Method for identifying an emotion in a subject from a saliva sample

Abstract

The invention relates to a method for identifying an emotion or a group of emotions felt by a subject after olfactory stimulation, from saliva samples of said subject implementing salivary biomarkers signature. In a second aspect, the invention also relates to a method for evaluating the capacity of an olfactory stimulus to provoke in a subject a desired emotion or group of emotions. The invention finally relates to the use of kit in the context of the present invention.

Description

METHOD FOR IDENTIFYING AN EMOTION IN A SUBJECT FROM A SALIVA SAMPLEThe invention relates to a method for identifying an emotion or a group of emotions felt by a subject after olfactory stimulation, from saliva samples of said subject implementing salivary biomarkers signature. In a second aspect, the invention also relates to a method for evaluating the capacity of an olfactory stimulus to provoke in a subject a desired emotion or group of emotions. The invention finally relates to the use of kit in the context of the present invention. Recent research has demonstrated that emotional states (positive and negative) can be discriminated by physiological responses. This physiological regulation is exerted by the action of biological molecules such as enzymes and hormones present in conventional biological fluids but also in saliva. The levels of these molecules represent the activity of the autonomic nervous system (SN), including the sympathetic (SNS) and parasympathetic (PNS) nervous systems.Among these biological molecules or biomarkers (BMs) which are present in the saliva, and which could be used for discriminating such emotional states, some of the known biomarkers associated to emotions which can be used to identify and predict emotions triggered by a myriad of stimuli can be listed (see Figure 1) :Phenylalanine, Tyrosine, Tryptophan, L-dopa, Dopamine, Noradrenaline, Adrenaline, Acetylcholine, Cortisol, Oxytocin, 5-HT, Melatonin, Substance P, Chromogranin A, DHEA, DHEA sulfate, Pre-ghrelin, Alpha amylase, BDNF, NPY and Peptide YY, Prostaglandins, Endorphins, Immunoglobulin A.Currently, even if some approaches have been published, in particular from blood samples, there are no reliable or effective biological tests, preferably inexpensive and easy to implement, to provide physiological data on the emotional state of an individual.Without limiting to these particular molecules present in saliva, the inventors have identified 4 biomarkers (BMs) among a large number of biomarkers present in saliva, the combination of which alone makes it possible to provide such data. The salivary biomarkers identified by the inventors are as follows:- Alpha-Amylase (ɑ-amylase; AA)- Cortisol- DeHydroEpiAndrosterone (DHEA) and its derivatives known to be present in the human body (such as its sulfated ester (SDHEA for DHEA sulfate)); and- Oxytocin (OXT).Nevertheless, it has been clearly demonstrated by the inventors that the circadian rhythm, physiological state od the tested person, stress, food intake, etc.., is trongly impact the salivary levels of the distribution of the basal biomarkers studied rendering difficult or even impossible reliable or effective biological tests for measuring the effect of olfactory stimulus on said Salivary biochemical markers collecting after stimulation (see for examples on figures Figures 6 and 7 the strong variation in the middle of the day could which induce a bias in the dosage of this MB during the olfactory stimulation study). Surprisingly, by calculating the ratio of the quantity and / or concentration of each of said BMs present in saliva samples before and after an olfactory stimulus, the inventors have demonstrated that it is possible, from saliva samples of a subject, to establish specific biological analysis signatures (named also herein specific biomarkers-based pattern or profiles or algorithms) for differentiating / identifying the emotion or group of emotions felt by the subject and provoked by said stimulus.Subject to the fact that these BM ratios have to be calculated, the inventors have demonstrated the possibility of identifying profiles of people responding more or less strongly biologically to an olfactory solicitation.According to the present invention, and in another aspect, it has been also demonstrated by the inventors that panelists or potential customers for particular olfactory stimuli such as perfumes / fragrances can be selected on the basis of measurement of these 4 BMs ratios and their specific combination from saliva samples. In a first step, the inventors have confirmed in particular the presence of these 4 BMs in saliva and verified on a first cohort of subjects the quality of the measurement of said selected 4 BMs in saliva, then in a second step the search for the specific biological signature by different methodological approaches, particularly by using other cohort of subjects, and based on standard statiscal methods wellknown by the silled person in the art .  In a first aspect, the present invention is directed to a method for measuring the quantity and / or the concentration of salivary biomarkers (BMs) in a subject stimulated by an olfactory stimulus, wherein said salivary BMs are selected from the group consisting of the following BMs:- Phenylalanine, Tyrosine, Tryptophan, L-dopa, Dopamine, Noradrenaline, Adrenaline, Acetylcholine, Cortisol, Oxytocin, 5-HT, Melatonin, Substance P, Chromogranin A, DHEA, DHEA sulfate, Pre-ghrelin, Alpha amylase, BDNF, NPY and Peptide YY, Prostaglandins, Endorphins and Immunoglobulin, and wherein said method comprises the steps of:1) collecting a first saliva sample (named S1) of said subject before said stimulation;2) stimulating the subject with said olfactory stimulus;3) collecting at least a second saliva sample (named S2);4) measuring the quantity and / or concentration of each of said BMs present in the S1 and S2 saliva samples;5) calculating the BM ratio’s (named S2 / S1) for each of said selected BMs. In a preferred embodiment of the method of the present inventio, said selected BMs are previously reliably measured on a statistical significant panel of saliva samples obtained with different and distinct fragrances or flagrances combinations to check whether no or weak correlations are observed between said selected BMs. This can be made for example but non limited to by the covariance analysis or by statistical methods wellknown by the person skilled in the art (see legends Figures 10A-10B and Example 8, figure 17) In a preferred embodiment of the method of the present invention, the quantity and / or concentration of each of the following BMs at step 4) are measured: alpha-amylase, cortisol, dehydroepiandrosterone (DHEA) and oxytocin. More preferred are methods of the present invention wherein the salivary sample S1 and S2 are collected at any time of the day, and preferably on morning, more preferably between 9:00 a.m. and 12:00 p.m..It is alsol preferred that the time interval (elapsed time) between S1 and S2 saliva samples collection is < 60 mn, preferably ≤ 30 mn or ≤ 20 mn.See for example Figure 9 where T1 = 5 mn before the stimulation and T2 = 5 mn after the stimulation. In a particular embodiment of the method of the present invention, and depending on the emotion felt after stimulation by the subject, a third S3 saliva sample is collected after S2 and at any time of the day, preferably on morning between 9:00 a.m. and 12:00 p.m., and wherein the the BM ratio’s (named S3 / S1) for each of said selected BMs is also calculated.See for example Figure 9 where T1 = 5 mn before the stimulation, T2 = 5 mn after the stimulation. And T3 = 20 mn after the stimulation. In another particular embodiment of the method according to the present invention, the saliva samples collected in step 1 and 3 before and / or after the stimulation are used freshly collected or stored at different temperatures, for measuring the quantity and / or concentration of each of said BMs present in said saliva samples, for example at room temperature (RT), +4 °C or -20 °C.(see Figure 3 demonstrated the possible storage og the saliva samples for a period of time of at least 7 days) In another preferred embodiment of the method according to the present invention the ratio value S2 / S1 calculated in step 4) for each of said BMs is a continue or a discretized variable, preferably discretized.By the wording “discretized”, it means that variable is a variable that takes on distinct, countable values. (see for example figure 31, where BMs ratios were discretized into three categories (i.e.<0.9 = decrease; ≥ 0.9 ≤1.1 = stable, and > 1.1 = increase).A continuous variable is a variable that takes on any value within a range, and the number of possible values within that range is infinite. In a second aspect, the present invention is directed to a methodfor determining the salivary BM signature associated to a given subject and to a given olfactory stimulus, from saliva samples of said subject, wherein said method comprises the steps of:A) measuring and calculating the BM’ ratio’s S2 / S1 for each of said selected BMs by a method according to one of claims 1 to 4 of the present invention, said step A) resulting to the determination of the profile for these 4 BMs associated to said stimulus and said subject; andB) from the results obtained in step A), determining the salivary BM signature associated to said stimulus and said subject In a preferred embodiment, in step B), the emotional response felt by the subject is discretized for a specific emotion or fragrance, preferably the participants’ responses for each emotion or fragrance are discretized according for example to the terms but non limited to, “True” and “False” or equivalent terms (i.e . positive versus negative ; 1 (positive) versus 0 (Negative).For example but non limited to, for emotional responses said responses is “discretized” into two categories such as:- negative responses, corresponding to the absence / response of not feeling any emotion towards a fragrance, and- positive responses, grouping together responses from strong to moderate feeling towards the fragrance.(see Example 7, panel study). In the present description, the wording ”reference signature”, ”reference salivary signature” ,“reference BM patterns”, “reference salivary BM patterns” , “ reference profiles”, and reference salivary profiles” have the same meaning and can be used interchangeably. In a third aspect, the present invention is directed to a methodfor identifying / determining a reference signature associated to an emotion / emotional response and / or to a group of emotions and / or to a fragrance and / or to a group of fragrances felt by a subject after an olfactory stimulation of said subject, said method comprising the steps of :C) measuring and calculating the BM ratio’s S2 / S1 for each of said selected BMs by the method according the present invention;D) repeating the step C) for all the subjects / individuals of a statistical significant cohort / panel with the same stimulus and identifying / determining by standard statistical methods the reference signature associated to said emotion and / or group of emotions and / or fragrance and / or group of fragrances. In a preferred embodiment, step D) is conducted on distinct groups resulting from the partition of said cohort / panel using standard partitioning statistical methods and allowing the identification or the determination of reference signature for each of these distinct groups, preferably k-means, hk-means (hierarchical k-means) statistical classification methods. These standard statistical methods are wellknown by the skilled person in the art (see but no limited to, the statistical methods cited and used in the concerned examples). In a preferred embodiment, in step B), said standard statistical methods allowing said partitioning are selected from one or a combination of methods selected from Principal component analysis (PCA), k-means, hk-means (hierarchical k-means), CART methods (“Classification and Regression Trees” methods). Are also preferred, the methods according to the present invention, wherein said individual measurement of BM ratio’s following said olfactory stimulus is carried out simultaneously with questionnaires on emotions felt by the subject. For the general concept of the present invention and in a preferred embodiment, among the emotions / emotional responses and / or groups of emotions and / or a fragrances and / or group of fragrances felt by a subject after an olfactory stimulation of said subject, and which is desired to identify / determine, associate to reference salivary signature :salivary BM pattern / reference profiles, the following emotions or fragrances can be cited, but not limited to : addict, elegant, confident, happy, energized, unique, sensual, comforted, relaxed, or combination thereof. In a preferred embodiment, the method of the present invention comprises an implicit test performed on a statistical significative cohort of subjects that scent the same fragrance or group of fragrances used to develop said biomarker-based emotional profiles. In a preferred embodiment, said step B) of the method of the present invention comprises a step wherein the cohort / panel participants are partitioned or clusterised in distinct group and nested together based on salivary biomarker’s ratio profiles or signature using standard statistical analysis method wellknown by the skilled person, such as, but non limited to, k-means and / or hierarchical k-means (hk-means) methods.See legends of Figures 20A-20D, Figure 21, Figures 22A to 22F and Figure 23. In a fourth aspect, the present invention is directed to a method for predicting, identifying or classifying the emotion or group of emotions felt by a subject after stimulation by an olfactory stimulus, from saliva samples from said subject, the method comprising the following steps of :E) measuring and calculating the BM ratio’s S2 / S1 from saliva samples of said subject for each of the selected BMs by the method according to one of claims 1 to 5 of the present invention and resulting to the determination of a salivary BMs signature for these 4 BMs for said stimulus;F) comparing the salivary signature obtained in step E) with reference salivary signatures obtained for these 4 BMs, said reference salivary signatures being identified to be associated with an emotion or group of emotions; andG) based on this comparison, predicting, identifying or classifying the emotion or group of emotions felt by said subject after stimulation with said olfactory stimulus,. In a fifth aspect, the present invention is directed to a method for determining the ability of an olfactory stimulus to provoke / elicit in a subject a desired emotion or group of emotions from saliva samples of said subject, the method comprising the following steps of:H) measuring and calculating the BM ratio’s S2 / S1 for each of said selected BMs by the method according to the present invention, resulting to the determination of the salivary BMs signature associated to said stimulus;I) identifying reference salivary BMs signature associated to the emotion or group of emotions which is desired to be provoked / elicited after stimulation by an olfactory stimulus, said reference signature being obtained by the method of claim 6 according to the present invention, J) comparing the salivary BMs signature obtained in step H) with said reference BMs signature identified in step I); andK) from this comparison, determining whether said tested stimulus is capable to provoke / elicit the desired emotion or group of emotions. In another aspect, the present invention is directed to the use of at least one, 2, 3, 4, or at least one for each different reference BM pattern / reference profile depicted in Table 3 for predicting, identifying or classifying the emotion or group of emotions felt by a subject after stimulation with a particular olfactory stimulus, from saliva samples of said subject. In another aspect, the present invention is also directed to the use of a kit for predicting, evaluating or identifying the emotion or group of emotions felt by a subject stimulated with an olfactory stimulus from saliva samples of said subject, said kit comprising:a) a means for collecting separatively at least 2 saliva samples from said subject; andb) a kit for measuring the quantity / concentration of each of the following 4 BMs present in said saliva samples: alpha-Amylase, Cortisol, DHEA and Oxytocin; andc) optionally an instruction leaflet to apply the method according to the present invention (one of claims 1 to 14), or containing at least one, preferably at least , 2, 3, 4, or at least one for each different reference BM pattern / reference profile depicted in Table 3. In another aspect, the present invention is also directed to the use of a kit for evaluating the capacity for an an olfactory stimulus to provoke / elicit in a subject a desired emotion or group of emotions from saliva samples of said subject, said kit comprising:a) a means for collecting separatively at least 2 saliva samples from said subject;b) a kit for measuring the quantity / concentration of each of the BMs present in said saliva samples, particularly the following 4 BMs present in said saliva samples: alpha-Amylase, Cortisol, DHEA and Oxytocin; andc) optionally an instruction leaflet to apply the method according to the present invention or containing at least one, preferably at least , 2, 3, 4, or at least one of the reference salivary BMs signature depicted in Table 3. In a preferred embodiment the use of kit according to the present invention is characterized in that the measurements of the quantity / concentration of each of said BMs present in said saliva samples are carried out by standard analysis biochemical assays wellknown by the skilled person such as but no limited to ELISA, LCMS or GCMS, (Liquid or Gas Chromatography Mass Spectrometry), Enzymatic assay, encapsulated nanaparticule enzymes assay. The inventors have also demonstrated that by measuring at least 3 BMs, preferably 4 BMs among the BMs listed in figure 1, more preferably , said 4 BMs alpha-amylase, cortisol, DHEA and oxytocin present in a saliva sample after stimulation in a subject and then combining them in a specific combination with a specific ratio for each BM (S2 / S1 type ratio) , it is possible:- to determine the emotion or group of emotions felt by the subject; or- to evaluate said stimulus tested for its ability to provoke / arouse the desired emotion or group of emotions,each of these at least 3 BMs, preferably 4 BMs among the BMs listed in figure 1, more preferably, said 4 BMs taken independently being able to provide statistically significant information with a gain in performance for the desired diagnosis, in particular compared to a more restricted combination of BMs, said combination aiming to obtain the best performance for the smallest number of BMs. Thus, preferably, in the method of the invention, the BMs measured can be combined with each other by a logistic function. In this logistic function, it is possible, for example, but not limited to:- use the direct measurements obtained for these BMs (isolated or simple measurements that have not undergone any modification before their introduction into the logistic function, or preferably,- use indirect measurements obtained for these BMs, for example relative measurements of these BMs reported to the measurements of these BMs in said subject before stimulation. The scores are obtained, for example, by a specific combination of said selected BMs. These specific combinations can be, for example, but not limited to, obtained by a statistical method of the binary logistic regression type in which, for example:- firstly, the measurements of the BMs, which may be independent, were tested in univariate analysis; and- secondly, the measurements of these 4 BMs identified by the inventors as significant in univariate analysis are then tested in multivariate analysis by binary logistic regression, this logistic regression making it possible to obtain a specific analysis profile or the formula of a score in the form for example:Score = a0 + a1.X1 + a2.X2 + a3.X3+ a4. X4 where the coefficients a1 to a4 are constants and the variables X1 to X4 are the values ​​measured for each of the 4BMs. This score allows:- either to determine the emotion or the group of emotions felt by the subject; or- either to evaluate whether or not the stimulus tested is capable of provoking / eliciting the desired emotion or the desired group of emotions, The person skilled in the art who wishes to use these specific analysis profiles or scores or algorithms in the context of the present invention is capable of determining them. It is then necessary to have a sufficient database including in particular:- the absolute or relative measurements of the 4 BMs identified by the inventors according to the stimulus tested; or- a population of subjects known to feel the desired emotion or the group of emotions sought for a given stimulus. Preferably, the method according to the present invention is characterized in that the stimulus is a sensory stimulus and more preferably an olfactory stimulus. Preferably and without limitation, the olfactory stimulus is chosen from perfumes or fragrances or any compound or support, material, plant, animal capable of emitting an odor likely to provoke an emotion.More preferably, the olfactory stimulus is a fragrance or an odor. By perfume or fragrance in the present invention, (these wording can be used interchangeably), we mean here but not limited to an odor or more often a more or less persistent odorous composition naturally emitted by a plant, an animal, a fungus or an environment, or even a consumer good such as a manufactured product intended for an end consumer. Which can include in particular beauty, household, hygiene products, etc. (creams, shampoo, laundry, etc.). Fragrances can be used to produce perfumes (fine fragrance) or consumer goods. In nature, perfumes are often chemical and biochemical messages, and in particular pheromones or phytohormones.By perfume or fragrance, we also mean here any emanation of a natural substance (for example a flower extract) or created (substance resulting from a chemical preparation process) or recreated from different aromas, solvents and fixatives intended for cosmetic use or to perfume objects, animals or indoor air. Perfume can be made from plant essences and / or synthetic molecules.Perfume or fragrance is also understood to mean a particular olfactory composition, generally pleasant and which can be more or less concentrated. This composition can be offered packaged and at a more or less strong olfactory concentration (we can cite here as an example in cosmetics an eau de Cologne, an eau de toilette, an eau de parfum or perfume which is generally distinguished by their concentration in natural or essential oil or by the quantity of active perfumed ingredient present in the product).Perfumes or fragrances can thus be made up of different compounds such as natural oils, alcohol and water. The combination of these ingredients creates a specific perfume that can be personalized according to the preferences of the user. Perfumes or fragrances are thus often designed to reflect the personality of the person wearing them, while giving them positive emotions.Perfumes or fragrances can also contain other ingredients such as fixatives or solvents which help to extend the longevity of the perfume.By perfume or fragrance, or perfume fragrance, is also meant here to designate any combination of several different aromatic materials that mix to produce a unique perfume. Preferably, the ratio between the measurement of their quantity / concentration present in said saliva sample obtained between 1 min and 30 min and advantageously between 5 min and 20 min after stimulation and that obtained approximately between 1 min and 30 min, preferably between 5 min and 10 min before stimulation, is calculated for each of the 4 BMs. Also preferably, the method according to the present invention is characterized in that in step 3, the saliva samples taken after and / or before stimulation are previously frozen, preferably at a temperature less than or equal to 20°C before determining the quantity / concentration of said BMs. Preferably, the method according to the present invention is characterized in that said emotion or said group of emotions is chosen:- from the following emotions felt but not limited to: sensuality, comfort, relaxation, elegance, confidence, addiction, energy, happy, specific (unique);and- from the following groups of emotions felt but not limited to: comfort and relaxation, elegance and sensuality. Also preferably, the method according to the present invention is characterized in that the measurement of the quantity / concentration of said selected BMs is carried out using biochemical or spectrometric measurement. In another aspect, the present invention relates to a kit comprising:a) a means for collecting saliva samples for the measurement of each of the 4 biomarkers (BMs) alpha-amylase, cortisol, DHEA and, optionally its derivatives, and oxytocin, in said samples;b) a kit for measuring the quantity / concentration of each of said 4 BMs present in said saliva samplesc) an instruction leaflet. Preferably, the kit according to the present invention is characterized in that the measurements of the quantity / concentration of each of said selected BMs present in said saliva samples are carried out, for the most part, by ELISA. In another aspect, the present invention relates to the use of a kit according to the invention for evaluating or identifying the emotion or group of emotions felt after stimulation by a subject from a saliva sample of said subject. In yet another aspect, the present invention relates to the use of a kit according to the invention for evaluating the capacity for a stimulation to provoke / arouse / elicit in a subject a desired emotion or group of emotions from a saliva sample of said subject. Preferably for these uses according to the invention, said stimulation is an olfactory stimulation and / or said emotion or said group of emotions is chosen:- from the following emotions felt, but not limited to: sensuality, comfort, relaxation, elegance, confidence, addiction, energy, happy, specific (unique);and- from the following groups of emotions felt, but not limited to: comfort and relaxation, elegance and sensuality.Although the present disclosure and its advantages have been described in detail, it should be understood that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the disclosure as defined by the appended claims. Moreover, the scope of the present application is not intended to be limited to the particular embodiments of the methods and steps described in the specification. As one of ordinary skill in the art will readily appreciate from the disclosure presently existing or later to be developed that perform substantially the same function or achieve substantially the same result as the corresponding embodiments described herein may be utilized according to the present disclosure. Legends of the Figures Figure 1: Molecular biomarkers of emotionsFigure 2: Stability of BMs as a function of storage temperature for 48 hours.Figure 3: Stability of BMs studied over 7 days.Figures 4.1A to 4.1D and 4.2: Diagrams of the study of circadian variations in BM levels.Figures 5A to 5D: Concentrations of salivary biomarkers according to the circadian cycle.Figure 6: Summary diagram of variations in salivary biomarkers expressed as a percentage compared to the measurement taken at 9 a.m. (100%).Figure 7: Schematic representation of the sampling process.Figures 8A-8E: Kinetics of biomarker concentration after stimulation with a fragrance.Figure 9: Diagram of the panel process.Figures 10A-10B: Correlation of variances.Figures 11A to 11D and 11A’ to 11D’: Comparison of biomarker concentration averages measured with olfactory stimulations.Figures 12A-12B: Representation of the variations in biomarker concentrations measured with fragrance stimulation according to emotions.Figure 13: Example of individual representation of emotions and biomarker evolution profiles.Figure 14: Example of individual representation for the emotions sensual and comforted.Figures 15-1A-15-1C and 12-2: K-means partitioning of individuals from biomarker profiles into 6 groups.Figure 16: Study design and objective.Figure 17: Salivary biomarkers’ ratios association.Figure 18: Principal component analysis (PCA) on biomarkers ratios (S2 / S1 and / or S3 / S1). Figures 19A-19C: ANOVA analysis on salivary BM ratios among different groups of emotion valence (0 to 2) in the questionnaire.Figures 20A-20D: Participants were nested together based on salivary biomarker’s ratio profiles using hierarchical k-means (hk-means).Figure 21: Participants were nested together based on salivary biomarker’s ratio profiles using hierarchical k-means (hk-means).Figures 22A to 22F: Salivary biomarkers’ ratio profiles for each cluster nested by hk-means performed on the total cohort (59 participants x 4 fragrances: 236 unique combinations) using the biomarkers’ ratio S2 / S1.Figure 23: Temporal evolution of individual salivary biomarkers’ ratios (S1 / S1; S2 / S1; S3 / S1) in hk-means clusters performed on the total cohort (59 participants x 4 fragrances: 236 unique combinations) using the biomarkers’ ratios S2 / S1.Figures 24A to 24F: Principal component analysis (PCA) was performed on discretized binary emotions (questionnaire response O = negative response to a given emotion; questionnaire response 1 and 2 were nested together = positive response to a given emotion) used as quantitative variables and correlations to different variables were analyzed (B – G)Figure 25: Clusters formed by hk-means using salivary BMs S2 / S1 ratio and using data by fragrances separately.Figure 26: Different salivary BM profiles correspond to different emotion profiles.Figure 27: Salivary BM profiles are shared between clusters formed by independent hk-means performed by fragrances separately and they present similar emotional response profilesFigure 28: Individual case studies. Participant E-044.Figure 29: Individual case studies. Participant E-078.Figure 30: Salivary Oxytocin and Alpha-Amylase levels associate with valuationFigure 31: Discretization of salivary biomarker ratios (S2 / S1) for new clustering approachesFigure 32: Classification of salivary biomarkers’ ratio (S2 / S1) profiles for emotion identification / predictionFigure 33: Classification of salivary biomarkers’ ratio (S2 / S1) profiles for emotion identification / prediction.Figure 34: Implicit test-based emotional profiles induced by fragrances      EXAMPLESIntroductionThe aim of this description is to demonstrate the feasibility of discovering biological signatures of emotions from salivary samples. In order to confirm this feasibility, biochemical kits for assaying in saliva the different biomarkers identified by the inventors (Alpha-Amylase, Cortisol, DHEA and Oxytocin) were studied and used, different parameters related to the sample were also studied and selected, finally, different approaches for discovering signatures were studied and provided satisfactory results. Example 1: Validation of emotionsThe target emotions were identified and a bibliographic search on the selected emotions was conducted. Example 2: Choice of reference kits for biomarker assay in salivaThe results given in the following examples confirmed not only the presence in saliva of the 4BMs identified by the inventors, but also their link with emotions. For each of the four biomarkers, different assay kits were tested. The technology used for some of the kits is ELISA. Several analytical performance criteria are examined and allow the selection of a kit by biomarker. In order to illustrate the performances, the recovery rate is used. The recovery rate represents the ability of the kit to detect a controlled increase of the biomarker in saliva. These performances are those of the commercial kits used in this first study. They are compatible with a proof of concept phase but can of course be improved. Example 3: Approach to discover biological signatures of emotionsA biological signature of an emotion can be defined as a combination of one or more biomarkers allowing to significantly identify an emotion. The approach to discovering this signature takes place in two main phases. First, preliminary checks to ensure the quality of the biomarkers measurements. This phase is carried out on a reduced cohort. Second, the search for the signature by different methodological approaches. This phase is carried out on a complete cohort. These two phases are described below.  Preliminary verification phaseFor the preliminary verification phase, different experimental parameters are tested in order to study their influence on the measurement and therefore to determine the final configuration of the tests.a) Question: Will the biomarkers measurements provide usable information?b) Verification of the repeatability of biomarkers measurementsc) Verification of the stability over time of biomarkers in the samplesd) Verification of the influence of the environment on the level of biomarkers (circadian rhythm)e) Verification of the measurement parameters (time after exposure), andf) Response: Reliable tests and protocols for signature discovery At the end of this phase, the various experimental parameters were established (measurement kit, ability to freeze samples if necessary, sampling time, sampling period after exposure to the fragrance). Signature research phaseFor the signature research phase, the cohort is subjected to the fragrance and must answer the questionnaire related to emotions or to all other methodologies of assessing emotions known by the skilled person in the art. Identifying the signature of an emotion means being able to separate the following two populations based on biological measurements: the one that showed an emotion (respondent), the one that did not express an emotion (non-respondent).The different stages of the signature research are set out below:a) Question: Is there a biological signature that can identify an emotion?b) Verification of the independence of the four biomarkers by correlation studiesc) Pre-selection and understanding of the mechanisms- Search for general trends in the entire population independently of the questionnaires- Search by biomarker and by emotion expressed for a difference between respondents and non-respondents- Search by emotion profile for a biological signature (based on the 4 biomarkers)d) Search for a biological signature by using advanced statistics In this case, the exercise is made more complex by the subjective side of emotions and the response to questionnaires. Thus, the arbitrary choice of 9 emotions of interest in the context of olfactory stimulation was made. In addition, to be carried out strictly, this phase requires significant cohorts that were not available at this level of the study. Thus, the approach that was favored with the use of advanced statistics consisted of correlating biological signatures to a set of emotions that are consistent with each other. Advanced statistics highlighted biomarker variation profiles characteristic of the “relaxed” emotion, as well as a “relaxed + comfortable” set and an “elegant + sensual” set. Example 4: Salivary samplesDuring this study, several parameters of the salivary sample were validated. Confirmation of the compatibility of the salivary sample with current practices in terms of panel was obtained. The salivary sample was thus confirmed using the 4 BMs identified by the inventors as a tool for identifying an emotion or group of emotions after an olfactory stimulation of a subject, in particular during the response to exposure to a fragrance / perfume. Example 5: Choice of reference kits for the dosage of the selected biomarkersObjective: Determine the reference kits for the dosage of BMsMeasurements of molecular biomarkers in saliva can be biased by many physiological and technical aspects. In this section, we describe in particular the research work to define and qualify a reference kit for the measurement of DHEA (and its sulfated ester SDHEA) and OXT. The choices were made based on the analysis of 2 biochemical kits by BM. The kits, their performance and their relevance as reference kits were evaluated. MethodologyThree steps allowed us to establish the choice of reference kits for the measurement of biomarkers in saliva:- Evaluation and comparison of 2 biochemical kits used for the measurement of each BM in the buffers recommended by the suppliers- Evaluation of two artificial salivary matrices (one rather fluid and one very viscous)- Quantification of salivary samples from volunteer donors (n=6). All saliva samples analyzed with the 2 kits for the quantification of each BM were collected in the morning (before 10:00 a.m.), then aliquoted and stored at -80°C until use. A - Definition of the reference test for CortisolCortisol (and its derivative cortisone),Knowledge related to cortisolCortisolCortisol is a hormone or glucocorticoid produced from cholesterol in the adrenal cortex in response to the adrenocorticotropin hormone (ACTH) secreted by the pituitary gland. Cortisol has different actions on the body. Among others, it promotes the catabolism of lipids and proteins, inhibits the action of insulin and promotes glycogenesis in the liver. It also has an anti-inflammatory and anti-allergic action, inhibiting the production of pro-inflammatory cytokines (e.g. IL-12, IFN-ɣ, IFN-ɑ and TNF-ɑ) and upregulating IL-4, IL-10, IL-13 by Th2 cells, thus stimulating the release of serotonin and histamine.Cortisol secretion is controlled by the hypothalamic-pituitary (HPA) axis through the action of ACTH and CRH (corticotropin-releasing hormone) Cortisol and emotionThe limbic center composed of the thalamus and the amygdala, among others, is the center of emotion in the central nervous system and can, depending on the emotion, trigger the HPA axis and release cortisol into the body (6). In a context of depression or chronic stress, hyperactivation of the HPA axis can occur, causing constant hypercortisolemia leading to a worsening of the psychological state and a significant deregulation of the individual's metabolism.Increased salivary cortisol has also been shown in the Trier Social Stress Test (TSST), in response to a state of acute stress and anxiety in the individuals tested."Positive" mental states tend to moderate the impact of stress sources on health. They probably act on the same metabolic pathways through which stressors exert their neuroendocrine effects. Cortisol and CortisoneIn saliva, the degradation of cortisol is linked to non-enzymatic reactions (hydrolysis, redox), bacterial metabolism (intracellular enzymes and exo-enzymes) and the endogenous metabolism of cortisol into cortisone by the action of the enzyme 11-beta-hydroxy-steroid-dehydrogenase (11b-HSD). Storing saliva samples at 4°C limits this degradation. According to the literature, cortisol levels remain stable for up to 5 days in human saliva samples stored at room temperature or at 4°C (3).   Cortisol assay methodsSalivary cortisol assay correlates well with its bioactive serum level. Normal salivary cortisol values range from 1 to 11.3 ng / mL. Salivary cortisol is commonly measured using enzyme-linked immunosorbent assays (ELISA). Other assay methods can also be used, such as liquid chromatography coupled with mass spectrometry detection (LC-MS). However, immunoassays remain the most widespread and their advantages lie in their ease of use and high degree of automation. Cortisone detection and assayFree cortisol diffuses into saliva, where it is rapidly converted into cortisone by the enzyme 11bHSD (11β-Hydroxysteroid dehydrogenase) type 2 secreted by the salivary glands. Normal values for salivary cortisone have not been established, but studies have shown that it varies between 10 ng / mL and 100 ng / mL. Salivary cortisone is commonly measured using enzyme-linked immunosorbent assays such as ELISA. Other assay methods can also be used, such as LC-MS. However, immunoassays remain the most widespread and their advantages lie in their ease of use and high degree of automation. Cortisone levels are between 5 and 10 times those of cortisol, and small variations in salivary cortisol levels caused by olfactory stimuli can be seen in a magnified way on salivary cortisone. B- Definition of the Gold Standard for Alpha-AmylaseAlpha-Amylaseα-Amylase is an enzyme present in pancreatic and salivary secretions that specifically hydrolyzes α(1→4) glycosidic bonds. Its role is essential in the digestion of starch into simple sugars that are more easily assimilated by the body. The catalytic activity of amylase can be measured in several biological fluids including blood and saliva. The secretion of salivary α-amylase is increased during episodes of stress and is also correlated with the release of norepinephrine. Quantifying salivary α-amylase activity could therefore indirectly assess the activity of the two main nervous systems controlling stress: the hypothalamic-pituitary-adrenal (HPA) axis and the SNS. Therefore, salivary amylase may represent a biomarker of interest to assess certain psycho-physiological states in humans.In addition to its physiological link with emotional states, salivary α-amylase has good stability that is compatible with sample collection and storage. Indeed, salivary samples can be stored without loss of activity for 4 days at room temperature and at least 6 months at -20°C. The effect of successive freezing / thawing cycles has a negligible impact below 5 cycles (7). Methods for measuring salivary α-amylaseThe measurement of α-amylase is mainly done by measuring its enzymatic activity. To do this, there are several technological approaches as well as different definitions of the enzymatic unit. Monitoring of α-amylase activity can be performed by spectrophotometry, electrochemistry, potentiometry, photochemical immobilization, among others. Approaches by direct dosage of α-amylase concentration by immunofluorescence also exist. As a result, comparison between methods is very difficult. There is no reference value for human salivary alpha amylase level, but according to one study, its activity is estimated at approximately 100 to 200 U / mL. C - Definition of the reference test for DHEADHEA-related knowledgeDHEADeHydroEpiAndrosterone (DHEA) is a hormone produced from cholesterol, mainly by the cortex of the adrenal gland in response to adrenocorticotropin hormone (ACTH), secreted by the pituitary gland. DHEA circulates mainly in the form of sulfated DHEA. DHEA is the circulating precursor of testosterone, estrogens and progesterone. It is also associated with immune function and is also involved in cellular aging (1).DHEA concentration begins to increase in the body around 6-8 years of age and increases steadily. After the age of 30, this molecule decreases in a quasi-linear manner with age and presents circadian variations similar to those observed for cortisol. Serum DHEA concentration varies with age and gender and ranges from 60 to 9000 ng / mL. DHEA levels measured in blood and saliva are highly correlated and salivary DHEA concentration is not affected by enzymes or salivary turnover (2). DHEA and emotionDHEA is also produced directly in the nervous system and serves as a neuroactive and neuroprotective factor. In the face of acute psychosocial stress, its production will increase. Decreased DHEA concentration is correlated with deteriorations in cognitive abilities and mood. Restoring blood DHEA concentrations in patients aged 40 to 70 years to levels comparable to those of young adults showed significant improvements in their well-being at 3 months: better quality of sleep, better mood, more energy and better stress management (4). Several studies also suggest other protective roles of this molecule on physical and mental well-being, muscle strength, bone density, fat mass, skin elasticity, allergies, etc... Methods for measuring DHEA and resultsLike salivary cortisol, DHEA is present in saliva and can be measured using enzyme-linked immunosorbent assays (ELISA). Other methods of measurement can also be used, such as liquid chromatography coupled with mass spectrometry detection (LC-MS). However, immunoassays remain the most widespread and their advantages lie in their simplicity of use and their high degree of automation. D - Definition of the reference test for oxytocinOXT-related knowledgeOxytocinOxytocin (OXT) is a 9-amino acid neuropeptide produced mainly in the hypothalamus. It is involved in many physiological processes and has an important role in parturition and lactation (5). It is also associated with socio-emotional processes and reproduction. OXT acts locally in the central nervous system (CNS) but also systemically on several organs. The action of OXT is mediated by its binding to its receptor, expressed mainly in the mammary glands, uterus and CNS. Oxytocin is degraded by multiple aminopeptidases (oxytocinases) expressed by different cell types. Oxytocin and emotionOxytocin modulates the integration of emotional information and interacts with the reward pathway. In the brain, OXT acts as a neurohormone or neuromodulator. It is released during positive social interactions and can downregulate stress, heart rate, and blood pressure. OXT released by the posterior pituitary is presumed to be the major source of salivary OXT . Thus, neuropeptides such as OXT, synthesized primarily in the CNS, can be used as indirect measures of brain function and are therefore more likely to be directly related to emotions. OXT measurement methodsThe standard analytical techniques for oxytocin quantification are radioimmunoassay, enzyme-linked immunosorbent assay, and liquid chromatography-mass spectrometry.Oxytocin is a difficult molecule to measure. There are no reference values ​​for salivary oxytocin because there is currently no validated and standardized test. To date, the most widely used method is the enzyme-linked immunosorbent assay (ELISA). Several studies involve sample pretreatment steps (centrifugation, column extraction, etc.), however, these methodscan result in a significant loss of measurable oxytocin. Finally, the proteolytic enzymes present in saliva can break down peptides such as oxytocin. B – Study of the stability of biomarkers as a function of temperature and storage timeStudy of the stability of BMs 48 hours at different temperaturesWe evaluated the stability of salivary biomarkers under different preservation and storage conditions. First, saliva samples (N=4) were stored at different temperatures (-20°C, 4°C, room temperature, 37°C, 42°C and 60°C) for 48 hours. The concentration or enzymatic activity of the four MBs was quantified on the day of collection (D0) to determine the reference values ​​(100%), then at 48 hours (see Figure 2). Figure 2: Stability of BMs as a function of storage temperature for 48 hours.The results represent the averages of the measured values ​​of BMs in saliva samples (N=4). The results are expressed in % relative to the concentration measured at D0 which represents 100%. The assays at D0 were carried out on fresh saliva. The average CVs are represented by bars. RT: room temperature.After 48 hours of storage, cortisol and DHEA remain stable and at levels comparable to D0 for temperatures from -20°C to 42°C. Alpha-amylase is stable at -20°C and 4°C, with no loss of activity detected. Storage at room temperature results in a significant decrease in measured activity and for higher temperatures its activity is not quantifiable. As for oxytocin, all storage temperatures tested seem to impact its stability. Storage at -20°C and 4°C, result in a minimum loss of 26% of its measured concentration. For the other temperatures studied, oxytocin after 48 hours of storage is not quantifiable. For storage at 60°C, no quantification was possible for the 4 BMs studied. Conclusion: For the rest of the stability study, the storage temperatures chosen for the different biomarkers are: -20°C for oxytocin, +4°C for alpha-amylase and room temperature (RT=21°C) for DHEA and cortisol. Study of the stability of BMs over one weekBased on the previous data, we evaluated the impact of storing saliva samples at a defined temperature for 7 days on the stability of each BM. The storage temperature chosen for each biomarker is the storage temperature beyond which the BM loses its stability, and therefore its dosage is less efficient. The experiment was also carried out on four saliva samples (see Figure 3). Figure 3: Stability of the BMs studied over 7 days.The results represent the averages of the measured values ​​of the BMs in the saliva samples (N=4). The results are expressed in % relative to the concentration measured at D0 which represents 100%. The assays at D0 were carried out on fresh saliva. The average CVs are represented by bars. RT: room temperature.All the BMs show satisfactory stability in saliva over 7 days. The variations observed are of the same order of magnitude as the variations observed during the robustness study (+ / - 20%). Surprisingly, oxytocin here shows remarkable stability at -20°C over 7 days with an increase in its concentration measured at D2, D3 and D7. Conclusion: All the results describe a stability of BMs in saliva for at least 7 days provided that the predetermined storage temperatures for each BM are respected. This temperature is 4°C for α-amylase, room temperature (RT) for cortisol and DHEA and -20°C for oxytocin. C- Study of the temporal variation of endogenous levels of biomarkersa) Distribution of basal levels of salivary biomarkers.(See Figures 4.1A to 4.1D) Figures 4.1A to 4.1D: Distribution of basal levels of salivary biomarkers.Distribution of basal levels (S1) of salivary biomarkers ((A) Alpha amylase - AA; (B) Cortisol - CORT; (C) DHEA; (D) Oxytocin - OXT) shows inter-individual variations displaying considerable amplitude (AA = 79.1±57 (U / mL, mean ± SD), 268.1 (MAX), 12.2 (MIN); CORT = 2559.2±1534.6 (pg / mL, mean ± SD), 10825.2 (MAX), 562.12 (MIN); DHEA = 745.3±460.3 (pg / mL, mean ± SD), 2467.2 (MAX), 55.1 (MIN); OXT = 96.1±75.1 (pg / mL, mean ± SD), 406.5 (MAX), 14.7 (MIN). Lowest amplitude was observed for cortisol = 19.2 times (MAX / MIN) and the highest for DHEA = 44.8 times (MAX / MIN). Basal levels were analysed using the S1 sampling before olfactory stimulation.b) Study of the salivary biomarkers level according to the circadian and the inter-day cycle.(See Figure 4.2)The objective of this study is to evaluate the impact of the circadian variation on the endogenous levels of the salivary biomarkers studied. The intra- and inter-individual variations of 8 donors were evaluated. For this, saliva samples were taken at 3 times (9 a.m., 12 p.m. and 3 p.m.) during the same day and for 3 consecutive days (see Figure 4.2). Figure 4.2: Diagram of the study of intra-day variations in BM levels.The analysis was performed on 8 individuals over 3 consecutive days (D1: Day 1; D2: Day 2 and D3: Day 3) at three different times (9 am, 12 pm and 3 pm). Results:The results of the temporal variations in the level of salivary biomarkers are presented in Figures 4.1A-1C and 4.2. In general, intra-day variations are more significant than inter-day variations and these follow the data already described in the literature. Figures 5A-5D: Concentrations of salivary biomarkers according to the circadian cycle.Box plot representation of the concentrations of BMs quantified in 8 salivary samples taken at 3 different times (9 AM, 12 PM, 3 PM), 3 consecutive days (D1-D3). A: Salivary alpha-amylase activity (U / mL); B: Salivary cortisol concentration (ng / mL); C: Salivary DHEA concentration (pg / mL); D: Salivary oxytocin concentration (pg / mL). D1: Day 1; D2: Day 2 and D3: Day 3. The median is represented by a line and the mean by a cross. Alpha-amylaseAlpha-amylase activity increases during the day in the 8 volunteers studied (Figure 5A). The median (measurement over 3 days) at 9 a.m. is 70 U / mL, at 12 p.m. it is approximately 90 U / mL and at 3 p.m. it is 133 U / mL. We observe variability between individuals, less important for the samples taken at 9 a.m. and 12 p.m. compared to the samples taken at 3 p.m. Indeed, the distribution of individual data is more dispersed at 3 p.m. These results are equivalent for the 3 days of the experiment. Cortisol and DHEAThe concentrations of salivary cortisol and DHEA also vary during the day but in the opposite direction to alpha-amylase (Figures 5B and 5C). These 2 molecules tend to decrease between early morning (9 a.m.) and mid-afternoon (3 p.m.). The concentrations of cortisol and DHEA are highest at 9 a.m. A greater inter-individual variation is observed in the population at this time. Then, a sharp decrease in concentrations is measured at 12 p.m. and less significantly at 3 p.m. These results are reproduced over the 3 days of the experiment.OxytocinThe concentration of salivary oxytocin appears to decrease very slightly during the day (Figure 5D). A high intra- and inter-individual variability is observed, with rates ranging from 28 pg / mL to 129 pg / mL for the same individual (different days, same time) and from 38 pg / mL to 1167 pg / mL between two individuals on the same day at the same times. If the data for the same day are analyzed, a trend emerges with a progressive decrease in salivary oxytocin during the day (between 9 a.m. and 3 p.m.). Figure 6: Summary diagram of the variations in salivary biomarkers expressed as a percentage compared to the measurement taken at 9 a.m. (100%).Each point represents the averages obtained for the measurements taken on the 3 days at each time studied.In the context of a panel study, it is preferable to carry out saliva samples in the morning, taking into consideration: • For 3 biomarkers (cortisol, DHEA and oxytocin), the median value at 9 am is in the middle of the standard range of the reference kits.Mean median at 9 am:o Cortisol: 1830 pg / mL; Range of the kit range: 50 to 3200 pg / mL.o DHEA: 505.21 pg / mL; Range of the kit range: 20 to 2160 pg / mL.o Oxytocin: 183.31 pg / mL; Range of the kit range: 15.6 to 1000 pg / mL.For the 12 noon and 3 p.m. times, the concentrations obtained are close to the low values ​​of the standard range. If an olfactory stimulus induces a significant decrease in these BMs, we could not quantify them because the levels would be at the lower limit of quantification of the kit. • The concentration of α-amylase varies from 10 to 20% in the morning (between 9 a.m. and 12 p.m.) but by + 80% between 12 p.m. and 3 p.m. This strong variation in the middle of the day could induce a bias in the dosage of this BM during the olfactory stimulation study where several samples will be collected over a period of 30 min to 1 hour. • Samples taken in the morning allow technical processing and experimental analysis on the same day. This is not the case for technical and logistical reasons if the samples are taken after 2 p.m. Indeed, the reference kit for the oxytocin dosage requires an overnight incubation, which would extend the storage time of the samples before dosage (in this case the samples would be dosed more than 30 hours after collection).In conclusion, the circadian rhythm impacts the salivary levels of the biomarkers studied. Salivary concentrations of cortisol and DHEA decrease during the day. Conversely, the measured activity of alpha-amylase increases more strongly after 12 p.m. Salivary oxytocin levels seem to be less impacted by the circadian rhythm, with however a concentration decreasing throughout the day in a moderate way. All of these results are in agreement with the data in the literature. The envisaged panel tests will have a duration of approximately 30 min. The variations over 30 min of the biomarkers, observed in the morning between 9 a.m. and 12 p.m., are sufficiently low (around 6%) to be compatible with analyses during a panel. D – Preparation of the Panel conditions: Kinetics of biomarker measurement after Fragrance stimulation (pre-panel)A pre-panel was organized on a small number of volunteers. This pre-panel had 2 main objectives:- Evaluate the logistical aspect: the organization of this panel with a reduced cohort allowed the verification and planning of the panel's progress in line with the study objectives.- Evaluate the kinetics of biomarkers following olfactory stimulation. A kinetics after stimulation was carried out and the samples were taken at different times post-stimulation. This allowed the determination of the collection times of salivary samples after stimulation (see Figure 7)  Figure 7: Schematic representation of the sampling process.S1 to S5: sampling 1 to sampling 5; FT1 to FT5: Fragrance Time 1 to Fragrance Time 5.For the pre-panel, the participants were summoned in the morning at two different slots, either at 9 a.m. (n=2) or at 10 a.m. (n=3) to carry out the olfactory stimulation test (Figure 7). The fragrance provided by MANE and used for the pre-panel is the same as that of the future panel (No. 267). A sample was taken before stimulation (S1, 5 minutes before stimulation). Four samples were taken after stimulation (S2-S5) at short times (5 and 15 min after stimulation) and at long times (30 and 60 min after stimulation). The kinetics results are shown in Figure 7.  Figures 8A-8E: Kinetics of biomarker concentration after stimulation with a fragrance.Participants (C and D) were called for the test at 9 am (n=2) and participants (A, B and E) were called for the test at 10 am (n=3). A sample was taken 5 minutes before stimulation (T1) and 4 samples were taken after stimulation (5 min; 15 min; 30 min; 1 h) (T2-T5). The ratio represents the values ​​measured for T2, T3, T4 and T5 reported to the T1 measurement. The stimulation is indicated by the vertical red dotted bar.The evaluation of the kinetics of biomarker measurement after olfactory stimulation showed that:• Olfactory stimulation does indeed generate a significant variation in salivary biomarkers linked to emotions and this in the minutes following the stimulation.• The results of the people summoned at 9 am and 10 am are comparable in terms of variations in the quantities or activity of the BMs following the olfactory stimulus.• We chose 2 short times, 5 min and 20 min after the stimulus, to measure the 4 BMs during the panel. We mainly want to measure the effect of olfactory stimulus on the quantitative variation or on the activity of target BMs and not a return to a basal state whose duration can be variable. E – POC Panel: Study on a number of subjects allowing preliminary statistical analysesFollowing the completion of the pre-panel, we organized a panel on a real scale with 30 participants in order to measure the 4 target biomarkers for a proof of concept. The following were studied:- Measurement of the response of salivary BMs following olfactory stimulation with a fragrance (provided by MANE on a solid support) and a carrier (solid support without fragrance provided by MANE).- Conducting a study on a number of subjects allowing preliminary statistical analyses. - Individual tests carried out on the 4 target BMs. Regulatory authorization for research involving humans (RIPH)The study is a category 3 research involving humans (RIPH) corresponding to a non-interventional study without risk and constraint for people. To carry out this research on humans, we created an administrative file that included:- Registration of the study on the ANSM (National Agency for the Safety of Medicines and Health Products) website in order to obtain a national RCB (Research and Biological Collections) number. The research thus bears the number 2022-A02008-35.- Performance of a personal data impact analysis by the CNRS data protection department that brought together:• A systematic description of the processing operations envisaged and the purposes of the processing, including, where applicable, the legitimate interest pursued by the data controller;• An assessment of the necessity and proportionality of the processing operations with regard to the purposes;• An assessment of the risks to the rights and freedoms of the persons concerned;• The measures envisaged to address the risks, including guarantees, measures and security mechanisms aimed at ensuring the protection of personal data and providing proof of compliance with the regulation.- Declaration of compliance of the processing with the CNIL (National Commission for Information Technology and Civil Liberties). This declaration stipulates that the processing of data in the laboratory is strictly compliant with the MR-003 reference methodology. This is self-declarative and was carried out on September 26, 2022 (CNIL Ref: 2227634 v 0).- Submission of a file to a CPP (Committee for the Protection of Persons) with a favorable opinion. Organization of the panelThe objective is to define a methodology and sampling protocols that will satisfy both the scientific and technical constraints but also acceptability by the panelists.Given that several sources of stress, independent of the test itself, can occur in the first stages, we chose to study the effect of a carrier stimulus (support without fragrance). So in total 2 olfactory stimuli were evaluated during this panel: a fragrance stimulation (No. 267) and a stimulation carrier (No. 943).A total of 30 subjects were tested. Each subject's participation was spread over 2 days: day 1 including carrier stimulation and day 2 fragrance stimulation or vice versa. This is equivalent to testing 60 subjects.The chosen fragrance has already been evaluated; the questionnaire results are available. The chosen support: Air Freshener beads in a glass jar.The panel was organized over 4 days (15 subjects per day with only one type of stimulation).The subjects invited participated in a regulatory information meeting on the project. After signing the consent to participate in the research, each subject went to a room to perform the test. Three saliva samples were collected: the first, 5 minutes before stimulation (S1), the second, 5 min after stimulation (S2) and the third, 20 min after stimulation (S3).The samples were stored at +4°C until the time of experimental analysis.All subjects were tested with the 2 stimuli over 2 days. All biomarker assays could also be performed on all samples (see Figure 9). Figure 9: Diagram of the panel procedure.Sampling S1 to S3: sampling 1 to sampling 3; FT1 to FT3: Fragrance Time 1 to Fragrance Time 3. CT1 to CT3: Carrier Time 1 to carrier Time 3. Conclusion: The salivary samples are therefore compatible with use in a panel with olfactory stimulation. For the purposes of this study, one sample before and two after olfactory stimulation were taken. F - Analysis of the response of BMs to olfactory stimulationThe dosage of the 4 BMs was carried out according to the protocols previously defined.Each measurement was performed in triplicate. The data collected in the form of concentration or enzymatic activity were supervised by 2 people. For each BM, the concentration per individual was determined in fragrance and carrier stimuli conditions. All data was analyzed with statistical tests appropriate to this type of data.Before the analysis, several steps were previously carried out to ensure that the data followed a normal distribution:- Use of the Shapiro test to ensure that the transformed measurements followed a normal distribution. - Box-Cox transformation to transform the measurements so that they had a distribution close to that of a normal distribution. Overall analysis of correlationsThe first analysis carried out was to study the correlation between the data obtained with the 4 BMs following stimulation with the fragrance or the carrier (Figures 10A-10B).   Figures 10A-10B: Correlation of variances.T1: collection 5 min before stimulation; T2: collection 5 min after stimulation; T3: collection 20 min after stimulation. The color scale indicates the degree of correlation between 2 measurements: 1 = maximum correlation; -1 = maximum anti-correlation. Correlations are calculated using the Spearman method.The dark blue diagonal indicates a strong correlation expected between measurements for the same biomarker before and after olfactory stimulation between the 3 samples (T1, T2, T3).During a fragrance stimulus, the data obtained for cortisol are weakly correlated with those of DHEA for all times studied. This is also the case during stimulation with the carrier, but to a lesser extent. This corroborates the data in the literature that describe similar (but not identical) variations in salivary cortisol and DHEA.The measurements of alpha-amylase after stimulation with the fragrance are anti-correlated with those of cortisol and to a lesser extent with DHEA. This anti-correlation of measurements is not observed with the carrier stimulus.Oxytocin is not correlated with other BMs.In conclusion, weak correlations between cortisol and DHEA measurements are observed after stimulation with fragrance or carrier. The measurement of alpha-amylase is anti-correlated with that of cortisol and DHEA only for the fragrance stimulus. Oxytocin provides very different information from other BMs. These results are observed for the 3 measured times (T1, T2 and T3). All of these results show that there are differences in correlations depending on the type of olfactory stimulation and are consistent with the data in the literature when they exist. This observation leads to the relevance of retaining the 4 biomarkers to determine the signature. Global analysis of the variation of biomarkers following olfactory stimuliThe data were then analyzed for each BM and each stimulus at the level of the population studied. Transformed measurements were compared between T1 (before stimulation), T2 (5 min after stimulation) and T3 (20 min after stimulation) with a paired Student T test (see Figures 11A to 11D and 11A’ to 11D’ and Table 1). Figures 11A to 11D and 11A’ to 11D’: Comparison of mean biomarker concentrations measured with olfactory stimulations. Each graph represents a biomarker: alpha-amylase (A, A’), cortisol (B, B’), DHEA (C, C’) and oxytocin (D, D’). For each biomarker, measurements were transformed with the Box-Cox transformation using a different lambda per biomarker (0.35 for alpha-amylase, 0.13 for cortisol, 0.23 for DHEA and -0.15 for oxytocin). Each graph shows for each time (T1 = 5min before the stimulus, T2 = 5 min after, T3 = 20 min after) the distribution of the data with the quantiles, the mean (point), the median (line) and the representation of the p-values ​​obtained with paired Student T tests by comparing the means of the different samples (T1, T2 and T3). Ns: p > 0.05; *: p <= 0.05; **: p <= 0.01; ***: p <= 0.001; ****: p <= 0.0001.The overall distribution of biomarker measurements across the entire population studied in the panel (N=30) varies according to the time of sampling and the stimulus (fragrance or carrier).Table 1 summarizes all the p-values ​​obtained during the comparisons between the 3 times studied (T1, T2 and T3) for the same BM and the same stimulus.In the case of the fragrance, the means of alpha-amylase, cortisol and DHEA vary very significantly in at least 2 comparisons. This is not the case for the stimulation with the carrier except for cortisol.   FragranceCarrierT2 vs T1T3 vs T1T2 vs T3T2 vs T1T3 vs T1T2 vs T3α-Amylase4.90E-050.000350.790.0110.180.058Cortisol0.0241.90E-050.00250.0750.00210.0002DHEA0.00440.000170.000380.440.0120.0051Ocytocine0.0270.520.140.030.650.0084 Table 1: Comparison of mean concentrations for each BM represented by p-values.Means were compared between the 3 times T1, T2 and T3 with a paired Student T test. Comparisons highlighted in gray have a p-value < 0.01. T: sample collection times (T1, T2 and T3). Conclusions: These results indicate that even if there are clear differences depending on the stimulation, an analysis at the population level is relevant at this level. The sample size N=30 may induce a lack of power. In addition, each person may have a different feeling and therefore a different biological response. Subgroup or even individual analysis are therefore adopted for the rest of the proof-of-concept study with a systematic link with the emotions reported by the subjects.  Global analysis of BM variations in relation to emotionsThe panel participants answered a questionnaire, prepared in collaboration with MANE, just after the olfactory stimulation. Subjects were able to answer the 10 questions while continuing to smell the olfactory stimulus for 5 min and before the T2 saliva sample.Discretization of responses to emotion-related questions: The questionnaire included a question related to emotions with 9 categories: addict, elegant, confident, happy, energized, unique, sensual, comforted, relaxed. For each emotion, 5 responses were possible. Responses considered negative or neutral (“totally disagree”, “rather disagree” and “neither agree nor disagree”) were grouped under the term False. Responses “rather agree” and “totally agree” were grouped under the term True.Participants’ responses for each emotion were discretized according to the terms True and False. All analyses of the variation of biomarkers according to each of the discretized emotions are represented by box plots The results obtained with the measurement of biomarkers for each discretized emotion (False or True) are different depending on a fragrance or carrier stimulation.It should be noted that the vast majority of responses to emotions are rather negative (False). No emotion collects a majority of positive responses (True) regardless of the stimulation. For the carrier, there are 7 emotions (out of 9 in total) for which the number of False is greater than 24 individuals. It is therefore difficult to analyze the response of BMs associated with an emotion for the carrier. For fragrance, three categories of emotions can be defined based on the number of True responses:1- About 50% of True responses: relaxed, comforted, sensual 2- About 35% of True responses: elegant, energized, happy 3- About 20% of True responses: addict, confident, unique During fragrance stimulation, for the 4 emotions “addict, comforted, relaxed and unique”, the average concentrations of BMs vary for the different times studied (Figure 12A). The evolution of biomarkers at times T2 and T3 may be different. This indicates that the variations over time of biomarkers may be dependent on the emotion felt after stimulation. For example, we observe after stimulation a rapid variation (T2 / T1) of the AA biomarker for the emotion “energized”, and for “happy”, the DHEA biomarker varies later (T3 / T2) (Figure 12B).  Figures 12A and 12B: Representation of the variations in biomarker concentrations measured with fragrance stimulation as a function of emotions.Four emotions were represented as an example. A) BMs vary regardless of the time ratio studied; B) BMs vary very quickly (T2 / T1) or late (T3 / T2) after stimulation. Each point represents an individual salivary sample. The ratios were calculated with the raw values ​​of BM concentrations. Each graph shows for the three ratios studied (T2 / T1, T3 / T1 and T3 / T2), the distribution of the data with the quantiles, the mean (point) and the median (line). False: response considered negative or neutral with respect to the emotion; True: response considered positive with respect to the emotion. Conclusion: Emotions provoked by an olfactory stimulus generate different biomarker variation profiles over time. Different olfactory stimuli (e.g. fragrance or carrier) lead to different biomarker variation profiles. Individual analysis of biomarker variations related to emotionsThe data collected during the panel allow us to perform analyses of individual biological responses following olfactory stimulations and to compare them to the emotions expressed.For each individual (N=30) and the olfactory stimulations carried out, we have indications of the feeling for 9 emotions (rated from 1 to 5) and measurements of the variation of 4 biomarkers over time.These data are expressed individually (Figure 13). Figure 13: Example of individual representation of emotions and biomarker evolution profiles.The numbers E-0XX correspond to the subjects participating in the study. E-016 and E-024 have biomarker profiles that vary little. E-004, E-007, E-017 and E-026 present biomarker profiles that vary more strongly after olfactory stimulation.It can be noted that for 15 subjects out of 30 (E001, E003, E005, E008, E012, E013, E014, E015, E016, E018, E020, E021, E024, E09, E030), the responses to the 9 emotions after stimulation by the carrier are systematically expressed in a negative way (score 1) or neutral (score 3) (Appendix). This never happens with stimulation by fragrance. However, for some of them, variations in the concentration of biomarkers are observed.We observe that after olfactory stimulation, some people (example E016, E024) present profiles of biological biomarkers that do not vary or vary very little with more or less felt emotions (Figure 13). Conversely, several people (example E007, E017, E026, E030, etc.) are very reactive from the point of view of biomarkers and the questionnaire. This suggests that there would be people more sensitive to olfactory stimulations than others. This point is interesting because it allows us to identify people who are sensitive or not very sensitive to olfactory stimuli (this phenomenon is known for many sensory stimulations). After olfactory stimulus, the profiles of individual responses to emotions as well as the biological responses are very varied. However, certain emotions (example sensual or comforted) seem to correspond to particular profiles of biomarkers (Figure 13). For example, for the emotion “sensual”, the biomarker profile would seem to correspond to a decrease then an increase in alpha-amylase between T1 and T2 then between T2 and T3 and to a continuous decrease after the olfactory stimulation for cortisol and / or DHEA. For “comforted”, alpha-amylase would decrease and oxytocin would seem to decrease (T2 / T1) after the stimulus and then increase (T3 / T1). Figure 14: Example of individual representation for the emotions sensual and comforted.The emotions “sensual” and “comforted” were represented as a radar rated from 1 to 5. For “sensual”, subjects E-004, E-017, E-019 and E-011 present similar profiles of BM variations and for “comforted”, subjects E-011, E-013, E-027 and E-007 present similar profiles of BM variations. Conclusions: Individual analysis of emotion-focused biomarker variations shows that some emotions seem to give off particular biomarker profiles. Similarly, these data suggest that it will be possible in the future to assess a person’s level of sensitivity to olfactory stimuli. This preliminary study requires further analysis based this time on a classification focused on biomarker profiles. Classification of individuals based on their biomarker profile after olfactory stimulationA K-means classification analysis of individuals based on their biomarker profile was performed (Figure 14). K-means partitioning was used to separate the individuals in the panel based solely on the biomarker measurements at different times. The principle of this method is to start from k points, centroids, positioned in the space of our data (the biomarker measurements) by following an algorithm. The centroids are associated with the closest individuals and thus form the groups. Several algorithms exist to apply this method, in our case we used the Hartingan-Wong algorithm. The results presented below are those obtained by partitioning our individuals into k = 6 groups.The objective is to objectify the preliminary observation made during the individual analysis (certain emotions could correspond to particular biomarker profiles). In addition, we leave ourselves the possibility here of associating groups of emotions with these profiles. The choice of the 6 groups was made in particular in connection with the sample size (30 subjects) which allows both to have groups of sufficient size while maintaining an acceptable degree of finesse. Figures 15-1A to 15-1C and 15-2: K-means partitioning of individuals from biomarker profiles into 6 groups.Figures 1A to 1C: Overall representation of the analyses.(A) The number of individuals for each group. (B) Representation of the percentage of individuals per group having felt each emotion. (C) Representation of the biomarker profiles with a different color per group.Each curve represents the mean of the BM concentration for the individuals in the group and the standard error of the mean. 2) Representation of the biomarker profiles per individual for groups 2, 4 and 5.6 profiles were identified (Figure 15-1). Groups 2, 4 and 5 were studied more specifically because of their size (the other groups had 4 people or less, which was considered insufficient to draw conclusions). These groups show a majority response to one or two emotions. Group 5 largely highlights the emotions “sensual” and “elegant” (75 and 62% respectively), while group 4 rather highlights the emotions “comforted” and “relaxed”. Group 2 very clearly indicates “relaxed” (80%). For these groups 2, 4 and 5, the variation profile of each biomarker after stimulation is different (Figure 15-2). Conclusion: The specific analysis of biomarker profiles for groups 2, 4 and 5 indicates that it is possible to identify biomarker profiles associated with an emotion or a group of emotions felt. General conclusions of the panelIn conclusion, this study has demonstrated that it is possible to measure the 4 biological biomarkers related to the emotional response in saliva, as part of a panel and in a non-medical context in a reduced sampling time (< 30 min).Our study shows that olfactory stimulations induce variations in salivary biomarkers. That two different stimulations generate different salivary biomarker responses.Individual measurement of BMs following an olfactory stimulus can be carried out simultaneously with classic questionnaires on emotions. Preliminary statistical analyses indicate that:- Emotions can generate particular biomarker profiles in saliva and this a few minutes after the olfactory stimulation.- Measuring BMs in saliva allows us to capture in a very sensitive and objective way the diversity of responses to an olfactory stimulus in a population.- The number of individuals tested (N=30) in this preliminary study makes it possible to identify trends but these must be tested by a larger study (N=120 individuals planned in the current EMJOY study).- It is possible to assess a person's sensitivity to olfactory stimuli; that is, people who systematically respond strongly or weakly to olfactory stimuli. Example 6: Saliva samplesDuring this study and in order to reduce experimental bias due to the sampling context, we established a saliva sampling protocol for individuals participating in a panel of olfactory tests outside a laboratory, in the field.Protocol:1- The saliva sample should preferably be taken in a seated position and after having rested for 5 to 10 minutes beforehand.2- The sample should preferably be taken 15 minutes after the last drink, food, cigarette / e-cigarette, candy, chewing gum, brushing your teeth or rinsing your mouth.3- Do not cough or clear your throat (otherwise you will get spit) but let the saliva come; salivate the necessary quantity, making sure to have liquid and not foam.4- The sample must be collected in a suitable and sterile tube, preferably cleaned after collection with a virucide / bactericide and then immediately placed at +4°C.5- The person sampled must not suffer from symptoms and signs of active oral inflammation, advanced periodontitis or severe gingivitis or have a proven or suspected chronic infectious disease.6- It is preferable to collect at least 2 saliva samples, one before and one after stimulation. Example 7: Study design and objective(See Figure 16)Study design and objective.a) Panel sudyFor the panel study, a questionnaire was used to collect participants' emotions during olfactory stimulation. In this questionnaire, emotions were defined by precise semantics using specific words corresponding to six emotions that were to be evaluated. Emotion responses were discretized into two categories: negative responses, corresponding to the absence / response of not feeling any emotion towards the fragrance, and positive responses, grouping together responses from strong to moderate feeling towards the fragrance. We then studied variations of the four biomarkers in relation to responses to the emotions collected in the questionnaire. Global and individual analyses were carried out. These analyses have enabled us to define a biomarker-based emotional profiles induced by fragrances. Example 8: Salivary biomarkers’ ratios association.(See Figure 17)The four biomarkers (AA, Cort, DHEA, OXT) were reliably measured (mean coefficient of variation <9%) for 708 saliva samples including 236 unique participant / fragrance combinations and 3 sampling time points (1, 2 and 3) for each combination. Weak correlations (between DHEA and cortisol) or no correlation were observed between BMs showing that each biomarker varies independently and corroborating the importance and the individual contribution of each BM to the molecular signatures of emotions. Example 9: Principal component analysis (PCA) on biomarkers ratios (S2 / S1 and / or S3 / S1)(see Figure 18)Principal component analysis (PCA) on biomarkers ratios (S2 / S1 and / or S3 / S1). PCAs were performed using either the data set of all participants that scented the four fragrances or by fragrance analyzed separately. No pattern distribution was observed for any of the analyzed emotion. PCA analysis on BM ratios does not allow to explain individual emotions. Example 10: ANOVA analysis on salivary BM ratios among different groups of emotion valence (0 to 2) in the questionnaire.(see Figures 19A-19C)ANOVA analysis on salivary BM ratios among different groups of emotion valence (0 to 2) in the questionnaire. Groups responding to different levels of emotional valence were compared regarding their biomarkers ratios. ANOVA analysis was performed using each emotion separately or the six emotions together as either continuous or discrete variables on the entire cohort (59 participants x 4 fragrances = 236 unique combinations). Results on the above figures are examples of some biomarker’s ratios that can be predicted by all 6 emotions together in discrete ANOVA analysis when looking at all fragrances together. A: DHEA S3 / S1 ratio for the Comforted emotion; B: DHEA S3 / S1 ratio for the Dynamised emotion; and C: DHEA S2 / S1 for the Comforted emotion. Table shows significant p-values (p<0.1) found for each emotion and the corresponding biomarker’s ratio. Example 11: Participants were nested together based on salivary biomarker’s ratio profiles using hierarchical eans (hk-means).(see Figures 20A-20D)This approach allows to 1) compute the hierarchical clustering, 2) cut the tree into k-clusters, 3) compute the center (i.e the mean) of each cluster, 4) perform k-means by using the set of cluster centers (defined in step 3) as the initial cluster centers and 5) optimize the clustering.This means that the final optimized partitioning obtained at step 4 might be different from the initial partitioning obtained at step 2. We performed k-means and hk-means using either 8 BMs ratios (4 BMs x 2 sampling ratios S2 / S1 and S3 / S1) or 4 BMs ratios (4 BMs using only S2 / S1sampling ratio). We also performed hk-means using different thresholds for the hierarchization step in an attempt to obtain different nested architectures and further we also performed sub-clustering applying hk-means clustering on groups formed by a former hk-means.Examples of some hk-means performed on each fragrance separately using the biomarkers ratios S2 / S1 are: (A) fragrance J3 nested into five clusters composed of 21, 22, 7, 9 and 1 individuals, respectively; (B) fragrance J4 nested into four clusters composed of 16, 25, 16 and 2 individuals, respectively; (C) fragrance J5 nested into five clusters composed of 24, 24, 4, 3, and 4 individuals, respectively and; (D) fragrance J6 nested into four clusters composed of 28, 25, 3 and 3 individuals, respectively. Example 12: Participants were nested together based on salivary biomarker’s ratio profiles using hierarchical k-means (hk-means)(see Figure 21)This approach allows to 1) compute hierarchical clustering, 2)cut the tree in k-clusters, 3) compute the center (i.e the mean) of each cluster, 4) do k-means by using the set of cluster centers (defined in step 3) as the initial cluster centers and 5) optimize the clustering.This means that the final optimized partitioning obtained at step 4 might be different from the initial partitioning obtained at step 2. This example shows the hk-means clustering performed on the total cohort (59 participants x 4 fragrances: 236 unique combinations) using the biomarkers ratios 2 / 1. Participants were nested into five clusters composed of 69, 46, 18, 95 and 8 individuals. Example 13: Salivary biomarkers’ ratio profiles for each cluster nested by hk-means performed on the total cohort (59 participants x 4 fragrances: 236 unique combinations) using the biomarkers’ ratio S2-S1(see Figures 22A-22F)Spiderplots show the absolute ratios for each cluster (A: cluster 1 – 69 individuals; B: cluster 2 - 46 individuals; C: cluster 3 – 18 individuals, D: cluster 4 - 95 individuals and E: cluster 5 - 8 individuals). Comparison of the distribution of biomarkers ratios (S2 / S1) among groups in violinplots. Light blue: cluster 1; green: cluster 2; violet: cluster 3; orange : cluster 4 and dark blue: cluster 5. Example 14: Temporal evolution of individual salivary biomarkers’ ratios (S1 / S1; S2 / S1; S3 / S1) in hk-means clusters performed on the total cohort (59 participants x 4 fragrances: 236 unique combinations) using the biomarkers’ ratios S2 / S1(see Figure 23)Curves represent individual biomarkers’ ratios in different sampling (S1, S2 and S3) normalized to S1 biomarkers’ concentrations of each participant. Curves are colored by fragrance and the mean of each cluster is represented by the bold black line. Standard deviation is represented by dotted black lines. Example 15: Questionnaire-based emotional profiles in hk-means clusters(See Table 2):  Chi2 pval DynamisedHappyRelaxedComfortedSensualConfidentALL>0.5>0.5>0.5>0.5>0.50.23J30.3>0.50.31>0.5>0.50.23J4>0.50.2>0.50.34>0.50.23J5>0.5>0.5>0.5>0.5>0.50.29J6>0.5>0.5>0.50.080.140.27 Table 2: Questionnaire-based emotional profiles in hk-means clusters. Differential emotional responses were evaluated among groups formed by hk-means performed either on the whole cohort (236 unique combinations participants / fragrances) or by fragrance separately using Chi2 test. Data of the p-values are shown in the table. Emotional responses were discretized (questionnaire response O = negative response to a given emotion; questionnaire response 1 and 2 were nested together = positive response to a given emotion). Example 16: Principal component analysis (PCA) was performed on discretized binary emotions (questionnaire response O = negative response to a given emotion; questionnaire response 1 and 2 were nested together = positive response to a given emotion) used as quantitative variables and correlations to different variables were analyzed (B – G)(see Figures 24A to 24G)(A) The figure shows PCA performed using data from the whole cohort (236 unique combinations). The clusters projected were nested using the hk-means clusterization performed using whole data of all participants that scented the four fragrances. No pattern distribution was observed for any of the analyzed variable: (B) Hk-means clusters groups; (C) Evaluation; (D) Intensity; (E) Same as (B) with some random effects to visualize the presence of multiples measures with identical coordinates due to the limited number of variables (6 emotions) with only two possible values (0 ; 1). Large point correspond to true values; smaller points are randomly distributed around their real values); (F) Fragrances; (G) Gender. A similar analysis was also performed using hk-means clusterization performed by fragrance separately. Example 17: Clusters formed by hk-means using salivary BMs S2 / S1 ratio and using data by fragrances separately(see Figure 25)For each fragrance (J3 to J6) and for each cluster, the figure shows on the left-hand column the profile of emotional responses in radar plots (data is represented by the % of responding participants to the total number of individuals nested in each cluster); in the center and in the right-hand columns the log2 of mean BM ratios profiles (S2 / S1) are shown either in histograms (center) or spiderplots (right).   Example 18: Different salivary BM profiles correspond to different emotion profiles(see Figure 26)Clusters were formed by hk-means using BMs S2 / S1 ratio and using data by fragrances separately. Four clusters were formed using data from fragrance J6 (G1 to G4). For each cluster the figure shows on the left the profile of emotional responses in radar plots (data is represented by the % of responding participants to the total number of individuals nested in each cluster). In the center and in the right the log2 of mean BM ratios profiles (S2 / S1) are shown either in histograms (center) or spiderplots (right).Clusters were considered highly lowly Confident; G3 highly Happy, Relaxed, Comforted and Confident and lowly Dynamised and Sensual; G4 highly Happy, Relaxed, Dynamised and Confident and lowly Comforted. Clusters were considered highly responsive to a given emotion when more than 66% of the participants in each cluster responded positively to an emotion; moderately responsive when >33%<66% of the participants in each cluster responded positively to an emotion and lowly responsive when less 33% of the participants in each cluster responded positively to an emotion. Each cluster nested using only BM ratio (S2 / S1) data shows an unique emotional response profile, i.e. G1 mainly highly Happy, moderately Relaxed, Comforted and lowly Dynamised, Confident and Sensual; G2 highly Happy, Relaxed and Comforted, moderately Dynamised and Sensual. Example 19: Salivary BM profiles are shared between clusters formed by independent hk-means performed by fragrances separately and they present similar emotional response profiles(see Figure 27)Clusters were formed by hk-means using BMs S2 / S1 ratio and using data by fragrances separately. The figure shows cluster 2 (G2) formed using data from fragrance J5 (J5G2); cluster 2 (G2) formed using data from fragrance J6 (J6G2); cluster 2 (G2) using data from fragrance J3 (J3G2) and cluster 2 using data from fragrance J4 (J4G2). For each cluster the figure shows on the left the profile of emotional responses in radar plots (data is represented by the % of responding participants to the total number of individuals nested in each cluster). In the center and in the right the log2 of mean BM ratios profiles (S2 / S1) are shown either in histograms (center) or spider plots (right). Clusters were considered highly responsive to a given emotion when more than 66% of the participants in each cluster responded positively to an emotion; moderately responsive when >33%<66% of the participants in each cluster responded positively to an emotion and lowly responsive when less 33% of the participants in each cluster responded positively to an emotion.Clusters J5G2 and J6G2 present similar BM profiles (small decreased / stable levels of alpha-amylase, cortisol and DHEA concomitant with increased levels of oxytocin) and very close emotional response profiles (J5G2 highly happy, moderately relaxed, comforted and dynamised and lowly confident and sensual versus J6G2 highly happy and relaxed, moderately comforted dynamised and sensual and lowly confident). Example 20: Individual case studies.A)Participant E-044(see Figure 28)Salivary biomarkers’ ratio (S2 / S1) profiles are associated to specific emotions profiles at individual levels. After scenting the fragrances J3 and J4, participant E-044 presented the same biomarker ratio profile (low decreased / stable levels of alpha-amylase, high decreased levels of cortisol and DHEA and low increased / stable levels of oxytocin). This biomarker ratio profile was associated with an absent emotional responsiveness. Further, when the same participant was stimulated by fragrances J5 and J6 the participant presented the following biomarker ratio profile: high decreased levels of alpha-amylase, moderate / high increased levels of cortisol, DHEA and oxytocin. In this case, fragrances triggered an emotional response associate do moderate happy and high comforted and relaxed.Clusters J3G2 and J4G2 present similar BM profiles (accentuated decreased levels of alpha-amylase, cortisol and DHEA concomitant with low decreased / stable levels of oxytocin) and very close emotional response profiles (J3G2 highly happy and dynamised, moderately confident and lowly relaxed, comforted and sensual versus J6G2 highly happy and dynamised, moderately relaxed and comforted and lowly confident and sensual). B) Participant E-078(see Figure 29)Salivary biomarkers’ ratio (S2 / S1) profiles are associated to specific emotions profiles at individual levels. After scenting the fragrances J3 and J5, participant E-078 presented the similar biomarker profiles (for J3 low increased / stable levels of alpha-amylase, moderated / high increased levels of cortisol and DHEA and low / moderated decreased levels of oxytocin; for J5 low / moderate decreased levels of alpha-amylase, moderated / high increased levels of cortisol and DHEA and low / moderated decreased levels of oxytocin). These biomarker profiles were associated with an emotional response associated to highly relaxed and moderately happy and confident. After being stimulated by the fragrances J4 and J6, participant presented moderate to high decreased levels of alpha-amylase and oxytocin while stable to moderate decreased levels of cortisol and DHEA similar biomarker profiles. This biomarker ratio profile was associated to the emotions happy and relaxed for J4 and happy, relaxed and confident for J6. Conclusion Examples 18 to 20 (see figures 26 to 29)Salivary biomarkers’ ratios (S2 / S1 and S3 / S1), following the olfactory stimulations, were used to cluster together individual participants based exclusively on their biomarkers profiles. Clusterization was performed using either biomarkers’ ratios as continuous (hk-means, CART) or discrete variables (CART). Then, we analyzed profiles of emotional responses into the clusters.In the hk-means clusters (BM as continuous variables), the different salivary BM profiles correspond to different emotion profiles (see Figure 26)). In this case, each cluster nested using only BM ratio (S2 / S1) data shows an unique emotional response profile, i.e. G1 mainly highly Happy, moderately Relaxed, Comforted and lowly Dynamised, Confident and Sensual; G2 highly Happy, Relaxed and Comforted, moderately Dynamised and Sensual and lowly Confident; G3 highly Happy, Relaxed, Comforted and Confident and lowly Dynamised and Sensual; G4 highly Happy, Relaxed, Dynamised and Confident and lowly Comforted, Sensual. Further, the salivary BM profiles are shared between clusters formed by independent hk-means (see figure 27)) performed by fragrances separately. In this case, they present similar emotional response profiles as well. For example, Clusters J5G2 and J6G2 present similar BM profiles (small decreased / stable levels of alpha-amylase, cortisol and DHEA concomitant with increased levels of oxytocin) and very close emotional response profiles (J5G2 highly happy, moderately relaxed, comforted and dynamised and lowly confident and sensual versus J6G2 highly happy and relaxed, moderately comforted dynamised and sensual and lowly confident).Clusters J3G2 and J4G2 present similar BM profiles (accentuated decreased levels of alpha-amylase, cortisol and DHEA concomitant with low decreased / stable levels of oxytocin) and very close emotional response profiles (J3G2 highly happy and dynamised, moderately confident and lowly relaxed, comforted and sensual versus J6G2 highly happy and dynamised, moderately relaxed and comforted and lowly confident and sensual).Association of emotions responses to salivary biomarker profiles was also found at the individual level (See figures 28 and 29). For example, (1) scenting the fragrances J3 and J4, participant E-044 presented the same biomarker ratio profile (low decreased / stable levels of alpha-amylase, high decreased levels of cortisol and DHEA and low increased / stable levels of oxytocin). This biomarker ratio profile was associated with an absent emotional responsiveness. Further, when the same participant was stimulated by fragrances J5 and J6 the participant presented the following biomarker ratio profile: high decreased levels of alpha-amylase, moderate / high increased levels of cortisol, DHEA and oxytocin. In this case, fragrances triggered an emotional response associate do moderate happy and high comforted and relaxed. (2) Following olfactory stimulation with the fragrances J3 and J5, participant E-078 presented the similar biomarker profiles (for J3 low increased / stable levels of alpha-amylase, moderated / high increased levels of cortisol and DHEA and low / moderated decreased levels of oxytocin ; for J5 low / moderate decreased levels of alpha-amylase, moderated / high increased levels of cortisol and DHEA and low / moderated decreased levels of oxytocin). These biomarker profiles were associated with an emotional response associated to highly relaxed and moderately happy and confident. After being stimulated by the fragrances J4 and J6, participant presented moderate to high decreased levels of alpha-amylase and oxytocin while stable to moderate decreased levels of cortisol and DHEA similar biomarker profiles. This biomarker ratio profile was associated to the emotions happy and relaxed for J4 and happy, relaxed and confident for J6. Example 21: Salivary Oxytocin and Alpha-Amylase levels associate with valuation.(see Figure 30)Alpha amylase and Oxytocin ratio (S2 / S1) levels were associated to the valuation of the fragrances (the valuation parameter was evaluated for all fragrances together) by the participants. The mean ratios of alpha-amylase (AA) were lower in participants responding 5 or 6 in questionnaire for the valuation of the fragrances (low valuation) compared to those who responded 3 or 4 (medium valuation) and 1 or 2 (high valuation). On the opposite, oxytocin ratios were higher in participants responding 5 / 6 to the fragrance valuation compared to those who responded 3 / 4 or 1 / 2. This analysis was performed using data from the entire cohort (236 unique combinations of participants / fragrances). Example 22: Discretization of salivary biomarkers ratios (S2 / S1) for new clustering approaches.(see Figure 31)In order to try to optimize the signal to noise ration of biomarkers ratios inside clusters (as observed in clusters formed by k-means or hk-means - see figure “Temporal evolution of individual salivary biomarkers’ ratios (S1 / S1; S2 / S1; S3 / S1) in hk-means clusters ”, Figure 23) we discretized the biomarkers’ ratios (S2 / S1) taking into account our mean coefficient of variation (~10%). Biomarkers ratios were discretized into three categories (i.e.<0.9 = decrease; ≥0.9≤1.1 = stable and >1.1 = increase). Based on this approach, the 236 unique combinations of participants / fragrances can be classified into 81 putative categories by combining the discretized biomarkers’ ratios (e.g. category 1 = alpha-amylase decrease, cortisol decrease, DHEA decrease, oxytocin decrease; category 2 = alpha-amylase decrease, cortisol decrease, DHEA decrease, oxytocin increase, and so on). Example 23: Classification of salivary biomarkers’ ratio (S2 / S1) profiles for emotion identification / predictionA) for the emotion Dynamised(see Figure 32)We used the “Classification and Regression Trees” (CART) algorithm to classify biomarkers’ ratio profiles responding positively or negatively to the evaluated emotions. CART method was performed using either biomarkers’ ratios S2 / S1 as continuous or discrete variables. Using this method, participants were classified into leaves with either positive (green) or negative (blue) responses to a given emotion. The classification was based on biomarkers’ ratio (S2 / S1) profiles. The numbers and information inside leaves are explained in the top right-hand corner of the figure. An example of CART for the Dynamised emotion using discrete variables is given. The analysis of green leaves responding positively to one emotion allows the identification of patterns of biomarkers ratio profiles corresponding to a given emotion (on the down left). In the opposite, analysis of blue leaves responding negatively to one emotion allows the identification of discrete biomarkers ratio profiles corresponding to the absence of a given emotion (on the down left).B) For the emotions Happy, Confident, Comforted and Relaxed.(see Figure 33)Classification of salivary biomarkers’ ratio (S2 / S1) profiles for emotion identification / prediction.We used the “Classification and Regression Trees” (CART) algorithm to classify biomarkers ratio (S2 / S1) profiles responding positively or negatively to the evaluated emotions. Examples of some decision trees generated by CART for the emotions Happy, Confident, Comforted and Relaxed. The CART method was performed using the entire cohort data (236 unique combinations of individuals / fragrances) and biomarkers’ ratios S2 / S1 as discrete variables  Example 24: Implicit test-based emotional profiles induced by fragrances.(See Figure 34)Implicit test was performed on 80 volunteers in a separated cohort that scented the same fragrances (J3, J4, J5 and J6) used to develop our biomarker-based emotional profiles. Two parameters were measured with the implicit test: the GO / NO GO percent (the number of people who validated the emotion for each fragrance) and the strength of the association (the speed of response to validate the emotion for each fragrance). The GO / NO GO percent was represented by the angle (emotions were presented 3 times during the test and the % represents the number of clicks for each emotion). The strength of association (speed of clicks) was represented by the size and the color. An association is considered high when rapid responses are given by the participant. Dark green : excellent (>80 percentile, emotion shown in bold); green : good (>50 percentile); grey : low (<50 percentile).This implicit test made it possible to define an emotional profile for each of the four fragrances studied. The molecular biomarker-based and implicit test-based emotional profiles induced by fragrances were compared. Fragrances emotion profiles determined by BM patterns are similar to implicit association test.The study was a cross-sectional evaluation including 59 participants. Each participant scented four different fragrances (one fragrance per day of participation. Fragrances were named J3, J4, J5 and J6 hereafter) and the order of scent was randomized. The participation consisted in 1) a 5-minute sit-down rest before the first sampling of saliva (5 minutes before olfactory stimulation, 2) scent of the fragrance and response to the questionnaire Conclusions:Salivary biomarker’s patterns for emotion prediction / identification. Identification of patterns of biomarkers’ ratio (S2 / S1) profiles corresponding to some of emotions identified (Happy, Confident, Comforted and Relaxed). Patterns were identified from the decision trees generated by CART performed on the entire cohort data (236 unique combinations of individuals / fragrances) and biomarkers’ ratios S2 / S1 as discrete variables Example 25: Performances of salivary biomarker’s patterns for emotion prediction / identification(See Table 3)EmotionConditionAACORTDHEAOXTTrue Positive / NegativeTotalIncidenceHappyPositive I  546879%HappyPositive DS D324276%HappyPositive DSSIS344772%HappyPositiveDIDSDIIS426663%         HappyNegativeSDSDIIS91464%         ConfidentPositiveDS II162662%ConfidentPositiveS ISD4757%         ConfidentNegative   S476671%ConfidentNegativeI  DI151979%ConfidentNegativeDS DDI406166%ConfidentNegativeD ISD151979%ConfidentNegativeDS SI223858%         ComfortedPositive ISD 254063%ComfortedPositiveIISIS 91464%ComfortedPositiveDSISI111479%ComfortedPositiveDSSII5771%         ComfortedNegative D  498359%ComfortedNegativeDSISISDS305060%ComfortedNegativeDSSSI111765%ComfortedNegativeDSIII91464%         RelaxedPositive   DS8312765%RelaxedPositiveDSI I203165%         RelaxedNegativeDSDS I366655%RelaxedNegativeI  I91275% Table 3Table 3: Salivary biomarker’s patterns for emotion prediction / identification. Identification of patterns of biomarkers’ ratio (S2 / S1) profiles corresponding to some of emotions identified (Happy, Confident, Comforted and Relaxed). Patterns were identified from the decision trees generated by CART performed on the entire cohort data (236 unique combinations of individuals / fragrances) and biomarkers’ ratios S2 / S1 as discrete variables.  Example 26: Biomarker-based emotional profiles induced by fragrances(See Table 4)Biomarker-based emotional profiles induced by fragrances. Salivary biomarker-based patterns associated to emotional responses were used to identify the emotional profiles of the tested fragrances (J3, J4, J5 and J6) in the cohort. Biomarkers-based emotional responses for each fragrance correspond to % of participants presenting the identified biomarker-based emotional profiles from the total number of participants that scented a given fragrance. Patterns used in this approach were those identified from the decision trees generated by CART performed on the entire cohort data (236 unique combinations of individuals / fragrances) and biomarkers’ ratios S2 / S1 as discrete variables. Table 4: Biomarker-based emotional profiles induced by fragrances  Example 27: Performances on prediction accuracy(See Table 5)       DynamisedHappyConfidentComfortedRelaxedTrue positive1041622050103Total positive13516784115136Sensitivity77%97%24%43%76%True negative5291399945Total negative10169152121100Specificity51%13%91%82%45% Table 5: Performances of salivary biomarker’s patterns for emotion prediction / identification.We used the patterns of biomarkers to identify emotions of participants taking into account only the discrete biomarkers ratio (S2 / S1) profiles of participants and then we compared the biomarkers-based emotional responses to the questionnaire-based emotional responses. For each emotion, we calculated a sensibility (% True positives / Total positives) and a specificity (% True negatives / Total negatives) for the given emotions. Patterns used in this approach were those identified from the decision trees generated by CART performed using the entire cohort data (236 unique combinations of individuals / fragrances) and biomarkers’ ratios S2 / S1 as discrete variables. ConclusionWe reliably measured (mean cv<9%) 4 BMs (AA, Cort, DHEA, OXT) for 732 saliva samples including 244 unique combinations of participant / fragrance and 3 sampling time points for each combinationNo / weak correlation is observed between BMs corroborating the importance and the individual contribution of each BM to the WSWe used data from 59 subjects who scented all the four tested fragrances (236 unique combinations of participants / fragrances) to develop methods (including but not limited to PCA, k-means, hk-means, CART, etc) capable of identifying biomarker-based patterns associated to emotional responses.Data from participants consisted in the measured biomarkers (AA, CORT, DHEA, OXT) mainly their relative measures (either the ratio sampling 2 to 1; or sampling 3 to1; or the combination of sampling 2 to1 and sampling 3 to 1). Biomarkers ratios' can be used as either continuous or discrete variables.Fragrances trigger BM responses and the participants can be categorized in accordance with their BM profilesDifferent BM profiles correspond to different emotion profilesSome BM profiles are shared by different fragrances and they correspond to similar emotion profilesOXT and AA levels associate to appreciation / valuation of fragrances Discretization of BMs ratios allows a better classification of BM profilesDifferent patterns of BM profiles associate to positive and negative feeling of emotionsDifferent BM patterns can be used for emotion predictionFragrances emotion profiles determined by BM patterns are similar to implicit association test ConclusionThe present invention has highlighted the general feasibility of saliva sampling, as well as the discovery of biological molecular signatures (BMs signature) allowing to differentiate emotions. The prospects opened up by this invention in terms of precision of identification of emotions by a salivary biomarker signature are very encouraging. According to the present invention, it is also possible to identify profiles of people responding more or less strongly biologically to an olfactory solicitation. This perspective could open the way to a method of selecting panelists or customers on a biological measurement basis. Bibliography 1. Buford TW, Willoughby DS. Impact of DHEA(S) and cortisol on immune function in aging: a brief review. Appl Physiol Nutr Metab. 2008;33(3):429-33.2. Vining RF, McGinley RA. The measurement of hormones in saliva: possibilities and pitfalls. J Steroid Biochem. 1987;27(1-3):81-94.3. Liu Y, Papadopoulos V. BIOSYNTHESIS AND REGULATION OF DEHYDROEPIANDROSTERONE (DHEA) IN BRAIN. Endocrinologia molecular. 2005:135-45.4. Morales AJ, Nolan JJ, Nelson JC, Yen SS. Effects of replacement dose of dehydroepiandrosterone in men and women of advancing age. J Clin Endocrinol Metab. 1994;78(6):1360-7.5. McCormack SE, Blevins JE, Lawson EA. Metabolic Effects of Oxytocin. Endocr Rev. 2020;41(2).6. Poisson B. Perspective biopsychologique systémique des émotions de base1. Santé mentale au Québec. 2016;40(3):223-44.7. Granger DA, Kivlighan KT, el-Sheikh M, Gordis EB, Stroud LR. Salivary alpha-amylase in biobehavioral research: recent developments and applications. Ann N Y Acad Sci. 2007;1098:122-44.  

Claims

1. A method for identifying / determining a reference signature associated to an emotion / emotional response and / or to a group of emotions and / or to a fragrance and / or to a group of fragrances felt by a subject after an olfactory stimulation of said subject selected from a statistical significant cohort / panel of individuals, said method comprising the steps of : A) calculating the BM ratio’s (named S2 / S1) for each of the salivary biomarkers (BMs) alpha-amylase, cortisol, dehydroepiandrosterone (DHEA) and oxytocin by a method comprising the steps of:1) collecting a first saliva sample (named SI) of said subject before said stimulation;2) stimulating the subject with said olfactory stimulus;3) collecting at least a second saliva sample (named S2), wherein the salivary sample SI and S2 are collected on morning.4) measuring the quantity and / or concentration of each of said BMs present in the SI and S2 saliva samples;5) calculating the BM ratio’s (named S2 / S1) for each of said salivary BMs; and6) discretizing the variable S2 / S 1 obtained for each of said salivary BMs, B) from the results obtained in step A)6), determining the salivary BM signature associated to said stimulus and said subject ; C) ) repeating the step A) and B for all the subjects / individuals of said statistical significant cohort / panel with the same stimulus ; and  D) identifying / determining by standard statistical methods the reference signature associated to said emotion and / or group of emotions and / or fragrance and / or group of fragrances from the salivary BM signature associated to said stimulus and determined for all the subjects of said statistical significant cohort / panel..2) The method according to claim 1, wherein step D) is conducted on distinct groups resulting from the partition of said cohort / panel using standard partitioning statistical methods and allowing the identification or the determination of reference signature for each of these distinct groups by CART (classification and regression tree),k-means or hk-means (hierarchical k-means) statistical classification methods.3) The method according to one of claims 1 or 2, wherein said emotion / emotional response and / or groups of emotions and / or fragrance and / or group of fragrances felt by a subject after an olfactory stimulation is selected from the group of addict, elegant, confident, happy, energized, unique, sensual, comforted and relaxed emotion / emotional response, or combination thereof.4) The method according to one of claims 1 to 3, wherein said individual measurement of BM ratio’s S2 / S1 following said olfactory stimulus is carried out simultaneously with questionnaires on emotions felt by the subject.5) The method according to one of claims 1 to 4, wherein said method comprises an implicit test performed on said statistical significative cohort of subjects that scent the same fragrance or group of fragrances used to develop said biomarker-based emotional profiles. 6) A method for predicting, identifying or classifying the emotion or group of emotions felt by a subject after stimulation by an olfactory stimulus, from saliva samples from said subject, the method comprising the following steps of :A) calculating the BM ratio’s S2 / S1 for each of the salivary BMs alpha-amylase, cortisol, dehydroepiandrosterone (DHEA) and oxytocin by a method comprising the steps of:1) collecting a first saliva sample (named S1) of said subject before said stimulation;2) stimulating the subject with said olfactory stimulus;3) collecting at least a second saliva sample (named S2), wherein the salivary sample S1 and S2 are collected on morning.4) measuring the quantity and / or concentration of each of said BMs present in the S1 and S2 saliva samples;5) calculating the BM ratio’s (named S2 / S1) for each of said salivary BMs; and6) discretizing the variable S2 / S1 obtained for each of said salivary BMs,B) from the results obtained in step A)6), determining the salivary BM signature associated to said stimulus and said subject ; E) comparing the salivary signature obtained in step B) with reference salivary signatures obtained for these 4 BMs ratios by the method of claim 1, said reference salivary signatures being identified to be associated with an emotion or group of emotions; and F) based on this comparison, predicting, identifying or classifying the emotion or group of emotions felt by said subject after stimulation with said olfactory stimulus,.7) A method for determining the ability of an olfactory stimulus to provoke / elicit in a subject a desired emotion or group of emotions from saliva samples of said subject, the method comprising the following steps of : A) calculating the BM ratio’s S2 / S1 for each of the salivary BMs alpha-amylase, cortisol, dehydroepiandrosterone (DHEA) and oxytocin by a method comprising the steps of:1) collecting a first saliva sample (named SI) of said subject before said stimulation;2) stimulating the subject with said olfactory stimulus;3) collecting at least a second saliva sample (named S2), wherein the salivary sample S1 and S2 are collected on morning.4) measuring the quantity and / or concentration of each of said BMs present in the SI and S2 saliva samples;5) calculating the BM ratio’s (named S2 / S1) for each of said salivary BMs; and6) discretizing the variable S2 / S 1 obtained for each of said salivary BMs ; B) from the results obtained in step A)6), determining the salivary BM signature associated to said stimulus and said subject ; G) identifying reference salivary BMs signature associated to the emotion or group of emotions which is desired to be provoked / elicited after stimulation by an olfactory stimulus, said reference signature being obtained by the method according to one of claims 1 to 5 ; H) comparing the salivary BMs signature obtained in step H) with said reference BMs signature identified in step I); and I) from this comparison, determining whether said tested stimulus is capable to provoke / elicit the desired emotion or group of emotions.8) The method according to one of claims 1 to 7, wherein in step A)6), the variable S2 / S1 obtained for each of said salivary BMs is discretized as exemplified below :  a) if the ratio S2 / S1 is < inferior to 0.9, as “D” (for decreased); if the ratio S2 / S1 is superior to 0.9 and inferior or equal to 1.1 as “S” (for stable) ; and if the ratio S2 / S1 is superior to 1.1 as “I ” (for increased), or  b) “0” Negative response, and “1” (Positive response).9) The method according to one of claims 1 to 8, wherein the elapsed time between S1 and S2 saliva samples collection is inferior to 60 mn. 10) Kit comprising: a) means for collecting separately at least 2 saliva samples from a subject; and b) means for measuring the quantity / concentration of each of the following 4 BMs present in said saliva samples: alpha-amylase, cortisol, DHEA and oxytocin.

Images