FR3053238A1 - METHOD FOR DETECTING AT LEAST ONE HEART RHYTHM DISORDER - Google Patents

Abstract

The invention relates to a method for detecting at least one cardiac rhythm disorder of a human or animal subject from a time series of RR intervals representative of the subject's cardiac rhythm extracted from a physiological signal (S ). The method comprises a step of characterizing (S07) the time series with at least one descriptive variable (Vb) whose value is calculated from a derived series (Sj2, Sj3, Sj4, Sj5, Sj6, Sj7) of which the constituent element is a discrete derivative of order 1 or greater than 1 of the RR interval of the time series.

Description

Holder (s): UNIVERSITY OF AIX-MARSEILLE Public establishment, NATIONAL CENTER FOR SCIENTIFIC RESEARCH Public establishment, PUBLIC ASSISTANCE - HOSPITALS OF MARSEILLE Public establishment.

Extension request (s)

Agent (s): OMNIPAT Société anonyme.

METHOD FOR DETECTING AT LEAST ONE HEART RATE DISORDER.

FR 3 053 238 - A1 t © The invention relates to a method for detecting at least one heart rhythm disorder of a human or animal subject from a time series of RR intervals representative of the heart rate of the subject, extracted from a physiological signal (S). The method comprises a characterization step (S07) of the time series with at least one descriptive variable (Vb) whose value is calculated from a derived series (Sj ₂ , Sj3, Sj ₄ , Sj ₅ , Sj ₆ , Sj ₇ ) whose constituent element is a discrete derivative of order 1 or greater than 1 of the interval RR of the time series.

i

Method for detecting at least one heart rhythm disorder

The present invention relates to a method and a device for detecting a heart rhythm disorder of a human or animal subject, from a physiological signal containing information on the heart rate of the subject.

The present invention relates more particularly but not exclusively to the production of a portable device making it possible to ensure continuous monitoring of the cardiac activity of a subject.

To date there is a real need for a device making it possible to take a person's vital parameters in real time in order to prevent any potential health risks, in particular cardiac risks. There is in particular a real need for screening for atrial fibrillation. Atrial fibrillation is the most common cardiac arrhythmia, responsible for a large number of hospital admissions for rhythm problems. It is estimated that one in six people over the age of 40 will experience this arrhythmia, which currently affects several million people. Its prevalence in the population increases with age, reaching 1.5% beyond 50 years and more than 20% beyond 80 years. Atrial fibrillation is also responsible for 10 to 15% of all strokes and 25% of strokes after 80 years. In almost a third of cases, atrial fibrillation is asymptomatic, behaving like a real silent killer. Due to the aging population, screening for atrial fibrillation is a major public health issue.

There is no device on the market aimed at the general public and performing cardiac monitoring functions in a way that can be considered to be medically reliable. Various portable devices, such as watches equipped with means for acquiring a photoplethysmographic signal, make it possible to measure the heart rate but are not capable of diagnosing in real time the occurrence of a rhythm disorder. Their degree of medical relevance is therefore very low or even non-existent. In addition, specialized equipment such as the Home Health Hub marketed by the company Freescale or the Body Guardian ® Heart developed by the companies Preventice and STMicroelectronics, are intended for patients suffering from cardiac arrhythmias. Biometric data is transmitted to a doctor outside the hospital environment, thus allowing patients to exercise normal daily activity while benefiting from continuous monitoring.

On subjects at risk, an electrocardiogram or ECG reading is generally performed using an event recorder called Holter, or a portable version of it, called Holter-ECG. A Holter-ECG recording typically lasts 24 hours. It has the advantage of allowing the monitoring of a patient outside the hospital environment but cannot be used daily by the population.

The heart is a muscle that contracts at a relatively regular rate. Each normal beat is initiated by an electrical signal generated by the heart's tissue and carried by the heart's conduction system. The ECG signal consists of a succession of electrical depolarizations resulting in the appearance of P, Q, R, S, T, U waves, the shape of which is represented in FIG. 3. The P wave, which corresponds at depolarization of the atria, has a low amplitude. The PQ interval translates the atrioventricular conduction time. The QRS complex reflects ventricular contraction, and the T-wave reflects ventricular repolarization. In practice, the peak R is considered as a marker of the ventricular systole, that is to say of the heartbeat. The R wave constitutes the finest peak and of greatest amplitude of the ECG signal, and is generally used to mark the instant of the heart beat with very good precision.

To automate the detection and identification of a rhythm disorder and thus replace the trained eye of the doctor, various methods have been proposed such as those consisting in analyzing the morphology of the ECG signal, in particular of the P wave and of the complex. QRS. For example, atrial fibrillation observed on an electrocardiogram is characterized by the replacement of P waves by rapid oscillations (called fibrillation waves) of variable size, shape and frequency of occurrence, associated with an irregular ventricular rhythm often rapid when atrioventricular conduction is intact.

Other methods propose a characterization of the ECG by classification of the RR interval using artificial intelligence methods (Ref. 1). The detection of the peaks R in fact makes it possible to form time series whose constituent element is the interval RR, that is to say the time between two successive peaks R. To build a time series of RR intervals, the ECG signal is sampled, digitized and then analyzed to detect R waves. A time series of RR intervals is thus made up of a plurality of successive RR intervals (.. .RRi- 1, RRi, RRi + 1 ...) as illustrated in FIG. 3, each interval RR corresponding to the time interval separating two successive R waves from the ECG signal.

The article Automatic Detection of Atrial Fibrillation Using R-R Interval Signal (Ref. 2) suggests using the mathematical operators VAI and VLI, SD1 and SD2 to distinguish time series in sinus rhythm from time series with atrial fibrillation. A scatter plot, known as the Lorenz or Poincaré graph, whose abscissa is the RR interval (RRi) and the ordinate the following RR interval (RRi + 1) of the time series, makes it possible to represent successive RR intervals as point clouds in which we try to distinguish those attached to subjects with sinus rhythm and those attached to subjects with atrial fibrillation.

But it is also recognized that the analysis of time series alone does not allow reliable identification of a rhythm disorder. Thus, document US Pat. No. 7,941,207 proposes adding, to the detection of the RR intervals, a step of detecting ventricular contractions. Document US 8,155,735 proposes to predict the occurrence of atrial fibrillation by classification of RR intervals originating from an ECG signal as well as from additional characteristics such as the wavelet transform of the ECG signal and the morphology of the P wave. (width and height).

Still within the framework of a detection of atrial fibrillation by classification of RR intervals, the document US Pat. No. 8,019,407 proposes to compare the RR intervals and the dRR variations of the RR intervals by means of a dispersion graph in which these are placed. on the abscissa and on the ordinate. Such a graph makes it possible to show point clouds (clusters) of different geometries depending on whether the successive RR intervals are in sinus rhythm or have a rhythm disturbance. More particularly, this document suggests a multivariate analysis of the topography of these point clouds, with a view to their classification by means of an artificial neural network or a fuzzy logic method. The proposed classification criteria (metrics) are purely graphic and include the number NZCEL of RdR cells of the graph comprising at least one point, the number NZCOL of columns of RdR cells comprising at least one point, the NZROW number of rows of RdR cells comprising at least one point, the number NPEAK of cells comprising a number of points greater than a threshold, the distance of trajectory XDIST of the cloud of points along the axis RR, that is to say the maximum difference between two intervals RR, the distance of trajectory YDIST of the point cloud along the dRR axis, i.e. the maximum difference between two dRR variations, the maximum distance DGMAX from a global center of the point cloud, the maximum distance DPMAX from a positive center determined from of all dRRs of positive value, and the maximum distance DNMAX from a negative center determined from all dRRs of negative value.

It might however be desired to provide a method for detecting a heart rhythm disorder by classification of time series which is simple to implement and offers better accuracy than the known methods.

It may also be desired to provide a portable device for detecting a heart rhythm disorder which is easy to use and not very restrictive, and which can detect a rhythm disorder with good accuracy while offering the possibility of subsequently enabling more in-depth monitoring of the subject on which the rhythm disorder was detected, in a way that can be used by the medical profession.

Embodiments of the present invention relate to a method of detecting at least one heart rhythm disorder of a human or animal subject, from a time series of RR intervals representative of the heart rate of the subject extracted from a physiological signal, comprising the steps consisting in characterizing the time series with at least one descriptive variable whose value is calculated from the RR intervals of the time series, by means of the value of the descriptive variable, classifying the time series in a first class of time series associated with subjects in sinus rhythm or in at least a second class of time series associated with subjects having heart rhythm disorder, in which the value of the descriptive variable is calculated from a derived series whose constituent element is a discrete derivative of order 1 or greater than 1 of the RR interval of the time series.

According to one embodiment, the constituent element of the derived series is chosen from the group comprising: the rate of change of the RR interval, that is to say the discrete time derivative of the RR interval, reflecting the acceleration of the heart rate, the absolute value of the rate of change of the RR interval, the rate of change of the rate of change of the RR interval, either the discrete time derivative of the rate of change of the RR interval, or the second discrete time derivative of the RR interval, reflecting the jolts of the heart rate, the absolute value of the rate of change of the rate of change of the RR interval, the rate of change of the absolute value of the rate of change of the RR interval, and the absolute value of the rate of change of the absolute value of the rate of change of the RR interval.

According to one embodiment, the step of characterizing the time series also comprises the calculation of the value of at least one descriptive variable from a time series whose constituent element is chosen from the group comprising the interval RR, the variation of the RR interval, and the absolute value of the variation of the RR interval.

According to one embodiment, the value of the descriptive variable is calculated by means of a mathematical operator chosen from the group comprising: the mean value, the median, the standard deviation, the variance, the asymmetry coefficient or Skewness, the flattening coefficient or Kurtosis, the ultra low frequency power ULF, the power at very low frequency VLF, the power at low frequency LF, the power at high frequency HF, the ratio LF / HF, the total power, the power Normalized LF, normalized HF power, SD1 or dispersion of points along the minor axis of the ellipse of the Poincaré diagram, SD2 or dispersion of points along the major axis of the ellipse of the Poincaré diagram, the SD1 / SD2 ratio, VAI or the vector angle index, VLI or the vector length index, RMSSD or quadratic mean of the successive differences of the constituent elements of the series, SDSDD or standard deviation of the absolute value of the order differentiation 2 of the constituent elements tifs of the series, the recurrence rate of the recurrence diagram, Lmax or maximum length of the diagonal lines of the recurrence diagram, Lmean or average length of the diagonal lines of the recurrence diagram, determinism, the Shannon entropy of the distribution of lengths of the diagonal lines of the matrix of the recurrence diagram, the Shannon entropy, the ApEn entropy, the SampEn entropy, the correlation dimension.

According to one embodiment, the step of classification of the time series is executed by means of a classifier tool with supervised learning, and the method comprises a step of configuration of the classifier tool comprising the steps consisting in supplying the classifier tool a plurality of values of the descriptive variable calculated from series derived from a first set of time series from subjects in sinus rhythm, and provide the classifier tool a plurality of values of the descriptive variable calculated from series derived from a second set of time series from subjects with heart rhythm disorder.

According to one embodiment, the method comprises a step of continuous or quasi-continuous production of a plurality of time series from the physiological signal, and the determination in real time of the class to which each time series belongs.

According to one embodiment, the time series are produced according to a sliding time window, so that two successive time series can comprise common RR intervals.

According to one embodiment, the method comprises a step of detecting missing RR intervals and in which the production of the time series is interrupted during the periods including missing RR intervals.

According to one embodiment, the time series are extracted from a photoplethy smographi signal.

According to one embodiment, the method comprises a decision step by which the heart rhythm disorder is considered to be detected when a determined number greater than 1 of time series does not belong to the first class of time series.

Embodiments also relate to a device for detecting at least one heart rhythm disorder of a human or animal subject from a time series of RR intervals representative of the heart rate of the subject extracted from a physiological signal, the device being configured to characterize the time series with at least one descriptive variable, the value of which is calculated from the RR intervals of the time series, by means of the value of the descriptive variable, classifying the time series in a first class of series time associated with subjects in sinus rhythm or in at least a second class of time series associated with subjects having heart rhythm disorder, the device being configured to calculate the value of the descriptive variable from a derived series whose l the constituent element is a derivative of order 1 or greater than 1 of the interval RR of the time series.

According to one embodiment, the device is configured to calculate the value of the descriptive variable from a derived series, the constituent element of which is chosen from the group comprising the rate of change of the interval RR, namely the time derivative of the RR interval, reflecting the acceleration of the heart rate, the absolute value of the rate of change of the RR interval, the rate of change of the rate of change of the RR interval, i.e. the discrete time derivative of the rate of variation of the RR interval, i.e. again the second derivative of the RR interval, reflecting the jolts of the heart rate, the absolute value of the rate of change of the rate of change of the RR interval, the rate of change of the absolute value of the rate of change of the RR interval, and the absolute value of the rate of change of the absolute value of the rate of change of the RR interval.

According to one embodiment, the device is configured to also calculate the value of at least one descriptive variable from a time series whose constituent element is chosen from the group comprising the interval RR, the variation of the RR interval, and the absolute value of the variation of the RR interval.

According to one embodiment, the device comprises a classifier tool with supervised learning configured to classify the time series into the first or second classes of time series.

According to one embodiment, the device is configured to continuously or almost continuously produce a plurality of time series from the physiological signal, and to classify each time series in real time.

According to one embodiment, the device is configured to produce the time series according to a sliding time window, so that two successive time series can include common RR intervals.

According to one embodiment, the device is configured to detect missing RR intervals and to interrupt the production of time series of RR intervals during periods including missing RR intervals.

According to one embodiment, the device is configured to extract time series of RR intervals from a photoplethysmographic signal.

According to one embodiment, the device is configured to execute a decision step by which the heart rhythm disorder is considered to be detected when a determined number greater than 1 of successive time series does not belong to the first class of time series .

According to one embodiment, the device is configured to activate a channel for acquiring the subject's electrocardiogram when a heart rhythm disorder is considered to be detected.

According to one embodiment, the device is configured to detect at least one atrial fibrillation.

According to one embodiment, the device is configured to be fixed around the wrist.

According to one embodiment, the device comprises a channel for acquiring a photoplethysmographic signal, at least one channel for acquiring an electrocardiographic signal, and a processor configured to detect heart rhythm disorder in the photoplethysmographic signal and in the electrocar diographic signal.

According to one embodiment, the device is configured to, in a first operating mode, monitor only the photoplethysmographic signal to detect there the heart rhythm disorder, and after detection of the heart rhythm disorder in the photoplethysmographic signal, switch to a second operating mode including monitoring the electrocardiographic signal to detect heart rhythm disturbance.

According to one embodiment, the device is configured to, in the second operating mode, monitor both the photoplethysmographic signal and the electrocar diographic signal.

According to one embodiment, the device comprises a housing or fixing means receiving an electrode for acquiring the electrocardiographic signal that the subject must touch with a part of his body so that the electrocardiographic signal is captured by the acquisition channel of the electrocardiographic signal.

According to one embodiment, the device comprises a multiplexing circuit comprising a first input receiving the photoplethysmographic signal, a second input receiving the electrocardiographic signal, and an output connected to the processor, the device being configured to, in an operating mode, select alternately each of the two inputs of the multiplexing circuit with a high switching frequency in front of the subject's heart rate.

These objects and characteristics will be better understood on reading the following description of embodiments of the present invention, given without implied limitation in relation to the attached figures among which:

FIG. 1 represents in the form of blocks an embodiment of a device for detecting a heart rhythm disorder according to the invention,

FIG. 2 represents in block form another embodiment of a device for detecting a heart rhythm disorder according to the invention,

FIG. 3 shows an electrocardiographic signal which can be used in an embodiment of the device of FIG. 1 or that of FIG. 2,

FIG. 4 shows a photoplethysmographic signal which can be used in an embodiment of the device of FIG. 1 or that of FIG. 2,

FIG. 5 is a flowchart showing an embodiment of a method for detecting a heart rhythm disorder implemented by the device of FIG. 1 or that of FIG. 2,

FIG. 6 is a flowchart describing a first embodiment of a step of characterizing time series appearing in the method of FIG. 5,

- Figures 7, 8 and 9 illustrate the results of studies concerning the characterization of time series,

FIG. 10 is a flowchart describing a second embodiment of the step of characterizing time series appearing in the method of FIG. 5,

FIG. 11 is a flowchart describing an embodiment of a step of classification of time series appearing in the method of FIG. 5,

FIG. 12 is a flowchart describing a learning step of a classifier tool used for the classification of time series,

FIG. 13 represents in the form of blocks an embodiment of a portable device for detecting a heart rhythm disorder according to the invention,

FIGS. 14A and 14B are respectively top and bottom views of the device of FIG. 13,

FIG. 15 is a flowchart describing the functionalities of an embodiment of the device of FIG. 13,

FIG. 16 is a flow chart describing an embodiment of a method for detecting a heart rhythm disorder implemented by the device of FIG. 13,

FIG. 17 is a flow chart describing another embodiment of a method for detecting a heart rhythm disorder implemented by the device of FIG. 13,

FIG. 18 is a flowchart describing an embodiment of a step for producing RR intervals appearing in the method of FIG. 16 or that of FIG. 17,

FIG. 19 is a flowchart describing an embodiment of a step for producing time series appearing in the method of FIG. 16 or that of FIG. 17,

FIG. 20 is a flowchart describing an embodiment of a decision step appearing in the method of FIG. 16 or that of FIG. 17, and

FIG. 21 is a flowchart describing other functionalities of an embodiment of the device of FIG. 13.

FIG. 1 represents an embodiment of a device DV1 according to the invention. The device comprises a processor PI, a channel CH for acquiring a physiological signal S, a program memory Mil, a data memory M12, and a wireless communication interface Cil. The acquisition channel CH is connected to an AT terminal coupled to the body of a subject which can be of any known type, in particular a cardiac probe, skin electrodes or a photoplethysmography module comprising emitting and receiving diodes. The acquisition channel CH supplies the signal S in digitized form to the processor PI. The processor analyzes this signal to detect a heart rhythm disturbance therefrom, by means of algorithm programs PG1, PG2, PG3, PG4 provided in the memory Mil. A clock circuit CCT can be provided to supply a clock signal CK usable as a reference time base for the measurement of intervals RR.

ίο

The device DV1 can be of the implantable type, for example subcutaneous, of the portable type, or even of the fixed type that the subject uses temporarily during a phase of observation of his heart rate. Various other members that may be provided in the device DV1 are not shown, such as an electrical power source, voltage regulators, a display, physiological or posture data sensors, for example a temperature sensor, an accelerometer, a magnetometer ...

FIG. 2 represents another embodiment of a DV2 device according to the invention. The device comprises a processor P2, a program memory M21, an HD data storage means, for example a hard disk, a wireless communication interface CI2 and a circuit CCT supplying a clock signal CK. The program memory M21 comprises the aforementioned algorithm programs PG1, PG2, PG3, PG4 for the analysis of the physiological signal S and the detection of a heart rhythm disorder. The signal S can be prerecorded in the HD data storage means, or be received from the device DV1 via the communication interface CI2 to be analyzed in real time or after having been recorded in the HD storage means. The device DV2 can be a personal computer or the computer of a doctor, a work station in a medical laboratory, a medical server, and in general any device equipped with calculation means making it possible to implement the programs PG1 to PG4.

Figures 3 and 4 show two examples of physiological signals S capable of being analyzed by the device DV1 or DV2. The signal shown in Figure 3 is an ECG signal, while the signal shown in Figure 4 is a photoplethysmographic signal, or PPG signal. The ECG signal presents the QRS complex corresponding to the depolarization of the ventricles, from which the R peak can be extracted for the measurement of RR intervals (RRi-1, RRi, RRi + 1 ...). The PPG signal has a roughly sinusoidal appearance and has peaks denoted R by analogy with the electrocardiographic signal, the occurrence of which is representative of the subject's cardiac activity. At each heartbeat, the time of occurrence of an R peak of the PPG signal is correlated, with a slight phase shift, to the time of occurrence of the R peak of the ECG signal, because the PPG signal accounts for the pulsatile nature of the pulse. which results from the contractile mechanical activity of the heart, the rheological properties of the blood and the mechanical properties of the vessels. Thus the pseudo-period defined by the duration of the peak-to-peak interval of the PPG signal is correlated to the duration of the interval RR. The detection of the peaks R of the photoplethysmographic signal and the measurement of the time intervals between these peaks therefore makes it possible to produce time intervals which will be considered in the following as RR intervals. The P waves and the fibrillation waves of the ECG signal, on the other hand, are electrical events without macroscopic mechanical counterpart, and are not observable in the PPG signal.

FIG. 5 shows steps of the method for detecting a heart rhythm disorder implemented by the device DVI, DV2 by means of the programs PG1 to PG4. The process includes:

a step S03 of detecting peaks R (or the like) in the physiological signal S, and of calculating intervals RR, executed by the program PG1,

a step S05 of forming time series Sj of RR intervals from the RR intervals provided by the program PG1, executed by the program PG2;

a step S07 for characterizing the time series Sj supplied by the program PG2, executed by the program PG3, and

a step S09 of classification of the time series Sj by the program PG4, on the basis of their characterization by the program PG3.

In step S03, the measurement of the RR intervals by the program PG1 can include filtering aimed at identifying aberrant RR intervals due for example to false detections and / or non-detections of RR intervals. In step S05, the program PG2 divides the continuous flow of RR intervals into time series Sj of a minimum duration Te, each comprising a number of successive RR intervals which is a function of the subject's heart rate. In step S07, the program PG3 analyzes the intervals RR constituting the time series Sj in order to extract from it discriminating information necessary for their classification. In step S09, the program PG4 uses the discriminating information to classify the time series Sj in a first class C1 of time series from subjects with sinus rhythm or in a second class C2 of time series from subjects with disorder rhythm to detect, for example atrial fibrillation.

When a rhythm disturbance is detected in real time by the device DV1 by classification of one or more time series in class C2, the classification step S09 is followed by an action performed by the device DVI. This action can for example include the fact of storing in the data memory Ml2 the raw signal S and / or the time series Sj in which the rhythm disturbance has been detected, and / or transfer these data via the communication interface Cil . The device DV1 can for example establish a communication with the device DV2 and transfer to it the signal S, the time series Sj and their classification. The device DV2 is then able to carry out, in deferred time, a verification of the classification ensured by the device DV1, or even to subject the signal S to other types of analyzes aiming to confirm the validity of the classification retained by the device D VI.

The action following the classification of one or more time series in class C2 can also include a decision step preceding an action proper, aimed at validating the classification retained by the PG4 program.

It is known that a time series of RR intervals can be characterized, for classification purposes, by means of descriptive variables whose values are calculated by applying mathematical operators to the RR intervals of the series. The aforementioned article Automatic Detection of Atrial Fibrillation Using R-R Interval Signal (Ref. 2) suggests in particular the use of the operators VAI and VLI or SD1 and SD2 of phase space analysis or Poincaré diagram. These mathematical operators, as well as others described below, for example the mean value, the median, the standard deviation, the variance of the RR intervals, etc., make it possible to obtain discriminating information usable for the classification of the series. temporal.

FIG. 6 shows a first embodiment S07 (l) of the characterization step S07 implemented by means of the program PG3. Step S07 (l) comprises a preliminary step S070 of choosing a group of descriptive variables Val, Va2 ... Vak, corresponding here to the choice of a group of mathematical operators intended to be applied to time series d 'RR intervals. The step S070 is preferably carried out before the device DV1, DV2 is put into service and the PG3 program is written, through studies and tests aimed at determining the best combination of descriptive variables related to the disorder of the rhythm to detect. In one embodiment, the device DV1, DV2 is configured to detect several types of rhythm disorder and uses different groups of descriptive variables, each dedicated to the detection of a pathology. In this case, step S070 can include a step of selecting, in the program memory Mil, M12, of the appropriate group of descriptive variables, or a step of selecting a branch of the program PG3 using the group of descriptive variables dedicated to detection of the target rhythm disorder.

Step S07 then comprises a loop for calculating the values of each descriptive variable, which is initiated after reception, during a step S071, of a new time series Sj provided by the program PG2. The calculation loop comprises a step S072 of calculating the value Valj of the descriptive variable Val, a step S073 of calculating the value Va2j of the descriptive variable Va2, and so on until a step S07k of calculating the value Vakj of the descriptive variable Vak, the value of each variable being calculated from the RR intervals of the series Sj. The PG3 program then supplies the PG4 program with the values Valj, Va2j ... Vakj of the descriptive variables.

The effectiveness of this characterization process was evaluated from RR intervals provided by the Physionet site (http://www.physionet.org/) of the Massachussetts Institute of Technology (MIT), which makes it available to the public. ECG signal banks and RR interval time series, including the Sinus Rhythm RR Interval Database, MIT-BIH Normal Sinus Rhythm Database, MIT-BIH Atrial Fibrillation Database, AF Termination Challenge Database, and MIT-BIH Arrythmia Database. From these databases, a library of RR interval time series has been created, including:

- 56,964 reference time series of 60 s each including 56,711 in sinus rhythm and 253 in atrial fibrillation (Atrial Fibrillation), and

- 322,779 reference time series of 10 s each, including 297,269 in sinus rhythm and 25,510 in atrial fibrillation.

The study revealed a non-negligible percentage of false negatives when classifying reference time series by logistic regression or using a classifying tool with supervised learning. In other words, a significant number of time series in atrial fibrillation have been classified as time series in sinus rhythm.

It was therefore desired to improve the time series characterization process implemented in step S07, in order to obtain a more exact classification without involving descriptors based on the analysis of the morphology of the ECG signal, which would not be not usable with another type of signal, in particular the PPG signal. An improvement has been found, which will now be described in relation to FIGS. 7 to 11.

According to this refinement, step S07 of characterizing the time series includes steps of calculating derived series whose constituent elements are discrete derivatives, denoted d ^N RR / dt ^N , intervals RR of the initial time series, N being an integer at least equal to 1. We have indeed highlighted the fact that such derived series contain discriminating information which can be added to that which can be extracted from the initial time series, or even replace it, with a view to their classification.

To fix the ideas, the table of figure 7 illustrates with shades of gray a degree of correlation between derivatives of intervals RR of increasing order N. This degree of correlation is measured on a scale from 0 to 1 using the Pearson coefficient. The table in Figure 7 is reproduced in Annex 1, Table 1, the shades of gray being replaced by numerical values. The central diagonal from the upper left to the lower right corner of the table shows that, trivially, the RR intervals and their derivatives have a correlation coefficient with respect to themselves equal to 1. The diagonals close to the central diagonal show that the derivatives of RR intervals of adjacent orders (for example a derivative of order 2 and a derivative of order 3) have a degree of correlation between them less than 1, for example of l '' order of 0.8. This degree of correlation decreases to tend towards 0 as the difference in rank between the derivatives increases. Likewise, there is a low degree of correlation between the RR intervals and their derivatives, starting with the first-order derivative. These results testify to the existence, in the derivative series, of discriminating information that is not found in the initial time series.

Thus, from an initial time series to be characterized for classification, the method includes the following steps:

a step of calculating at least one derivative series of order N whose constituent element is the discrete time derivative d ^N RR / dt ^N of the interval RR of the initial time series, and

a step of characterization of the derived series by means of one or more descriptive variables whose values are calculated by application of a mathematical operator to the constituent elements of the derived series, and are considered to form discriminating information for the classification of the initial time series.

For the sake of clarity of the description, the following will be called:

- derived descriptive variable Vb: a descriptive variable obtained by applying a mathematical operator to a derived series,

- primitive series: a time series of RR intervals or a time series whose constituent element is the variation of the RR interval or the absolute value of this variation, and

- primitive descriptive variable Vc: a descriptive variable obtained by applying a mathematical operator to a primitive series.

Thus, the notion of primitive descriptive variable includes the notion of descriptive variable previously mentioned, which relates only to the application of a mathematical operator to a time series of intervals RR.

Furthermore, in one embodiment, the concept of derived series includes series whose constituent element is calculated from the absolute value of the elements of a derivative series of lower order, for example, as will be seen more far, the rate of change of the absolute value of the rate of change of the RR interval, or the absolute value of this rate of change.

Thus, if Nmax designates the maximum degree of derivation used to characterize a time series, ie the number of available derivative series, and NVa the number of mathematical operators which can be applied to the derived series, the number of derived descriptive variables can be used is equal to NVa * Nmax. By adding to the derived series those whose constituent element is calculated from the absolute value of the elements of a derived series of lower order, the number of derived descriptive variables that can be used is even greater.

Examples of mathematical operators and their corresponding formulas, which can be used to define primitive or derived descriptive variables, are described in Annex 2. We distinguish:

- statistical operators, such as:

- the average value,

- the median,

- the standard deviation,

- the variance,

- the asymmetry coefficient or Skewness,

- the flattening coefficient or Kurtosis,

- spectral analysis operators (see Ref. 3), such as:

- ULF power (Ultra Low Frequency Power) or ultra low frequency power [0-3 mHz],

- VLF Power (Very Low Frequency Power) or very low frequency power [3 mHz - 40 mHz],

- LF Power (Low Frequency Power) or low frequency power [40 mHz

- 150 mHz],

- HF Power (High Frequency Power) or high frequency power [150 mHz

- 400 mHz],

- the LF / HF Ratio,

- total power,

- the normalized LF power [40 mHz - 150 mHz],

- the normalized HF power [150 mHz - 400 mHz],

- operators of phase space analysis (Poincaré diagram, Cf. Ref. 2) such as:

- SD 1, or dispersion of the points along the minor axis of the ellipse,

- SD2, or dispersion of the points along the major axis of the ellipse,

- the SD1 / SD2 ratio,

- VAI, or Vector Angular Index,

- VLI, or Vector Length Index,

- RMSSD, or Quadratic mean of the successive differences of the constituent elements of the series to be characterized,

- SDSDD, or Standard deviation of the absolute value of the order differentiation 2 of the constituent elements of the series to be characterized,

- analysis operators of the RP (Plot Recurrence) recurrence diagram such as:

- the Recurrence Rate,

- Lmax, or Maximum length of the diagonal lines of the RP recurrence diagram,

- Lmean, or Average length of the diagonal lines of the RP recurrence diagram,

- Determinism,

- the Shannon Entropy of the distribution of the lengths of the diagonal lines of the matrix of the RP diagram,

- operators to analyze the regularity of the series to be characterized, such as:

- Shannon's Entropy,

- ApEn Entropy (Approximate Entropy),

- SampEn Entropy (Sample Entropy), and

- operators for analyzing the chaotic and fractal properties of the series to be characterized, such as the Correlation Dimension.

The characterization of a time series according to the process described here can be implemented according to several variants:

- by using exclusively one or more descriptive variables derived Vb, or

- by using one of several descriptive variables derived Vb and one or more primitive descriptive variables Vc.

Studies have been carried out to evaluate the best combinations of descriptive variables derived Vb and of primitive descriptive variables Vc making it possible to obtain good classification accuracy without prohibitively complexifying the characterization step, and this from the perspective of '' an implementation of the method in a device of a reduced cost price, offering an average computing power. These studies show that it may be preferred, in certain embodiments, to use only primitive or derived descriptive variables formed from different mathematical operators, in other words not to apply the same mathematical operator to different primitive or derived series. It also appears that the use of primitive descriptive variables may not be necessary with certain derived descriptive variables, while with other derived descriptive variables the addition of one or more primitive descriptive variables may improve the accuracy of the classification. . Finally, it appears that this choice also depends on the rhythm disorder to be detected.

By way of example, the table in FIG. 8 illustrates the result of a study aimed at identifying the best combination of primitive and derived descriptive variables using as mean mathematical operators the mean value (mean) and the standard deviation (σ ). The table in Figure 8 is also reproduced in Annex 1, Table 2, the shaded areas being replaced by crosses. The study was conducted under the following conditions

- by fixing the maximum degree of derivation at 4 (Nmax = 4), either by using derived series of order 1, 2, 3 and 4,

- by using the mean value and the standard deviation to form combinations of primitive or derived descriptive variables comprising from 1 to 10 variables,

- within the framework of a classification of time series in a class of time series associated with subjects in sinus rhythm (class NSR) or in a class of time series associated with subjects having atrial fibrillation (class AF),

- by using a logistic regression function to distinguish the NSR and AF classes, and

- using the aforementioned reference time series, generated from Physionet data, whose membership class is known.

Each column of the table corresponds to a combination of variables comprising from 1 to 10 variables. The gray cells (or, in Annex 1, those with crosses) indicate the best choice of variables. When two variables are used (column 2), the pair of variables offering the best classification accuracy is (d ^ RR \ fdRR \ -], meanl—}, that is, the standard deviation of the components of the second derivative and the mean value of the components of the first derivative With these two descriptive derived variables, a classification accuracy of 99.9% was obtained, without using primitive descriptive variables From column 5 (ie 5 variables used) l standard deviation of RR intervals in time series is part of the optimal combination of variables. From column 9 (ie 9 variables used) the average value of RR intervals is also part of the optimal combination of variables. primitive descriptors were therefore preferred to certain derived descriptive variables, for example σ (yyr) ·

The study of all possible combinations of primitive and derived descriptive variables showed that the accuracy of the classification could vary between 69.6% and 99.9% depending on the combination chosen, hence the importance of the choice of variables . With a result as satisfactory as 99.9% obtained with only two carefully chosen derived descriptive variables, it appears that the choice of a large number of primitive and / or derived descriptive variables may not be necessary. Thus, the gray cells in columns 3 to 10 of the table include the two derived descriptive variables retained in column 2 of the table, which alone offer a classification accuracy of 99.9%, so that the addition of the other variables designated in these columns does not improve the accuracy of the classification.

Other studies have been conducted to assess the influence of the duration of the observation window, that is to say the minimum duration Te of each time series (step S05, Fig. 5) on the relevance of the classification. Figure 9 illustrates one of the results of these studies and shows a curve C1 corresponding to the variation in the accuracy of the classification, expressed as a percentage, as a function of the duration of the observation window, expressed in seconds, in using the pair of derived descriptive variables l · ⁷ (~ â [p) ' ^mean } · ^' other curves C2, C3 .... Cn appearing in figure 9 correspond to the variation of the accuracy of the classification for d other combinations of variables. Curve C1 shows that the duration of the observation window can be reduced up to 5 seconds, which represents only 5 heart beats with a subject whose heart rate is 60 beats per minute, while maintaining an accuracy of classification greater than 95%. However, the reduction in the duration of the observation window makes it possible to reduce the classification time of each time series and thus to increase the reactivity of the DVI device, DV2 during the appearance of a rhythm disorder, in the framework for real-time monitoring of the subject's cardiac activity.

It will be noted that the analyzes, conclusions and results of studies which have just been described are purely indicative and should not be considered as reducing the scope of implementation of the process presently described.

FIG. 10 shows an embodiment S07 (2) of the characterization step S07 according to this method. Step S07 (2) comprises a preliminary step S0700 of choosing at least one derived descriptive variable Vb. As before, this choice is preferably made before the commissioning of the device DV1 or DV2, by means of preliminary studies, and according to the rhythm disturbance to be detected. This is for example the average value of the first derivative and the standard deviation of the second derivative, the discriminating performances of which have been highlighted above in relation to FIG. 8. Optionally, at least one variable primitive descriptive Vc can also be used. Finally, within the framework of the detection of different rhythm disorders, this choice can be made dynamically by the device DV1, DV2 according to the rhythm disorder to be detected, by selecting a group of variables from several predetermined groups or of branches of the PG3 program configured to use these variables.

After reception of a time series Sj during a step S0701, the program PG3 executes all or part of the following steps:

step S0702: calculation of a series Sjo whose constituent element is the variation Vi of the interval RR of the series Sj,

- step S0703: calculation of a series Sji whose constituent element is the absolute value | Vi | the variation of the RR interval,

step S0704: calculation of a derivative series Sj2, the constituent element of which is the rate of variation Ai of the interval RR (acceleration of the heart rate),

- step S0705: calculation of a derived series Sj3 whose constituent element is the absolute value | Ai | the rate of variation Ai of the RR interval (absolute acceleration of the heart rate),

step S0706: calculation of a derived series Sj4 whose constituent element is the rate of change ACi of the rate of change Ai of the interval RR (jerk of the heart rate),

- step S0707: calculation of a derived series Sjs whose constituent element is the absolute value | ACi | the rate of change ACi of the rate of change Ai of the RR interval (jerk in absolute value of the heart rate),

- step S0708: calculation of a derived series Sjô whose constituent element is the rate of change Gi of the absolute value | Ai | the rate of change Ai of the RR interval,

- step S0709: calculation of a derived series Sj7 whose constituent element is the absolute value | Gi | of the rate of change Gi of the absolute value | Ai | the rate of change Ai of the RR interval.

Annex 3 provides examples of formulas for calculating the various derived series which have just been mentioned. Once these calculation steps have been performed, or as they are executed, the program PG3 performs the following steps:

step S0720: determination of the value Vclj of at least one primitive descriptive variable Vcl calculated from the series Sj,

step S0730: determination of the value Vc2j of at least one primitive descriptive variable Vc2 calculated from the series Sjo,

step S0740: determination of the value Vc3j of at least one primitive descriptive variable Vc3 calculated from the series Sji,

- step S0750: determination of the value Vbj of at least one descriptive variable derived Vb from a derived series Sj ₂ to Sj7.

Steps S0720, S0730 and S0740 are optional and are not executed if the choice made in step S0700 does not include a primitive descriptive variable. Thus, the method may include only two of these steps, only one of these steps, or none of these steps. Likewise, the number of derived descriptive variables whose values are calculated in step S0750 depends on the choice made in step S0700. Finally, some of the steps S0704 to S0709 may not be executed if the chosen derived descriptive variables do not require the calculation of the corresponding derivatives.

The program PG3 then supplies to the program PG4 the value (s) Vbj of one or more derived descriptive variables Vb, as well as optionally the value (s) Vcj of one or more primitive descriptive variables Vc, and returns to step S0701 to wait to receive a new time series Sj.

In one embodiment of the detection method, the PG4 program is a classifier tool comprising an artificial neural network (RNA) with vector quantization and supervised learning, called LVQ (Leaming Vector Quantization). It has a hidden layer called the competitive layer, followed by a classification layer. The competitive layer contains hidden neurons, the classification layer contains output neurons, each of which is representative of a vector membership class of a learning base. The results provided by this classification tool architecture are in the order of 100% in learning and 99.9% in cross-validation. The robustness of the model has also been tested in the presence of various special cases such as the presence of ectopic beats, a sinus rhythm presenting a strong respiratory sinus arrhythmia, etc.

FIG. 11 shows an embodiment of the classification step S09. Before being put into service, the classifier tool PG4 is the subject of a configuration step S0900 by supervised learning by means of two sets NI, N2 of reference time series. Once learning is complete, the program is able to classify time series into two classes C1, C2 of time series corresponding respectively to the sets NI and N2. In one embodiment, the classifier PG4 is configured to distinguish a class C1 of time series associated with subjects in sinus rhythm and a class C2 of time series associated with subjects in atrial fibrillation. In a step S091, the classifier PG4 receives one or more values Vbj of descriptive variables derived Vb calculated by the program PG3 in step S07 (2) for a time series Sj, and optionally one or more values Vcj of descriptive variables primitive Vc. In a step S092, the classifier tool PG4 evaluates, from these variables, the membership class of the series Sj. The classifier PG4 provides this classification and returns to step S091 to receive new values of these variables associated with a following time series.

Figure 12 shows an embodiment of the configuration step of the PG4 classifier tool. This can be executed using a version of the classifier tool installed on a workstation and aims to obtain a set of MC4 configuration matrices. It includes the following preparatory stages:

a step S0901 of preparation of a set NI of time series originating from subjects in sinus rhythm, for example abovementioned reference time series of one minute each originating from subjects in sinus rhythm and produced from the database of the Massachussetts Institute of Technology,

a step S0902 for characterizing the time series of the set NI. The same variables are used here as those which will be used during the characterization step S07 (2). This step therefore comprises, for each series Sj, the calculation of values Vbj of descriptive variables derived Vb, and if necessary the calculation of values Vcj of primitive descriptive variables Vc,

a step S0904 of preparing a set N2 of time series originating from subjects having the rhythm disorder to be detected, for example the abovementioned reference time series of one minute each originating from subjects with atrial fibrillation and produced from the Massachussetts Institute of Technology database,

a step S0905 for characterizing the time series of the set N2. The same variables are also used as those which will be used during the characterization step S07 (2). This step therefore comprises, for each series Sj, the calculation of values Vbj of derived descriptive variables Vb, and if necessary the calculation of values Vcj of primitive descriptive variables Vc.

The values Vbj, Vcj of the variables Vb, Vc calculated for the time series of the set NI are then supplied to the classifier tool PG4 with the status class C1 during a step S0903. Similarly, the values Vbj, Vcj of the variables Vb, Vc calculated for the time series of the set N2 are supplied to the classifier tool PG4 with the class status C2 during a step S0906. From all the values of variables provided for the class Cl and those provided for the class C2, the classifier tool learns, during a step S0907, to distinguish the classes Cl and C2. This is supervised learning since the classes are imposed on the classifier tool when the values of the variables are supplied in steps S0903 and S0906.

The configuration of the classifier tool resulting from the learning phase, taking the form of a set of configuration matrices MC4, is then saved during a step S0908. During a step S0909, the configuration matrices MC4 are loaded into the memory Mil, M21 of the device D VI, DV2 (FIGS. 1 and 2).

It will be clear to those skilled in the art that the process of characterization and classification of time series which has just been described is capable of various variants, embodiments and applications. Although the method was originally designed to monitor the cardiac activity of a human subject, it is also applicable to the detection of cardiac disorders in animals.

In addition, the PG4 classifier tool can be configured to provide a classification of rhythm disorders in several classes C2 (l), C2 (2), C2 (3) ... each corresponding to a specific rhythm disorder (atrial fibrillation) , atrial flutter, tachyarrhythmia, supraventricular tachycardia, sinus tachycardia, atrial or ventricular extrasystole, ventricular tachycardia, ventricular fibrillation ...). As a variant, several classifying tools, each configured to distinguish two classes, namely the class Cl and one of the classes C2 (l), C2 (2), C2 (3) ..., can be provided and intervene one after the others during classification step S09. A classification conflict arbitration algorithm can be provided in the case where several classification tools provide different classifications.

Also, the classification can be implemented with other types of classifier tools than that previously described, in particular a classifier tool with logistic regression or other types of neural networks under supervision such as those designated LVQ2, LVQ3, OLVQ3 in the litterature. Specific training for an individual at risk of a rhythm disorder may also be provided. In this case, the configuration of the classifier tool includes first supervised learning from an existing database relating to a population, such as that described above, then personalized learning in relation to the subject, under medical supervision .

It will also appear to those skilled in the art that the derived series formulas provided in Annex 3 are not limiting and result from a choice consisting in a direction of observation, here on the right, corresponding to the direction of flow of the time. Thus, from an instant Ti, we observe what happens at an instant Ti + n, n going from 1 to the maximum degree of derivation chosen. Conversely, one could, in another embodiment, study what happened before the instant Ti or at the instant Ti-n, the derivatives on the right (in i + n) not being equal to the derivatives on the left (in). This results from the fact that the observed time series are not continuous functions and that the derivatives considered here are discrete time derivatives. Other embodiments can include both the forward-looking derivatives (right derivatives) and the forward-looking derivatives (left derivatives), for further information.

It will also appear to those skilled in the art that the calculation steps which have just been described can be implemented from the instantaneous heart rate which is calculated by means of the formula 1 / RRi (number of beats per second) or the formula 60 / RRi (number of beats per minute), the RRi interval must then be expressed in seconds. In this case, the RR interval series are replaced by instantaneous heart rate series, and the RR interval series derivatives are replaced by derivatives of instantaneous heart rate series, which are calculated in the same way as that indicated in Annex 3 by replacing the RR interval with the instantaneous frequency. Thus, in the present description and in the claims, the term RR interval must be understood as also designating the instantaneous heart rate, the term RR interval time series must be understood as also designating the series of instant heart rates, and the term derived series should be understood as also designating a series derived from a series of instantaneous heart rates.

FIG. 13 represents an embodiment of a DV3 device according to the invention, intended to be worn by a user. The device comprises a channel CH1 for acquiring a photoplethysmographic signal SI, a channel CH2 for acquiring an electrocardiographic signal S2 (ECG) and optionally other channels CHn for acquiring signals Sn, for example a second channel acquisition of an electrocardiographic signal, a temperature acquisition channel, a signal acquisition channel supplied by an accelerometer, and a signal acquisition channel supplied by a magnetometer.

The channel CH1 is coupled to at least one light-emitting diode ED and at least one photodiode PD. The CH2 channel is coupled to two dry electrodes E1, E2. The signals SI, S2 ... Sn supplied by the different acquisition channels are applied to the inputs of a multiplexer MUX whose output is connected to a processor P3 via an analog-digital converter ADC. A bypassable LPF low-pass filter can be provided between the output of the multiplexer and the ADC converter, in order to remove parasitic components which may be present in the signals S1 and S2 or in one of these signals.

The device DV3 also includes a program memory M31, a data memory M32, a wireless communication interface CI3, a display DS and a clock circuit CCT supplying a clock signal CK forming a time base for the measurement of 'RR intervals. The multiplexer MUX receives a selection signal SEL supplied by the processor P3 and transfers to it the corresponding signal SI, S2, ... Sn selected on one of its inputs. Other peripheral means of the processor P3, shown diagrammatically by a PD block, may include a battery or any other source of energy, a power management circuit, regulators supplying different bias voltages, a USB port, a piezoelectric beeper (buzzer), a vibrator, an inertial microcentral, etc.

FIGS. 14A and 14B show the device DV3 respectively by a top view and a bottom view. The device is mounted in a protective box 10 fitted with a strap 11. The top of the box 10 receives the DS display. The underside of the housing 10 receives the electrode El, an auxiliary electrode 12 (reference potential electrode), as well as a photoplethysmography micromodule 13. The photoplethysmography module 13 comprises for example three light-emitting electrodes ED and a photodiode PD. The electrode E2 is arranged here on an external face of the bracelet 11, FIG. 14A, but could also be arranged on one face of the case 10.

For the acquisition of the ECG signal S2, the electrode El being in permanent contact with the skin, the user must touch the electrode E2 with a finger of the hand opposite to that which receives the DVI device, or with any part of this hand, for example the top of the hand. The difference in bioelectric potential for the acquisition of the electrocardiographic signal is thus optimal since it is measured between two ends of the body.

In other embodiments of the device DV3, the electronic means shown in FIG. 13 can be integrated into a bracelet, the assembly then being devoid of a case. These electronic means can also be integrated in the case of a watch, or in the strap of a watch. In a variant, the electrodes E1, E2 are skin electrodes connected to the device DV3 by wires and microconnectors. This embodiment can be provided for subjects who are already known to have a rhythm disorder, which must be closely monitored. In another embodiment, the electrodes E1, E2, instead of being connected to the device DV3, are connected to an electronic module configured to transfer the ECG signal S2 to the device DV3 via a wireless communication channel.

A general advantageous characteristic of the device DV3 is that it combines the ease of acquisition of the photoplethysmographic signal SI (no skin electrodes to be placed on the body of the user) while making it possible to acquire the electrocardiographic signal S2 when this s 'turns out to be necessary, here by a simple press on the electrode E2, otherwise by means of wired skin electrodes or connected to a wireless transmitter.

Figure 15 shows an operating configuration of the DV3 device based on this feature. The DV3 device has a continuous standby operating mode and an alert operating mode. In the continuous standby mode, the processor P3 activates the channel CH1 and continuously analyzes the photoplethysmographic signal SI. The processor switches to alert mode when a rhythm disturbance is detected in the photoplethysmographic signal. The processor then activates the channel CH2, selects the electrocardiographic signal S2 by means of the multiplexer MUX and requests the user, by any means provided for this purpose (DS display, ringtone, vibrator, voice message ...) to touch the electrode E2 for a determined time, for example for one minute. The electrocardiographic signal is then duly analyzed and if the rhythm disturbance is confirmed, the device can request urgent action from the user and / or carry out one or more actions aimed at protecting it.

In one embodiment of the alert operating mode, the processor P3 alternately selects the photoplethysmographic signals SI and electrocardiographic signals S2 by rapidly switching the inputs of the multiplexer MUX by means of the signal SEL, for example with a frequency of the order of 1000 Hz or plus, about 1,000 times the average heart rate of a person at rest. In this case, the multiplexer MUX supplies the converter ADC with signals SI, S2 in a pseudo-sampled form which is digitized by the converter ADC before being analyzed by the processor P3. Other signals Sn can also be simultaneously supplied to the processor in the alert mode. In particular a second acquisition channel of the ECG signal, using another acquisition circuitry to be chosen from various known circuits, connected to the electrodes E1, E2 or using other electrodes, can make it possible to reinforce the reliability of the acquisition and of analysis of the ECG signal by analysis of the two ECG signals and arbitration of the results obtained.

The analysis of the signal SI or S2 by the processor P3 is ensured by various program-algorithms loaded in the memory M31, in particular:

- a PG01 program for digital filtering and noise suppression in the SI signal,

- a program PG02 for digital filtering and noise suppression in the signal S2,

a program PG11 for detecting peaks R in the signal SI (peaks PPG) and for measuring intervals RR, equivalent to the program PG1 previously described,

a program PG12 for detecting peaks R in the signal S2 (QRS complex) and for measuring intervals RR, equivalent to the program PG1 previously described,

- the PG2 time series training program Sj already described,

the program PG3 for characterizing the time series Sj, preferably in its improved version using the descriptive variables derived Vb and primitives Vc,

- the PG4 program for classifying time series, for example the previously described classifier tool and its configuration matrices MC4.

A PG5 decision program is also planned. The PG5 program receives the time series classifications provided by the PG4 classifier and decides, in the light of these, whether it can be considered that a rhythm disturbance has been detected. Finally, a PAP application program is planned to manage the general functionalities of the device and its operating modes.

Figure 16 shows steps in a method of detecting a heart rhythm disorder performed by the DV3 device using the above programs. The method comprises a step SOI of selecting the channel CH1 and / or CH2 and steps specific to each of the signals SI and S2, namely:

a step S02 (l) of filtering the signal SI by the program PG01 followed by a step S03 (l) of detecting the peaks R (peaks PPG) and of measuring the intervals RR by the program PG11, and

a step S02 (2) of filtering the signal S2 by the program PG02 followed by a step S03 (2) of detection of the peaks R (QRS complex) and measurement of the intervals RR by the program PG12.

The method then comprises steps of processing the signal common to each of the signals S1, S2, but applied separately to these, in particular:

the step S05 previously described of forming time series Sj from the intervals RR provided by the program PG11 or PG12, executed by the program PG2,

the step S07 (2) previously described (FIG. 10) for characterizing the time series Sj, executed by the program PG3, and

- the step S09 previously described (FIGS. 5, 11) of classification of the time series Sj by the classifier tool PG4, in the class Cl (normal subject) or the class C2 (subject having a heart rhythm disorder).

Step S09 is followed by a decision step SI 1 executed by the program PG5, based on the classification information provided by the classifier tool PG4. In one embodiment, the program PG5 confirms the detection of a rhythm disturbance when a determined number D of successive time series have been assigned to the class C2. Thus, an isolated classification of a time series in class C2 is not considered sufficient to consider that the subject has a rhythm disorder, as well as several classifications in class C2 of temporal senna which are not successive.

In one embodiment, the PG5 program can, in an equivalent manner, be configured to confirm the detection of a rhythm disturbance when D successive time series have not been classified in class C1. This embodiment applies in particular when the classifier tool is configured to provide a classification of rhythm disturbances in several classes C2 (l), C2 (2), C2 (3) ... each corresponding to a heart rhythm disorder, the decision being made without looking to see if the disorders detected have been assigned to the same class among all the C2 classes available.

Finally, when the decision step SI 1 indicates that a rhythm disorder has been detected, the method comprises a step S12 consisting in initiating a specific action, carried out by the PAP application program, aimed at protecting the subject and / or collection of information allowing the medical profession to diagnose the event.

FIG. 17 shows another embodiment of a method for detecting a rhythm disorder executed by the device DV3. This embodiment differs from the previous one in that, when the electrocardiographic signal S2 is analyzed, the characterization step S07 (2) of the time series is replaced by a characterization step S07 (3) carried out by a program PG32 which replaces the PG3 program. This step S07 (3) includes, in addition to the characterization of the time series in the manner previously described, a step of characterization of the morphology of the ECG signal, which includes for example the analysis of the shape of the P wave and of the QRS complex, for example their amplitude and duration. In this case, the classification step S09 is replaced by a classification step S09 (2) carried out by a second classifier tool PG42 which replaces the program PG4. The PG42 program has MC42 configuration matrices obtained at the end of a learning phase based on both the characterization of the time series and the characterization of the morphology of the S2 signal.

Thus, when a rhythm disturbance has been detected by analysis of the photoplethysmographic signal SI, the switching to the alert mode allows the device DV3 to acquire the electrocardiographic signal S2 and to conduct a precise analysis of it, allowing the program PG5 to provide very reliable confirmation of the presence of a rhythm disorder. It will however be noted that the improvement proposed above concerning the characterization of the time series by means of the descriptive variables derived Vb, already makes it possible to obtain a reliable detection based on the sole observation of the photoplethysmographic signal SI.

The algorithm executed by the processor P3 under the control of the program PG11 or PG12 in step S03 (l) or S03 (2) is preferably designed to detect and eliminate the aberrant RR intervals and thus further improve the reliability of the process rhythm disturbance detection. FIG. 18 shows an embodiment of this algorithm, which can also be used to implement step S03 of FIG. 5. The algorithm comprises two steps S030 and S031 executed in background tasks and a loop for calculating the RRi intervals. Step S030 consists in the reception of the discrete values of the digitized signal SI or S2 and step S031 consists in the analysis of this signal for the detection of peaks R or assimilated (peaks PPG).

The calculation loop is initiated after detection, at a step S0302, of a peak Ri (Ti), Ti designating the instant when the peak Ri is detected. It comprises steps S0300 of initialization of a loop variable i (i = 0) and S0301 of initialization in the low state of an IFR interrupt flag. After detection of the peak Ri (Ti), the processor P3 checks during a step S0303 whether a previous peak Ri-1 (Ti-1) of instant Ti-1 has been stored. If so, the processor determines, during a step S0304, the value of the interval RRi, such that RRi = Ti- (Ti-1), then verifies during a step S0305 if the interval RRi is less than a threshold Tmin. If not, the processor stores the peak Ri (Ti) during a step S0306, then determines during a step S0307 if the interval RR is greater than a threshold Tmax. If not, the processor supplies the program PG2 with the interval RRi during a step S0308. The RRi interval is accompanied by the IFR interrupt flag, which may be in the low or high state depending on the steps previously executed. At a next step S0309, the processor forces the IFR interrupt flag into the low state, then conducts an optional step S0310 for adjusting or resetting the thresholds Tmin, Tmax of admissibility of the intervals RRi. This step consists in redefining the thresholds Tmin, Tmax as a function of the increase or decrease in the heart rate of the subject linked to his activity, and involves detection of slow variations in the interval RRi. The processor then increments the loop variable i (i = i + 1) during a step S0311 and returns to step S0302 to await the occurrence of the next peak Ri.

When it appears in step S0303 that no peak Ri has been stored, the processor stores the current peak Ri (Ti) during a step S0312 and then goes to step S0311 to increment the index of loop before returning to step S0302. When it appears in step S0305 that the interval RRi is less than Tmin, the current peak Ri is considered to be aberrant and the processor returns directly to step S0301 to wait for a new peak, without storing the current peak. When it appears in step S0307 that the interval RRi is greater than Tmax, the processor considers that one or more previous peaks have not been detected due to an interruption in the reception of the signal SI, S2 or a peak nondetection. The current peak Ri is considered to be the first peak received after the supposed interruption and the processor goes to a step S0314 where it sets the IFR interrupt flag high, then erases the previous peak Ri-1 (Ti- l) during a step S0315. The processor then goes to step S0311 to increment the loop variable and returns to step S0302 to wait for the occurrence of a new peak.

Thus, the programs PG1, PG11, or PG12 provide intervals RRi devoid of outliers and accompanied by the IFR interrupt flag allowing the program PG2 to know, when this flag is in the high state, that the corresponding RRi interval is the first RR interval detected after a suspected interruption in the reception of the R peaks.

FIG. 19 shows an example of an algorithm executed by the processor P3 during step S05, under the control of the program PG2. The algorithm comprises a loop for forming a time series Sj which is initiated after each reception, in a step S0502, of an interval RRi, and is preceded by a step S0500 of initialization of a loop variable j (j = 0) and a step S0501 of initialization in the low state of an IFS interrupt flag distinct from the IFR interrupt flag.

The loop for forming a time series Sj comprises a step S0503 during which the processor checks whether the interrupt flag IFR associated with the interval RRi received is in the high state. If not, the processor adds the interval RRi to the series Sj during a step S0504, then goes to a step S0505 where it determines whether the cumulative duration of the intervals RRi of the series is greater than or equal to a threshold Te corresponding to the minimum duration of the time series previously mentioned. If not, the processor returns to step S0502 to wait for a new interval RRi. If the cumulative duration of the intervals RRi is greater than or equal to the threshold Te, the processor goes to a step S0506 where it supplies the series Sj to the characterization program PG3 with the interrupt flag IFS in the low or high state as a function of the previously performed steps. The processor then increments the loop variable j during a step S0507 (j = j + 1) then goes to a step S0508 where it initializes a new series Sj by removing the first Np intervals RRi from the previous series to form the new Sj series. A sliding window for forming the time series is thus defined, each time series comprising intervals RRi present in the previous time series. The processor then forces the IFS interrupt flag to low during a step S0509, then returns to step S0502 to wait for a new interval RRi.

When the IFR interrupt flag is found in the high state at step S0503, the processor goes to a step S0510 where it increments the loop variable j (j = j + 1) then goes to a step S0511 where it initializes a new empty Sj series marked with the IFS interrupt flag in the high state, then adds the current RRi interval to the new Sj series during a step S0512 and returns to step S0502 to wait for a new RRi interval . Thus, the new time series formed from step S0511 is accompanied by the flag IFS in the high state indicating that the time series Sj is formed after a presumed interruption in the reception of the signal SI, S2 or of a non peak detection. This series will be supplied to the program PG3 in step S0506 after having received a sufficient number of intervals RRi so that the time Te is reached, if no other interruption occurs in the meantime.

The IFS flag is used during the decision step S11 executed by the processor P3 under the control of the program PG5, and makes it possible to know whether the time series supplied by the program PG2 are successive or not. In an embodiment shown in FIG. 20, the decision step S11 includes a step SI 11 during which the processor waits to receive a classification of a series Sj supplied by the classifier tool PG4 (or PG42, FIG. . 17), this step being preceded by a step SI 10 for initializing a counting variable d (d = 0). After receiving the result of the classification, the processor determines during a step SI 12 if the series Sj is attached to the class C2, namely if it has a heart rhythm disorder. If not, the processor returns to step SI 10 to reset the counting variable d to 0, then goes to step SI 11 to wait for the classification of the next time series Sj. When it appears in step SI 12 that the series Sj is attached to class C2, the processor goes to a step SI 13 where it checks whether the interrupt flag IFS is in the high state. If so, the processor sets the counting variable d to 1 during a step SI 17 then returns to step SI 11 to wait for the classification of the next time series Sj. If not, the processor goes to a step SI 14 where it increments the counting variable by 1 (d = d + l), then goes to a step SI 15 where it checks whether the counting variable d is equal to a threshold D greater than 1. If the threshold D is not reached, the processor returns to step SI 11 to wait for the classification of the next time series Sj. If the threshold D is reached, the processor goes to a step SI 16 where it indicates to the application program PAP that a rhythm disturbance has been detected. The threshold D is determined in such a way that the duration of a global observation window encompassing D successive time series, which is less than the sum of the respective durations of the time series because they are generated according to a sliding time window, is short enough according to the possible urgency of the management of the detected rhythm disorder, and long enough for the rhythm disorder to be medically relevant. In one embodiment, D is equal to 20, the duration of the time series is of the order of 10 seconds and the sliding window is regenerated with each heartbeat. In another embodiment, a count of a time corresponding to the desired duration of the observation window is provided instead of counting the number of successive series.

In a variant of the method, the processor, in step SI 12, determines whether the classification received consists in an attachment of the time series Sj to the class Cl of normal subjects. If the answer is positive, the processor returns to step SI 10, otherwise goes to step SI 13.

As indicated above, the device DV3 is initially placed in the continuous standby mode of operation in which the processor has activated the channel CH1 and selected it by means of the multiplexer MUX, and the application program PAP switches it into the mode of alert operation when the PG5 decision program indicates that a rhythm disturbance has been detected in the SI signal. FIG. 21 shows, purely by way of nonlimiting example, an example of configuration of the device DV3 in the alert mode. This configuration includes an initial step S120 during which the processor P3:

- archive in the data memory M32 the previous signal SI which caused the changeover to the alert mode, as well as the corresponding time series Sj (Sl) and their classification by the classifier tool PG4,

- asks the user to place their finger on the E2 electrode,

- activates the CH2 channel and selects it via the MUX multiplexer, or activates the two CH1, CH2 channels and alternately selects them via the MUX multiplexer,

- analyzes signal S2 and optionally signal SI,

- if the channel CH1 is active, archives in the memory M32 the signal SI received, the corresponding time series Sj (Sl) and their classification by the classifier tool PG4,

- archive in the memory M32 the signal S2, the corresponding time series Sj (S2) and their classification by the classifier tool PG4.

During a step S121, the processor awaits confirmation of a detection, in the signal S2, of the rhythm disturbance which caused the switching to the alert mode. It checks during a step S122 that a time Ta elapsed since the changeover to the alert mode has not reached a threshold Tmax. If the threshold Tmax is reached without a rhythm disturbance having been found in the signal S2, or without the signal S2 having been received (if the user has not responded to the request which has been addressed to touch the electrode E2), the processor returns to the continuous standby mode of operation, to possibly switch back a few moments later into the alert mode if the rhythm disturbance is again detected in the signal SI.

If, on the contrary, the rhythm disturbance is confirmed by the analysis of the signal S2 before the time Tmax is reached, the processor goes to a step S122 during which it attempts to connect to an SRV server and / or to a workstation WS shown in FIG. 13, via the wireless communication interface CI3 and a computer or telephone network NTW. If the connection is established, the processor transmits an alert to the remote device and then transfers archived data to it. The user can be informed of the success of the data transfer, for example via an information display.

In a variant, the step of connection to an SRV server and / or to a workstation WS is initiated as soon as the switch to alert mode, and the aforementioned data are transferred to the remote device without waiting for the confirmation provided for in step S121. The DV3 device can then return to the continuous standby operating mode or continue to manage the alert operating mode in any way conceivable by those skilled in the art, for example by taking into account the number of times the rhythm disturbance has been detected, user behavior in the face of requests for acquisition of the electrocardiographic signal, etc.

Also, the user can be offered the possibility of asking the DV3 device to leave the alert mode until further notice, if contact has already been made with the medical profession, or to switch to a silent alert operating mode where the DV3 device, while remaining connected to the remote device and / or while continuing to record data from the SI signal, no longer requests the user to capture the electrocardiographic signal. In the variant of the device DV3 equipped with skin electrodes, the signal S2 can be captured and stored or transferred throughout the duration of the silent alert mode. The user can also be offered the possibility of initiating a cardiographic recording himself even if the device has not detected a rhythm disturbance by analysis of the PPG signal.

It will be clear to those skilled in the art that the devices DV1, DV2 and DV3 which have just been described are susceptible of various variants, embodiments and applications. In particular, although it has been indicated in relation to FIGS. 16 and 17 that the device DV3 detects a rhythm disturbance in the photoplethysmographic signal by characterization and classification of time series, the general functionalities of the device, in particular the prediction of a continuous standby operating mode where the photoplethysmographic signal is used for monitoring the heart rate, and an alert operating mode where the electrocardiographic signal is captured with or without the contribution of the user, are independent of the process implemented for detect a rhythm disturbance in each of these signals.

References cited:

Ref. 1: M.G. Tsipouras, D.I. Fotiadis, and D. Sideris, “An arrhythmia classification System based on the RR-interval signal,” Artificial Intelligence in Medicine, vol. 33, pp.

237-250,2005.

Ref. 2: Automatic Detection of Atrial Fibrillation Using R-R Interval Signal Xiuhua Ruan, Changchun Liu, Chengyu Liu, Xinpei Wang, Peng Li, School of Control Science and Engineering, Shandong University, Jinan, Shandong Province, P.R. China,

250061, 2011, 4th International Conference on Biomedical Engineering and Informatics (BMEI).

Ref. 3 Heart rate variability, Standards of measurement, physiological interpretation, and clinical use. Task Force of The European Society of Cardiology and

The North American

Society of Pacing and Electrophysiology. European Heart Journal (1996) 17, 354-381.

Annex 1 (an integral part of the description)

Table 1 (correlation between series and derived series)

RR dRR / dt d ² RR / d t ² d ³ RR / d t ³ d ⁴ RR / d t ⁴ d'RR / d t ⁵ d ^b RR / d t ⁶ d'RR / d t ⁷ d ⁸ RR / d t ⁸ d ⁹ RR / d t ⁹ d ^w RR / dt ¹⁰ RR 1 0.3333 1129 0.1594 1908 0.0683 035 0.0282 7863 0.0132 6783 0.0087 0103 0.0080 3672 0.0086 1886 0.0093 5475 0.0097 4555 dRR / d t 0.3333 1129 1 0.8527 1786 0.5948 8729 0.3715 9902 0.2183 708 0.1251 0525 0.0725 3127 0.0449 9524 0.0318 8493 0.0263 8423 d ² RR / dt ² 0.1594 1908 0.8527 1786 1 0.8982 4406 0.6885 514 0.4756 8752 0.3056 8512 0.1866 5303 0.1110 1735 0.0669 6255 0.0434 1496 d ³ RR / dt ³ 0.0683 035 0.5948 8729 0.8982 4406 1 0.9216 0191 0.7462 7833 0.5495 709 0.3760 1513 0.2436 5287 0.1530 9965 0.0962 4648 d ⁴ RR / dt ⁴ 0.0282 7863 0.3715 9902 0.6885 514 0.9216 0191 1 0.9363 0232 0.7854 2371 0.6046 3801 0.4358 6099 0.3003 5827 0.2020 6407 d'RR / dt ⁵ 0.0132 6783 0.2183 708 0.4756 8752 0.7462 7833 0.9363 0232 1 0.9460 3016 0.8137 7784 0.6503 4949 0.4927 9573 0.3603 9693 d ⁶ RR / dt ⁶ 0.0087 0103 0.1251 0525 0.3056 8512 0.5495 709 0.7854 2371 0.9460 3016 1 0.9534 7374 0.8385 4757 0.6942 5413 0.5502 2555 d'RR / dt ⁷ 0.0080 3672 0.0725 3127 0.1866 5303 0.3760 1513 0.6046 3801 0.8137 7784 0.9534 7374 1 0.9604 2702 0.8623 0748 0.7358 8922 d * RR / dt ⁸ 0.0086 1886 0.0449 9524 0.1110 1735 0.2436 5287 0.4358 6099 0.6503 4949 0.8385 4757 0.9604 2702 1 0.9667 526 0.8827 5198 d ⁹ RR / dt ⁹ 0.0093 5475 0.0318 8493 0.0669 6255 0.1530 9965 0.3003 5827 0.4927 9573 0.6942 5413 0.8623 0748 0.9667 526 1 0.9716 3667 d ^lu RR / dt ¹⁰ 0.0097 4555 0.0263 8423 0.0434 1496 0.0962 4648 0.2020 6407 0.3603 9693 0.5502 2555 0.7358 8922 0.8827 5198 0.9716 3667 1

Table 2

Number of primitive and / or derived descriptive variables Variables 1 1 2 3 4 5 6 7 8 9 10 meanÇRR) X X a (RR) X X X X X X fdRR \ mean —-— V dt) X X X X X X X X X (dRR \ X X X X X X X X fd ² RR \ mean, _n V ^dt2 J X X X X X X X fd ² RR \ ^a {d _t2 ) X X X X X X X X X X fd ³ RR \ mean, V ^dt3 J X X X X fd ³ RR \ ^a { _dt ³ ) X X X X X (d * RR \ ^mean {dt *; X X X fd * RR \ ^a { _d t *) X

APPENDIX 2 (an integral part of the description) Mathematical operators

Statistical operators Formulas Average II Median m | carctfJO = card (E ₂ ) Where: E _± = [XG u | X <m} and E ₂ = {X eu | X> m} Standard deviation £ II _N ίν-ιΣ ⁽ Τ- *) ² 7 = 1 Variance NOT J = 1 Coefficient of asymmetry or Skewness γ = E (Al Flattening coefficient or Kurtosis β = E C; T

Analysis operators spectral (Cf. Ref. 3) Formulas ULF (Ultra Low Frequency Power) or ultra low frequency power [0-3 mHz] r3.10 ~ ³ ULF = 1 psd _x (f) .df Where: psd _x is the power spectral density of the series of resampled X elements

37 VLF power (Very Low Frequency Power) or very low frequency power [3 mHz 40 mHz] /40.10 ^-3 VLF = I psd _x (f) .df J3.10- ³ Where: psd _RR is the power spectral density of the series of resampled X elements LF power (Low Frequency Power) or low frequency power [40 mHz - 150 mHz] ^ 150.10 ^-3 LF = 1 psd _x (f) .df ^ 40.10 ^-3 Where: psd _x is the power spectral density of the series of resampled elements X HF power (High Frequency Power) or high frequency power [150 mHz - 400 mHz] , -400.10 ^-3 HF = 1 psd _x (f) .df - / 150.10- ³ Where: psd _x is the power spectral density of the series of resampled X elements LF / HF ratio LF ^r ~ hf Total power P _tot = ULF + VLF + LF + HF Normalized LF power [40 mHz - 150 mHz] LF LF _norm - 100 p _tot _ _ULF _ _VLF Standardized HF power [150 mHz - 400 mHz] HF HF = 100 Ptot - ULF - VLF

Analysis operators phase space (Poincaré diagram) (Cf. Ref. 2) Formulas SD1 or dispersion of points along the minor axis of the ellipse 1 SD1 = —std (kX) V2 Where: std is the standard deviation

SD2 or dispersion of points along the long axis of the ellipse SD2 = J 2σ% —SD l ² SD1 / SD2 ratio SD1 Ratio = ——— SD2 Nl 1 v ^-1 VAI or Angle Index TZ Λ J - \ lû / HERE ' ¹ Nl / -,' ¹ Vector (Vector Angular Index) i = i N-l VLI = AT / V-iZ ^{Gî Ό} i = l VLI or Vector Length Index Or : N-l 1 τ—, L -Does it ¹ 'i = i RM S SD or Average Nl 1 x ^-1 quadratic of differences RMSSD = \ (V V A 2 w-ιζΛ ¹ successive elements X y 7 = 1 SDSDD or Standard Deviation of the absolute value of the 1 SDSDD = —std (] AJ [) \ 2 order differentiation 2 of Or : elements X 7 = | ΔΧ |

Analysis operators RP recurrence diagram (Plot Recurrence) Formulas Recurrence Rate N-m + l REC = —-- y V RP (j, k) (Vm + 1) ² 2_i ^{7 7} j, k = l Where:

39 Lmax or Maximum length of the diagonal lines of the RP recurrence diagram RP (j = ^d (. ^U P ^u k) ^ ^r 10, Otherwise 1 Div = - lmax Lmean or Average length diagonal lines of the RP recurrence diagram _lN j _ Î — imin ^mean ”Σίτ ^N i ^L ~ ^L min Determinism ^ Imax _lN Î — imin Shannon entropy of the length distribution of the diagonal lines of the matrix of the RP diagram lmax ShanEn = - n _t In n _t l ~ lmin Where: M,. Mu ^LL min

Analysis operators regularity of series Formulas Shannon's Entropy "= - / pWln (PW) ^ Or : p (x) is the probability density of the elements X

40 ApEn Entropy (Approximate Entropy) ApEn (m, r, N) = - 0 ^{m + 1} (r) Where: N-m + l to ^m (r) = —-- y InC ^ Cr) N - m + 1 £ -1 ^{7 7} 7 = 1 nbr tu _k \ d (u _h u ^ <r] C, ^m (r) ₌ <V 7 'yi d (uj, ii _k ) = max {\ Xj _{+ n} - A _{k + n} || n = 0 , ..., m - l] SampEn Entropy (Sample Entropy) SampEn (m, r, N) = Zn (C ^m (r) / C ^{m + 1} (r)) Where: N-m + l C ^m (r) = --- V C. ^m (r) N - m + 1 £ -i ^J j = l nbr ÎUiAd (ui, Uk) <r] - "I l 1 · -i _vk ^ j ^J N - m

Analysis operator chaotic properties and series fractals Formulas Correlation dimension log C ^m (r) D ₂ (m) = Hm lim -: - r-> o log r Where: N-m + l C ^m (r) = --- V C. ^m (r) j = l nbr ϊτιΑάίιΐι, ιΐϊ) <r] C, ^m (r) ₌ < ^{fel V 7} 7 V / t ⁷ N - m + 1 m d (u _Jt u _fc ) = ^ (u / Z) - u _k (Z)) ² J ^{i = 1}

With:

X: constituent element of the series considered, u = (X _x ,% 2>> X _N ),

N = nhre (Xj): length of vector u, hist ₁₀ (u): histogram of 10 sections of vector u,

Uj (fij> Xj + i> · · ·> Xy + mi)> j 1 <2, ..., N m -f 1, m: subdivision dimension of the vector u, fixed at 2, r: Euclidean distance tolerated between 2 points (X ,, X _{j + 1} ) and (Xj, Xj + i), i and j E [1,7V], E: expectation.

Annex 3 (an integral part of the description)

Primitive series:

Sj: RR interval

RRi = T (Ri) - T (Ru)

Sjo: variation Vi of the RR interval:

Vi = dRRj = RR _{i + 1} - RRj

Sji: absolute value | Vi | of the variation Vi of the RR interval:

| Vi | = IdRRjl

Derived series:

Sj ₂ : rate of change (Ai) of the RR interval (acceleration of the heart rate):

zdRR \ RRj + i - RRj RRj ^Ai \ dt / j RRi + i ¹ RRi + i

Sj3: absolute value (| Ai |) of the rate of change (Ai) of the RR interval (acceleration to 20 absolute value of the heart rate):

Sj4: rate of change (ACi) of the rate of change (Ai) of the RR interval (sudden heart rate):

RRi RRi + ₂ RRf + i

RR ³ + 1RRi ₊₂

Sjs: absolute value (| ACi |) of the rate of change (ACi) of the rate of change (Ai) of the RR interval (suddenly in absolute value of the heart rate):

| RRjRRj ₊ 2 ^- RR? + I |

RR ^{I 2} + 1RRj ₊₂ 'd ² RR', dt ² , | ACi | =

Sjô: rate of change (Gi) of the absolute value (| Ai |) of the rate of change (Ai) of the RR interval:

Sj7: absolute value (| Gi |) of the rate of change (Gi) of the absolute value (| Ai |) of the rate of change (Ai) of the RR interval:

| Gil =

I -i _ RRj + il _ I -i _ RRi I I RR1 + 2I I RRj + J

RRi + i

Claims (20)

Claims 1. Method for detecting at least one heart rhythm disorder of a human or animal subject from a time series of RR intervals representative of the heart rate of the subject extracted from a physiological signal (S, SI, S2 ), including the steps of: - characterize (S07, S07 (l), S07 (2)) the time series with at least one descriptive variable (Vb, Vc) whose value is calculated from the RR intervals of the time series, - using the value of the descriptive variable, classify the time series in a first class (Cl) of time series associated with subjects in sinus rhythm or in at least a second class (C2) of time series associated with subjects having heart rhythm disorder, a process characterized in that the value of the descriptive variable (Vb) is calculated from a derived series (Sj2, Sj3, Sj4, Sjs, Sjô, Sj ₇ ) whose constituent element is a discrete derivative of order 1 or greater than 1 of the RR interval of the time series. 2. Method according to claim 1, in which the constituent element of the derived series is chosen from the group comprising: - the rate of change (Ai) of the RR interval, i.e. the discrete time derivative of the RR interval, reflecting the acceleration of the heart rate, - the absolute value (| Ai |) of the rate of change of the RR interval, - the rate of change (ACi) of the rate of change (Ai) of the RR interval, either the discrete time derivative of the rate of change of the RR interval, or again the second discrete time derivative of the RR interval, reflecting heartbeat, - the absolute value (| ACi |) of the rate of change (ACi) of the rate of change (Ai) of the RR interval, - the rate of change (Gi) of the absolute value (| Ai |) of the rate of change of the RR interval, and - the absolute value (| Gi |) of the rate of change (Gi) of the absolute value (| Ai |) of the rate of change of the RR interval. 3. Method according to one of claims 1 and 2, wherein the step of characterizing the time series also comprises the calculation of the value of at least one descriptive variable (Vc) from a time series of which l the constituent element is chosen from the group comprising: - RR interval, - the variation (Vi) of the RR interval, and - the absolute value (| Vi |) of the variation of the RR interval. 4. Method according to one of claims 1 to 3, in which the value of the descriptive variable is calculated by means of a mathematical operator chosen from the group comprising: the mean value, the median, the standard deviation, the variance, the asymmetry coefficient or Skewness, the flattening coefficient or Kurtosis, the ultra low frequency power ULF, the very low frequency power VFF, the low frequency power FF, the high frequency power HF, the ratio FF / HF, the total power, the normalized FF power, the normalized HF power, SD1 or dispersion of the points along the minor axis of the ellipse of the Poincaré diagram, SD2 or dispersion of the points along the major axis of the ellipse of the diagram of Poincaré, the SD1 / SD2 ratio, VAI or the vector angle index, VFI or the vector length index, RMSSD or quadratic mean of the successive differences of the constituent elements of the series, SDSDD or standard deviation of the abs value olue of the order 2 differentiation of the constituent elements of the series, the recurrence rate of the recurrence diagram, Fmax or maximum length of the diagonal lines of the recurrence diagram, Fmean or average length of the diagonal lines of the recurrence diagram, determinism , the Shannon entropy of the length distribution of the diagonal lines of the recurrence diagram matrix, the Shannon entropy, the ApEn entropy, the SampEn entropy, the correlation dimension. 5. Method according to one of claims 1 to 4, wherein the step of classifying the time series is carried out by means of a classifier tool (PG4) with supervised learning, and comprising a step (S0900) of configuration of the classifier tool comprising the steps of: - supply (S0903) to the classifier tool a plurality of values (Vbj) of the descriptive variable (Vb) calculated from series derived from a first set (Nl) of time series from subjects in sinus rhythm, and - provide (S0906) to the classifier tool a plurality of values (Vbj) of the descriptive variable (Vb) calculated from series derived from a second set (N2) of time series from subjects with heart rhythm disorder . 6. Method according to one of claims 1 to 5, comprising a step (S05) of continuous or quasi-continuous production of a plurality of time series from the physiological signal (S, SI, S2), and the determination in real time of the class (Cl, C2) to which each time series belongs. 7. Method according to claim 6, in which the time series are produced (S05, S0508) according to a sliding time window, so that two successive time series can comprise common RR intervals. 8. Method according to one of claims 6 and 7, comprising a step (S0307) of detection of missing RR intervals and in which the production of the time series is interrupted (S0315, IFS) during the periods including missing RR intervals. 9. Method according to one of claims 6 to 8, in which the time series are extracted from a photoplethysmographic signal (SI). 10. Method according to one of claims 6 to 9, comprising a decision step (Sll) by which the heart rhythm disorder is considered to be detected when a determined number (D) greater than 1 of time series does not belong to the first class of time series. 11. Device (DV1, DV2, DV3) for detecting at least one heart rhythm disorder of a human or animal subject from a time series of RR intervals representative of the heart rate of the subject extracted from a signal physiological (S, SI, S2), the device being configured to: - characterize (S07, S07 (l), S07 (2)) the time series with at least one descriptive variable (Vb, Vc) whose value is calculated from the RR intervals of the time series, - using the value of the descriptive variable, classify the time series in a first class (Cl) of time series associated with subjects in sinus rhythm or in at least a second class (C2) of time series associated with subjects having heart rhythm disorder, device characterized in that it is configured to calculate the value of the descriptive variable (Vb) from a derived series (Sj ₂ , Sj3, Sj4, Sjs, Sjô, Sj?) of which l the constituent element is a derivative of order 1 or greater than 1 of the interval RR of the time series. 12. Device according to claim 11, configured to calculate the value of the descriptive variable (Vb) from a derived series whose constituent element is chosen from the group comprising: - the rate of change (Ai) of the RR interval, i.e. the discrete time derivative of the RR interval, reflecting the acceleration of the heart rate, - the absolute value (| Ai |) of the rate of change of the RR interval, - the rate of change (ACi) of the rate of change (Ai) of the RR interval, either the discrete time derivative of the rate of change of the RR interval, or again the second derivative of the RR interval, reflecting the to - heart rate cuts, - the absolute value (| ACi |) of the rate of change (ACi) of the rate of change (Ai) of the RR interval, - the rate of change (Gi) of the absolute value (| Ai |) of the rate of change of the RR interval, and 5 - the absolute value (| Gi |) of the rate of change (Gi) of the absolute value (| Ai |) of the rate of change of the RR interval. 13. Device according to one of claims 11 and 12, configured to also calculate the value of at least one descriptive variable (Vc) from a time series of which 10 the constituent element is chosen from the group comprising: - RR interval, - the variation (Vi) of the RR interval, and - the absolute value (| Vi |) of the variation of the RR interval. 14. Device according to one of claims 11 to 13, comprising a classifying tool (PG4) with supervised learning configured to classify the time series in the first or second classes of time series. 15. Device according to one of claims 11 to 14, configured to produce (S05) of 20 continuously or almost continuously a plurality of time series from the physiological signal (S, SI, S2), and classify each time series in real time. 16. Device according to claim 15, configured to produce (S05, S0508) the time series according to a sliding time window, so that two time series 25 successive can include common RR intervals. 17. Device according to one of claims 15 and 16, configured to detect (S0307) missing RR intervals and to interrupt (S0315, IFS) the production of RR interval time series during periods including missing RR intervals. 5 18. Device according to one of claims 15 to 17, configured to extract time series of RR intervals from a photoplethysmographic signal (SI). 19. Device according to one of claims 15 to 18, configured to execute a decision step (Sll) by which the heart rhythm disorder is considered to be detected 10 when a determined number (D) greater than 1 of successive time series does not belong to the first class of time series. 20. Device (DV3) according to one of claims 11 to 19, configured to activate a channel (CH2) for acquiring the subject's electrocardiogram when a rhythm disorder 15 cardiac is considered detected. 21. Device according to one of claims 11 to 20, configured to detect at least one atrial fibrillation. 22. Device (DV3) according to one of claims 11 to 21, configured to be fixed around the wrist. 1/9