χ〇 ··原始時域訊號 X:原訊號轉換至頻域之頻譜 20 Ν :輸入訊號點數 - J · ~ 1 - η=0, 1, ...N-l 原訊號序列xG長度視每次量測脈搏血壓時間而異， 9 568768 本實施例中每秒係擷取16點，對x〇序列做1024點之快 速傅立葉轉換，以獲得原訊號31轉換至頻域之頻譜X。 經本步驟13之頻域轉換分析後，由於充氣泵馬達運轉之 干擾訊號與人體脈動訊號，於頻域分布上有重疊現象而無 5 法濾除，故須先經前段訊號去除步驟12將馬達運轉停止 前之前段脈動訊號311捨去，而僅將馬達運轉停止後之後 段脈動訊號312進行頻域轉換處理。 能量頻譜轉換步驟14係將頻域轉換步驟13所獲得後 段脈動訊號312之頻譜X，與其共軛值conj(X)相乘，以 10 獲得X之功率頻譜密度（power spectrum density)S，功 率頻譜密度S即代表訊號於頻域中能量分布狀況，而可以 下述[式二]表示。第三a及三b圖即分別舉例說明一量測 所得之後段脈動訊號312，及由該後段脈動訊號312經步 驟13、14處理所得之能量頻譜圖。 15 S=X*conj(X) / η [式二] 其中，n : X之長度 為評估受試者之脈動訊號是否受其狀態影響，本發明 經研究比較發現，運動後之脈動雖然加快、能量加高、血 20 壓變化，然就心臟指數（容下詳述）而言，由於其訊號特徵 仍然存在甚至更為明顯，故可透過訊號值之正規化處理， 而無須等待受試者於靜止狀態下休息一段時間後再行量測 ，使本發明不論受試者處於何種狀態皆可隨時量測應用， 而不影響計算結果。 25 本實施例中正規化步驟15包含一能量正規化步驟 10 568768 j51及一頻率正規化步驟152。能量正規化步驟Μ〗係自 第二b圖所示之功率頻譜密度圖中，選出最大基本波之振 幅，其對應之頻率定義為第一主頻，通常位於iHz位置左 右（圖中GHz之訊號僅代表未歸零之均值，無關振動）， 5而其他出現振幅較小者稱為譜波，依序定義為第二主頻及 2三主頻。將第—主頻之振幅定為1_ (即本實施例能 里正規化之標準），而將各諧波之振幅除以第一主頻振幅 ’則可清楚觀察能量經正規化之振幅比例關係。 —本實施例中頻率正規化步驟152係將標準設定於心搏 每:1里8G下’亦即每秒h 33下，並如前述本實施例係採 用母秒可摘測16點脈動訊號之電子血壓計，故即擷取 21.28脈動㈣點。將電子血壓計量測之每分鐘脈搏數除 以頻率正規化標準8M4，乘上時間序列，即可顯示頻率 正規化後之脈動情況。 15 帛四至六圖即用以舉例說明經正規化步驟15之效果 第四a圖中包含一正常受試者於靜止狀態心搏別下/分 量測所得之後段脈動訊號312(上圖），及一由該後段脈動 « 312經步驟13、ι4處理所得之能量頻譜圖（下圖）， 該能量頻譜圖顯示之第—主頻振幅約為2.㈣5。第四b 2〇 圖則為將第四a圖緩招卜 圃厶正規化，調整為心搏80下/分且第一 主頻振幅為5x105之結果。 同理第五a圖中則包含第四a、b圖之同-正常受 4者於運動過後心搏12〇下/分量測所得之後段脈動訊號 312(上圖）及一月頻譜圖（下圖），該㉟量頻譜圖顯示之 568768 第一主頻振幅約為6. lxl05。第五b圖則為將第五a圖經 與第四b圖採相同正規化標準，即調整為心搏8〇下/分且 · 第一主頻振幅為5xl05後之結果。 第六a圖中包含一心肌病變之心臟病患者於靜止狀態 5 心搏65下/分量測所得之後段脈動訊號312(上圖）及一能 里頻谱圖（下圖），該能量頻譜圖顯示之第一主頻振幅約為 8· 9x1 〇5。第六b圖則為將第六&圖經與第四b、五b圖採 相同正規化標準，即調整為心搏8〇下/分且第一主頻振幅 _ 為5x1 〇5後之結果。於第六b圖中顯示，該心肌病變患者 籲 1〇 之各主頻位置較第四b、五b圖所示之正常人難以辨識， 且各主頻間非零能量頻譜出現頻繁。 本發明經大買I測所得之分析結果顯示，一般無心臟 疾病者之脈動能量頻譜圖中，可明顯讀出四至五個主要能 量分布（主頻），依個人心跳速率而異，1 Hz附近通常為最 大月b里發生處’往高頻處則能量漸減。除主要四、五個頻 譜之外，無其他明顯易見之頻譜分布，在頻率軸上值多為 零。 Φ 而心臟疾病患者量得之脈動訊號經能量頻譜分析後， 於可分辨之四至五個主頻之外，常有明顯可見然能量大小568768 发明 Description of the invention [Technical field to which the invention belongs] The present invention relates to an analysis method, in particular to a pulsation signal and a heart that process a pulsation signal obtained by blood pressure measurement in any active state as a judgment indicator of cardiac function Functional correlation analysis method. 5 [Previous technology] Heart disease is known as the "hidden number one killer". However, when a patient is diagnosed with a heart disease at the hospital, it is often near the end of the period and treatment is not easy. In fact, as far as current global medical technology is concerned, it is often difficult to detect in the early and middle stages of heart disease. This-pending medical blind spot has led to a significant proportion of cases of sudden death from heart disease. Compared with the prevention of heart disease, in recent years, the price of sphygmomanometers has been declining and becoming cheaper, which has allowed them to go deeper. Generally, families have become one of the basic home health care equipment ', so that everyone can prevent high blood pressure. Among the sphygmomanometers, the easy-to-use electronic sphygmomanometer is particularly popular among families. In addition to measuring systolic and diastolic blood pressure, the electric 15-type sphygmomanometer also measures pulse signals such as energy and pulse number for reference. However, because the pulse signal is only an auxiliary function provided by the electronic sphygmomanometer, it is usually ignored without further analysis. [Summary of the invention] 20 ® Therefore, the first object of the present invention is to provide a correlation analysis method of a pulsation signal and a cardiac function using a pulsation signal measured by a sphygmomanometer as a cardiac function judgment index. A secondary object of the present invention is to provide a correlation analysis method of pulsation signals and cardiac function applicable to any active state of a subject. 7 Therefore, the correlation analysis method of the pulsation signal and the heart function of the present invention is to process a pulsation signal of a subject in a certain time interval measured by a blood pressure measurement in any active state as the The reference index of the ancient and dirty functions, the method includes: (丨) performing the energy spectrum conversion of the pulsation signal; (2) normalizing an energy spectrum obtained by the conversion; and (3) the normalized energy The spectrum calculates a predefined cardiac index, and the cardiac index is used as a reference index for judging the subject's cardiac function. The present invention discloses a method for determining cardiac function using pulsation signals, including: (1) performing energy spectrum conversion on a subject's pulsation signals; (2) normalizing an energy spectrum obtained by the conversion; and The normalized energy spectrum calculates a predefined cardiac index; and (4) determines the subject's cardiac function status based on the cardiac index. [Embodiment] The foregoing and other technical contents, features, and advantages of the present invention will be made clear in the following detailed description of a preferred embodiment with reference to the accompanying drawings. Firstly, as shown in the first figure, the preferred embodiment of the method for analyzing the correlation between the pulsation signal and the cardiac function of the present invention mainly includes: a pulsation measurement step 11, a block of signal removal step 12, and a frequency domain conversion step 13. An energy spectrum conversion step 14, a normalization step 15, a heart index definition step 16 and a heart index calculation step ι7. The pulsation measurement step in this embodiment uses a pulse signal that can detect 16 points per second The electronic sphygmomanometer measures the pulse of the subject, and it reads 568768 and 70 whole original signal 31 patterns as illustrated in the second figure. In the figure, the dotted line represents the pressure. At the beginning, the pulse pressure band of the sphygmomanometer is caused by the inflation of the air pump. The pressure rises to the south. After reaching the high point, the pump's variable frequency motor stops running and gradually deflates. The measured pulsation signal, the front pulsation signal 5 311 contains more noise due to the mechanical operation interference of the pump motor. After the pump motor stops without interference, the pulsation signal 312 at the rear section decreases and shows obvious and regularity. Facilitate subsequent analysis. From the original time-domain signal 31 to the frequency domain, the spectrum distribution of the pulsating signal presented will help to further understand and interpret the signal characteristics and the potential significance of Lu 10. In this embodiment, after the previous-stage signal removing step 12 is removed from the previous-stage pulsating flood number 311 (the reason is described in detail below), the frequency-domain conversion step 13 uses a fast Fourier transform (FFT) algorithm, and the calculation formula is as follows The following [Formula 1] is shown, but other conventional methods that can convert the time domain signal 31 to the frequency domain can also be applied. χ (η) = Σχ〇 ^) Κ [Formula 1] Wherein,% = 3 ® χ〇 ·· Original time domain signal X: The original signal is converted into the frequency domain spectrum 20 Ν: Input signal points-J · ~ 1 -η = 0, 1, ... Nl The length of the original signal sequence xG varies depending on the pulse blood pressure measurement time. 9 568768 In this embodiment, 16 points are acquired per second, and 1024 points are fast for the x〇 sequence. Fourier transform to obtain the spectrum X converted from the original signal 31 to the frequency domain. After the analysis of the frequency domain conversion in this step 13, because the interference signal of the air pump motor operation and the human pulsation signal overlap in the frequency domain distribution and cannot be filtered out, the motor must be run through the previous signal removal step 12 The pulsation signal 311 before the stop is discarded, and only the pulsation signal 312 after the motor is stopped is subjected to the frequency domain conversion process. Step 14 of the energy spectrum conversion is to multiply the spectrum X of the latter-stage pulsation signal 312 obtained in the step 13 of the frequency domain conversion by the conjugate value conj (X) to obtain a power spectrum density S of X, the power spectrum. The density S represents the energy distribution of the signal in the frequency domain, and can be expressed by the following [Formula 2]. The third a and three b diagrams respectively exemplify a measured pulsation signal 312 in the subsequent stage, and the energy spectrum obtained by processing the pulsation signal 312 in the subsequent stage in steps 13 and 14. 15 S = X * conj (X) / η [Equation 2] Wherein, the length of n: X is to evaluate whether the subject's pulsation signal is affected by its state. According to the research and comparison of the present invention, although the pulsation after exercise is accelerated, Increased energy and changes in blood pressure. However, as far as the heart index (more details below) is concerned, its signal characteristics still exist or are even more obvious, so it can be processed through the normalization of the signal value without waiting for the subject to After resting for a period of time in a stationary state, measurement is performed, so that the present invention can be measured and applied at any time regardless of the state of the subject, without affecting the calculation result. 25 In this embodiment, the normalization step 15 includes an energy normalization step 10 568768 j51 and a frequency normalization step 152. The energy normalization step M is the maximum fundamental wave amplitude selected from the power spectrum density diagram shown in Figure 2b, and the corresponding frequency is defined as the first main frequency, which is usually located around iHz (GHz signal in the figure) It only represents the mean value without returning to zero, irrespective of vibration), while the others with smaller amplitudes are called spectral waves, which are sequentially defined as the second main frequency and two third main frequencies. Set the amplitude of the first main frequency to 1_ (that is, the standard for normalization in this embodiment), and divide the amplitude of each harmonic by the first main frequency amplitude 'to clearly observe the normalized amplitude ratio relationship of energy . — In this embodiment, the frequency normalization step 152 is to set the standard at a heart rate of 8G per 1 ', that is, h 33 per second, and as described above, this embodiment uses the mother seconds to measure 16 pulse signals. Electronic sphygmomanometer, so the 21.28 pulse point was captured. Divide the pulse rate per minute of the electronic blood pressure measurement by the frequency normalization standard 8M4, and multiply it by the time series to display the pulsation after the frequency normalization. Figures 14 to 6 are used to illustrate the effect of the normalized step 15. Figure 4a contains a normal subject's pulsation signal 312 (above) in the resting state / component measurement. And an energy spectrum chart (below) obtained by the post-pulsation «312 through steps 13 and ι4. The energy spectrum chart shows that the first-frequency amplitude is about 2.㈣5. The fourth b 2o chart is the result of normalizing the fourth step a. It is adjusted to a heart rate of 80 beats / minute and the first main frequency amplitude is 5x105. In the same way, the fifth a includes the same as the fourth a and b-the normal subject's heart rate after exercise is 120 beats / component measured after the pulse signal 312 (top) and the month spectrum ( The figure below shows that the magnitude of the first main frequency of 568768 is about 6. lxl05. The fifth graph b is the result after the fifth a graph and the fourth b graph adopt the same normalization standard, that is, adjusted to a heart rate of 80 beats / minute and the first main frequency amplitude is 5xl05. Figure 6a contains a heart disease patient with a cardiomyopathy at rest 5 heartbeat 65 beats / component measured after the pulsation signal 312 (upper) and an energy spectrum (lower), the energy spectrum The first main frequency amplitude shown in the figure is about 8.9x1 05. The sixth b picture is the same as the fourth and fifth b pictures with the same normalization standard, that is, adjusted to a heart rate of 80 beats / minute and the first main frequency amplitude_ 5x1 〇5 result. It is shown in the sixth graph that the main frequency positions of the patient 10 of the cardiomyopathy are harder to identify than normal people shown in the fourth and fifth b graphs, and the non-zero energy spectrum appears frequently between the main frequencies. The analysis results obtained by the present invention measured by the big buy I show that in the spectrum of pulsating energy of people without heart disease, four to five main energy distributions (dominant frequencies) can be clearly read, which varies depending on the individual's heart rate, around 1 Hz Usually, the energy decreases gradually at the place where the maximum b occurs. Except for the main four or five frequency spectrums, there is no other clearly visible spectrum distribution, and the value on the frequency axis is mostly zero. Φ After the energy spectrum analysis of the pulsation signals measured by patients with heart disease, there are often clearly visible energy levels beyond the distinguishable four to five main frequencies.
20 T L 不如主頻能量之若干頻譜出現。該等非零值頻譜之出現越 廣，代表擷取之脈動訊號越不規則，而含有越多快速變動 之不穩定訊號，更嚴重者甚至不規則訊號能量極大，而無 法分辨出主頻位置。是故，心臟疾病患者能量頻譜圖中， 除主頻之外，頻域上出現能量非零值數量多寡，與心血管 12 568768 患病與否有極大關聯。 瞀是故，藉由下述心臟指數定義步驟16及心臟指數計 异步驟17,而將脈動訊號能量頻譜分析結果與心臟機 間之關聯判斷準則（criteHa)量化及公式化，使本發明可 5撰寫為一程式軟體型式，配合—館存有該程式軟體之儲存 媒體i如磁碟片、光碟片或硬碟）及—可執行該程式軟體I 電子6又備(如一電腦、個人數位助理機或醫用儀器），而自 · 動迅速計算判讀，供量測域者即時參考而無須等待盆社 ^ ° 八。 1〇 本實施例心臟指數定義步驟16中心臟指數（hean index)係定義如下。首先於步驟15所得能量頻错密度圖 中’將0點至第一主頻定義為第一區間，第一主頻至第二 主頻間定義為第二區間，第二主頻至第三主頻間定義為第 二區間’第三主頻後仍有第四甚至第五主頻，明顯程度因 u人而異’此處定義為第四區間。配合第七圖所示，當各段 區間内之能量頻譜分布狀況一旦符合下述[式三]之=件^ 即疋義於i點發生一具有意義之棘波。 鲁 若[s(i)-s(i-i)>v〇 或 s⑴—s(i—2)>v〇] 20 ^[S(i)-S(i + l)>V0S(i)-S(i+2)>V〇] [式三] 其中，Vq-Pi / 50 ’ Pi為第一主頻振幅。而第一至第 四區間内所得之棘波數總合，即定義為心臟指數。 而後於心臟指數計算步驟丨7，將步驟丨6獲得之定義 13 568768 用以計算前述第四b、五b及六b各圖之心臟指數，可獲 知第四b、五b圖中之心臟指數皆為〇。易言之，同_名 雙試者於運動前、後，其經正規化後之心搏及頻譜能量值 雖不同，然其心臟指數則相同，故心臟指數不因受試者於 5測®時之狀態所影響。至於第六b圖中心臟疾病患者經正 規化後其心臟指數為23,而與第四b、五b圖中正常受試 者明顯有異，故心臟指數確可作為心臟疾病之一量化客觀 之判斷指標。 然需指出者，心臟指數並非以上述[式三]之定義為 1〇 限，且如參數Vq大小亦可調整，舉凡以上述能量頻譜密 度圖中之棘波或其他波形變化情形作為心臟指數之計算依 據者，皆屬本發明實質範疇。 以下為將本發明實際應用於國内某醫院之臨床實測 及分析結果。受試者包括於該醫院心臟内科門診中求診人 15數201人，其中經醫師診斷判定為心臟病患者53人，非 心臟病患者148人。配合上述心臟指數之計算及臨床實際 診斷結果，可將心臟指數依以下分級，供作受試者心臟^ 能正常與否之參考依據·· 心臟指數7以上··屬於心臟異常跳動頻率頻繁而明 20 顯，心肌收縮不規則，此為多種心臟病之特徵。 心臟指數4〜6 ··代表心臟具有異當k 兴⑦跳動，屬於心臟病 高危險群或已罹患心臟疾病之特徵；及 心臟指數3以下··代表量測時心磁、心心t t u贓跳動頻率無異狀 ，心臟健康者心臟指數多為0。 14 568768 若再以、醫師判定為心臟病患者之53名受試者分析 ，其中包括瓣膜性疾病13例、心肌病變（包括心肌病變及 心肌火）12例、心律不整7例、冠狀動脈疾病21例，其 5 各種疾病患者與其對應心臟指數之相關統計如以下表一; 示0 &臟指數 種類 瓣膜性疾病 —(百分比） 4〜6 (百分比） 7以上 (百分比） 1 ( 7.69%) 11 ( 84.62%) 7.69%)20 T L is not as good as some of the frequency spectrum energy. The wider the appearance of such non-zero frequency spectra, the more irregular the captured pulsation signals, and the more unstable signals that contain more rapid changes. The more severe the irregular signals are, the greater the energy is, and the main frequency position cannot be resolved. Therefore, in the energy spectrum of heart disease patients, in addition to the main frequency, the amount of non-zero energy in the frequency domain appears, which is greatly related to the cardiovascular disease. Therefore, by using the following cardiac index definition step 16 and cardiac index calculation step 17 to quantify and formulate the correlation judgment criterion (criteHa) between the energy spectrum analysis result of the pulsation signal and the cardiac machine, the present invention can be written as 5 It is a type of program software that cooperates with—the library stores the program software i such as a magnetic disk, CD-ROM, or hard disk—and—executable program software I and 6 (such as a computer, personal digital assistant, or Medical instruments), and automatically calculate and interpret quickly, for those in the field of measurement and reference without having to wait for the pot society ^ ° eight. 10 The heart index in step 16 of the heart index definition step of this embodiment is defined as follows. First, in the energy frequency error density map obtained in step 15, 'zero point to the first main frequency is defined as the first interval, the first main frequency to the second main frequency is defined as the second interval, and the second main frequency to the third main frequency is defined. The frequency interval is defined as the second interval. 'There is still a fourth or even a fifth frequency after the third main frequency, and the degree of obviousness varies with u people.' Here is defined as the fourth interval. As shown in the seventh figure, when the energy spectrum distribution in each section meets the following [Formula 3] = ^, it means that a meaningful spike occurs at point i. Lu Ruo [s (i) -s (ii) > v〇 or s⑴-s (i-2) > v〇 20 ^ [S (i) -S (i + l) > V0S (i) -S (i + 2)> V〇] [Equation 3] where Vq-Pi / 50 'Pi is the first main frequency amplitude. The total number of spikes in the first to fourth intervals is defined as the cardiac index. Then in the heart index calculation step 丨 7, the definition 13 568768 obtained in step 丨 6 is used to calculate the heart index of each of the fourth b, fifth b and six b, and the heart index of the fourth b, five b can be obtained. All are 0. In other words, the normalized heart rate and spectral energy values of the same test subjects before and after exercise are different, but their heart indices are the same, so the heart index is not affected by the subject's 5 test ® Affected by the state of time. As for the heart disease patient in Figure 6b after normalization, his heart index is 23, which is obviously different from the normal subjects in Figures 4b and 5b, so the heart index can indeed be used as one of the quantitative and objective indicators of heart disease. Judgment indicators. However, it should be pointed out that the heart index is not defined as the 10 limit according to the above-mentioned [Formula 3], and the parameter Vq can also be adjusted. For example, the spike or other waveform changes in the energy spectrum density chart are used as the heart index. Those who calculate the basis belong to the essential scope of the present invention. The following are the clinical measurement and analysis results of applying the present invention to a hospital in China. The subjects included 15 201 patients who were consulted in the cardiology department of the hospital, of which 53 were diagnosed by the physician as heart disease patients and 148 were non-heart disease patients. With the calculation of the above-mentioned heart index and clinical diagnosis results, the heart index can be classified according to the following levels, which can be used as a reference for the subject's heart ^ Whether the heart index is 7 or higher. 20, irregular myocardial contraction, which is a characteristic of many heart diseases. Cardiac index 4 ~ 6 ·· Represents that the heart has abnormal k-beats, which belongs to a high-risk group of heart disease or has characteristics of heart disease; and a heart index of 3 or lower ·· Represents the cardiac magnetic and cardiac ttu frequency There is no abnormality, the heart index of healthy people is mostly 0. 14 568768 According to the analysis of 53 subjects who were diagnosed as heart disease by the physician, including 13 cases of valvular disease, 12 cases of cardiomyopathy (including cardiomyopathy and myocardial fire), 7 cases of arrhythmia, and coronary artery disease 21 For example, the relevant statistics of patients with various diseases and their corresponding cardiac indexes are shown in Table 1 below; Shows 0 & dirty index types of valve disease— (percent) 4 ~ 6 (percent) 7 or more (percent) 1 (7.69%) 11 (84.62%) 7.69%)
總計 10* 18.87%) 14 ( 26.42%) 29 ( 502%-)Total 10 * 18.87%) 14 (26.42%) 29 (502%-)
43 ( 81.13%)'43 (81.13%) '
表一各種疾病患者與其對應心臟指數之相關統計 由表一得知，於所有樣本中，瓣膜性疾病共13例， 其異常現象偵測率達84.62%，加上疑似異常偵測率7 69% 10 ，共達92.31% 。心肌病變之11個樣本中，異常現象偵 測率66.67%，加上疑似異常偵測率25%，可測得之比例達 91.67% ，亦屬容易測得之狀況。心律不整樣本數共7例 ，其中疑似異常比例為57. 14%、異常現象偵測率為 28.57% ，共85.71%。冠狀動脈疾病之樣本數共21例。 15 568768 其中異常現象偵測率38〗n〇/ ^ m/。，疑似異常偵測率28.57〇/0, 總和6 6 · 6 7 % ，遺漏率齡古品、去。n 年季又回而達33.33% ，其原因容後詳 述。 於此並需敘明者，由於總樣本數僅53例，母體樣本 5量小，以此母體計算偵測率，其獲得之百分比數字 尚不足 以作為必然之定論，而僅供參考。 以下則就上述各種心臟疾病與其對應心臟指數之關 聯及可能原因依序分析： 首先以心臟瓣膜性疾病而言，其可詳分為主動脈瓣 10 S常、二尖瓣異常、肺瓣膜異常、三尖瓣異常、心房或心 室中隔缺損等，不同原因導致不同之脈動頻譜分布。舉例 而言，若主動脈瓣因先天畸形或鈣化造成組織纖維化，導 致主動脈瓣狹窄，則當心臟收縮時，主動脈瓣之狹窄將阻 礙左〜至中之血液流入主動脈，視狹窄程度可能導致不同 15 嚴重程度。此時心臟將試圖彌補血液循環失調情況（代償 作用），以期使左心室流至主動脈中之血液量不致過低。 因此心臟即有幾種應變方式： 一、增加左心室收縮力量，此現象將導致左心室肌 壁增厚’因此可有更大力量推送血液；及 20 一、延長心臟收縮時間，心臟收縮時間較正常時間 長，期望可使由左心室進入主動脈之企液量多。 以上述方式增加左心室工作量、增強左心室血壓， 使心臟不斷有力且快速收縮，然脈搏可能仍屬微弱，患者 則有疲倦、呼吸困難、頭暈、胸痛等症狀。而在如第八圖 16 568768 所示其中一瓣膜性疾病患者心跳之擷取信號及能量頻譜分 析圖顯示，能量頻譜圖中可能因血液由左心室流入主動脈 時受到異常阻礙，流動遭受干擾而產生額外頻率信號；另 因心肌收縮之時間或力量改變，亦可由此分析中觀察出異 5 常，而其心臟指數為22。 第九圖則顯示其中一心肌病變患者心跳之擷取信號 及能量頻譜分析圖，其心臟指數為22。 如第十圖所示一心律不整患者心跳之擷取信號及能 里頻谱分析圖，其心臟指數為5。心律不整之個案，其心 10臟指數大多落在4至6間，百分比為57%，心臟指數在7 以上者佔28.57% ，顯示心律不整雖可於本分析方法中偵 知異常，然能量頻譜分析圖僅出現幾個額外頻率分布，不 似瓣膜性疾病或心肌病變患者之案例有明顯混亂之頻率分 布，使心臟指數高達1 〇或2〇以上。心律不整個案中，遺 15漏百分比雖高達14·29% ，然因總樣本數少，此數值不具 有太大統計意義。遺漏之案例為一 75歲男子，其信號特 別微弱，於此分析中未能找出異狀。 如第十一圖所示一冠狀動脈性疾病患者心跳之擷取 信號及能量頻譜分析圖，其心臟指數為8。冠狀動脈内血 20液流動之障礙，大多由動脈粥樣硬化引起，動脈血管内膜 上沉積膽固醇、類脂質等，使管道變窄，冠狀動脈疾病中 亦包含少數先天性畸形或冠狀動脈異常擴張。當冠狀動脈 發生病變，無法正常供應心肌氧氣與營養，便發生缺血性 心臟病。然缺血性心臟病患平日休息時，往往觀察不到， 17 568768 Z病患心臟卫作量增加時（如運動、環境驟變及情緒起 仇），此時心肌可能因缺氧而導致心絞痛。 冠狀動脈性疾病之個案中，偵測到之百分比只達 疑似異㊉為28· 57% ，未偵測到者高達33· 33% 。由此數字顯示，冠狀動脈機能異常病患中，約六成以上 會產生異常錢收縮頻率，而有三成以上於平日休息狀況 下則無法得知。 ίο 至於前述受試者非心臟病患者之148人中，高血壓 病患佔大多數而共11G W，高血壓及非高血壓患者與其對 應心臟指數之相關統計如以下表二所示。Table 1 The relevant statistics of patients with various diseases and their corresponding cardiac indexes are shown in Table 1. Among all the samples, there were 13 valvular diseases. The abnormality detection rate was 84.62%, and the suspected abnormality detection rate was 7 69%. 10, a total of 92.31%. Among the 11 samples of myocardial disease, the abnormality detection rate was 66.67%, and the suspected abnormality detection rate was 25%. The measurable ratio was 91.67%, which is also an easily detectable condition. The number of arrhythmia samples was 7 in total, of which 57.14% were suspected abnormalities, and the abnormality detection rate was 28.57%, a total of 85.71%. There were 21 samples of coronary artery disease. 15 568768 Among them, the abnormality detection rate is 38〗 n〇 / ^ m /. The suspected anomaly detection rate is 28.57〇 / 0, which is 66.67% in total. The n season returns to 33.33%, and the reason will be described in detail later. It is necessary to explain here, because the total number of samples is only 53, and the number of mother samples is small. Using this mother to calculate the detection rate, the percentage number obtained is not enough as an inevitable conclusion, but for reference only. The following is a sequential analysis of the correlations and possible causes of the above-mentioned various heart diseases with their corresponding heart indices: First, in terms of heart valve disease, it can be divided into 10 aortic valve abnormalities, mitral valve abnormalities, pulmonary valve abnormalities, Tricuspid valve abnormalities, atrial or ventricular septal defects, etc., cause different pulsation spectrum distributions for different reasons. For example, if the aortic valve is caused by fibrosis due to congenital malformation or calcification, resulting in stenosis of the aortic valve, when the heart contracts, the stenosis of the aortic valve will prevent the left to middle blood from flowing into the aorta, depending on the degree of stenosis May cause different severity. At this time, the heart will try to compensate for the blood circulation disorder (compensation effect), so that the amount of blood flowing from the left ventricle into the aorta will not be too low. Therefore, the heart has several strain modes: 1. Increase the left ventricular contractile force. This phenomenon will cause the left ventricular muscle wall to thicken so that it can push the blood with greater force; and 20 1. Prolong the systolic time. The normal time is long, and it is expected that the amount of enterprise fluid that can enter the aorta from the left ventricle is large. Increasing the workload of the left ventricle and increasing the blood pressure of the left ventricle in the above-mentioned manner keeps the heart constantly and rapidly contracting, but the pulse may still be weak, and the patient may have symptoms such as fatigue, dyspnea, dizziness, and chest pain. As shown in the eighth figure 16 568768, the heartbeat acquisition signal and energy spectrum analysis chart of a patient with valvular disease shows that the energy spectrum diagram may be abnormally blocked when the blood flows from the left ventricle into the aorta and the flow is disturbed. Generates extra frequency signals; and because of changes in the time or strength of myocardial contraction, abnormalities can also be observed from this analysis, with a cardiac index of 22. The ninth figure shows a heartbeat acquisition signal and energy spectrum analysis chart of one of the patients with cardiomyopathy, and the heart index is 22. As shown in the tenth figure, the heartbeat acquisition signal and the energy spectrum analysis chart of an arrhythmic patient have a heart index of 5. In cases of arrhythmia, the heart viscera index mostly falls between 4 and 6, with a percentage of 57%, and those with a heart index above 7 account for 28.57%, which shows that although an arrhythmia can be detected in this analysis method, the energy spectrum There are only a few additional frequency distributions in the analysis chart, which is not like the case of patients with valvular disease or myocardial disease, which has a significantly chaotic frequency distribution, which makes the heart index as high as 10 or more. In the case of arrhythmia, although the missing percentage is as high as 14.29%, this value is not of statistical significance because the total sample size is small. The missing case was a 75-year-old man whose signal was particularly weak, and no abnormalities were identified in this analysis. As shown in Figure 11, a heartbeat acquisition signal and energy spectrum analysis chart of a patient with coronary artery disease has a heart index of 8. The obstacles to the flow of 20 fluids in the coronary arteries are mostly caused by atherosclerosis. Cholesterol and lipids are deposited on the intima of the arteries to narrow the pipeline. Coronary artery diseases also include a few congenital malformations or abnormal expansion of the coronary arteries . Ischemic heart disease occurs when coronary arteries become diseased and myocardial oxygen and nutrition cannot be supplied normally. However, patients with ischemic heart disease are often not observed during weekdays. When the amount of cardiac work in patients with 17 568768 Z increases (such as exercise, environmental changes, and emotional revenge), the myocardium may cause angina due to hypoxia. . In the case of coronary artery disease, the percentage detected was only 28.57% of suspected foreigners, and 33.33% were detected. This figure shows that about 60% of patients with coronary artery dysfunction will have abnormal money contraction frequency, and more than 30% will not be known under weekday rest conditions. ο As for the 148 non-heart disease patients in the aforementioned subjects, hypertension patients accounted for the majority with a total of 11G W. The relevant statistics of hypertension and non-hypertension patients and their corresponding cardiac indexes are shown in Table 2 below.
5 ( 13.16%) 種類 高血壓患者5 (13.16%) type Hypertensive patients
(百分比） 2「( 23.64%)(Percentage) 2 "(23.64%)
非高血壓 非心臟病患者 31 (81.58%) 2 ( 5.26%) 7 ( 18.42%)' 以 (百分比） ΐΓ( 12.73%) 總計 mm m m mmmm ι 101 (68.24%) 31 ( 20.95%) 16 ( 10.81%)Non-hypertensive non-heart disease patients 31 (81.58%) 2 (5.26%) 7 (18.42%) '(percentage) ΐΓ (12.73%) Total mm mm mmmm 101 (68.24%) 31 (20.95%) 16 (10.81 %)
47 ( 31.76%) ΤΓ 表二高血壓及非高血壓患者與其對應心臟指數之統計 由表二可知，高血壓族群心臟指數大於4之比例為 15 36·37%，較非高血壓者18· 42%高。易言之，高血壓病患之 心肌收縮頻率疑似異常或異常之比例較一般人高。事實上 局血壓族群為心血管疾病罹患之高危險群之一，且其中少 數已有徵兆。第十二圖所示為其中一非心臟病且非高血壓 18 568768 病患心跳擷取信號及能量頻譜分析圖，其心臟指數為0， 其規則且清楚顯示之主要頻率分布與前述第八至十—圖之 混乱情形有明顯區別。 綜上所述’本發明較習知技術具有下述優點： 本發明直接將由非侵入式之血遷計量測所得而 -般未再予利用之脈動職，經簡單處理後作為判定受試 者心臟機能之參考依據，而無須其他任何額外甚至昂貴t - 脈動或心搏量測儀器設備，故可謂一經濟、安全、簡便而 . 了大量推廣深入普及至-般家庭，供使用者於日常量測血 # 10屋之際’ 一併測知其心臟機能狀態而作為預防保健或就診 參考。 二、以第四a、五a圖之正常受試者及第六a圖之心 肌病變患者脈動能量頻譜圖差異所顯示，該心肌病變患者 能量頻譜圖各主頻位置較正常人者難以辨識，且各主頻間 15非零能量頻譜出現頻繁，故可以脈動能量頻譜圖作為判定 受試者是否具心臟疾病之初步依據。 二、藉由將能量頻譜圖進一步經正規化處理，其呈現 春 之心臟指數特徵無論受試者係於休息或運動後狀態下皆為 一致’使本發明不論受試者處於何種活動狀態、時間、地 20 點皆可隨時量測應用而不影響結果，相較大多數習知技術 需於特定狀態下測量人體生理數據之限制，本發明顯具高 度使用彈性。 · 四、如前所述，本發明可以一程式軟體配合一電腦自 動迅速計算，供受試者即時參考而無須長時間等待其判讀 19 568768 結果’對受試者而言可及早預防或就診，對醫師而言亦可 即時採取對應之診斷治療程序，使受試者健康更獲保障。 五、 於實際應用準確性上，可由本發明偵測得心臟機 能異常之比例，視心臟病原因種類而異，並非所有心臟疾 5 病皆導致心肌收縮或脈動頻率異常，於前述樣本中獲得之 平均測出率則約80% ，足證本發明確具相當實用性及可 靠度。 六、 於上述樣本中，經醫師診斷為心臟正常者，藉本 發明所判定結果仍有近32%者予以質疑（特別為高企壓 1〇患&本發明偵測判定心臟病之標準似較醫師嚴格， 而可發揮預先警示病人可能罹患心臟病之效能。 惟以上所述者，僅為本發明之較佳實施例而已，當不 能以此限定本發明實施之範圍，即大凡依本發明申請專利 範圍及發明說明書内容所作之簡單等效變化與修飾，皆應 15 仍屬本發明專利涵蓋之範圍内。 【圓式簡單說明】 ▲第一圖為本發明脈動訊號與心臟機能相關性分析方法 較佳實施例之步驟流程圖； 20 完整原始訊號圖； 第二圖為該較佳實施例之一 脈動量測步驟所獲得之一 一置測所得之後段脈動訊號圖 能量頻譜轉換步驟所得之能量 第三a及三b圖分別為一量 ，及由該後段脈動訊號經一能量 頻譜圖； 第四a圖為一 正常受試者於靜止狀態 心搏80下/分量 20 568768 測所得之後段脈動訊號及能量頻譜圖，第四b圖為第四a 圖經頻率正規化及能量正規化之結果； 第五a圖為該正常受試者於運動過後心搏12〇下/分 量測所得之後段脈動訊號及能量頻譜圖，第五b圖為第五 5 a圖經頻率正規化及能量正規化之結果； 第六a圖為一心肌病變之心臟病患者於靜止狀態心搏 65下/分量測所得之後段脈動訊號及能量頻譜圖，第六b ' 圖為第六a圖經頻率正規化及能量正規化之結果； · 第七圖為該較佳實施例之一心臟指數定義步驟各參數 鲁 10 示意圖； / 第八圖為一瓣膜性疾病患者心跳之擷取信號及能量頻 $普分析圖； 第九圖為—心肌病變患者心跳之擷取信號及能量頻譜 分析圖； 15 ^ $十圖為一心律不整患者心跳之擷取信號及能量頻譜 分析圖； 第十一圖為一冠狀動脈性患者心跳之擷取信號及能量 · 頻譜分析圖；及 第十二圖為一非心臟病患者心跳之擷取信號及能量頻 2〇譜分析圖。 21 568768 【圚式之主要元件代表符號簡單說明】 11…… …·脈動量測步驟 12…… …·前段訊號去除步驟 13…… …·頻域轉換步驟 14…… …·能量頻譜轉換步驟 15…… …正規化步驟 151… …能量正規化步驟 152… …頻率正規化步驟 16…… ......臟指數定義步驟及 17…… …心臟指數計算步驟 31…… …原始訊號 311… …前段脈動訊號 312 ... …後段脈動訊號 2247 (31.76%) ΤΓ Table 2 Statistics of hypertension and non-hypertensive patients and their corresponding cardiac indices From Table 2, it can be seen that the proportion of patients with a cardiac index greater than 4 is 15 36 · 37%, which is 18 · 42 compared with non-hypertensive patients. %high. In other words, the frequency of myocardial contractions in patients with hypertension is suspected to be abnormal or higher than in the general population. In fact, the local blood pressure group is one of the high-risk groups of cardiovascular disease, and a few of them have already shown signs. The twelfth figure shows a heartbeat acquisition signal and energy spectrum analysis of one of the non-heart disease and non-hypertensive patients. The heart index is 0. The regular and clearly displayed main frequency distribution is the same as the eighth to eighth. Ten-the chaos of the picture is clearly different. In summary, the present invention has the following advantages over the conventional technology: The present invention directly uses the non-reusable pulse function directly measured by non-invasive blood migration measurement, and is used as a judgement subject after simple processing. The reference basis for cardiac function without any additional or even expensive t-pulsation or heartbeat measurement equipment, so it can be described as economical, safe, simple and convenient. A large number of promotions have been popularized to ordinary families for users to measure daily测 血 # 10 屋 之 'also measures their heart function status as a reference for preventive health care or consultation. 2. As shown by the differences in the pulse energy spectrum of normal subjects in the fourth a and fifth a and the myocardial disease patients in the sixth a, the main frequency positions of the energy spectrum of the myocardial disease patients are more difficult to identify than normal people And 15 non-zero energy spectrums appear frequently between each main frequency, so the pulsation energy spectrum can be used as a preliminary basis for judging whether the subject has heart disease. Second, by further normalizing the energy spectrum, it shows the characteristics of the heart index of spring, regardless of whether the subject is at rest or after exercise. The 20 points can be measured and applied at any time without affecting the results. Compared with the limitation that most conventional technologies need to measure human physiological data in a specific state, the present invention is highly flexible. · 4. As mentioned above, the present invention can automatically and quickly calculate with a program software and a computer for the subject's immediate reference without waiting for a long time to read 19 568768 results. 'It can be prevented or treated early for the subject, For physicians, the corresponding diagnostic and treatment procedures can be taken immediately, so that the health of the subjects can be more protected. 5. In terms of actual application accuracy, the proportion of abnormal heart function that can be detected by the present invention varies depending on the type of heart disease. Not all heart diseases cause cardiac contraction or abnormal pulse frequency. The average detection rate is about 80%, which proves that the present invention is quite practical and reliable. 6. Among the above samples, those diagnosed by the doctor as having a normal heart are still questioned by nearly 32% of the results determined by the present invention (especially high-pressure patients with 10 diseases & the standard for detecting and determining heart disease by the present invention seems to be more The physician is strict, but can exert the effect of warning patients in advance that they may suffer from heart disease. However, the above are only the preferred embodiments of the present invention. When the scope of the implementation of the present invention cannot be limited in this way, that is, those who apply according to the present invention The simple equivalent changes and modifications made to the scope of the patent and the contents of the invention description should all be within the scope of the invention patent. [Circular brief description] ▲ The first figure shows the correlation analysis method of the pulsation signal and cardiac function of the invention The flowchart of the steps of the preferred embodiment; 20 The complete original signal diagram; The second diagram is the energy obtained by the energy spectrum conversion step of the subsequent pulse signal diagram obtained from one of the pulse measurement steps of the preferred embodiment. The third a and three b graphs are a quantity, respectively, and an energy spectrum diagram is passed by the post-pulsation signal; the fourth a graph is a normal subject in a resting state. Pulse 80 beats / component 20 568768 The pulsation signal and energy spectrum of the latter segment are measured. Figure 4b is the result of frequency normalization and energy normalization in Figure 4a. Figure 5a is the normal subject ’s exercise. The subsequent pulse signal and energy spectrum graph after the heartbeat is measured at 120 beats / component. The fifth b graph is the result of frequency normalization and energy normalization of the fifth 5 a graph. The sixth a graph is the result of a myocardial disease. Heartbeat patients' pulse signal and energy spectrum at the next resting heartbeat of 65 beats / components. The sixth b ′ is the result of frequency normalization and energy normalization in the sixth a; One of the preferred embodiments is a schematic diagram of the parameters of the heart index definition step; / Figure 8 is a diagram of the heartbeat acquisition signal and energy frequency analysis of a patient with valvular disease; Figure 9 is- Acquired signal and energy spectrum analysis chart; 15 ^ $ 10 is an acquired signal and energy spectrum analysis chart of a heartbeat of an arrhythmia patient; Fig. 11 is an acquired signal and energy spectrum analysis of a heartbeat of a coronary patient Figure; and The second figure is a 20-spectrum analysis signal of the heartbeat acquisition signal and energy frequency of a non-heart patient. 21 568768 [Simplified explanation of the representative symbols of the main components of the formula] 11 ……… · Pulse measurement step 12 ……… · Previous paragraph Signal removal step 13 ... frequency conversion step 14 ... energy spectrum conversion step 15 ... normalization step 151 ... energy normalization step 152 ... frequency normalization step 16 ... .Dirty index definition steps and 17 ......… Heart index calculation step 31 ……… Original signal 311… Front stage pulsation signal 312 ...… Post stage pulsation signal 22