CN1113739A - Function testing system for vegetative nervous system and its method - Google Patents

Abstract

The present invention relates to a system for testing the functions of vegetative nervous system of human body and its method, including: irritating sympathetic nervous system and/or parasympathetic nervous system, sectionally making R wave analysis of electrocardiosignal (ECG) according to the difference in irritation and using autoregression (AR) model to make power spectrum analysis for it so as to obtain the quantitiative analytic results on the excitabilities of subject's sympathetic nervous system and parasympathetic nervous system, strong or weak, and their extent.

Description

Function testing system for vegetative nervous system and its method

The present invention relates to a kind of system and method thereof that human body autonomic nervous system function is tested, more specifically, relate to a kind of being used for by heart rate variability is carried out analysis of spectrum, and then the system that the function status of sympathetic nerve and parasympathetic nerve is detected, and the method that the function status of sympathetic nerve and parasympathetic nerve is detected and analyzes.

Among people's health, comprise the autonomic nervous system that sympathetic nerve and parasympathetic nerve are formed,, controlling inside of human body physiology and biochemical movable just like the controller in the automatic control system, adjust the activity of nonvoluntary organ, and then reach the balance of the internal and external environment of health.So to the detection of autonomic nervous system function is to the activity of the cardiovascular system of human body and even the important content that the whole body internal organs irritability power that is subjected to autonomic nervous system control is estimated and diagnosed.

In the such control system of human body, be positioned at the parasympathetic nerve maincenter of oblongata and directly controlling the unify active state of some nonvoluntary organs of cardiovascular system in the other sympathetic chain of spinal cord and nerve fiber thereof.In general, symptoms such as palpitating speed, nervous night sweat, diarrhoea, face flushing, emotion agitation can appear in the patient that sympathetic excitability is strong excessively; And opposite symptom then often appears in the strong excessively patient of parasympathetic nerve irritability.But, up to the present, also do not find a kind of method or instrument in medical circle, can be directly to the functional status of sympathetic nerve and parasympathetic nerve, promptly the irritability power is carried out qualitative and even quantitative analysis and measurement.Because except in laboratory, doing zoopery, can not only surgical incision experimenter's sympathetic and parasympathetic nerve detects it for test purpose.At present can accomplish, be according to above-mentioned mechanism, inquire the main suit who suffers from the patient of autonomic nervous system dysfunction under a cloud by the doctor, allow patient do comprehensive inspection of each tract then.After the organic change of having got rid of each internal organs, it is autonomic nervous system disorder (or being called the autonomic nervous system disorder) with subjective mind.So and that the sympathetic nerve that arrives still is the parasympathetic nerve irritability is strong, then only by supposition.This makes, and interval between diagnosis prolongs, the diagnostic fees costliness, and this diagnosis that draws on the basis of exclusion method, and its accuracy rate also is unfavorable.

It has been recognized that the dancing speed of human heart is subjected to the excitatoty control of sympathetic nerve and parasympathetic nerve.Sympathetic excitability is strong, can cause heart rates to be accelerated; And the parasympathetic nerve irritability is strong, then can cause slowing down of heart rates.So can be by heart rates as a window, the functional status of sympathetic nerve and parasympathetic nerve is detected and judges.Particularly in recent years, the institute's target paying attention to and study that the variability of heart rate (Heart Rate Variability(HRV)) more and more becomes people.It is relevant with the physiology and the pathological state of heart that the variability of heart rate (HRV) is considered to, be cardiovascular system internally, the reaction that changes of external environment and to a performance of heart disease.Therefore, the spectral analysis technology of the analysis of heart rate variability, particularly development in recent years for estimating physiology, pathological state, especially the cardiovascular autonomic nervous system function of cardiovascular system, provides a kind of noninvasive research method and important quantitative index.

Early stage in laboratory the analysis to heart rate variability mainly adopt statistical method, the frequency rectangular histogram that occurs as the average of R-R interval and standard deviation, all lengths R-R interval etc.These methods can be differentiated the variation that autonomic nerve is movable total, but can not distinguish the interaction situation between sympathetic and the parasympathetic nerve.Therefore, in recent years, people develop into the employing power spectrum analysis method again, comprise the spectrum method of estimation of fast Fourier transform (FFT) method, periodogram and autoregression model.Because for short time data, the getable frequency resolution of autoregression model method has better estimation effect than classical FFT and period map method, therefore now adopts the autoregression model method to carry out the power spectrumanalysis of heart rate variability more.But this research also all only rests on the level of academic discussion, does not also have a kind of instrument and equipment that detects the autonomic nervous system function according to the heart rate variability analysis of spectrum to be made and is used among the clinical medicine.

Therefore, the objective of the invention is to, a kind of system and method that can not have wound ground testing human autonomic nervous system functional status is provided, thoroughly changes in the prior art and to get rid of by the detection of each physiological system being carried out cumbersome complexity and infer the method for diagnosing.

Further aim of the present invention is, provide a cover complete autonomic nervous system functional test equipment, it can be by the test to the heart rate variation tendency, and the interaction between differentiation experimenter's sympathetic nerve and the parasympathetic nerve, and it is strong to provide in sympathetic nerve and the parasympathetic nerve which irritability, and what qualitative and quantitative diagnostic data by force.

The present invention more further purpose be, provide a cover in function testing system for vegetative nervous system, to move, make the software of the quantitative property conclusion of interaction between experimenter's sympathetic nerve and the parasympathetic nerve, functional status by the test data of heart rate variability being carried out analyzing and processing.

According to the present invention, provide a kind of by to the collection of electrocardiosignal and the system and method for heart rate variability being analyzed and then detected the autonomic nervous system function.System of the present invention comprises an electrocardiosignal memory, the experimenter's that collected within a certain period of time electrocardiogram is stored in this memory, this memory can be portable or SCM type, it can have an infrared transmitter, be used to export the electrocardiosignal of being stored, correspondingly, this system can also comprise that one is used for receiving and by interface circuit, the electrocardiosignal of described output being sent into the infrared remote receiver that computer is handled; A pico computer is used for electrocardiosignal is carried out power spectrumanalysis, to draw the qualitative report after the testing of experimenter's sympathetic nerve and parasympathetic nerve system; An output device is used to print the check and analysis result.System of the present invention also provides and instructs the experimenter to move, adjust different position states so that gather the voice unit and the speaker of electrocardiosignal under the state that sympathetic and/or parasympathetic nerve system is carried out needed stimulation simultaneously.

Ultimate principle of the present invention is, utilize on the one hand position and breathing state different, utilize the not same-action of medicine (not administration, administration, give different medicines) on the other hand, respectively sympathetic nervous system and/or parasympathetic nerve system are stimulated, detect the heart rate degeneration before and after stimulating, utilize this window, carry out a series of signal controlling reasons that will touch upon and analyze, thereby draw in the human body sympathetic nervous system the unify interaction of parasympathetic nerve system and the diagnosis of functional status.

According to the present invention, utilize Zymography to analyze heart rate variability.At first be exactly from primary ECG signal, to extract each R-R interval, form the R-R interval series.Usually adopt first-class intervening sequence, its range value equals time value between R-R, also can adopt interpolation method to carry out the constant duration sampling.Once the someone proved that these two kinds of methods were equivalent, and for the jumping of gushing of chance, premature beat and other interference can be carried out pretreatment by ways such as filtering, interpolation.

Aim at heart rate variability signal at random, can be similar to usually and think a stationary random process, therefore, can regard white noise W(n as by spectrum density=No) a certain impulse response of excitation is h(n) linear system produced.

X _n＝W _n＊h _n

H wherein _nBe total sharp response, * is a convolution, W _nBe white noise.

Z＝e ^jw

S then _x(e ^Jw)=S _w(e ^Jw) | H(e ^Jw) | ²

＝No｜H（e ^jw）｜ ²

S wherein _w, S _xBe respectively the power spectrum of input and output, H is the frequency characteristic or the transfer function of system, therefore, as long as H(Z) and as can be known, the power spectrum S of stochastic signal _x(e ^Jw) just can ask.H(Z) promptly be the parameter model of stochastic signal.What the present invention adopted is autoregression (AR) model (claiming all-pole modeling again), and its expression formula is as follows:

X（n）＝- $\underset{K=}{\overset{M}{\mathrm{\Σ}}}$ a（k）X（n-k）+W（n）

(X（Z）)/(W（Z）) ＝H（Z）＝ $\frac{1}{1+\underset{K=1}{\overset{M}{\mathrm{\Σ}}}a\left(K\right)Z^{\·k}}$

Wherein M is the AR model order, can be according to some criterion, determine as AIC or FPE criterion and signal properties.By the model parameter estimation algorithm, can try to achieve the parameter a(k of AR model), and white noise spectrum density No, so can pass through the density of following formula estimated power spectrum:

$S_X\left(f\right)=\mathrm{No\&Delta;t}/{|1+\underset{K+1}{\overset{M}{\mathrm{\&Sigma;}}}{a}_{M,K}\exp (-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="941075710\_DRIMG10.png,941075710\_DRIMG11.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;fk\&Delta;t})|}^2$

Utilize system and method for the present invention, within a certain period of time, the continually varying electrocardiosignal (ECG) under posture (put down and crouch and stand) that collection is different with the record experimenter and the different breathing states (freely breathing and controlled breathing).At first these ECG signals are carried out the identification of R ripple, adopt the self adaptation criterion, discern the R ripple, find out the R-R interval, thereby provide R wave train signal with first derivative.The ECG signal is carried out pretreatment adopt the low pass difference, try to achieve the first derivative of ECG, its formula is:

d（t）＝ECG（t）-ECG（t-4）

Carry out the single order low-pass filtering then, its first derivative through low-pass filtering is:

f（t）＝f（t-1）+d（t）-d（t-6）

In the electrocardiogram self study of (for example 5 seconds), after promptly experimenter's general ECG being learnt and adapting to, to above-mentioned f(t through after a while) find out maximum value PK and two R-R at interval.For example with H ₀(O)=0.7*PK as the meansigma methods of the threshold value of identification R ripple, two R-R as the average R-R initial value of (RRAV) at interval.Sequentially calculate f(t), find out PK(n)＞H ₀(n) point, and judge whether R involves R ripple position for it.Here, the self adaptation recurrence of threshold value is:

Through above pretreatment and according to the difference of above-mentioned position and breathing state, signal is carried out segmentation.Carry out autoregression (AR) model spectra estimation for every segment signal, in the hope of the AR model parameter.Estimation to the AR model parameter comprises model coefficient a ₁-a _MEstimation, the estimation of white noise spectrum density No, and the estimation of the order M of model, from time domain, transform to frequency domain, power spectrum in the hope of the tested ECG signal of each section, find out each eigenvalue, final through the statistical theory analysis, draw comparison and conclusion to sympathetic nerve under the different conditions and parasympathetic nerve system irritability situation.

The present invention has changed existing means to the autonomic nervous system functional test, make and academicly utilize the power spectrum method of heart rate variability (HRV) to combine up to now, for the first time heart rate variability (HRV) power spectrum analysis method is successfully applied among the test to the autonomic nervous system function clinically with clinical research.For the test of autonomic nervous system function provides more directly and accurate means.

Utilize the present invention, not only can understand the trend of changes in heart rate, can also distinguish the interaction between certain sympathetic nerve and the parasympathetic nerve, changed to utilize in the prior art and got rid of the deduction method, judge it is sympathetic nerve or the strong state of parasympathetic nerve irritability by detecting, utilize test macro of the present invention to carry out the autonomic nervous system functional diagnosis, only need 15 minutes, just can obtain the quantitative diagnostic result of sympathetic nerve and parasympathetic nerve functional status, for example obtain sympathetic nervous system or parasympathetic nerve system irritability is strong, strong what etc. parameter; And whether take a favorable turn after the treatment, the improvement degree how, also can be at an easy rate from test curve and numeral acquisition conclusion.

By below in conjunction with the detailed description of accompanying drawing to embodiment, the above and other purpose of the present invention, feature and advantage can become clearer and more definite.In the accompanying drawings:

Fig. 1 is the sketch map according to function testing system for vegetative nervous system of the present invention.

Fig. 2 is the sketch map of an embodiment of the test process of system and a method according to the invention.

The software flow pattern of Fig. 3 in function testing system for vegetative nervous system according to the present invention, moving.

Fig. 4 is according to electrocardiosignal memory electrical schematic diagram in the test macro of the present invention.

Fig. 5 is the schematic block diagram of infrared remote receiver in the system shown in Figure 1.

Fig. 6-Figure 12 is respectively in the control experiment of not administration resulting HRV curve under four kinds of states, wherein:

Fig. 6 is the curve of under the free breathing state of prostration position (be among Fig. 2 in the A section);

Fig. 7 is the curve of under the controlled breathing state of prostration position (be among Fig. 2 in the B section);

Fig. 8 is the curve of under the free breathing state of the position of standing (be among Fig. 2 in the C section);

Fig. 9 is the curve of under the controlled breathing state of the position of standing (be among Fig. 2 in the D section);

Figure 10 freely breathe when being whole prostration position and the state of controlled breathing under the (curve among Fig. 2 in the A+B=E section;

Figure 11 be whole stand freely breathe during position and the state of controlled breathing under the curve of (among Fig. 2 in the C+D=F section);

Figure 12 be comprised prostration position in whole 15 minutes in the test process and the position of standing under freedom and controlled breathing state under the curve of (being A+B+C+D=E+F=R section among Fig. 2).

Figure 13-the 18th, the contrast figure of each section time domain X letter corresponding and curve behind the filtering direct current composition and the frequency domain power spectrum signal S behind the process AR model transferring with Fig. 6-12 difference; Wherein

Figure 13 be with not administration contrast experiment shown in Figure 6 in the contrast figure of X signal and power spectrum signal S in the flat A section of freely breathing that crouches;

Figure 14 is the X signal corresponding with B section shown in Figure 7 and the contrast figure of S signal;

Figure 15 is the X signal corresponding with C section shown in Figure 8 and the contrast figure of S signal;

Figure 16 is the X signal corresponding with D section shown in Figure 9 and the contrast figure of S signal;

Figure 17 is the X signal corresponding with E section shown in Figure 10 and the contrast figure of S signal;

Figure 18 is the X signal corresponding with F section shown in Figure 11 and the contrast figure of S signal;

Figure 19 is the three-dimensional curve diagram of the A, the B that make for the Z axle with the fragmentation state, C, four curves of D.

Figure 20 is the three-dimensional curve diagram of the A, the B that make for the Z axle with the fragmentation state, C, D, E, six curves of F.

Figure 21 is according to the quantitative conclusion report instance of the position that provides after the test macro analytical calculation of the present invention to the vegetative nerve function influence.

Figure 20 is according to the quantitative conclusion report instance of the breathing that provides after the test macro analytical calculation of the present invention to the vegetative nerve function influence.

Referring to Fig. 1, show schematic diagram according to function testing system for vegetative nervous system of the present invention.In system of the present invention, comprise a SCM type ECG signals collecting storing apparatus 10, collect different position states in certain cardiac diagnosis lead mode on one's body from the experimenter and be stored within the electrocardiosignal memory 11 with ECG signal under the different breathing states (Fig. 2).Fig. 4 shows the electrical schematic diagram of electrocardiosignal memory, wherein: U1 is a central processor CPU, U2 is read-only memory ROM, U3 is an address latch, and U4 is a random memory ram, and U5 is a modulus converter A/D, U6 is a line driver, U7 is two input four nor gates, and U8 is the ecg signal amplifier of being made up of four operational amplifiers, and LEAD is the electrocardiosignal input.The electrocardiosignal that collects by this end is delivered to A/D converter U5 after through ecg amplifier U8, delivers to CPU(U1) handle, be stored among random memory (RAM) U4 and read-only memory (ROM) U2.

In the present embodiment, gather and store 15 minutes electrocardiosignaies.Referring to Fig. 2, because flat position that crouches and the controlled breathing faster than free style respiration rate stimulate the parasympathetic nerve system respectively, position of standing and free style are breathed then stimulates sympathetic nervous system respectively.So the present invention was divided into such four stages of respectively sympathetic and/or parasympathetic nerve system being carried out different stimulated with 15 minutes: in first section time A, the experimenter takes prostration position (stimulation parasympathetic nerve) freely to breathe (stimulation sympathetic nerve); Second section time B, experimenter's position still crouch (stimulation parasympathetic nerve) for flat, but make controlled breathing into, promptly breathe under a guiding faster than the beat of freely breathing that is provided by voice unit (stimulation parasympathetic nerve); In the 3rd section time C, experimenter's position changes stand (stimulation sympathetic nerve) into by flat crouching, and free style is breathed (stimulation sympathetic nerve); The 4th section time D breathed (stimulation parasympathetic nerve) for the experimenter for the controlled formula of stand (stimulation sympathetic nerve).In the present embodiment, the A section is 3 minutes; The B section is 4 minutes; The C section be 4 minutes and wherein the signal in first minute to be human body become signal in the position process of standing, special treatment in processing by prostration position; The D section also is 4 minutes.The electrocardiosignal memory was tested these the four kinds electrocardiosignaies under the different position breathing patterns continuously in 15 minutes.By an infrared transmitter 12(referring to Fig. 4), the electrocardiosignal of being stored is transferred to microcomputer 23.Microcomputer 23 is to be connected with infrared remote receiver 21 by interface 22, and then receives the electrocardiosignal (ECG) that is received by infrared remote receiver 21.

Fig. 5 shows the schematic diagram of the infrared remote receiver 21 that can be used for system of the present invention.Wherein the signal that is received through infrared receiving tube D1 carries out computing by operational amplifier IC1 and amplifies, and is anti-phase by inverter ic 3 comparator IC2 compares computing after, again via line drive IC4 driving and export this output signal D _OutDeliver to the RS232 mouth of microcomputer 23 through socket S2.S1, S2, S3 are socket among the figure.Certainly, also can not adopt infrared transmission and adopt the mode of wire transmission to realize that the signal between electrocardiosignal memory and the pico computer transmits, at this moment can utilize to drive chip and link to each other with the serial communication interface of microcomputer.

In test macro of the present invention, also provide a voice unit and a speaker by microcomputer 23 control, in order to send commander experimenter's action (from clinostatism to stand, from freely breathing controlled breathing etc.) order.Certainly, this linguistic unit also can be done with storing apparatus 10 with the portable cardiac signals collecting with speaker and be in the same place.

Can comprise sound card, multifunction card, display card hard disk, floppy drive, mainboard, power supply etc. in the microcomputer 23.Also have an output device 24, in order to the curve of the S signal of getting time domain X signal in the different sections and frequency domain and analysis report (referring to Fig. 6-Figure 22).

Fig. 3 shows MICROCOMPUTER PROCESSING ECG signal, carries out power spectrumanalysis, thereby makes the software flow pattern of the quantitative result of sympathetic nerve, parasympathetic nerve systemic-function state.

Referring to Fig. 3, after infrared transmission, send into pico computer 32 by the electrocardiosignal that ecg signal acquiring and storing apparatus 10 are sent.In pico computer 32, for example electrocardiogram self study in 5 seconds of at first carrying out certain hour enters R ripple identification step S12, adopts the self adaptation criterion, discerns the R ripple with first derivative, sends R wave train signal (i.e. time value between the R-R that gathers sequentially).It comprises the first derivative of trying to achieve ECG with the low pass difference

d（t）＝ECG（t）-ECG（t-4）

First derivative to the ECG signal is carried out low-pass filtering, obtains:

f（t）＝f（x-1）+d（t）-d（t-6）

Then to f(t) find out maximum value PK and two R-R at interval.With H ₀(O)=and 0.7*PK is as the threshold value of identification R ripple, and the meansigma methods of two R-R intervals is as the average R-R initial value of RRAV at interval.Process order computation f(t) find out PK(n)＞H ₀(n) point involves the position of R ripple so that determine whether R.Here, the recurrence formula of threshold adaptive is:

Then, at step S13 resulting R-R interval time-domain signal is advanced segmentation and pretreatment.Promptly resulting R-R interval time-domain signal (R wave train signal) is intercepted and handles according to segmentation shown in Figure 2 (the E section of A, B, C, D and the whole prostration position of representative and represent the F section of the whole position of standing), the R wave train signal of here each section being told is called the X signal, and it is sectional R-R interval time-domain signal.Figure 16-the 18th, the R-R interval waveform of the control experiment of when the experimenter is not applied medicine, being done record.Wherein Figure 18 is whole R-R interval oscillogram, and Figure 16-17 then is respectively the R-R interval waveform of A, B, C, D, E, F section.The symbol of indicating in the drawings " XA1200A ", " XB1200A " ... in, second letter is represented the branch segment labeling that this curve is taken from, last alphabet shows the character of experiment, and not administration is promptly arranged.Here A represents the experiment of not administration.As: indicate among Figure 17 that " XB1200A " represents that this curve is promptly flat crouch controlled respiration phase, the not control experiment of administration of B section; It is that whole flat crouching freely breathed the control experiment of the not administration that adds controlled breathing section for the E section that " XE1200A " that and for example indicates among Figure 10 then represents this curve, or the like, the rest may be inferred by analogy for it.

To it should also be noted that the accuracy in order testing, to prevent omission, also provide back checking functions in the treatment step of system, when R-R＞1.8RRAV, inspection is returned in the possible omission in centre, the threshold value of returning inspection is:

H ₁（n）＝0.5＊H ₀（n）

After the R-R interval is carried out above-mentioned analyzing and processing,, utilize autoregression (AR) model to compose estimation from step S21.

The following describes the ultimate principle of autoregression of the present invention (AR) model spectra estimation.

According to the Wold decomposition theorem, one flat section stochastic signal X(n) can regard white noise W(n as by spectrum density=No) a certain impulse response of excitation is h(n) linear system produced, that is:

X _n=W _n* h _n(*-convolution)

Z＝e ^jw

S then _x(e ^Jw)=S _w(e ^Jw) | H(e ^Jw) | ²

＝No｜H（e ^jw）｜ ²

S wherein _w, S _xBe respectively the power spectrum of input and output, H is the frequency characteristic or the transfer function of system.Therefore, as long as H(Z) known, the power spectrum S of stochastic signal _x(e ^Jw) just can ask.H(Z) promptly be the parameter model of stochastic signal.

Autoregression (AR) model claims all-pole modeling again, and its expression formula is as follows: $\mathrm{Xn}=-\underset{K=1}{\overset{M}{\mathrm{\&Sigma;}}}{a}_K{X}_{n-k}+\mathrm{Wn}$ （1）

Make transform:

(X（Z）)/(W（Z）) ＝H（Z）＝ $\frac{1}{1+\underset{K=1}{\overset{M}{\mathrm{\&Sigma;}}}{a}_K{Z}^{\&CenterDot;K}}$ （2）

After trying to achieve AR model parameter and white noise spectrum density No, just be easy to the power spectral density of estimated signal

$S_X\left(f\right)=\mathrm{No\&Delta;t}/{|1+\underset{K=1}{\overset{M}{\mathrm{\&Sigma;}}}{a}_{M,K}\exp (-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="941075710\_DRIMG10.png,941075710\_DRIMG11.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;fk})|}^{:2}$

Wherein f is frequency (Hz), and △ t is the sampling interval.

The estimation of AR model parameter comprises: model coefficient a ₁-a _M, white noise spectrum density No and model order M.

Estimation approach is not only a kind of, only illustrates the two-way LS(least square that adopts in the program here) prediction the MARPLE method.This kind method is typical B usg algorithm relatively, has higher frequency resolution, do not compose the composition division and observe, and frequency drift is less, has equal extent on amount of calculation, with square being directly proportional of AR.Rudimentary algorithm is as follows:

Be provided with the flat section of full limit stochastic process, extend out from (1) formula forward direction and back forecast error are shown below respectively:

The forward prediction error:

$f_{M,K}=X_{K+M}+\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}{a}_{M,i}{X}_{K+M-i};1\&le;K\&le;N-M$ （3）

$b_{M,K}=\underset{i=0}{\overset{M}{\mathrm{\&Sigma;}}}{a}_{M,i}{X}_{K+i};1\&le;K\&le;N-M$ （4）

a _{M, 0}=1; The M:AR model order; The N:X data length

According to the lowest mean square criterion, make forward and backward forecast error sum e _MFor minimum is tried to achieve AR model parameter, a _M1, a _M2..., a _{M, M},

（5）

Differentiate:

Here: $r_M(i,j)=\underset{K=1}{\overset{N-M}{\mathrm{\&Sigma;}}}(X_{K+M-j}{X}_{K+M-i}+X_{K+j}{X}_{K+i})0\&le;i,j\&le;M$ （7）

By (6), (7) are released:

$e_M=\underset{j=0}{\overset{M}{\mathrm{\&Sigma;}}}{a}_{\mathrm{Mj}}{r}_M(O,j)$ （8）

By formula (6), (8) can form the matrix of (M+1) * (M+1).

R _MA _M＝E _M（9）

Wherein

Obviously, only require R _M, use Linear algebra method, can obtain the AR model parameter by (9) formula:

A _M＝R -1 _ME _M（11）

But directly push away computing method if adopt, required amount of calculation is big, utilizes R here _MThe Hermite of matrix and anti-Hermite characteristic are derived.

At first, set up two error amount em ', em ":

(12)

$e_M^{\&prime;}=\underset{K=1}{\overset{N-M-1}{\mathrm{\&Sigma;}}}[{\left|f_{M,K}\right|}^2+{\left|b_{M,K+1}\right|}^2]$ （13）

According to the lowest mean square criterion, with aforesaid e _MThe same processing, that is:

R _M’A _M’＝E _M’（14）

R _M”A _M”＝E _M”（15）

$r_M^{\&prime;}(i,j)=\underset{K=1}{\overset{N-M-1}{\mathrm{\&Sigma;}}}[X_{K+M+l-j}{X}_{K+M+l-i}+X_{K+j}{X}_{K+i}]$ （17）

$r_M^{\&prime;}(i,j)=\underset{K=1}{\overset{N-M-1}{\mathrm{\&Sigma;}}}[X_{K+M-j}{X}_{K+M-i}+X_{K+l+j}{X}_{K+l+i}]$ （18）

Derive A at last _M' time area update formula:

A _M’＝a _M（A _M+β ₁C _M+γ ₁D _M）（19）

And AR model coefficient iterative formula:

a _M+1，i＝a’ _M，i+a _M+1，M+1（a’ _M，M+1-i），1≤i≤m（20）

In formula (19):

Formula (20) has with the same form of Levinson iteration, can try to achieve the AR model parameter thus, and white noise spectrum density No.

e _M+1＝e′ _M（1-｜a _M+1，M+1｜ ²）（23）

M rank AR model is obtained its parameter at last:

a ₁＝a _M，1，a ₂＝a _M，2，…a _M＝a _M，M

No＝e _m，

Then power density is:

$S_X\left(f\right)=e_M/{|1+\underset{K=1}{\overset{M}{\mathrm{\&Sigma;}}}{a}_k\exp (-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="941075710\_DRIMG10.png,941075710\_DRIMG11.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;fk\&Delta;t})|}^2$

In model parameter estimation, the problem of another one key is exactly the selection of model order M, and this is an important but still unresolved good problem.Order is low excessively, and the spectrum estimation of making can make in esse spectrum peak thicken; Order is too high, can produce false details again.In program of the present invention, the inventor is through repetition test, and calculating, comparison, combination are found to be chosen in the order that can satisfy the theory of information criterion between the 11-19 rank, have obtained satisfied result.Here the information-theoretical criterion of indication is:

AIC _M＝lne _M+ (α（M+1）)/(N)

According to aforementioned, the mean square error e of prediction _mThe increase with order M always descends, and α (M+1)/N then increases with M and rises, and so just might obtain minimum under a certain M value.The M value of this minimal point just is chosen as best order.

After trying to achieve AR model parameter and white noise spectrum density No, just be easy to utilize following formula to come the power spectral density (PSD) of estimated signal at step S22:

$S_X\left(f\right)=\mathrm{No\&Delta;t}/{|1+\underset{K=1}{\overset{M}{\mathrm{\&Sigma;}}}{a}_{M,K}\exp (-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="941075710\_DRIMG10.png,941075710\_DRIMG11.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;fk\&Delta;t})|}^2$

Wherein f is frequency (Hz), and Δ t is the sampling interval.

Figure 13 to 18 show one not the control experiment of administration each X signal waveform and through the contrast of the power spectrum S signal curve of each section correspondence of obtaining after the above AR model transferring.Wherein Figure 13 is for getting the power spectral density of the stage A that clinostatism freely breathes corresponding to the experimenter, Figure 14 is for getting the power spectral density of the stage B of the controlled breathing of clinostatism corresponding to the experimenter, Figure 15 is for getting the power spectral density of upright stage C that freely breathes corresponding to the experimenter, Figure 16 is for getting the power spectral density of the stage D of the upright controlled breathing in position corresponding to the experimenter, freely breathes and the power spectral density of the stage E of controlled breathing by clinostatism in order to get corresponding to the experimenter for Figure 17, and Figure 18 gets upright power spectral density of freely breathing with the stage F of controlled breathing corresponding to the experimenter.

At step S23, according to above-mentioned power spectral density results, calculate each eigenvalue, provide as Figure 21 with by the report of the conclusion shown in 22, for the foundation of doctor as diagnosis.Symbolic significance used in table and in each bar curve is as follows:

T-R-R interval meansigma methods (ms)

VARI(is designated as V sometimes)-R-R interval mean square deviation (ms ²)

P-general power (equaling mean square deviation after returning-changing) (ms ²)

PL-low frequency power (0.02-0.15Hz) (ms ²)

PH-high frequency power (0.15-0.40Hz) (ms ²)

LF-low frequency mid frequency (Hz)

HF-high frequency mid frequency (Hz)

RPLH-PL/PH low frequency power/high frequency power

RSLH-SL/SH PSD low frequency peak value/PSD high frequency peaks

The PSD-power density spectrum

Referring to Figure 19 and 20, the present invention has also made with each section power spectral density comparison diagram of segmentation as the Z axle.Can know the irritability situation of finding out sympathetic nerve and parasympathetic nerve under various stimulation states intuitively.In two figure, first group of spike of curve shows sympathetic and the total activity intensity of parasympathetic nerve, and second group of spike then is the intensity of the independent excitation time of parasympathetic nerve.This provides to patient's autonomic nervous system functional status intuitively and wherein sympathetic and interactional direct evidence of parasympathetic nerve and report for the doctor.

More than each blood processor and process can finish by computer software programs, in the present invention, provide a kind of application software of in function testing system for vegetative nervous system of the present invention, moving of being specially adapted to.

In order to understand the neuro function of human body, the present invention compares following four kinds of experimental results:

Experiment 1: control experiment.It is human body not to be applied extraneous medicine, i.e. the neuro situation of not disturbance;

Experiment 2: injecting atropine (atropine), to stimulate the irritability of sympathetic nervous system;

Experiment 3: take practolol (propranolol), to stimulate the irritability of parasympathetic nerve system;

Experiment 4: while administration atropine and practolol, carry out disturbance to sympathetic and parasympathetic nerve system.

Table 1 be to 20 examples experimentize and obtain about the statistical result of position motion and medicine to the influence of R-R interval variability.Here the meaning that reaches alphabetic character used in the follow-up table is identical with aforementioned regulation.

Table 1

State T (ms) V (ms ²) PL (%) LF (Hz)

1. not administration:

E. flat (m) 891 3,992 30.0 0.099 that crouch

(sd) 123 2431 10.6 0.019

F. (m) stands ^{* *}727 3400 ^{* *}46.4 ^*0.083

(sd) 96 1506 11.9 0.015

2.Atropine:

E. flat (m) 937 that crouch ^*5662 ^*28.3 0.094

(sd) 145 3496 14.0 0.020

F. (m) stands ^{* *}644 ^{* * *}2926 ^*37.7 0.084

(sd) 114 1841 16.7 0.021

3.Propranolol:

E. flat (m) 955 that crouch ^*5,417 35.1 ^*0.103

(sd) 133 3582 10.7 0.017

F. (m) stands ^{* *}828 ^{* * *}3794 ^*43.5 ^{* *}0.084

(sd) 105 2445 9.7 0.016

4. in above 2 and 3

Medicine give simultaneously:

E. flat (m) 1026 that crouch ^{* *}5817 ^*32.2 0.094

(sd) 124 2791 13.1 0.022

F. (m) stands ^{* *}795 ^{* *}3,515 37.2 ^*0.090

(sd) 126 2326 11.4 0.022

Table 1 (continuing)

State PH (%) HF (Hz) PL/PH SL/SH

1. not administration:

E. flat (m) 27.0 0.280 1.388 1.535 that crouch

(sd) 11.9 0.062 0.987 1.463

F. (m) stands ^{* *}11.1 0.259 ^{* *}4.999 ^{* *}9.862

(sd) 5.3 0.065 2.220 6.407

2.Atropine:

E. flat (m) 20.6 that crouch ^*0.275 1.986 ^*3.664 ^*

(sd) 13.4 0.065 1.587 5.272

F. (m) stands ^{* *}7.3 ^*0.248 ^{* *}5.702 ^*12.336

(sd) 3.1 0.066 2.825 10.036

3.Propranolol:

E. flat (m) 22.7 0.274 1.970 1.738 that crouch

(sd) 9.9 0.066 1.352 1.668

F. (m) stands ^{* *}11.7 0.253 ^{* *}4.190 ^*8.149

(sd) 4.6 0.070 1.593 9.262

4. in above 2 and 3

Medicine give simultaneously:

E. flat (m) 21.6 0.277 1.897 1.592 that crouch

(sd) 11.2 0.074 1.249 1.053

F. (m) stands ^{* *}10.3 0.241 ^{* *}4.365 ^{* *}7.512 ^*

(sd) 5.1 0.063 2.175 5.395

Wherein, the asterisk of indicating before the data represents that freely breathing of E(prostration position under each state adds the controlled section of breathing) add the controlled section of breathing with F(freely the breathing of position of standing) comparison (being that the situation that position changes is added up); Asterisk after the data is illustrated respectively in 2(administration atropine), 3(administration practolol) and 4(give atropine and practolol simultaneously) state and not administration of contrast state 1(state) comparison (being the situation statistics of drug influence).One of them asterisk " * " expression statistical value P＜0.05; Two asterisks " * * " expression P＜0.01, three asterisk " * * * " expression P＜0.001, they show the intensity that significant difference is arranged.

At first see the influence of position to the heart rate spectrum:

From the statistical result of table 1 as can be seen, utilize this method, when human body during from clinostatism to upright position, the variation on its heart rate spectrum can detect significantly.In control experiment 1 (not medication), when clinostatism, two main frequency components are arranged: low frequency (～0.099Hz) 30.0 ± 10.6%, high frequency (～0.280Hz) 27.0 ± 11.9%; PL/PH is 1.4 ± 1.When upright position, and low frequency (～0.083Hz) 46.4 ± 11.9%, high frequency (0.259Hz) 11.1 ± 5.3%; PL/PH is 5.0 ± 2.2.As seen, the normal person is during from clinostatism to upright position, represent sympathetic and the coefficient low frequency component of parasympathetic nerve increases, and represent the independent high fdrequency component that acts on of parasympathetic nerve to reduce.When showing from clinostatism to upright position, sympathetic activity increases, and the parasympathetic nerve activity weakens relatively.From the R-R interval, parasympathetic nerve also reduces, i.e. increased heart rate.See that statistically the P value of several variations illustrates difference clearly all less than 0.001.

In addition, under the situation of its excess-three group medication, from clinostatism to upright position, R-R average (T), PL, PH, PL/PH, SL/SH also have significant difference, just its variation is different with control experiment, and this provides such prompting, and promptly the influence that position is composed heart rate among the normal person is bigger, under general less medicine disturbance, still can show.And had influence among the neural patient in myocardial infarction and diabetes, to upright position, its low frequency component does not then have obvious increase from clinostatism.

Table 2 provided to 20 examples experimentize and obtain about breathe and medicine to the statistical result of the influence of R-R interval variability.

Table 2

A/C: free respiratory frequency B/D: controlled respiratory frequency (0.33Hz)

State T (ms) V (ms ²) PL (%) LF (Hz)

1. not administration:

A. flat (m) 901 3,424 41.8 0.100 that crouch

(sd) 119 1982 12.4 0.021

B. flat (m) 887 2956 that crouch ^*31.3 0.095

(sd) 128 2064 12.8 0.019

C. (m) 728 3,699 55.0 0.088 stands

(sd) 95 2108 12.9 0.026

D. (m) 721 stands ^*2,642 50.8 0.086

(sd) 102 1371 13.3 0.016

2.Atropine:

A. flat (m) 956 that crouch ^{△ △}5076 ^{△ △}42.0 0.096

(sd) 154 3256 12.3 0.026

B. flat (m) 922 that crouch ^*3469 ^*31.7 0.091

(sd) 147 2596 9.7 0.021

C. (m) 667 stands ^{△ △}3,144 51.0 0.079

(sd) 112 2117 14.3 0.017

D. (m) stands ^{* *}625 _{△ △ △} ^{* *}1690 ^{△ △}47.8 0.083

(sd) 115 1265 17.1 0.023

3.Propranolol:

A. flat (m) 961 that crouch ^△5145 ^△43.3 0.098

(sd) 135 3823 9.2 0.017

B. flat (m) 956 that crouch ^△4,061 39.2 ^{△ △}0.102

(sd) 136 3097 12.4 0.021

C. (m) 830 stands ^{△ △ △}3,943 54.0 0.090

(sd) 110 2610 11.6 0.024

D. (m) 828 stands _{△ △ △} ^*2957 ^*45.4 0.081

(sd) 109 1957 11.3 0.017

4. give in 2 and 3 simultaneously

Two kinds of medicines:

A. flat (m) 1036 that crouch ^{△ △ △}6031 ^△42.1 0.094

(sd) 115 3590 10.7 0.022

B. flat (m) 1017 that crouch _{△ △ △} ^*3664 ^*34.4 0.084

(sd) 133 2683 13.5 0.025

C. (m) 801 stands ^{△ △}3,520 46.4 0.090

(sd) 125 2117 12.6 0.017

D. (m) 787 stands ^△2,902 37.3 ^{△ △}0.086

(sd) 128 2567 13.6 0.021

Table 2 (continuing)

State PH (%) HF (Hz) PL/PH SL/SH

1. not administration:

A. flat (m) 26.6 0.237 1.958 2.172 that crouch

(sd) 12.7 0.062 1.129 1.771

B. flat crouch (m) ^*35.1 ^{* *}0.336 ^*1.235 1.376

(sd) 15.7 0.017 1.054 1.976

C. (m) 10.1 0.217 6.352 9.685 stands

(sd) 4.3 0.059 2.984 9.917

D. (m) stands ^*13.6 ^{* *}0.336 ^*4.655 9.586

(sd) 7.1 0.020 2.500 7.730

2.Atropine:

A. flat (m) 21.7 0.249 3.024 that crouch ^△4.104

(sd) 12.5 0.076 2.666 6.473

B. flat crouch (m) ^*31.9 ^{* *}0.333 ^*1.306 ^*1.769

(sd) 16.2 0.017 0.795 2.567

C. (m) 8.7 0.255 stands ^△7.557 16.047

(sd) 4.9 0.051 4.144 16.804

D. (m) 9.3 stands _△ ^{* *}0.332 6.180 ^△18.921

(sd) 3.9 0.047 3.469 19.135

3.Propranolol:

A. flat (m) 25.3 0.249 2.262 2.067 that crouch

(sd) 9.3 0.053 1.987 1.775

B. flat (m) 28.2 that crouch ^{* *}0.337 2.022 1.764

(sd) 14.2 0.012 1.952 1.516

C. (m) 10.1 0.235 6.124 8.691 stands

(sd) 4.9 0.069 2.371 6.679

D. (m) stands ^*16.1 ^{* *}0.330 ^*3.614 7.605

(sd) 7.8 0.019 2.114 7.734

4. in above 2 and 3

Medicine give simultaneously:

A. flat (m) 22.2 0.258 2.628 2.746 that crouch

(sd) 11.6 0.063 1.718 2.812

B. flat crouch (m) ^*29.3 ^{* *}0.338 ^*1.396 1.864

(sd) 11.8 0.017 0.875 2.295

C. (m) 10.3 0.246 5.561 7.337 stands

(sd) 4.8 0.072 3.245 6.320

D. (m) 13.1 stands ^{* *}0.335 4.068 8.408

(sd) 8.8 0.027 3.074 8.786

Wherein, the freely section of breathing/controlled breathing section of A/B, C/D(represented in the asterisk before the data) comparison; Triangle after the data number is represented 2(administration atropine respectively), 3(administration practolol) and 4(administration atropine and practolol simultaneously) state and not administration of control experiment 1() comparison during state.One of them asterisk " * " expression P＜0.05; Two asterisks " * * " expression P＜0.01; Three asterisks " * * * " expression P＜0.001, they show significant specificity; Triangle number " △ " expression P＜0.05; Two triangle number " △ △ " expression P＜0.01; Three triangle number " △ △ △ " expression P＜0.001, they show significant specificity.

Can see the influence of breathing referring to table 2 to the heart rate spectrum:

The influence that breathing is composed heart rate:

Generally speaking, the mid frequency of high fdrequency component is consistent with respiratory frequency.In experiment, the influence of heart rate spectrum is contrasted freely breathing with controlled breathing (20 times/min, i.e. 0.33Hz).In control experiment 1, controlled breathing makes the mid frequency of high fdrequency component move on to 0.33Hz by 0.22～0.24Hz, and is identical with respiratory frequency.At clinostatism and upright position state, low frequency component is reduced, high fdrequency component increases, thereby when making PL/PH become 1.2 ± 1.0(clinostatism by 2.0 ± 1.1 respectively), 6.4 ± 3.0 when becoming the upright position of 4.7 ± 2.5(), and the P value is all less than 0.05, and there were significant differences on the statistics.

Controlled breathing is carried out in this explanation, and respiratory movement is strengthened, and stimulates and pays sympathetic excitement, makes the active increase of parasympathetic nerve, and sympathetic activity weakens relatively.Coming to the same thing of this result and theoretical research.

Let us is looked at the influence of medicine to the heart rate spectrum again:

Carrying out the research of heart rate spectrum under the medicine condition, be for the reflection to sympathetic nerve, parasympathetic nerve system activity of the characteristic quantity of confirming the heart rate spectrum, and this analytical method is to sensitivity sympathetic, the parasympathetic nerve activity change.

From table 1 and table 2 we as can be seen, medicine is for the influence of heart rate spectrum.

Behind intramuscular injection atropine 1.0-1.5mg, under clinostatism and upright position state, all make high fdrequency component that remarkable decline is arranged, when dropping to 20.6 ± 13.4(clinostatism by 27.0 ± 11.9 respectively), when reducing to the upright position of 7.3 ± 3.1() by 11.1 ± 5.3.In the zoopery of this result and foreign study, pay sympathetic (fan walks) nerve, the unanimity as a result that makes high fdrequency component disappear or reduce with heavy dose of atropine blocking-up.This explanation heart rate spectrum medium-high frequency component reflection is really paid sympathetic (fan walks) neural activity, can be used as an index clinically.

Behind oral 40mg practolol, as can be seen from Table 1,, make when clinostatism becomes upright position because practolol is to orthosympathetic blocking effect, the increase that control experiment 1 is compared in the increase of low frequency component is few.In the control experiment 1: 30.0 ± 10.6 become 46.4 ± 11.9, and in the practolol experiment (3): 35.1 ± 10.7 become 43.5 ± 9.7.Thereby obtain such prompting: practolol has certain blocking-up for the change of position to orthosympathetic enhancing stimulation, but because oral practolol, major part is fallen by metabolism, and individual variation is bigger, so upward difference is not remarkable for statistics.

From the situation of following special case, the method that also can prove this heart rate spectrum is a kind of effective research method to autonomous (plant) function of nervous system.As doing in the Electrocardiographic object, wherein No.4 number, from its electrocardiogram, diagnose out sinus bradycardia, but still belong to the normal ECG scope in former examples.In control experiment (1), his each eigenvalue and 20 routine statistical values contrast as following table 3:

Table 3

State T (ms) V (ms ²) PL (%) LF (Hz)

E. flat for sleeping in 1,016 4,376 23.3 0.059

20 routine mean 891 3,992 30.0 0.099

sd 123 2431 10.6 0.019

F. stand 889 5,239 40.5 0.065

20 routine mean 727 3,400 46.4 0.083

sd 96 1506 11.9 0.015

Table 3 (continuing)

State PH (%) HF (Hz) PL/PH SL/SH

E. flat for sleeping in 43.5 0.339 0.535 0.718

20 routine mean 27.0 0.280 1.388 1.535

sd 11.9 0.062 0.987 1.463

F. stand 15.9 0.319 2.555 4.892

20 routine mean 11.1 0.259 4.999 9.862

sd 5.3 0.065 2.220 6.407

From the contrast of he and 20 routine statistical values, his parasympathetic nerve activity obviously is better than other people, and sympathetic nerve a little less than.The electrocardiogram of this and its sinus bradycardia matches, thereby proves that this method is effectively and sensitive.

From at present heart rate being composed existing research, some disease clinically is as myocardial infarction, chronic heart failure, diabetes, hypertension etc.; Change to the heart rate spectrum is much bigger more than sinus bradycardia, therefore, the research by to the heart rate spectrum is the classification of diagnosing clinically, the prediction of disease and the observation of curative effect, particularly assess cardiac patient's autonomic nervous system provides a kind of effective means.

Though above in conjunction with embodiments of the invention and the experiment invention has been described, those skilled in the art should understand, within the spirit and scope of the present invention, can make various modifications and change.

Claims (23)

1, a kind of function testing system for vegetative nervous system is characterized in that comprising:

Portable cardiac signals collecting and storing apparatus are used for gathering continuously and write down the experimenter and are in the electrocardiosignal that some different conditions produce so that sympathetic nervous system and/or parasympathetic nerve system are stimulated in a scheduled time;

The voice unit that contains a speaker is used for instructing the experimenter to be in more described different conditions to obtain needed stimulation to sympathetic nervous system and/or parasympathetic nerve system mutually in phase with the described scheduled time;

Pico computer, electrocardiosignal corresponding segment ground when being used for each different conditions to the storage of described ecg signal acquiring and storing apparatus carry out power spectrumanalysis so as to obtain sympathetic nerve and the parasympathetic nerve system to the response of institute's irriate, and then obtain the conclusion of their irritability situations; It further comprises:

R ripple recognition device is used for the R ripple of ECG signal is discerned, and finds out the R-R interval, with output R wave train signal;

Segmentation and pretreatment unit carry out segmentation and pretreatment corresponding to more described different conditions to R wave train signal, send sectional R wave train signal under each different stimulated state as the X signal;

Autoregression (AR) model spectra estimation device is used to find out the model parameter that comprises model coefficient, white noise spectrum density and model order M, the described X signal of time domain is carried out transform, so that it is carried out the power Spectral Estimation of frequency domain;

The power spectral density accountant is used for calculating according to described model parameter the power spectrum of each section electrocardiosignal; And

Output device is used for relevant figure and the power spectrumanalysis and the relatively report of printout electrocardiosignal.

2, according to the system of claim 1, it is characterized in that described ecg signal acquiring and storing apparatus further comprise an infrared emission apparatus, described pico computer further is connected with an infrared remote receiver device, in order to finish the electrocardiosignal transmission between described ecg signal acquiring and storing apparatus and the pico computer.

3, according to the system of claim 1, more wherein said different stimulated states are respectively following four sections different conditions: first section A, in period, the experimenter is in prostration position and free breathing state at this section, in order to stimulate parasympathetic nerve system and sympathetic nervous system; Second section B, in the period, the experimenter is in prostration position and controlled breathing state at this section, in order to stimulate the parasympathetic nerve system; The 3rd section C, in the period, the experimenter becomes stand position and free breathing state at this section, in order to stimulate sympathetic nervous system; The 4th section D, in the period, the experimenter is in stand position and controlled breathing state at this section, in order to stimulate sympathetic and the parasympathetic nerve system; Wherein the frequency of controlled breathing is higher than the frequency of freely breathing.

4, according to the system of claim 3, it is characterized in that the described scheduled time is 15 minutes, wherein said A section is 3 minutes, and the B section is 4 minutes, and the C section is 4 minutes, and the D section is 4 minutes.

5, according to the system of claim 4, it is characterized in that described R ripple recognition device and segmentation pretreatment unit further comprise the low pass differential attachment and the single order low-pass filter that is used to obtain through the ECG-of low-pass filtering order derivative that is used to obtain the electrocardiosignal first derivative, its function is respectively

D(t)=ECG(t)-ECG(t-4) and

f（t）＝f（t-1）+d（t）-d（t-6）;

Described R ripple recognition device also comprises electrocardiogram self study device, be used for electrocardiosignal is carried out after the self study of a period of time, to f(t) find out maximum value PK and two R-R intervals, these two R-R meansigma methods at interval is as average R-R initial value at interval; Wherein the R ripple recognition threshold of the device employing of R ripple identification is:

H ₀（O）＝0.7＊PK

Its threshold adaptive recursive function is:

6, according to the device of claim 5, it is characterized in that described system also comprises the result who is used for according to R ripple recognition device, make the device of the R wave train signature tune line chart of R-R interval, and by the described signature tune line chart of described output device printout.

7, according to the device of claim 5, it is characterized in that described system also comprises the result who is used for according to R ripple recognition device and segmentation pretreatment unit, make the device of the R wave train signature tune line chart of each section (A, B, C, D and E=A+B, F=C+D section and R=E+F section), and by described output device printout.

8, according to the system of claim 5, it is characterized in that described R ripple recognition device also further comprises checking device one time, when R-R＞1.8RRAV, this section is carried out the identification of R ripple once more, its time inspection threshold value is:

H ₁（n）＝0.5＊H ₀（n）

9,, it is characterized in that described autoregression (AR) model spectra estimation device and the following transform of power spectrum accountant utilization carry out analysis of spectrum according to the system of claim 1 and 5:

X _n＝W _n＊h _n

Wherein: h _nBe impulse response; * be convolution, W _nBe white noise;

If Z=e ^Jw

S then _x(e ^Jw)=SW(e ^Jw) | H(e ^Jw) | ²

＝No｜H（e ^jw）｜ ²

Wherein the expression formula of AR model is:

X（n）＝

a（k）X（n-k）+W（n）

Fading to frequency domain has:

(X(Z))/(W(Z)) =H(Z)=1/ $(1+\underset{K=1}{\overset{M}{\mathrm{\&Sigma;}}}a\left(K\right)Z\&CenterDot;k)$

The model coefficient a(k that is wherein found) be:

a ₁＝a _M，1，a ₂＝a _M，2，…a _M＝a _M，M;

The white noise spectrum density that is found is:

N ₀＝e _M;

Model order M is then according to the theory of information criterion

AIC _M＝lne _M+ (α（M+1）)/(N)

M value when getting minimal point is as best order value, and described order M selects between 11-19; And

The power spectral density of described power spectral density accountant is calculated and is carried out according to following formula: $S_X\left(f\right)=\mathrm{No\&Delta;t}/{|1+\underset{K=1}{\overset{M}{\mathrm{\&Sigma;}}}{a}_{M,k}\exp (-j2\mathrm{\&pi;fk}\stackrel{\&perp;}{\mathrm{\&Delta;}})|}^2$

Wherein f is frequency (Hz), and △ t is the sampling interval.

10,, it is characterized in that described system also comprises the conversion estimation result according to autoregression (AR) model, makes the curve chart that corresponding each section and whole R wave train signal and its power spectrumanalysis signal contrast respectively according to the system of claim 7 and 9.

11,, it is characterized in that each included described device of described pico computer finished by software according to the system of claim 10.

12, a kind of method that detects neuro function comprises step:

A. in a scheduled time, when experimenter's sympathetic nervous system and/or parasympathetic nerve system are stimulated, gather and storage experimenter's core signal ECG;

B. ECG is carried out the identification of R ripple, find out the R-R interval, provide R wave train signal;

C. according to the difference that among the step a sympathetic nervous system and/or parasympathetic nerve system is stimulated, segmentation intercepts described R wave train signal as the X signal;

D. utilize autoregression (AR) model, described X signal is made transform, so that it is done frequency-domain analysis;

F. calculate the power density spectrum of electrocardiosignal;

G. obtain eigenvalue, thereby provide the unify quantitative analysis result of parasympathetic nerve system of sympathetic nervous system.

13,, it is characterized in that among the described step a experimenter's sympathetic nervous system mode that the parasympathetic nerve system stimulates of unifying is divided into following four sections according to the method for claim 12:

ⅰ. the experimenter is in prostration position and free breathing state A;

ⅱ. the experimenter is in the state B of prostration position and controlled breathing;

ⅲ. the experimenter becomes the state C that the position and carrying out of standing is freely breathed; And

ⅳ. the experimenter is in the state D that stands position and carry out controlled breathing;

Wherein the frequency of controlled breathing is higher than the frequency of freely breathing.

14,, it is characterized in that the described scheduled time is 15 minutes, wherein state A(ⅰ section according to the method for claim 13) continue 3 minutes; State B(ⅱ) section continues 4 minutes; State C(ⅲ section) continues 4 minutes; And state D(ⅳ section) continues 4 minutes.

15,, it is characterized in that described stimulation to sympathetic nervous system and/or parasympathetic nerve system comprises to the experimenter to impose specific medicine according to the method for claim 14.

16,, it is characterized in that described specific medicine is atropine (Atropine) and/or practolol (Propranolol) according to the method for claim 15.

17,, it is characterized in that described step b adopts the self adaptation criterion to carry out, and further comprises step according to the method for claim 14:

B1. utilize formula

d（t）＝ECG（t）-ECG（t-4）

Obtain the electrocardiosignal first derivative;

B2. utilize formula

f（t）＝f（t-1）+d（t）-d（t-6）

First derivative to electrocardiosignal is carried out low-pass filtering;

B3. in a period of time, carry out the electrocardio self study, then to f(t) find out maximum value PK and two R-R intervals, the latter's meansigma methods is as the initial value of average R-R interval RRAV, and wherein R ripple recognition threshold is:

H ₀（O）＝0.7＊PK

B4. order computation f(t), find out PK(n)＞H ₀(n), and determine whether to involve for R the position of R ripple, wherein the threshold adaptive recursive function is:

18, according to the method for claim 17, it is characterized in that described step b also comprise discern the R ripple once more return the inspection step, its decision condition is, when R-R＞1.8RRAV, it returned inspection, time examines threshold value to be:

H ₁（n）＝0.5＊H ₀（n）。

19,, it is characterized in that the transformation for mula that described steps d and f make analysis of spectrum is according to the method for claim 12 and 17:

X _n＝W _n＊h _n

Wherein: h _nBe impulse response, * is a convolution, W _nBe white noise;

If Z=e ^Jw

S then _x(e ^Jw)=S _w(e ^Jw) | H(e ^Jw) | ²

＝No｜H（e ^jw）｜ ²

Wherein the function of AR model is

Xn＝- $\underset{K=}{\overset{M^{\&prime;}}{\mathrm{\&Sigma;}}}$ a _kX _n-k+W _n

After making transform:

(X（Z）)/(W（Z）) ＝H（Z）＝ $\frac{1}{1+\underset{K=1}{\overset{M}{\mathrm{\&Sigma;}}}{a}_k{Z}^{\&CenterDot;K}}$

Model coefficient a wherein _kFor:

a ₁＝a _M，1，a ₂＝a _M，2，…a _M＝a _M，M;

White noise spectrum density N ₀=e _M;

Wherein said steps d also comprises the step that model order M is estimated, promptly utilizes formula

AIC _M＝lne _M+ (α（M+1）)/(N)

Get its M value when minimum and be best order; Described M value is selected between 11-19.

20,, it is characterized in that the formula that calculates the electrocardiosignal power spectral density among the described step f is according to the method for claim 18

$S_X\left(f\right)=\mathrm{No\&Delta;t}/{|1+\underset{K=1}{\overset{M}{\mathrm{\&Sigma;}}}{a}_{M,K}\exp (-2\mathrm{\&pi;fk\&Delta;t})|}^2$

Wherein f is frequency (Hz), and Δ t is the sampling interval.

21,, it is characterized in that further comprising that the test result to the described different stimulated mode of many cases compares and add up and provide the step of comparison and statistical result according to the method for claim 12-16.

22,, it is characterized in that wherein said each step finished by software according to the method for claim 12 and 20.

23,, it is characterized in that wherein said each step finished by software according to the method for claim 21.

Images