METHOD FOR RECORDING ARTERIAL PARAMETERS OF A HUMAN BEING AND DEVICE FOR CARRYING OUT THE METHOD

DE502016016788D1Active Publication Date: 2024-11-28WSH ENG SERVICES GBR VERTRETUNGSBERECHTIGTER GESELLSCHADTER HIERONYMI ANDREAS
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
DE502016016788
Authority / Receiving Office
DE · DE
Patent Type
Patents
Current Assignee / Owner
Priority Date
2015-12-08
Filing Date
2016-12-07
Publication Date
2024-11-28
Estimated Expiration
2036-12-07

AI Technical Summary

Technical Problem

Current methods for determining arterial compliance, conduction resistance, and blood flowability are hindered by measurement uncertainties, limited reproducibility, and the need for invasive procedures, with pulse contour-based methods often failing to clearly separate these parameters due to superimposed waves and tissue interference.

Method used

A method using a trained neural network to analyze pulse curves, combined with an electro-hydraulic artery model, allows for non-invasive determination of arterial compliance, line resistance, and blood inertia by simulating the arterial tree and applying correction factors to improve measurement accuracy and reproducibility.

Benefits of technology

This approach enables reliable and reproducible measurement of vascular parameters, reducing measurement uncertainties and fluctuations, and allows for precise determination of arterial stiffness and impedance, enhancing diagnostic accuracy for cardiovascular diseases.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader
Need to check novelty before this filing date? Find Prior Art

Description

[0001] The investigation of hemodynamic properties of the human vascular system is usually performed using pulse contour analysis, identifying and evaluating characteristic features of the pulse waveform, such as points of specific slope or curvature. Often, surrogate parameters, such as pulse velocity as a measure of general arterial stiffness or ejection fraction as a measure of cardiac performance, are determined from the position of these characteristic points, such as the onset of the excitation wave, its maximum, and the incisura, and used to diagnose cardiovascular diseases. Direct measurement of such surrogate parameters is only possible invasively, while non-invasive measurement is always hindered by the immediate vascular environment (muscle and adipose tissue or bone mass) and sometimes requires correction.

[0002] US patent 2006 0235323 A1 discloses a method for determining a cardiovascular parameter that corresponds to or can be derived from cardiac output. For this purpose, a current pressure waveform dataset corresponding to the arterial blood pressure over a current pressure cycle is input, and defined parameters of an assumed input flow waveform are determined.

[0003] US Patent 2010 0274102 A1 discloses a method for processing pulse oximeter data using a dynamic state-space model to extract an estimated value for at least one physiological parameter not measured by the pulse oximeter. The reference Westerhof et al.: "Analog studies of the human systemic arterial tree", Journal of Biomechanics, Pergamon Press, New York, NY, US, Vol. 2, No. 2, May 1, 1969 (1969-05-01), page 121, XP026262071, ISSN: 0021-9290, DOI: 10.1016 / 0021-9290(69)90024-4 discloses an arterial model with cascaded RLC resonant circuits as Windkessel elements. In this document, the arterial elements originally based on centimeter segments are combined into areas of equal arterial segments, and substitute values ​​for R, L, and C are established for the direct series connection of resistors and inductors, respectively.This applies to the parallel connection of capacitors, but not to the series connection of RLC resonant circuits.

[0004] However, when using these pulse contour-based methods, a clear distinction is usually not made between arterial compliance (the resistance to blood flow) and blood flow properties. Blood flow properties are understood as a combination of inertia, viscosity, and similar characteristics of the blood. To improve the diagnosis and treatment of cardiovascular diseases and pathological conditions such as hypertension, arteriosclerosis, stenoses, aneurysms, or heart failure, it is desirable to make a reliable distinction between vascular compliance, vascular resistance, and blood flow properties, as specific disease patterns can be associated with these characteristics.For example, generally reduced arterial compliance is associated with age-related hypertension, while increased conduction resistance is an indication of atherosclerotic and / or arteriosclerotic changes, and reduced blood flow can indicate, for example, increased cardiac workload or an increased risk of stroke. Blood flow is currently usually measured by determining the hematocrit value of blood samples in centrifuges.

[0005] Solutions for oscillometric measurement of arterial compliance are described, for example, in US 5,447,163 and US 6,733,461. However, these methods usually require cumbersome calibrations and recalibrations. The resulting measurement uncertainties and limited reproducibility of the results can hinder the widespread use of such technologies. Other methods for measuring vascular impedances, such as those described in US 5,211,177, rely on calculations based on the so-called Windkessel model. However, these methods require the measurement of additional physiological parameters, such as cardiac output and mean arterial pressure (MAP), and also operate using pulse contour analysis, i.e., by determining and evaluating characteristic features of the pulse waveform.Anyone familiar with the visual estimation or automatic detection of such characteristic features will agree that there are often cases where the exact position and amplitude cannot be clearly determined, for example, due to the superposition of the excitation wave and the reflected wave. Therefore, such measurement results are subject to sometimes significant fluctuations in the case of repeated measurements.

[0006] The present invention describes a method and a device for carrying out the method, which allow a reliable and reproducible determination of arterial compliance, arterial conduction resistance and the inertia of the blood, avoiding the disadvantages listed above.

[0007] The first-mentioned problem is solved according to the invention by a method according to claim 1. Further embodiments are found in dependent claims 2-4.

[0008] The parameters to be fed into the human simulation model could, for example, be determined iteratively from the pulse curve. However, according to an advantageous embodiment of the invention, the process can be significantly accelerated if the pulse curve measured in the human is fed into a suitable, trained neural network, which then outputs the desired parameters using its stored knowledge of the relationships between the pulse curve shape and the corresponding parameters. The calculated parameters can then be used in the human simulation model to examine pulse curves at various points in the simulated arterial tree.

[0009] Particularly precise parameters for arterial stiffness or vascular impedance can be easily obtained according to an advantageous embodiment of the invention by measuring the pulse waveform. This is achieved by increasing the cuff pressure until blood flow ceases due to occlusion of the arteries, maintaining this cuff pressure for a predetermined number of cardiac cycles, measuring the pressure profile of several pulse waveforms within the cuff, and generating an averaged pulse waveform from these multiple pressure profiles. This allows for the determination of pulse wave velocity in the aorta, aortic blood pressure (also called central pressure), the augmentation index, or left ventricular ejection fraction by briefly and completely occluding, for example, the wrist arteries.This complete occlusion allows the determined pulse curves to be sufficiently freed from disturbing influences of local blood flow turbulence and distal reflections, both of which can lead to distorted measurement results.

[0010] The second problem mentioned, namely the creation of a device for carrying out the method, is solved according to the invention by a device according to claim 5. Further embodiments are found in dependent claims 6-8.

[0011] A particularly fast determination of vascular and circulatory parameters can be achieved according to another advantageous embodiment of the invention if the calculation unit is a neural network instead of the optimization algorithm.

[0012] This design allows the neural network to recognize the pattern of the input pulse curve and directly output the parameters. In practice, such pattern recognition is significantly faster than an optimization algorithm for determining the parameters.

[0013] According to another advantageous embodiment of the invention, the human simulation model allows for the inference of arterial parameters of the human being from a pulse curve measured at the arm or leg, provided that the arteries in the human body simulation model are divided into individual arterial segments and these arterial segments are approximated as an equivalent electrical circuit using resistances, capacitances, and inductances. Thus, the electro-hydraulic simulation model, which replicates an arterial model, forms an analogue with which, via a non-invasive measurement at the arm or leg, calculations can be performed back to the central pressure, aortic stiffness, and aortic blood flow.

[0014] According to an advantageous embodiment of the invention, the in-vivo recording of the course curve of a hemodynamic process is particularly simple if a device for measuring the pulse curve is designed to include a pressure cuff intended for application to the human body and a pressure source for generating a pressure that enables blood flow to stop by occlusion of the arteries, in particular the wrist arteries, and for maintaining this pressure for a predetermined number of cardiac cycles. Beschreibung der Zeichnungen

[0015] Fig. 1 shows the electrical equivalent circuit diagram of a wind chamber element. Fig. 2 shows an electro-hydraulic artery model (analog), source: [1] Fig. 3 shows a possible amplitude profile of the electrical source of an analogue Fig. 4 shows the presentation of amplitudes at the inputs of a neural network according to the invention during its training. Fig. 5 shows the components of a device for measuring arterial parameters Fig. 5a shows an overview illustration of the method according to the invention. Fig. 6 shows the representation of arterial parameters in a radar diagram. Fig. 7 shows an example of a simulation of an in-vivo determined trend curve. Beschreibung der Erfindung

[0016] It is already known that flow and pressure conditions in the arterial vascular system can be artificially replicated using so-called electro-hydraulic analogues. These involve the use of several so-called windkessel elements ( Fig. 1 ) in series. By connecting several of these wind chamber elements, consisting of capacitors C, ohmic resistances R and inductors L, in series ( Fig.1 ) and the application of an electrical current or voltage source to the input of the series circuit with corresponding waveforms Fig.3 , which replicates the pumping function of the heart, flow or pressure values ​​can be measured at any point in the body model thus formed, hereinafter referred to as analogue (2), ( Fig. 2 ) are calculated. The capacitances C, resistances R and inductances L of the individual wind chamber elements ( Fig. 1 The parameters represent arterial compliance (C), arterial conduction resistance (R), and blood flow rate (L) in the considered vessel segment. The calculated electrical currents represent arterial blood flow, and the electrical voltages represent arterial pressure.

[0017] The purpose of the described vascular model is to simulate hemodynamic processes. However, the present invention is based on the reverse of this principle: The medically relevant arterial parameters compliance, conduction resistance, and blood flow are to be determined from existing hemodynamic profiles. To solve the problem described in the invention, an electro-hydraulic arterial model is therefore first created as a simulation model. This simulation model is also called an analogue. Such an electro-hydraulic analogue (2) can be created, for example, by means of computer-aided analysis of electrical networks, as described under the term SPICE ( Simulation Program with Integrated Circuit Emphasis It is common knowledge that these models can be created without significant equipment. From this, currents and voltages can be calculated at any point in the model.

[0018] It is also known that models based on this technique can be flawed because, due to their complexity, they inevitably contain simplifying assumptions. For example, the model artificially divides arteries into individual segments, whereas in the bodies of living organisms, the diameter and elasticity of the arteries change continuously. Furthermore, nonlinear relationships based on fluctuating pressure are not taken into account. The simulation results obtained with this technique therefore repeatedly show deviations from hemodynamic profiles determined in vivo. For example, the inventors noticed that the C, R, and L values ​​for 121 Windkessel elements ( Fig. 1 ) with arterial segment lengths of 1 cm, but in the subsequent simulation, due to the high level of equipment required at that time to replicate analogue 2, several (up to 8) of these windkessel elements ( Fig. 1 ) were combined by forming parallel equivalent capacitances as well as series equivalent resistances and inductances. This was due to the neglect of the dynamic vibration behavior of the individual air chamber elements ( Fig. 1 The resulting errors lead to significant deviations of the simulated results from real hemodynamic conditions. Therefore, in the embodiment of analogue 2 according to the invention, instead of the 121 segments, 711 segments with R, L, and C values ​​recalculated to a segment length of 1 cm are now considered. Taking into account that the influence of the correction network, consisting of R0 and R2 to Rm as well as L2 to Lm, on the simulation result is small, the following is obtained in Fig. 1 The Windkessel element is shown. Empirically determined correction factors were also introduced to replicate the hemodynamic pulse curves as accurately as possible. These correction factors ranged from 0.3 to 1.0 and were calculated for the arm. Various Windkessel models are known from the scientific literature, differing in the number and arrangement of the capacitors C, resistors R, and inductors L used, as well as in their advantages and disadvantages. Several, from the one in Fig. 1 shown as an exemplary embodiment ( Fig. 1 ) and the one in Fig. 2 Therefore, variations of the exemplary analogue (2) shown are also conceivable for solving the problem within the meaning of this invention. This concerns, for example, the number, value, and position of the components of each individual wind boiler element ( Fig. 1 Examples include 2-, 3- or 4-element models, as well as symmetrical and asymmetrical wind chamber models, and the specific arrangement and combination of the individual wind chamber elements ( Fig. 1 ) within the analogue (2).

[0019] Using a suitable analogue (2), the simulation result for a specific vessel segment should theoretically correspond to the course of an in vivo curve determined for the same vessel segment. In practice, however, deviations between the two curves often exist. These deviations are based on previously described model deficiencies, on individual, general in vivo deviations of arterial compliance, conduction resistance, or blood flow from the model's average values, and on deviations in the values ​​of individual segments, e.g., due to stenoses and aneurysms. Deviations attributable to such pathological changes can be avoided by measuring in areas where such changes rarely occur, such as the arm arteries.

[0020] The previously described individual, general deviations in arterial compliance, conduction resistance, or blood flow arise because the component values ​​for C, R1, and L1 were calculated for a healthy average subject with a height of 175 cm, a weight of 75 kg, etc. Therefore, to determine the individually varying arterial parameters compliance, conduction resistance, and blood flow to be determined according to the invention, an additional correction factor α(C), α(R), and α(L) is introduced as a representative for each of these parameters. To simulate the hemodynamic profiles with the analogue (2), each of the component values ​​C, R1, and L1 is multiplied by its correction factor. In contrast to the segment-specific component values ​​C, R1, and L1 listed in [1], the correction factors α(C), α(R), and α(L) assume the same value for all segments.By manually adjusting the three correction factors α(C), α(R), α(L), the simulation result can now be adjusted so that the previously described individual deviations disappear and the simulation result matches the in-vivo determined trend curve (. Fig. 7 The evaluation of measurement results from a large number of subjects revealed that a certain range exists for all correction factors. A so-called radar diagram is proposed below for the graphical representation of these individual correction factors. For this purpose, the respective range was scaled to 100%, thus ensuring the comparability of the correction data for all subjects. The influence of deviations based on model deficiencies is also reduced by applying these correction factors.

[0021] To obtain matching curve profiles, it is also necessary to replicate the behavior of the electrical source (3) so that it corresponds to the individual behavior of the heart at the time of in-vivo recording of the hemodynamic profiles. In the present embodiment, an electrical source (3) is used for this purpose, the time course of which of the output variable corresponds to the course in Fig. 3 This corresponds to HMaxTime, where the maximum of the output variable (pressure or flow curve) is reached. The time of aortic valve closure, HMinTime, is characterized in the hemodynamic course by a local minimum with a flow of amplitude HMinValue in the reverse direction and is often also referred to as the incisura, left ventricular ejection time, or systole duration, and is approximately 250 ms in healthy individuals. The maximum of the hemodynamic curve, HMaxTime, is normally reached in humans after approximately 100 ms of the period of a complete cardiac cycle. The period of the complete cardiac cycle, Cycle length, can be read from the available in vivo curve and can be manually transferred to the parameterization of the source (3) of the analog (2).To simulate real-world conditions as accurately as possible, this example introduces a second, positive maximum at time HRetTime, whose amplitude corresponds to 10% of the main maximum. For further refinement, the model can be extended to include an HPreShut value, which can be both positive and negative ( Fig. 3 ) and by which the decay of the excitation wave can be varied. The individual, straight segments of the in Fig. 3 The depicted amplitude profiles can also consist of curved curve segments for a more realistic representation of the heart. In analogue (2), this can be achieved approximately, for example, by using a so-called piecewise linear source (3) with a corresponding number of support points, or continuously by an electrical low-pass filter connected downstream of the source (3) in analogue (2).

[0022] Since the previously described manual determination of the correction factors α(C), α(R), α(L) and the source parameters HMaxTime, HMinTime, HMinValue, HRetTime, and HPreShut for individual in-vivo determination is too time-consuming, a method according to the invention will now be described with which this can be automated. This method can consist of creating software that allows the desired parameters to be determined directly. In this invention, this method consists of a genetic algorithm, but can also be implemented by any other optimization algorithm. Since the optimization algorithms are very time-consuming, the desired parameters can also be determined with the help of a suitably trained neural network. The training of the algorithm described in the method initially involves the in-vivo acquisition of a sufficient number of hemodynamic curves (3).These curves can represent a so-called pressure, volume, or flow pulse, as the described electro-hydraulic analogy refers to these quantities. Similarly, existing curves, e.g., from databases, can also be used for training, provided these curves are based on quantities to which the electro-hydraulic analogy is applicable.

[0023] The invention employs an evolutionary optimization algorithm, although other optimization methods can also be used. The evolutionary optimization algorithm used here defines the model parameters to be determined, P1...Pn, as "genes." The entirety of the genes can be referred to as a "chromosome." The algorithm first generates an initial population of size 30, consisting of individual "individuals," each with randomly assigned chromosomes. For each of these 30 individuals, the resulting simulated pulse contour is determined using the simulation model. A "fitness function" then determines, for each individual, the sum of squared errors between the measured pulse contour and the pulse contour of the individual. This allows the population to be sorted according to fitness, i.e., the accuracy of the simulations. The following sequence is then executed until a termination criterion is met: Selection: Choosing the best 40% of individuals for recombination. Recombination: Combining parts of the chromosomes of the selected individuals to produce offspring. Mutation: Random change in the offspring. Evaluation: Determining the fitness of the offspring. Classifying the new offspring (the new "generation") into the population according to the quality of their simulations.

[0024] A predefined fitness, i.e., a fixed sum of squared errors between the simulation contour of the best individual and the measured pulse contour, is used as a termination criterion. Additionally, if the measured pulse contour cannot achieve a high degree of agreement with the best individual due to strong artifacts, a fixed number of generations produced since the last population improvement without the emergence of a new, improved individual is used as a termination criterion.

[0025] A significant advantage of the present invention is that the analogue (2) can be adapted to the specifics of various measurement methods. For example, for suprasystolic brachial measurements, in which the hemodynamic course is recorded under full occlusion of the brachial artery, a terminating resistance ∞ (interruption) can be provided behind the segment of the brachial artery (see Figure 2). Fig. 2 ) are placed, with the simultaneous elimination of all other distal segments. Likewise, it is conceivable to use the present invention for oscillometric or tonometric measurements in which a hemodynamic course is taken under partial occlusion of an artery, if, for this purpose, the value of the conduction resistance R at the windkessel element is determined in analogue (2) ( Fig. 1 ) of the relevant vessel segment is increased accordingly.

[0026] The hemodynamic profiles are subsequently processed by signal processing so that they are available in the form of temporally digitized amplitude values ​​and are then segmented according to cardiac cycles, so that individual pulsations (10, Fig. 4 ) can be represented. By averaging, an average pulsation (10) can also be calculated from several individual pulsations (10) to increase the robustness of the method, e.g. against arrhythmic variances or repeated measurements.

[0027] To train the neural network (15) described above as an example, the first five hundred digitized amplitude values ​​of a pulsation (10) thus obtained are presented to a corresponding number of inputs of a backpropagation network, according to the present embodiment. This enables supervised learning with random initialization of the neuron weights, an input unit with five hundred inputs, a hidden unit, and an output unit with eight outputs. The result is the parameters α(C), α(R), α(L), HMaxTime, HMinTime, HMinValue, HRetTime, and HPreShut. In the present embodiment, a practical number of five hundred amplitude values ​​was chosen because, on the one hand, the number of input data points should be kept as low as possible, while on the other hand, all essential features of a cardiac cycle, taking into account the minimum and maximum heart rates, must be fully represented.Five hundred amplitude values ​​correspond to a period of 500 ms at a signal sampling rate of 1 ms. However, other sample counts are also possible, and it is also conceivable, in accordance with the present invention, to simultaneously present amplitude values ​​of several successive individual pulsations (10) of a hemodynamic course to the inventive method at its inputs.

[0028] A second aspect of this invention aims at creating a device with which arterial compliance, arterial conduction resistance and blood flow can be measured automatically, reliably and reproducibly.

[0029] Fig. 5 and Fig. 5a Figure 2 illustrates the components of the present embodiment of a device according to the invention. Hemodynamic profiles are recorded from an individual by means of a sensor (20) and fed to a signal processing chain (21). These hemodynamic profiles can be, for example, a so-called pressure, volume, or flow pulse, since the electro-hydraulic analogy described above refers to these quantities.

[0030] The hemodynamic profiles thus obtained are subsequently processed by signal processing (21) so that they are available in the form of temporally digitized amplitude values ​​and then segmented according to cardiac cycles, thereby mapping individual pulsations (22). By averaging, an average pulsation (22) can also be calculated from several individual pulsations (22) to increase the robustness of the method, e.g., against arrhythmic variances or repeated measurements. Furthermore, higher-frequency artifacts in the measurements can be eliminated using special analysis methods (e.g., Fourier analysis). In the course of further processing, the amplitude values ​​of the pulsation (22) thus obtained are applied to the corresponding number of inputs of a device (15) ( Fig. 5 ) supplied, at whose outputs the desired values ​​of the parameters are then obtained. α(C) as a measure of compliance; α(R) as a measure of arterial conduction resistance; α(L) as a measure of blood inertia; HMaxTime as the time of maximum pressure or flow cycle; HMinTime as the time of aortic valve closure; HMinValue; HRetTime; HPreShut be available.

[0031] A significant advantage of the invention is that the desired parameters can be determined very precisely in this way. The in Fig. 5 or Fig. 5a The components shown can be housed within a mechanically coupled apparatus or distributed across separate sensor, computing, and display systems. The parameter values ​​at the output of the device (15) can be stored in the short or long term and thus displayed at any time using a display device (30). Fig. 6 shows the advantageous representation of the determined arterial parameters compliance α(C), conduction resistance α(R) and mass inertia of the blood α(L) in a so-called network or radar diagram, where the so-called arterial stiffness is shown as the reciprocal of compliance 1 / α(C).

[0032] The hemodynamic pressure, volume, or flow-pulse profiles determined in vivo can be recorded invasively or non-invasively using sensors (20) without deviating from the fundamental idea of ​​this invention. A previously described, significant advantage of the invention is that the design, based on the specific configuration of analogue 2, is adaptable to various measurement methods. Thus, a device according to Fig. 5 and Fig. 5a for example, for invasive, but also for oscillometric, suprasystolic brachial measurements, photoplethysmography or other oscillometric or tonometric methods that are based on a partial occlusion of different vessel segments, provided the Windkessel elements ( Fig. 1 ) of analogue 2 were adapted according to the physiological effects resulting from the in vivo application of the sensor (20).

[0033] In Fig. 5a Both evaluation methods using a neural network (15) or an optimization algorithm (15a) are presented without further discussion of the sensor technology (20) and signal processing (21). The application of the optimization algorithm offers the advantage that the results are very precise and achieved without prior training of a neural network. This means that, in this case, no training data is required for using the method according to the invention.

[0034] Figur 7The output shows a diagram in which a measured pulse pressure curve (1) and the corresponding simulated pulse pressure curve (2) are superimposed.

Claims

1. Software-based method for detecting arterial parameters of a human being, wherein a pulse curve is measured on the human being, wherein an electro-hydraulic arterial model, an analogue (1), is formed as a simulation model, wherein the pulse curve is compared with simulated pulse curves of a simulation model of the analogue (1) and various parameters of the simulation model are varied until the pulse curve of the simulation model matches the measured pulse curve to an intended approximation, and wherein the arterial parameters are output, characterised in that the simulation model is formed by connecting multiple Windkessel elements consisting of capacitors C, ohmic resistors R and inductors L in series, wherein 711 of these Windkessel elements are used with R, L and C values back-calculated to a 1 cm Windkessel element segment length.

2. Method according to claim 1, characterised in that the measured pulse curve is fed to a calculation unit which detects the curve shape, and in that the calculation unit sets the parameters of a simulation model.

3. Method according to claim 1 or 2, characterised in that the pulse curve is measured by increasing a cuff pressure until blood flow stops by occluding the arteries, in that this cuff pressure is maintained for a predetermined number of cardiac cycles, and in that the pressure curve of multiple pulse curves is measured in the cuff, and in that an averaged pulse curve is generated from the pressure curve of multiple pulse curves.

4. Method according to any one of claims 1-3, characterised in that the compliance, the arterial conduction resistance and the mass inertia of the blood can be calculated for each arterial section proximal to the measuring point with the aid of the simulation model for the human body.

5. Device for carrying out the method according to any one of the preceding claims, comprising a calculation unit having: an input interface for inputting the data from a pulse curve, an optimisation apparatus for determining the parameters of a simulation model that enable an optimal match between the pulse curve and a simulated pulse curve, and has an output interface for outputting the parameters of the next stored pulse curve.

6. Device according to claim 5, characterised in that the calculation unit is a neural network.

7. Device according to claim 5 or 6, characterised in that in the simulation model of the human body the arteries are divided into individual arterial sections and the arterial sections are approximated as an electrical equivalent circuit diagram by means of resistors, capacitors and inductors.

8. Device according to any one of claims 5 to 7, characterised in that a device for measuring the pulse curve is formed of a pressure cuff intended to be placed on the human body and a pressure source for generating a pressure that allows the blood flow in the arteries, in particular the wrist arteries, to stop and for maintaining this pressure for a predetermined number of cardiac cycles.