A device for developing an indicator of the regularity of an individual's respiratory signal, intended for use in determining the individual's condition.

Abstract

The device comprises a sensor (1) capable of generating a respiratory signal (S) from an individual, and a respiratory signal processing module (2). Furthermore, the processing module (2) includes a determination element (3) for a regularity indicator (Id) of at least a portion (Sp) of the respiratory signal (S).

Description

1 Device for developing an indicator of the regularity of an individual's respiratory signal intended for use in determining the individual's state. Field of the Invention [1] The field of the invention relates to data concerning an individual, particularly in the field of self-measurement, and more specifically to a state of the individual such as a mental state, in particular a state of stress. Prior Art [2] Many devices today offer the ability to collect data relating to an individual such as the individual's heart rate, respiratory rate, or respiratory frequency. This data is generally then used in connected health applications, thus enabling, for example, the remote monitoring of elderly people, or the improvement of personal safety or well-being.

[003] To determine an individual's state of stress,Physiological parameters such as heart rate, skin conductance measurement, and electroencephalography measurements are used. [4] In addition to physiological measurements, behavioral indicators such as gestures, blinks, gaze direction, and facial expressions are widely used. However, these indicators require the use of cameras, which limits the range of possible applications to those where the user remains within the camera's field of view. [5] The document "Objective measures,"Sensors and Computational Techniques for Stress Recognition and Classification: A Survey" by Nandita Sharma et al., published in "Computer Methods and Programs in Biomedicine 108 (2012) 1287-1301," describes various ways to detect an individual's stress state. [6] This document specifically mentions the use of respiration, based on measurements of respiratory rate and volume, as an indicator. Respiratory rate classically corresponds to the number of respiratory cycles (inhalations and exhalations) per minute, and respiratory volume corresponds to the volume of air inhaled. [7] The indicators using a respiratory signal and described above, while functional in some cases, are not perfect. There is a need to find a new, more reliable indicator based on a respiratory signal. [8] An object of the present invention is the determination of a new indicator,developed from a respiratory signal, and more indicative and more relevant than the 10 prior art indicators based on the respiratory signal. [9] This goal is achieved by means of a device comprising a sensor capable of generating a respiratory signal from an individual, and a respiratory signal processing module. The processing module includes an element for determining a regularity indicator for at least a portion of the respiratory signal. 15

[0010] According to one embodiment, the processing module includes a respiratory signal cutting element configured to cut said respiratory signal into segments such that each segment corresponds to a portion of the respiratory signal, said regularity indicator determination element being configured to determine the regularity indicator for at least two segments,and 20 in particular for each segment.

[0011] The respiratory signal cutting element can be configured such that two successive segments do not overlap.

[0012] Advantageously, the regularity indicator determination element includes a regularity indicator processing function taking as input 25 said at least a part of the respiratory signal, and determining an autocorrelation signal of said at least a part of the respiratory signal, said processing function outputting the regularity indicator which then corresponds to the maximum value of said autocorrelation signal.

[0013] According to one embodiment, the processing module includes an element for determining an individual's state, in particular a mental state indicating a state of stress in the individual,said individual state determination element being configured to use said regularity indicator in order to determine said individual state.

[0014] According to one improvement, the device comprises a probe capable of measuring the individual's skin conductance, the individual state determination element 3047654 3 being configured to further use a measurement of the individual's skin conductance from said probe in order to determine said individual state.

[0015] According to another improvement, notably combinable with the one above, the device comprises a heartbeat measurement element, the individual state determination element being configured to further use a measurement of heart rate, or of heart rate variability,of the individual from the measuring device in order to determine said state of the individual.

[0016] The invention also relates to a method for developing an indicator intended to be used to determine a state of an individual, said method comprising the following steps: a step of providing a respiratory signal from the individual; a step of determining the indicator from said respiratory signal, said determined indicator being an indicator of the regularity of at least a part of said respiratory signal from the individual.

[0017] In particular, the step of determining the indicator may include determining an autocorrelation signal of said at least a part of the respiratory signal and determining a maximum of said determined autocorrelation signal.

[0018] According to one embodiment,The method may include a step of slicing the respiratory signal into a plurality of segments, each corresponding to a time window applied to said respiratory signal, and the step of determining the indicator 20 from the respiratory signal is implemented for different segments.

[0019] In particular, said time window is such that each segment contains data relating to several respiratory cycles of the individual.

[0020] For example, the duration of the time window is between 20 seconds and 1 hour, and in particular equal to 1 minute.

[0021] The invention also relates to a method of determining a state of an individual, in particular a state of stress; such a method includes at least one step of implementing the method for developing the indicator as described to determine a regularity indicator.and a step of using said regularity indicator to determine said state of the individual. 30

[0022] The step of using said regularity indicator may include a step of comparing said regularity indicator with a reference threshold.

[0023] For example, the step of using said regularity indicator includes injecting said regularity indicator into a transfer function that outputs said state of the individual. 3047654 4

[0024] In particular, the transfer function takes into account at least one of the parameters from the following signals to develop said state of the individual: - a measurement of the individual's skin conductance, - a measurement of the individual's heart rate, - a measurement of the individual's heart rate variability. Summary description of the figures

[0025] The invention will be better understood upon reading the following description, given solely by way of non-limiting example,and made with reference to drawings 10 on which: - figure 1 illustrates a representation of a respiratory signal given in mV as a function of time in seconds, - figure 2 illustrates respiratory volume values ​​given in mV (defined as the amplitude, i.e., the difference between the minimum and maximum of the signal on the segment of the respiratory signal) as a function of a segment number of a duration of one minute of the respiratory signal, the segments being successive and without overlap, - figure 3 illustrates respiratory rate values ​​given in number of respiratory cycles per second as a function of a segment number of a duration of one minute of the respiratory signal, the segments being successive and without overlap, - figure 4 illustrates a representation of a respiratory signal given in mV as a function of time in seconds centered on a moment of transition from a non-stressed state to a stressed state,25 - Figure 5 schematically illustrates a device according to an embodiment of the invention; Figure 6 illustrates values ​​of the regularity indicator determined according to an embodiment of the invention as a function of a segment number of the respiratory signal; 30 - Figure 7 illustrates a method for developing the indicator according to an embodiment of the invention; Figure 8 illustrates a method for determining the state of an individual using the development method of Figure 7. 3047654 5

[0026] In these figures, the same reference numerals are used to designate the same elements. Description of embodiments 5

[0027] The device and methods described below differ from the prior art, in particular, in that the developed indicator allows for better discrimination of an individual's state. In particular,This state of the individual is a mental state, such as a state of stress. A state of stress can be chosen, for example, from a stressed state and a non-stressed state. In the case of a stressed state, this can be established on a scale quantifying the importance of the stress.

[0028] In the present description, a respiratory signal corresponds to a set of values ​​measured by a sensor capable of measuring said respiratory signal, each value being associated with a time instant at which it was measured. For example, the respiratory signal can correspond to the evolution of a voltage in mV, as shown in Figure 1, where the high peaks correspond to a maximum of the thoracic volume during inspiration and the low peaks correspond to a minimum of the thoracic volume during expiration.

[0029] In the present description, a segment of the respiratory signal corresponds to a part of the respiratory signal. In particular,This part of the respiratory signal is associated with a time window during which the respiratory signal is to be studied in order to determine the indicator.

[0030] The sensor capable of measuring the respiratory signal can be configured to generate the voltage value mentioned above as a function of the movements of the individual's thorax or abdomen over time. Such a sensor can be integrated into a garment, such as a T-shirt, which, when worn close to the body, can be representative of the movements of the individual's rib cage. For example, a piezoresistor integrated into the T-shirt can measure the stretching of tissues due to respiratory movements.

[0031] In this description, an individual can be a human.

[0032] To validate the indicator developed according to the invention, which will be described below,The following protocol was implemented: an individual whose stress state (specifically chosen from a non-stressed state and a stressed state) over time was known was monitored. The individual's non-stressed state corresponds to a resting phase of the individual, while the stressed state was induced by placing the individual behind the wheel of a vehicle in situations that provoked stress.

[0033] According to this protocol, indicators according to the prior art were determined and are shown in Figures 2 and 3.

[0034] Figure 2 shows the values ​​of the individual's respiratory volume as a function of a tested segment number of the respiratory signal in Figure 1 measured on the individual. The vertical line li represents the change in the individual's state (for reasons of consistency,This line li is also shown in Figure 1). The points to the left of the vertical line li correspond to the indicators representing respiratory volume determined when the individual is not stressed, and the points to the right of the vertical line li correspond to the indicators representing respiratory volume determined when the individual is stressed. It can be deduced from Figure 2 that it is difficult to obtain good discrimination based solely on respiratory volume as an indicator.

[0035] Figure 3 shows the respiratory rate values ​​as a function of a tested segment number of the respiratory signal from Figure 1.The vertical line li represents the change in the individual's state. The points located to the left of the vertical line li correspond to the indicators determined when the individual is not stressed, and the points located to the right of the vertical line li correspond to the indicators determined when the individual is stressed. It can be deduced from this Figure 3 that the faster the respiratory rate, the greater the chance of correctly determining a state of stress from a simple threshold. However, it is also noted that many points established when the individual was stressed fall within a range of respiratory rates consistent with the respiratory rate when the individual is not stressed. 25

[0036] The indicator used in the present invention differs from the indicators above in that it represents, more specifically, the regularity of at least a part of the respiratory signal. By "regularity of at least a part of a signal,"We understand that we can observe a signal that repeats itself in terms of amplitude, frequency, and shape. Thus, the regularity indicator of at least part of a respiratory signal 30 allows us to determine whether this signal is regular or varies over time. Changes in respiratory signal regularity as a function of stress level can, for example, be better understood from Figure 4, which represents a respiratory signal in mV as a function of time in seconds, centered on a change in the individual's state. The vertical line li represents the transition from the unstressed state (to the left of line 3047654 7 li) to the stressed state (to the right of line li). In this Figure 4, the high peaks are represented by P1 and the low peaks by P2. As we will see later,Such an indicator allows for better discrimination of an individual's state because the more irregular the breathing, the more likely the individual is to be in a stressed state, and conversely, the more regular the breathing, the more likely the individual is to be in a non-stressed state.

[0037] As illustrated in Figure 5, a device according to an embodiment of the invention may include a sensor 1 capable of generating a respiratory signal S from an individual. This sensor 1 may be an accelerometer, piezoresistor, radio signal sensor representing body movements related to breathing, etc., and capable of generating a respiratory signal of the type shown in Figure 1. Furthermore, the device may include a respiratory signal processing module 2. This processing module 2 may be connected to the sensor 1,By "connected," we mean that all or part of the data from sensor 1 can be transmitted to the processing module 2. The processing module 15 2 then includes the hardware, and where applicable software, means to process at least a part of the respiratory signal from sensor 1 in order to develop the regularity indicator of said at least a corresponding part of the respiratory signal.

[0038] It is understood that the processing module 2 includes a determination element 3 of the regularity indicator Id of said at least a part Sp of the respiratory signal S.

[0039] More specifically, the determination element 3 of the regularity indicator includes a function for developing said regularity indicator, taking as input said at least a part Sp of the respiratory signal, and determining an autocorrelation signal of said at least a part Sp of the respiratory signal. Finally,The 25th function for generating the regularity indicator outputs the regularity indicator Id, which then corresponds to the maximum value of said autocorrelation signal.

[0040] The determination of an autocorrelation signal from a base signal (here the respiratory signal, and more specifically the respiratory signal portion) will not be described in detail as it is well known to those skilled in the art. For example, the autocorrelation signal C(m) can be given by the following equation: C(m) = ERR(n+m)-E(R))*(R(n)-E(R)Vstd2(R) Where E represents the expected value, std the standard deviation, and R corresponds to the respiratory signal.and with n and m the sample values ​​of the signal varying over the relevant part of the respiratory signal. The calculation of the maximum value of C(m) is within the scope of those skilled in the art and will not be detailed here.

[0041] The functions of the determination element 3 of the regularity indicator can be implemented by using a processor of the device capable of performing the associated calculations and executing corresponding computer code.

[0042] The processing module 2 may further include a determination element 4 of an individual's state, in particular a mental state indicating a state of stress in the individual. Said determination element 4 of the individual's state is configured to use said regularity indicator Id in order to determine said individual's state. In other words,The individual's state determination element 4 can be linked to the regularity indicator Id's determination element 3. By "determination element 4 linked to determination element 3," we mean that the latter two are configured such that indicator Id can be retrieved by the individual's state determination element 4. In particular, this individual's state determination element 4 includes a transfer function that takes the regularity indicator Id as input and outputs the individual's state. This transfer function can, for example, compare the regularity indicator Id to a reference threshold that allows the state to be classified.particularly between a stressed mental state and a non-stressed mental state.

[0043] The functions of the individual state determination element 4 can be implemented by using a device processor capable of performing the associated calculations and executing corresponding computer code.

[0044] In the context of health applications, particularly monitoring vital signs, it can be advantageous to track the evolution of the regularity indicator Id over time for different parts of the respiratory signal S.

[0045] It is with this monitoring in mind that the processing module 2 can include a respiratory signal segmentation element 5 configured to segment said respiratory signal into segments such that each segment corresponds to a part of the respiratory signal. The regularity indicator determination element 3 is then configured to determine the regularity indicator for at least two segments.and in particular for each segment. As a result, it is then possible to monitor the regularity indicator over time and therefore monitor the individual's condition over time. It is then understood that the respiratory signal segmentation element 5 can be connected, on the one hand, to sensor 1 to retrieve the respiratory signal S, and connected, on the other hand, to the regularity indicator determination element 3 to provide the latter with different segments, each corresponding to a part Sp of the respiratory signal for which a regularity indicator is to be determined.

[0046] The respiratory signal segmentation element 5 can be configured such that two successive segments do not overlap, or that there is partial overlap between successive segments. In the event of non-overlap between segments,This allows for the temporal segmentation of the moments when it is desired to establish the individual's state. In the event of overlapping segments, this allows for a more precise determination of the individual's state over time. The functions of the indicator's cutting element 5 can be implemented by using a processor 10 of the device capable of performing the associated calculations and executing corresponding computer code.

[0047] Figure 6 illustrates the evolution of the regularity indicator determined according to the present invention as a function of the segment number, using the same protocol and the same respiratory signal data from Figure 1 as for Figures 2 and 3. The respiratory signal is divided here into successive, non-overlapping segments, each lasting one minute.and established from the respiratory signal illustrated in Figure 1. The vertical line li shown in Figure 6 illustrates the moment when the individual's state changed from a normal (non-stressed) state to a stressed state. The points to the left of this vertical line li are points associated with a non-stressed state of the individual, and the points 20 to the right of this vertical line li are points associated with a stressed state of the individual. It can be seen here that it is possible, from all these regularity indicator values, to identify a reference threshold that allows for a more reliable separation of the stressed state from the non-stressed state than in the context of Figures 2 and 3. This reference threshold in Figure 6 is illustrated, for example, by the horizontal line 25.

[0048] For example, the discrimination of the individual's state can be done by simply comparing the determined regularity indicator with a reference threshold,for example, established in light of the results in Figure 6. This discrimination can then use the transfer function mentioned previously, that is, a function 30 capable of translating the determined indicator Id into a state of the individual and implemented by the element determining the individual's state 4. An example of such a transfer function might be that the individual is considered stressed if the determined indicator is below the reference threshold or considered unstressed if the determined indicator is above the reference threshold. 3047654 10

[0049] The value of the reference threshold can depend on the amplitude of the respiratory signal, the individual, and the intensity of stress considered to imply that the person is detected as stressed. Thus, the value of the reference threshold can be adjusted by learning,for example, from a database 5 representative of the targeted mental states established on one or more reference individuals. Alternatively, or in combination, the device can be used by the individual wearing it during a calibration phase in such a way as to enrich a device database that can subsequently be used to determine said reference threshold. This calibration phase may include a rest period corresponding to a non-stressed state 10. It may also include a stress period corresponding to a stressed state. The values ​​of the regularity indicator can be calculated for each of the time segments of the rest period and the stress period. The values ​​thus collected will make it possible to determine the threshold by choosing, for example, the one that minimizes the number of segments misclassified by the comparison rule to the threshold 15. For a calibration phase containing only a rest period,A threshold slightly lower than the value of the lowest regularity indicator calculated over the rest period can be taken, such that all segments of the rest period are classified as "non-stressed" by comparison to the threshold.

[0050] The transfer function described above is said to have discrete values, that is to say, it only differentiates a stressed state from a non-stressed state. However, this is by no means limiting, as the transfer function can be more elaborate to, for example, allow for the continuous evaluation of the individual's state on a scale, notably from 0 to 10, where 0 indicates the individual is not stressed and 10 indicates the individual is in a highly stressed state. In the case where the transfer function is more elaborate, it can be determined by learning from a database. For example,The transfer function can be a linear or polynomial function of the individual's characteristics, a neural network, a Bayes classifier, a support vector machine (SVM), etc.

[0051] For a given individual, when calibration allows for the determination of a personalized reference threshold, the success rate of classifying the individual's mental state can exceed 93%. Furthermore, based on calibration of the reference threshold according to a given population not including the individual wearing the device, tests have shown that the success rate of classifying the individual's state can exceed 88%.

[0052] The success rates mentioned above are higher than those using the indicators in Figures 2 and 3. To verify this,It is possible to implement a fifth methodology using classification results by employing a Bayes classifier and a selection algorithm, specifically a "forward selection" type algorithm, for indicators (chosen here, in particular, from respiratory volume as in Figure 2, respiratory rate as in Figure 3, and the indicator according to the invention, i.e., representative of regularity as in Figure 6). The indicators are then added one by one to the transfer function based on the classification results they provide. The classes used by the classifier can simply be "stress" and "non-stress," and the database used to implement the classification can correspond to a set of measurements taken on several vehicle drivers. Using this methodology,The indicator according to 15 of the invention appears first because it is more relevant than those of the prior art.

[0053] To increase the reliability of determining the individual's condition, the regularity indicator of the respiratory signal portion can be combined with other measurements, particularly physiological ones, which, when fed into the transfer function described above, can further improve the success rate of discriminating the individual's condition. 20 Preferably, these other measurements can be a measurement of the individual's skin conductance taken during a time period coinciding with the corresponding respiratory signal portion, or a measurement of heartbeats, for example, a heart rhythm, taken during a time period coinciding with the corresponding respiratory signal portion. 25

[0054] In other words,The device may include (Figure 5) a probe 6 capable of measuring the skin conductance Cskin of the individual. The individual state determination element 4 is then configured to also use the individual's skin conductance measurement from said probe 6 in order to determine said individual state. For this purpose, probe 6 can be connected to the individual state determination element 4. In this case, for the consistency of the individual state determination, the individual state determination element 4 is preferably configured to provide an individual state that takes into account the conductance measurement and the regularity indicator associated with a portion of the respiratory signal having a time window including a time range (preferably, the time range is of the same duration as the time window) in which said conductance measurement is performed. As with respiration,Characteristics such as the average signal, the number of peaks, their amplitude, etc., are extracted.

[0055] Alternatively, or in combination with the individual's skin conductance measurement 5, the device may include a heartbeat measurement component 7. The individual state determination element 4 is configured to further use a heart rate (HR) measurement, or heart rate variability (HRV) measurement, of the individual from the measurement component 7 in order to determine said individual state. For this purpose, the measurement component 7 may be connected to the individual state determination element 10. In this case, for consistency in determining the individual state, the individual state determination element 4 is configured to provide an individual state that takes into account the heart rate measurement, or heart rate variability measurement.and the regularity indicator associated with a portion of the respiratory signal having a time window including a time range during which the heart rate measurement, or the heart rate variability measurement, is performed. This time range may be of the same duration as the time window.

[0056] It is understood from all that has been said above that the described device makes it possible to develop, for an individual, a regularity indicator of at least a portion of the individual's respiratory signal, and in particular to use such a regularity indicator to determine the individual's condition. The device may include means for being worn by an individual; such means may be a T-shirt.

[0057] The invention also relates to a method for developing an indicator intended to be used to determine the individual's condition. In particular,Such a method comprises, as illustrated in Figure 7, the following steps: a step E1 of supplying a respiratory signal from the individual; and a step E2 of determining the indicator from said respiratory signal, said determined indicator being a regularity indicator of at least a part of said respiratory signal from the individual.

[0058] The step E1 of supplying the respiratory signal can be implemented by equipping the individual with a suitable sensor, and in particular the device described above.

[0059] The step E2 of determining the indicator can comprise a determination E2-1 of an autocorrelation signal of said at least a part of the respiratory signal and the determination E2-2 of a maximum of the determined autocorrelation signal. This maximum then corresponds to the desired regularity indicator (i.e., the one to be determined). In the context of using the device as described above,The determination of the autocorrelation signal and its maximum can be implemented by the determination element 3.

[0060] In relation to the method used to implement monitoring, the processing method 5 may include a step E3 (represented by dashed lines as it is optional) of cutting the respiratory signal into a plurality of segments, each corresponding to a time window applied to said respiratory signal. In this case, the step of determining the indicator E2 from the respiratory signal can be implemented for different segments, and in particular for each segment. 10

[0061] The time window, whether within the processing method or the duration of each segment of the respiratory signal within the device,is such that each segment contains data relating to several respiratory cycles of the individual. The use of several respiratory cycles makes it possible, in particular, to determine the regularity of the respiratory signal. A respiratory cycle can include an inspiration and an expiration. Furthermore, this window or duration is such that it is small enough to allow discrimination of the individual's state with sufficient temporal resolution to avoid mixing two very different states. For this reason, a temporal window or segment duration of between 20 seconds and 1 hour, and specifically equal to 1 minute, is preferably chosen. In adults, the respiratory rate corresponds to the number of respiratory cycles per minute and is generally between 12 and 20 cycles. In Figure 3,It is represented by a number of cycles per second. Such a number of cycles is particularly suitable for verifying the regularity of breathing while not being so long as to include several very different mental states on the same respiratory segment.

[0062] It will have been understood from what has been said previously that the invention also relates to a method for determining a state—in particular a mental state—of an individual, notably a state of stress. As illustrated in Figures 7 and 8, such a method comprises at least one implementation step E100 of the method for developing the indicator as described to determine a regularity indicator. Of course, this method for determining the state includes a use step E101 of said regularity indicator to determine said state of the individual.

[0063] In particular, the use step E101 includes a comparison step E101-1 of said regularity indicator with a reference threshold,or more generally 3047654 14 the injection of the determined indicator into the transfer function giving the individual's state as output.

[0064] Furthermore, as mentioned previously, the method for determining the individual's state can also take into account other data relating to the individual 5 such as a measurement of the individual's skin conductance, and / or a measurement of the individual's heart rate, and / or a measurement of heart rate variability, to determine said state. This or these data then complement each other to improve the determination of the individual's state, and have notably been measured during the time window corresponding to the part of the respiratory signal associated 10 with said regularity indicator. In other words,The transfer function may take into account at least one (and in particular several or all) of the following parameters from the signals to determine the individual's state: a measurement of the individual's skin conductance; a measurement of the individual's heart rate; a measurement of the individual's heart rate variability. Here, the transfer function may, in particular, be of the type described above.

[0065] It should be noted that the individual's determined state is simply a parameter used to observe the individual's condition. In other words, the process of determining the individual's state is not a diagnosis with a curative purpose in the strict sense, representing a phase of deductive decision-making in human medicine. Indeed, the individual's determined state can be used to trigger simple alerts, for example, for an isolated elderly person.In the event of detection of a prolonged state of stress, an alert could be triggered to emergency services so that they can intervene to see what is happening. Another example is a device that will allow the user to better understand themselves, and in particular to better understand their stress reactions to different situations. The device could display a graph of the evolution of their stress level throughout the day, or alert them when a high stress level lasts too long.

[0066] In particular, the method for developing the indicator can of course use the device as described above, from which it follows in particular that: - the respiratory signal provided comes from sensor 1, 30 - the regularity indicator comes from the determination element 3 of the processing module, - the individual's state is provided by the determination element 4 of the individual's state.

Claims

Claims 1. Device comprising a sensor (1) capable of generating a respiratory signal (S) of an individual, and a processing module (2) for processing the respiratory signal, the processing module (2) comprising an element (3) for determining a regularity indicator (Id) of at least one part (Sp) of the respiratory signal (S), characterized in that the element (3) for determining the regularity indicator (Id) comprises a function for developing said regularity indicator taking as input said at least one part of the respiratory signal, and determining an autocorrelation signal of said at least one part of the respiratory signal, said development function giving as output the regularity indicator (Id) which then corresponds to the maximum value of said autocorrelation signal.

2. Device according to claim 1, characterized in that the processing module (2) comprises a cutting element (5) of the respiratory signal configured so as to cut said respiratory signal into sections such that each section corresponds to a part of the respiratory signal, said determination element (3) of the regularity indicator being configured so as to determine the regularity indicator for at least two sections, and in particular for each section.

3. Device according to the preceding claim, characterized in that the cutting element (5) of the respiratory signal is configured such that two successive sections do not overlap.

4. Device according to any one of the preceding claims, characterized in that the processing module (2) comprises an element (4) for determining a state of the individual, in particular a mental state indicating a state of stress of the individual, said element (4) for determining the state of the individual being configured to use said regularity indicator (Id) in order to determine said state of the individual.

5. Device according to the preceding claim, characterized in that it comprises a probe (6) capable of measuring the conductance of the skin of the individual, the element for determining (4) the state of the individual being configured to further use a measurement of the conductance of the skin of the individual from said probe (6) in order to determine said state of the individual.

6. Device according to claim 4 or 5, characterized in that it comprises a member (7) for measuring heartbeats, the element (4) for determining the state of the individual being configured to further use a measurement of the heart rate (HR), or of a variability of the heart rate, of the individual from the measuring member (7) in order to determine said state of the individual.

7. A method of developing an indicator for use in determining a state of an individual, said method comprising the following steps: "a step of providing (And) a respiratory signal from the individual, ® a step of determining (E2) the indicator from said respiratory signal, said determined indicator being an indicator of regularity of at least part of said respiratory signal of the individual, characterized in that the step of determining (E2) the indicator comprises a determination (E2-1) of an autocorrelation signal of said at least one part of the respiratory signal and the determination (E2-2) of a maximum of said determined autocorrelation signal.

8. Method according to claim 7, characterized in that it comprises a step of cutting (E3) the respiratory signal into a plurality of sections each corresponding to a time window applied to said respiratory signal, and in that the step of determining (E2) the indicator from the respiratory signal is implemented for different sections.

9. Method according to the preceding claim, characterized in that said time window is such that each section includes data relating to several respiratory cycles of the individual.

10. Method according to one of claims 8 or 9, characterized in that the duration of the time window is between 20 seconds and 1 hour, and in particular equal to 1 minute.

11. Method for determining a state of an individual, in particular a state of stress, characterized in that it comprises at least one step (E100) of implementing the method for developing the indicator according to one of claims 7 to 10 to determine a regularity indicator, and in that it comprises a step of using (E101) said regularity indicator to determine said state of the individual.

12. Method according to claim 11, characterized in that the step of using (E101) said regularity indicator comprises a step of comparing (E101-1) said regularity indicator with a reference threshold.

13. Method according to claim 11 or 12, characterized in that the step of using (E101) said regularity indicator comprises the injection of said regularity indicator into a transfer function giving as output said state of the individual.

14. Method according to the preceding claim, characterized in that the transfer function takes into account at least one of the parameters from the following signals to develop said state of the individual: ® a measurement of the individual's skin conductance, ® a measurement of the individual's heart rate, "a measure of an individual's heart rate variability.

Images