Methods and systems for determining and monitoring a blood-related individual anaerobic threshold, method for training a corresponding data model, device, computer program, computer readable medium

Abstract

The invention relates to a method for determining a blood-related, individual anaerobic threshold (IAT) of a test subject (10). For this purpose, first data (D1) is obtained comprising performance values (P) describing a performance of the test subject (10), time values (T) describing a respective associated detection time and respective associated saliva lactate concentration values (KS). The first data (D1) exceeds a maximum saliva lactate concentration value (KS). The blood-related, individual anaerobic threshold (IAT) is obtained on the basis of the first data (D1) and using a data model (M). The invention also relates to a method for monitoring a blood-related, individual anaerobic threshold (IAT) of a test subject (10). The invention further relates to a method for training a data model (M) for determining a blood-related, individual anaerobic threshold (IAT). The invention further relates to a data-processing device (22), to a computer program (32), and to a computer-readable medium (30). Furthermore, the invention relates to a system (16) for determining a blood-related, individual anaerobic threshold (IAT) and to a system for monitoring a blood-related, individual anaerobic threshold (IAT).

Description

[0001] IVOCLAR VIVADENT AG Munich, November 28, 2025

[0002] Our reference number: I10395EP / SE

[0003] IVOCLAR VIVADENT AG Bendererstrasse 2, 9494 Schaan, Liechtenstein

[0004] Methods and systems for determining and monitoring a blood-related, individual anaerobic threshold, methods for training a data model to determine a blood-related, individual anaerobic threshold, data processing device, computer program, computer-readable medium

[0005] The invention relates to a method for determining a blood-related, individual anaerobic threshold of a subject.

[0006] Furthermore, the invention relates to a method for monitoring a blood-related, individual anaerobic threshold of a subject.

[0007] Furthermore, the invention relates to a method for training a data model to determine a blood-related, individual anaerobic threshold.

[0008] Furthermore, the invention relates to a device for data processing, a computer program and a computer-readable medium.

[0009] The invention also relates to a system for determining a blood-related, individual anaerobic threshold of a subject and to a system for monitoring a blood-related, individual anaerobic threshold of a subject.

[0010] In this context, the blood-related individual anaerobic threshold of a subject is a term from sports physiology that describes the subject's energy metabolism. The blood-related anaerobic threshold, which can also be referred to as the threshold value, can differ for each subject and is therefore individual. The blood-related individual anaerobic threshold is determined based on a lactate concentration.

[0011] SE:TOP is defined in the subject's blood, which is why this individual anaerobic threshold is also referred to as blood-related. In this context, lactate is produced in the subject's muscles as a result of physical exertion and from there enters the bloodstream. This means that lactate is produced when the subject exerts themselves, with generally more lactate being produced at higher intensities than at lower intensities. In addition to lactate production, the human body is also capable of breaking down lactate. The anaerobic threshold describes the highest possible intensity of exertion or power output that a subject can sustain while maintaining a state of equilibrium between lactate production and breakdown. At the anaerobic threshold, lactate production and breakdown are just barely in a state of dynamic equilibrium.This means that a test subject can maintain a performance level below the anaerobic threshold for a relatively long time without fatigue. Above the anaerobic threshold, lactate production, i.e., the accumulation of lactate in the blood, occurs faster than the human body can break it down. The anaerobic threshold is usually expressed as a power output or performance value. Lactate concentration, for example, is given in mmol / liter.

[0012] The blood-related individual anaerobic threshold is of great importance in performance diagnostics, i.e., in assessing the physical performance of a subject. Furthermore, the blood-related individual anaerobic threshold is frequently used for training control. The training goal can be to shift the blood-related individual anaerobic threshold towards higher values. For this purpose, training intensity can be specifically adjusted in relation to the blood-related individual anaerobic threshold, for example, as a percentage of the blood-related individual anaerobic threshold. The subject can thus individualize and efficiently control their training.

[0013] The blood-based, individual anaerobic threshold is typically determined using a so-called incremental exercise test. In this test, the subject performs a physical activity, for example, on an ergometer. The workload is increased incrementally at regular intervals. For example, the subject starts at a workload of 70 watts, which is increased by 30 watts every three minutes. For each workload increment, a blood sample, e.g., from the fingertip or earlobe, is taken by an assistant, and the lactate concentration is measured. The incremental exercise test ends when the subject has to stop due to exhaustion. By plotting the lactate concentrations measured at each workload increment against the workload increments and / or against time, the blood-based, individual anaerobic threshold can be determined based on the trend in lactate concentration.

[0014] It goes without saying that the performance level test can be unpleasant for the subject not only from a sporting point of view, but also that the necessary blood draws can be painful for the subject and cause effort for the support person during the performance level test.

[0015] The object of the present invention is therefore to simplify the determination of a blood-related, individual anaerobic threshold of a subject.

[0016] The task is solved by a procedure for determining a blood-related, individual anaerobic threshold of a subject, the procedure comprising:

[0017] - Obtaining initial data comprising performance values ​​describing the subject's performance, time values ​​describing a corresponding recording time, and corresponding salivary lactate concentration values, where the initial data extend beyond a maximum salivary lactate concentration value, and

[0018] Obtaining the blood-related, individual anaerobic threshold based on initial data and using a data model. In this context, obtaining initial data means determining or receiving the initial data. In contrast to the prior art, the method according to the invention uses salivary lactate concentration values, i.e., values ​​that describe the lactate concentration in the subject's saliva. Blood sampling and analysis are therefore unnecessary. This simplifies the process for both the subject and the support person. The initial data, comprising the salivary lactate concentration values, are converted into the blood-related, individual anaerobic threshold using a data model. This is based on the understanding that lactate can also be detected in the saliva of the subject during physical exertion.This means that the lactate produced due to exertion is also present in the saliva, and that lactate broken down through physiological degradation processes also leads to a decrease in the lactate concentration in the saliva. Under otherwise identical conditions, however, a salivary lactate concentration is always lower than a blood lactate concentration. Furthermore, the salivary lactate concentration and the blood lactate concentration are delayed, typically by 6 to 10 minutes. The data model used in the method according to the invention represents the relationship between the salivary lactate concentration and the blood lactate concentration. The fact that the initial data exceeds a maximum salivary lactate concentration value means that a maximum salivary lactate concentration is included in the initial data. The same applies to the power value associated with the maximum salivary lactate concentration and a corresponding time value.This takes into account that, as already mentioned, salivary lactate concentration lags behind blood lactate concentration. If the method according to the invention is used in conjunction with a performance-based exercise test, this means that the salivary lactate concentration must continue to be recorded even after the subject has reached their maximum performance, i.e., after the actual test has ended. This makes it possible to fully map the relationship between salivary lactate concentration and blood lactate concentration using the data model already mentioned. Overall, the method according to the invention allows for the precise and reliable determination of a subject's blood-related, individual anaerobic threshold, while simultaneously eliminating the need for blood sampling. This makes determining the blood-related, individual anaerobic threshold simpler and more convenient.

[0019] In connection with the inventive method for determining the blood-related, individual anaerobic threshold, salivary lactate concentration values ​​can be recorded, for example, using an intraoral sensor designed to detect salivary lactate concentration values ​​or a salivary lactate concentration. It is also possible to use an extraoral sensor in connection with the inventive method for determining the blood-related, individual anaerobic threshold. In this case, the subject must transfer saliva into a container to measure the salivary lactate concentration, for example, by spitting into the container. The salivary lactate concentration in the container is then measured.

[0020] The inventive method for determining the blood-related, individual anaerobic threshold can be executed locally on a data processing device located in the vicinity of the subject. Alternatively, the method for determining the blood-related, individual anaerobic threshold can also be executed on a data processing device located further away from the subject, for example, on a server, in particular a cloud server. Hybrid forms are also possible, i.e., parts of the process steps of the inventive method for determining the blood-related, individual anaerobic threshold can be executed locally, i.e., in the vicinity of the subject, and other parts of the process steps can be executed on a server.

[0021] The use of the data model can also include preprocessing the initial data. Such preprocessing can involve removing artifacts from the initial data. Alternatively or additionally, the initial data can be filtered or normalized. Furthermore, it is possible to subject the initial data to resampling during preprocessing. Another option is to align and / or segment the initial data in time. It is also possible to mask or attribute the initial data. Additionally, features can be extracted during preprocessing. The initial data can also be augmented and / or versioned, taking the data source into account. Overall, preprocessing improves the quality of the data used in conjunction with the data model.This allows the blood-related, individual anaerobic threshold to be obtained with high accuracy and reliability.

[0022] The data model can include a trained data model that uses the initial data as input and is trained to output the blood-related, individual anaerobic threshold. Training data for such a trained data model thus includes training data elements that encompass a blood-related lactate concentration and a saliva-related lactate concentration, each with associated power and time values. For example, the trained data model includes a trained artificial neural network. The artificial neural network can be implemented, for example, as a convolutional neural network (CNN), a recurrent neural network (RNN), or a transformer. Alternatively, the trained data model can include a support vector machine (SVM), a probabilistic model, or hybrid models.It is also possible to use trained data models obtained through transfer learning, fine-tuning, or distillation. Such a data model is particularly well-suited to at least implicitly representing relationships between blood lactate concentration and salivary lactate concentration. Furthermore, such a trained data model can be easily and reliably executed in the procedure for determining the blood-related, individual anaerobic threshold. The blood-related, individual anaerobic threshold can therefore be determined quickly and precisely.

[0023] In one example, the trained data model is specific to a performance level test, a sport, and / or a fitness level. This means that the data model has been specifically trained for the relevant performance level test, sport, and / or fitness level. Performance level tests can differ, for example, in the size of the time and workload increments used. Therefore, the trained data model can be particularly precise for that specific performance level test, sport, and / or fitness level. Consequently, it is also possible to have multiple trained data models, each specific to a performance level test, sport, and / or fitness level. Thus, when performing the procedure to determine the blood-related, individual anaerobic threshold, the appropriate trained data model can always be selected and applied.The data model used is always the one trained for the relevant performance level test, sport, and / or fitness level. For example, different data models can be used for swimming, running, cycling, and rowing. Similarly, different data models can be used for the fitness levels of non-athlete, recreational athlete, and professional athlete. This ensures that the method for determining the blood-related, individual anaerobic threshold works precisely and reliably in a wide variety of applications. Optionally, the trained data model can also use information about the subject's age, height, and / or weight. More generally, the trained data model can use medical history data describing the subject.In this context, medical history data can be used as additional parameters for a trained data model. This naturally requires that medical history data is also considered during the training of the data model. Alternatively, one or more trained data models can be provided that are specific to an age group, height range, and / or weight range of subjects. More generally, trained data models can be specific to certain groups of subjects, with these groups being defined based on medical history data. By utilizing these options, the method for determining the blood-related, individual anaerobic threshold can operate with particular precision and reliability in a wide variety of applications.

[0024] The data model can also include a regression model that describes the relationship between the blood-related individual anaerobic threshold and a saliva-related individual anaerobic threshold. Consequently, the saliva-related individual anaerobic threshold can be easily and reliably converted into a blood-related individual anaerobic threshold. The saliva-related individual anaerobic threshold can be determined based on the initial data. In this way, the blood-related individual anaerobic threshold can be determined precisely and reliably without the need for blood sampling or analysis. It is emphasized that the regression model concerns the anaerobic thresholds and not the lactate concentrations in blood and saliva.

[0025] It is understood that creating the regression model, which describes the relationship between the blood-related individual anaerobic threshold and a saliva-related individual anaerobic threshold, may require determining both the blood-related and saliva-related individual anaerobic thresholds in a large number of subjects. A test performed on a large number of subjects can also be called a cohort test. For this purpose, for example, a graded exercise test of the type described above can be used. The determination of the blood-related individual anaerobic threshold can be based on actual blood samples. Consequently, the relationship between the blood-related individual anaerobic threshold and the saliva-related individual anaerobic threshold can be determined based on the large number of subjects.This relationship is then independent of the individual subject, meaning it is valid at least for the group of subjects. Ideally, the relationship is even universally valid. This relationship can then be used in the form of a regression model to determine the blood-related, individual anaerobic threshold, eliminating the need for taking or analyzing blood samples.

[0026] According to one approach, it is possible to provide several regression models that describe a relationship between a blood-related, individual anaerobic threshold and a saliva-related, individual anaerobic threshold as alternatives, and to select the appropriate regression model during the procedure for determining the blood-related, individual anaerobic threshold. As already described above in connection with the trained data model, the different regression models can relate to different performance level tests, different sports, and / or different fitness levels.

[0027] Alternatively or additionally, the data model includes a regression model that describes the relationship between the blood-related, individual aerobic threshold and a saliva-related, individual aerobic threshold. Additionally or alternatively, the data model includes a regression model that describes the relationship between a maximum salivary lactate concentration value and a maximum blood lactate concentration value. In this context, the aerobic threshold describes the level of exertion at which the lactate concentration in the blood rises above the resting lactate concentration but can still be maintained in the aforementioned steady state by the body's own breakdown processes. In other words, the aerobic threshold describes the lowest exercise intensity or lowest power output of the subject at which the muscles are no longer working purely aerobically.The regression models used here therefore concern aerobic thresholds and / or maximum concentration values, but not lactate concentrations in blood and saliva. Using a regression model that describes a relationship between the blood-related individual aerobic threshold and a saliva-related individual aerobic threshold, the saliva-related individual aerobic threshold can be easily and reliably converted into a blood-related individual aerobic threshold. The saliva-related individual aerobic threshold can then be determined based on the initial data. Furthermore, using a regression model that describes a relationship between a maximum salivary lactate concentration and a maximum blood lactate concentration, the maximum salivary lactate concentration can be easily and reliably converted into a maximum blood lactate concentration.The maximum salivary lactate concentration can be determined based on the initial data. This allows for the precise and reliable determination of the blood-related individual aerobic threshold and / or the maximum blood lactate concentration without the need for blood sampling or analysis. Subsequently, the blood-related anaerobic threshold can be determined based on the maximum blood lactate concentration and / or the blood-related individual aerobic threshold. This can be done, for example, using a so-called D-max method. It may also be necessary to convert salivary lactate concentration values ​​that lie between the salivary individual aerobic threshold and the maximum salivary lactate concentration into blood lactate concentration values. The salivary lactate concentration values ​​can be converted individually into blood lactate concentration values.Optionally, a curve can then be fitted through the blood lactate concentration values, for example, during a fitting. Alternatively, a curve can be fitted through the salivary lactate concentration values, for example, during a fitting, and then the curve describing the salivary lactate concentration values ​​can be converted into a curve describing blood lactate concentration values. The conversion of salivary lactate concentration values ​​or a curve describing salivary lactate concentration values ​​into blood lactate concentration values ​​or a curve describing blood lactate concentration values ​​can be achieved, for example, by applying a first offset along a time axis and a second offset along a lactate concentration axis. The first offset accounts for the aforementioned time lag between the blood lactate concentration values ​​and the salivary lactate concentration values.The second offset takes into account that, as already mentioned, salivary lactate concentration values ​​are generally lower than blood lactate concentration values. Alternatively, the conversion of salivary lactate concentration values ​​or a curve describing salivary lactate concentration values ​​to blood lactate concentration values ​​or a curve describing blood lactate concentration values ​​can be performed using another data model, in particular another regression model. In detail, when carrying out the D-max method, a straight line is first created that extends between the maximum blood lactate concentration and the blood lactate concentration that occurs at the blood-related, aerobic individual threshold. Starting from this line, the blood lactate concentration value that is furthest away from this line is determined, with the distance being measured perpendicular to the line.The power output corresponding to this blood lactate concentration represents the blood-related anaerobic threshold. It follows that, in the D-max method, instead of the blood lactate concentration value that deviates most from the straight line, the blood lactate concentration value that has the same slope as the line can also be sought. The result is the same. Therefore, the individual blood-related anaerobic threshold can also be determined in this way without having to take or analyze blood samples.

[0028] It is understood that creating the regression model, which describes the relationship between the blood-related individual aerobic threshold and a saliva-related individual aerobic threshold, may require determining both the blood-related and saliva-related individual aerobic thresholds in a large number of subjects. As mentioned earlier, a test performed on a large number of subjects can also be called a cohort test. For this purpose, an exercise increment test of the type described above can be used. The determination of the blood-related individual aerobic threshold can be based on actual blood samples. Consequently, the relationship between the blood-related individual aerobic threshold and the saliva-related individual aerobic threshold can be determined based on the data from the large number of subjects.This relationship is then independent of the individual subject, meaning it is valid at least for the group of subjects. Ideally, the relationship is even universally valid. This relationship can then be used in the form of a regression model to determine the blood-related, individual aerobic threshold, eliminating the need for taking or analyzing blood samples.

[0029] According to one approach, it is also possible here to provide several regression models that describe a relationship between the blood-related, individual aerobic threshold and a saliva-related, individual aerobic threshold as alternatives, and to select the appropriate regression model during the procedure for determining the blood-related, individual anaerobic threshold. As already described above in connection with the trained data model, the different regression models can relate to different performance level tests, different sports, and / or different fitness levels.

[0030] It is further understood that creating the regression model, which describes the relationship between the maximum blood lactate concentration and the maximum salivary lactate concentration, may require determining both the maximum blood lactate concentration and the maximum salivary lactate concentration in a number of subjects. As mentioned previously, a test performed on a number of subjects can also be called a cohort test. For this purpose, for example, a graded exercise test of the type described above can be used. The determination of the maximum blood lactate concentration can be based on actual blood samples. Consequently, the relationship between the maximum blood lactate concentration and the maximum salivary lactate concentration can be determined based on the number of subjects. This relationship is then independent of the individual subject, i.e.,This is valid at least for the group of test subjects. Ideally, the relationship is even universally valid. This relationship can then be used in the form of a regression model to determine the maximum blood lactate concentration, eliminating the need for taking or analyzing blood samples.

[0031] According to one approach, it is also possible here to provide several regression models that describe a relationship between the maximum blood lactate concentration and the maximum salivary lactate concentration as alternatives, and to select the appropriate regression model during the procedure for determining the blood-related, individual anaerobic threshold. As already described above in connection with the trained data model, the different regression models can relate to different performance level tests, different sports, and / or different fitness levels.

[0032] In the present context, a regression model is generally understood to be a mathematical model that describes a relationship between two variables or parameters.

[0033] All the regression models described above can be linear, polynomial, or exponential. Optionally, Pearson correlation analysis can be used when creating these models. Such regression models allow for a simple and reliable description of the relationship between an individual's blood-related anaerobic threshold and their saliva-related anaerobic threshold, and / or between their blood-related and saliva-related aerobic thresholds, and / or between their maximum salivary lactate concentration and their maximum blood lactate concentration.

[0034] Creating a regression model can include a validation step for all the aforementioned types of regression models. In such a validation step, an output parameter obtained using the regression model is compared with an actually measured parameter. For example, a blood-related, individual anaerobic threshold obtained using one or more regression models can be compared with a blood-related, individual anaerobic threshold determined using actual blood samples. This ensures a high level of reliability for the regression models.

[0035] Creating a regression model—that is, determining the parameters of the mathematical model that describes a relationship between two variables or parameters—can be considered equivalent to training a trained data model. In this context, it is also possible to further develop a regression model over time by recalculating and, if necessary, adjusting the parameters of the underlying mathematical model. This can be done, for example, based on newly conducted participant tests, particularly cohort tests, as described above. In other words, the regression model can be refined over time, further increasing its accuracy and reliability.

[0036] According to one embodiment, obtaining the blood-related individual anaerobic threshold includes obtaining a saliva-related individual anaerobic threshold based on the initial data. This means that a saliva-related individual anaerobic threshold is calculated based on the initial data. This saliva-related individual anaerobic threshold is then used to obtain the blood-related individual anaerobic threshold. In essence, the lactate concentration only needs to be converted from saliva to blood. As mentioned earlier, a data model, such as a regression model, can be used for this purpose. Overall, the blood-related individual anaerobic threshold can be obtained simply and reliably.

[0037] Determining the salivary-related individual anaerobic threshold can involve determining both the salivary-related individual anaerobic threshold and the salivary-related individual aerobic threshold. The salivary-related anaerobic threshold is thus determined based on the salivary-related aerobic threshold. The salivary-related individual aerobic threshold can be defined as the performance value at which the ratio of salivary lactate concentration to the corresponding performance value is minimal. Alternatively, determining the salivary-related individual aerobic threshold can involve evaluating the rate of change of salivary lactate concentration values ​​over time. In the first alternative, a ratio of salivary lactate concentration to performance value is plotted against a performance value, either virtually or actually. The performance value at which this ratio is minimal represents the salivary-related individual aerobic threshold.In this way, the saliva-related, individual aerobic threshold can be calculated simply and reliably. According to the second alternative, the rate of change of salivary lactate concentration values ​​is evaluated. As already mentioned, the aerobic threshold is defined as the lowest power output at which the lactate concentration rises above the resting lactate concentration. In other words, the saliva-related, aerobic threshold can be determined as the power output at which the gradient of the lactate concentration first deviates significantly from zero. A significant deviation can be defined as the gradient deviating from zero by at least a predetermined value. Alternatively, it is possible to define a limit value or interval for the gradient, at the point where it is reached, thus defining the saliva-related, individual aerobic threshold.Even within this alternative framework, the saliva-related, individual aerobic threshold can be determined easily and reliably.

[0038] If, in the variant described above, threshold values ​​or intervals for the gradient are predefined, these threshold values ​​or intervals can be determined by, for example, measuring lactate concentrations in the saliva of a number of subjects across a certain performance range during an incremental exercise test. Simultaneously, the saliva-related, individual aerobic threshold is determined for each subject using established methods. Based on this, the gradient of the lactate concentration at the saliva-related aerobic threshold can then be determined for each subject. The gradients determined for different subjects can then be aggregated, for example, by calculating an average. Such aggregated values ​​can then serve as threshold values ​​and / or be used to define intervals.

[0039] In one example, determining the saliva-related, individual anaerobic threshold involves calculating the maximum distance between a straight line and a curve connecting the salivary lactate concentration values. This straight line extends from a minimum salivary lactate concentration value to the maximum salivary lactate concentration value, or from a salivary lactate concentration value associated with the individual aerobic threshold to the maximum salivary lactate concentration value.

[0040] Salivary lactate concentration value. The saliva-related, individual anaerobic threshold is obtained as a performance value at the point of maximum deviation. It is understood that, in the D-max method, instead of the salivary lactate concentration value that is furthest from the line, the salivary lactate concentration value with the same slope as the line can also be sought. The result is the same. In this example, a so-called D-max method is applied to the salivary lactate concentration values. The salivary-related, individual anaerobic threshold can also be determined simply and reliably in this way.

[0041] In one approach, determining the saliva-related, individual anaerobic threshold can involve evaluating the rate of change of salivary lactate concentration values ​​over time. As previously explained, the anaerobic threshold is defined by the point at which lactate is generated so rapidly that it can no longer be maintained in steady state by the body's own breakdown processes. This means that at the anaerobic threshold, a gradient of salivary lactate concentration begins to rise from a essentially constant value. This can be determined either using a virtual or actual curve describing lactate concentrations. Alternatively, a predefined threshold value for the gradient can be used, the reaching of which is considered to indicate that the anaerobic threshold has been reached.Such a threshold value can be determined, for example, by measuring lactate concentrations in the saliva of a number of subjects across a certain performance range during a performance-level test, and simultaneously determining the saliva-related, individual anaerobic threshold for each subject using established methods. Based on this, the gradient of the lactate concentration at the saliva-related anaerobic threshold can then be determined for each subject. The gradients determined for different subjects can then be aggregated, for example, by calculating an average. Such aggregated values ​​can then serve as threshold values.

[0042] Additionally, the task is solved by a method for monitoring a subject's blood-related, individual anaerobic threshold. The method includes:

[0043] - Obtaining second data, which includes salivary lactate concentration values ​​and time values ​​describing the corresponding measurement time,

[0044] - Obtaining the blood-related, individual anaerobic threshold of the subject, wherein the blood-related, individual anaerobic threshold of the subject is determined by means of the inventive method for determining a blood-related, individual anaerobic threshold of a subject, or obtaining a rate of change over time of the salivary lactate concentration values ​​and obtaining a salivary-related, individual anaerobic threshold based on the rate of change over time of the salivary lactate concentration values, and

[0045] - Providing an indicator that describes a relationship of the second data to the saliva-related individual anaerobic threshold or to the blood-related individual anaerobic threshold.

[0046] In this context, the indicator can, for example, indicate whether the saliva-related individual anaerobic threshold or the blood-related individual anaerobic threshold has been reached, fallen below, or exceeded. The indicator thus tells a subject how their performance relates to their saliva-related individual anaerobic threshold or their blood-related individual anaerobic threshold. In other words, the indicator tells the subject which dominant metabolic mechanisms they use to generate their performance. This allows the subject to precisely adjust their workload, i.e., the power output, for example, for training purposes. Alternatively or additionally, the subject can use this indicator to adjust their workload, i.e.,For example, his training should be documented in such a way that it is recorded over which periods or proportions of time the saliva-related, individual anaerobic threshold or the blood-related, individual anaerobic threshold was undershot, exceeded or reached.

[0047] As previously mentioned, salivary lactate concentration typically lags behind blood lactate concentration by 6 to 10 minutes. However, it has been shown that despite this time lag, it can be useful to control a subject's performance, for example, for training purposes, based solely on secondary data. This secondary data includes salivary lactate concentration values ​​and corresponding time values, as well as on a saliva-related, individual anaerobic threshold derived from this data. While the indicator can only show whether the blood lactate concentration falls below, exceeds, or is reached with a certain time delay, this information is nevertheless valuable for exercise control and documentation.This is especially true the less the workload fluctuates over the exercise period and the longer the exercise period. Put simply, the influence of the time lag, which is used to determine the saliva-related, individual anaerobic threshold, decreases the longer the subject experiences a substantially constant workload, i.e., exerts a substantially constant effort. For exercise documentation, the time lag is irrelevant, as complete exercise documentation is always available after a waiting period of typically 6 to 10 minutes.According to one variation of the procedure for monitoring a subject's blood-related individual anaerobic threshold, obtaining a rate of change over time of salivary lactate concentration values ​​and obtaining a salivary individual anaerobic threshold based on this rate of change can optionally include predicting or estimating the blood-related individual anaerobic threshold. In other words, the blood-related individual anaerobic threshold can be predicted or estimated based on the second set of data: the rate of change over time of salivary lactate concentration values ​​and / or the salivary individual anaerobic threshold. A data model can again be used for this purpose. This data model can include a regression model.Creating the regression model may require determining both the blood-related and saliva-related individual anaerobic thresholds for a number of subjects, as previously described. Alternatively, the regression model can be created based on the subject's historical exercise loads. A trained data model can also be used, either as an alternative or in addition to the regression model. Similar to the regression model, training can be performed using training data obtained from a number of subjects. Alternatively or in addition, the training data can be derived from the subject's historical exercise loads.

[0048] Furthermore, the task is solved by a method for training a data model to determine a blood-related, individual anaerobic threshold. The data model is trained using an annotated dataset comprising several training data elements. These training data elements include performance values ​​describing a given performance, time values ​​describing a corresponding measurement point, and corresponding salivary lactate concentration values. The training data elements are annotated with their corresponding blood-related, individual anaerobic thresholds. The training data elements thus correspond in content to the initial data described previously. The data annotation can be based on measurements, e.g., using clinically validated measuring instruments. Alternatively or additionally, the annotation can be performed by an expert.As previously explained, a blood-related, individual anaerobic threshold can be easily and reliably determined using initial data that includes performance values ​​describing the subject's performance, time values ​​describing the corresponding measurement point, and corresponding salivary lactate concentration values, without the need to take and / or analyze blood samples. In other words, such a trained data model allows a blood-related, individual anaerobic threshold to be obtained solely from saliva measurements.

[0049] The process for training the data model can also include validation of the trained data model. This means ensuring that the trained data model fulfills its purpose with the desired accuracy and reliability. A validation dataset can be used for this purpose.

[0050] Alternatively or additionally, the training process includes preprocessing the training data elements. During such preprocessing, artifacts can be removed from the training data elements. Alternatively or additionally, the training data elements can be filtered or normalized. Furthermore, alternatively or additionally, the training data elements can be resampled during preprocessing. Another option is to align and / or segment the training data elements temporally. It is also possible to mask or attribute the training data elements. In addition, features can be extracted during preprocessing. The training data elements can also be augmented and / or versioned, taking the data origin into account.Overall, preprocessing improves the quality of the training data elements. This has a positive effect on the quality of the trained data model. It is also possible that parameter tuning takes place during the execution of the data model training process to ensure that the training is efficient in terms of both time and / or training resources, while simultaneously maintaining a high quality of the trained data model.

[0051] The procedure for training the data model may also include anonymizing the training data elements.

[0052] According to one variant, the data model can include sub-models that can be trained separately.

[0053] Such a submodel can be trained to determine a blood-related, individual aerobic threshold. Therefore, a procedure for training a data model, more precisely the aforementioned submodel, to determine a blood-related, individual aerobic threshold is performed. The submodel is trained using an annotated dataset comprising several training data elements. These elements include performance values ​​describing a given exercise, time values ​​describing a corresponding measurement point, and corresponding salivary lactate concentration values. The training data elements are annotated with their respective blood-related, individual aerobic thresholds. The training data elements thus correspond in content to the initial data described previously.Using such a trained data model, as already explained, a blood-related, individual aerobic threshold can be easily and reliably obtained from initial data, which includes performance values ​​describing the subject's performance, time values ​​describing the corresponding measurement point, and corresponding salivary lactate concentration values, without the need to take and / or analyze blood samples. In other words, such a trained data model allows a blood-related, individual aerobic threshold to be obtained solely from saliva measurements. Another sub-model can be trained to determine a maximum blood lactate concentration value. Therefore, a procedure for training a data model, more precisely the aforementioned sub-model, to determine a maximum blood lactate concentration value is carried out.The submodel is trained using an annotated dataset comprising several training data elements. These elements include performance values ​​describing a participant's performance, time values ​​describing a corresponding measurement point, and corresponding salivary lactate concentration values. The training data elements are annotated with their corresponding maximum blood lactate concentration values. The training data elements thus correspond in content to the initial data described earlier. As explained, using such a trained data model, a maximum blood lactate concentration value can be easily and reliably obtained from initial data—which includes performance values ​​describing a participant's performance, time values ​​describing a corresponding measurement point, and corresponding salivary lactate concentration values—without the need to take and / or analyze blood samples.In other words, such a trained data model allows a maximum blood lactate concentration value to be obtained solely from measurements in saliva.

[0054] Another submodel can be trained to convert salivary lactate concentration values ​​into blood lactate concentration values. A procedure is therefore implemented to train a data model, specifically the aforementioned submodel, to convert salivary lactate concentration values ​​into blood lactate concentration values. The submodel is trained using an annotated dataset comprising several training data elements. These elements include performance values ​​describing a given performance, time values ​​describing a corresponding measurement point, and corresponding salivary lactate concentration values. The training data elements are annotated with their corresponding blood lactate concentration values. The training data elements thus correspond in content to the first data described previously.As previously explained, a data model trained in this way allows for the simple and reliable determination of blood lactate concentration values ​​based on initial data, which includes performance values ​​describing the subject's performance, time values ​​describing the corresponding measurement point, and corresponding salivary lactate concentration values. This is achieved without the need to take and / or analyze blood samples. In other words, such a trained data model allows blood lactate concentration values ​​to be obtained solely from salivary measurements.

[0055] It goes without saying that the trained sub-models, in particular, can be used in combination with the D-max method already explained, in order to determine, for example, a blood-related, individual anaerobic threshold.

[0056] In one example, the training of the data model, including any sub-models, is specific to a performance level test, a sport, and / or a fitness level. This means that the data model is trained specifically for the relevant performance level test, sport, and / or fitness level. In this context, performance level tests can differ, for example, in the size of the time and workload increments used. They can also differ in the power metrics and / or fatigue metrics used. This allows the trained data model to be particularly precise for that specific performance level test, sport, and / or fitness level.In this context, it is therefore possible to provide multiple trained data models, each specific to a performance level test, a sport, and / or a fitness level. Thus, when performing the procedure to determine the blood-related, individual anaerobic threshold, the appropriate trained data model can always be selected and applied. The data model used is always the one trained for the relevant performance level test, sport, and fitness level. For example, different data models can be provided for the sports of swimming, running, cycling, and rowing. Similarly, different data models can be provided for the fitness levels of non-athlete, recreational athlete, and professional athlete.Overall, it can therefore be achieved that the method for determining the blood-related, individual anaerobic threshold always works precisely and reliably in a large number of applications.

[0057] The problem is also solved by a data processing device comprising means for carrying out at least one of the methods according to the invention. In a first alternative, the data processing device thus comprises means for carrying out the method according to the invention for determining a blood-related, individual anaerobic threshold of a subject. In a second alternative, the data processing device comprises means for carrying out the method according to the invention for monitoring a blood-related, individual anaerobic threshold of a subject. In a third alternative, the data processing device comprises both means for carrying out the method according to the invention for determining a blood-related, individual anaerobic threshold of a subject and means for carrying out the method according to the invention for monitoring a blood-related, individual anaerobic threshold of a subject.The data processing device according to the invention allows for the precise and reliable determination of a subject's blood-related, individual anaerobic threshold, eliminating the need for blood sampling and analysis. This simplifies and improves the convenience of determining the blood-related, individual anaerobic threshold. Alternatively or additionally, the data processing device according to the invention can indicate how a subject's performance relates to their saliva-related, individual anaerobic threshold or their blood-related, individual anaerobic threshold. In other words, the indicator shows the subject which dominant metabolic mechanisms are responsible for their performance. This enables the subject to precisely adjust their workload, i.e., their power output, for example, for training purposes.

[0058] Furthermore, the problem is solved by a computer program comprising instructions that, when executed by a computer, cause it to perform at least one of the methods according to the invention. Thus, in a first alternative, the computer program can be used to perform the method according to the invention for determining a blood-related, individual anaerobic threshold of a subject. In a second alternative, the computer program can be used to perform the method according to the invention for monitoring a blood-related, individual anaerobic threshold of a subject. In a third alternative, the computer program can be used to perform both the method according to the invention for determining a blood-related, individual anaerobic threshold of a subject and the method according to the invention for monitoring a blood-related, individual anaerobic threshold of a subject.The computer program according to the invention allows for the precise and reliable determination of a subject's blood-related, individual anaerobic threshold, eliminating the need for blood sampling and analysis. This simplifies and improves the convenience of determining the blood-related, individual anaerobic threshold. Alternatively or additionally, the computer program can indicate how a subject's performance relates to their saliva-related, individual anaerobic threshold or their blood-related, individual anaerobic threshold. In other words, the indicator shows the subject which dominant metabolic mechanisms are responsible for their performance. This enables the subject to precisely adjust their workload, i.e., their power output, for example, for training purposes.

[0059] The problem is also solved by a computer-readable medium comprising instructions which, when executed by a computer, cause it to perform at least one of the methods according to the invention. The computer-readable medium can also be referred to as a computer-readable storage medium. Thus, in a first alternative, the computer-readable medium can be used to execute the method according to the invention for determining a blood-related, individual anaerobic threshold of a subject. In a second alternative, the computer-readable medium can be used to execute the method according to the invention for monitoring a blood-related, individual anaerobic threshold of a subject.In a third alternative, the computer-readable medium can be used to perform both the inventive method for determining a subject's blood-related, individual anaerobic threshold and the inventive method for monitoring a subject's blood-related, individual anaerobic threshold. Thus, the inventive computer-readable medium allows for the precise and reliable determination of a subject's blood-related, individual anaerobic threshold without the need for blood sampling. This makes determining the blood-related, individual anaerobic threshold simpler and more convenient. Alternatively or additionally, the inventive computer-readable medium can be used to indicate how a subject's performance compares to their saliva-related, individual anaerobic threshold or their blood-related, individual anaerobic threshold.In other words, the indicator tells the subject which dominant metabolic mechanisms are responsible for their performance. This allows the subject to precisely adjust their workload, i.e., the power output, for example, for training purposes.

[0060] In one example, the computer-readable medium or computer-readable storage medium is implemented as a component of a mobile electronic device. This mobile electronic device could be, for example, a mobile phone or a tablet. It could also be a laptop computer. It is also possible that the mobile electronic device is a smartwatch. Likewise, the computer-readable medium or computer-readable storage medium is implemented as a component of a stationary electronic device, for example, as a component of a desktop computer.

[0061] The problem is also solved by a system for determining a blood-related, individual anaerobic threshold of a subject. The system comprises a data processing device according to the invention, including means for carrying out the inventive method for determining a blood-related, individual anaerobic threshold of a subject. Furthermore, the system comprises a first sensor unit for detecting at least one salivary lactate concentration value. The system also comprises a second sensor unit for detecting at least one power value, which describes the subject's performance. Additionally, the system comprises a time unit for providing a time value. The first sensor unit, the second sensor unit, and the time unit are either communicatively coupled to the data processing device or are components of the data processing device.As mentioned previously, the first sensor unit can be designed as an intraoral or extraoral sensor unit. Using this system for determining a subject's blood-related, individual anaerobic threshold, a subject's blood-related, individual anaerobic threshold can be obtained precisely and reliably, while simultaneously eliminating the need for blood sampling and analysis. This makes determining the individual anaerobic threshold simpler and more convenient.

[0062] The problem is also solved by a system for monitoring a subject's blood-related, individual anaerobic threshold. The system comprises a data processing device, including means for carrying out the inventive method for monitoring a subject's blood-related, individual anaerobic threshold. Furthermore, the system comprises a first sensor unit for detecting at least one salivary lactate concentration value. The system also includes a time unit for providing a time value. Additionally, the system includes an output interface for outputting an indicator that describes the relationship between at least one salivary lactate concentration value, detected at a corresponding time point, and the salivary-related, individual anaerobic threshold or the blood-related, individual anaerobic threshold.The first sensor unit, the timer, and the output interface are either communicatively coupled to the data processing device or are integral components of the data processing device. As mentioned previously, the first sensor unit can be implemented as an intraoral or extraoral sensor unit. Thus, the system for monitoring a subject's blood-related, individual anaerobic threshold can indicate how their performance relates to their saliva-related, individual anaerobic threshold or to their blood-related, individual anaerobic threshold. In other words, the indicator shows the subject which dominant metabolic mechanisms they use to generate power. This allows the subject to precisely adjust their workload, i.e., their power output, for example, for training purposes.

[0063] Both the inventive method for determining a blood-related, individual anaerobic threshold of a subject, as well as the inventive method for monitoring a blood-related, individual anaerobic threshold, and the method for training the data model can be computer-implemented methods.

[0064] Alternatively or additionally to the methods already described, a procedure for determining training zones can also be carried out based on the initial data, which includes performance values ​​describing the subject's performance, time values ​​describing the corresponding recording time, and corresponding salivary lactate concentration values. This requires that a blood-related, individual aerobic threshold and a blood-related, individual anaerobic threshold of the subject have been determined, either implicitly or explicitly, based on the initial data. This has already been explained. At least three training zones can thus be determined, with the first training zone being characterized by a blood lactate concentration value below the subject's blood-related, individual aerobic threshold.A second training zone is characterized by a blood lactate concentration value between the subject's individual blood-related aerobic threshold and their individual blood-related anaerobic threshold. A third training zone is characterized by a blood lactate concentration value above the subject's individual blood-related anaerobic threshold.

[0065] Based on such training zones, a subject can individually and efficiently control and / or document their training. As an alternative to these three training zones, each directly characterized or defined by the subject's blood-related individual aerobic threshold and blood-related individual anaerobic threshold, further training zones can be defined. These can be characterized or defined, for example, by percentages of the performance corresponding to the subject's blood-related individual aerobic threshold and / or by percentages of the performance corresponding to the subject's blood-related individual anaerobic threshold. Hybrid forms are also possible.Both percentage values ​​and the individual blood-related aerobic threshold and / or the individual blood-related anaerobic threshold of the subject can be used to characterize or define training zones. This allows training zones to be defined or characterized according to application or need. This also contributes to individualized and efficient control and / or documentation of training.

[0066] In this context, five training zones can be defined. The first training zone is characterized or defined by a performance level of 70% or less of the subject's individual blood-related anaerobic threshold. This first training zone can also be referred to as the compensation zone. The second training zone can be characterized or defined by a performance level above 70% and up to and including 80% of the subject's individual blood-related anaerobic threshold. This second training zone can also be referred to as the first basic endurance zone. In this second training zone, the subject's movement or performance is primarily powered by aerobic energy production.A third training zone can be characterized or defined by a performance level above 80% and up to and including 90% of the subject's individual blood-related anaerobic threshold. This third training zone can also be referred to as the second basic endurance zone. In this third zone, the subject's movement or performance is still predominantly powered by aerobic energy production, although a small portion of energy production occurs anaerobically. A fourth training zone can be characterized or defined by a performance level above 90% and up to and including 110% of the subject's individual blood-related anaerobic threshold. This fourth training zone can also be referred to as the development zone.In the fourth training zone, the subject's movement or performance is enabled by both aerobic and anaerobic energy production. A fifth training zone can be characterized or defined by a performance level above 110% of the subject's individual blood-related anaerobic threshold. This fifth training zone can also be referred to as the peak zone. In this zone, the subject's movement or performance is predominantly enabled by anaerobic energy production. The subject can only sustain the required performance for a short period in this zone. In other words, the subject can only withstand the exertion for a short time. Based on these training zones, a subject can individually and efficiently control and / or document their training.Due to the large number of training areas, this can be done relatively accurately and in detail.

[0067] It goes without saying that the training areas can also be defined in other ways. Therefore, the number of training areas, as well as their definition or characterization, can be chosen to suit the desired control and / or documentation of the training.

[0068] It is understood that a data model, in particular a suitably trained data model, can be used to determine the training zones. In this context, it is also possible to determine the training zones directly, i.e., without having previously and separately determined the subject's blood-related individual aerobic threshold and blood-related individual anaerobic threshold. Such a data model can also be trained using a data model training method. The data model is trained using an annotated dataset comprising several training data elements. These elements include performance values ​​describing a given exercise, time values ​​describing a corresponding measurement point, and corresponding salivary lactate concentration values, with the training data elements being annotated with their associated training zones.The training data elements thus correspond in content to the initial data described previously. Using such a trained data model, as already explained, training zones can be determined simply and reliably based on initial data that includes performance values ​​describing the subject's performance, time values ​​describing the corresponding measurement point, and corresponding salivary lactate concentration values, without the need to take and / or analyze blood samples. In other words, such a trained data model allows training zones to be determined solely based on saliva measurements.

[0069] The invention is explained below with reference to various embodiments shown in the accompanying drawings. These show:

[0070] Figure 1 shows a test subject using a system according to the invention for determining a blood-related, individual anaerobic threshold, wherein a method according to the invention for determining a blood-related, individual anaerobic threshold of the test subject is carried out using the system.

[0071] Figure 2 shows a graphical representation of initial data used when carrying out the inventive method to determine a blood-related, individual anaerobic threshold of the subject.

[0072] Figure 3 illustrates a procedure step for determining a saliva-related, individual aerobic threshold; Figure 4 is a graphical representation of a data model that includes a

[0073] The regression model includes a relationship between a blood-related, individual anaerobic threshold and a saliva-related, individual anaerobic threshold.

[0074] Figure 5 is a graphical representation of a data model that is a

[0075] The regression model includes a relationship between a blood-related, individual aerobic threshold and a saliva-related, individual aerobic threshold.

[0076] Figure 6 is a graphical representation of a data model that is a

[0077] The regression model includes a relationship between a maximum blood lactate concentration value and a maximum salivary lactate concentration value.

[0078] Figure 7 is a graphical representation of a data model that a trained

[0079] The data model includes:

[0080] Figure 8 illustrates a training of the data model from Figure 7.

[0081] Figure 9 shows a subject using a system according to the invention for monitoring a blood-related, individual anaerobic threshold, wherein a method according to the invention for monitoring a blood-related, individual anaerobic threshold of the subject is carried out by means of the system, and

[0082] Figure 10 is an illustration of different training areas that are in

[0083] The dependence of a blood-related, individual aerobic threshold and a blood-related, individual anaerobic threshold are defined.

[0084] Figure 1 shows a subject 10 who is completing a performance level test. The subject 10 sits on a bicycle ergometer 12 and pedals 14, whereby the resistance of the pedals 14, which opposes the pedaling of the subject 10, can be adjusted.

[0085] Thus, during the power-level test, the resistance is increased incrementally. More precisely, the test starts at a predetermined, relatively low power level of subject 10, e.g., 70 watts, which corresponds to a predetermined resistance of the pedals 14. After a predetermined time interval, e.g., 3 minutes, the resistance of the pedals 14 is always increased by a predetermined increment, e.g., by 30 watts. Subject 10 must therefore deliver a power output increased by the corresponding power increment.

[0086] The resistance of the pedals 14, i.e., the power output of the subject 10, is increased incrementally by one power increment at a time until the subject 10 gives up due to exhaustion. From this point on, the subject can either rest or continue pedaling against a comparatively low resistance, e.g., 50 watts or 70 watts. It goes without saying that, to protect the subject 10, the power-level test can also be terminated at any time for medical reasons. However, this is the exception.

[0087] Subject 10 uses a system 16 in connection with the performance level test to determine his blood-related, individual anaerobic threshold ANS.

[0088] This system 16 comprises a first sensor unit 18 for detecting a salivary lactate concentration value KS. In the illustrated embodiment, the first sensor unit 18 is designed as an intraoral sensor unit, i.e., as an intraoral sensor. This means that the subject 10 carries the first sensor unit 18 in his mouth.

[0089] The first sensor unit 18 can continuously or quasi-continuously measure a salivary lactate concentration value KS. The system 16 also includes a second sensor unit 20. The second sensor unit 20 is designed to record a power output value P, which describes the power output of the subject 10. In the illustrated embodiment, the second sensor unit 20 is attached to the bicycle ergometer 12, so that the power output of the subject 10 while pedaling 14 can be measured using the second sensor unit 20.

[0090] Furthermore, the system 16 includes a device 22 for data processing.

[0091] In the illustrated embodiment, the data processing device 22 is designed as a mobile electronic device attached to the handlebars of the bicycle ergometer 12. This mobile electronic device could be, for example, a mobile phone or a tablet. It is also possible for the mobile electronic device to be a smartwatch. Likewise, the data processing device 22 could be a stationary electronic device, such as a desktop computer or laptop.

[0092] The first sensor unit 18 is coupled to the data processing device 22 via communication technology. In the illustrated embodiment, the communication technology coupling is a wireless coupling, for example Bluetooth. This is illustrated by a dashed line.

[0093] Similarly, the second sensor unit 20 is coupled to the data processing device 22 via communication technology. In the illustrated embodiment, this communication technology coupling is also a wireless coupling, for example Bluetooth. This is also illustrated by a dashed line.

[0094] System 16 further comprises a time unit 24, which is configured to provide a time value. In the illustrated embodiment, the time unit 24 is implemented as a component of the data processing device 22 and is therefore communicatively coupled to all other components of the data processing device 22. Alternatively, it would also be conceivable to provide the time unit 24 as a separate unit from the data processing device 22. In this alternative, the time unit 24 must be communicatively coupled to the data processing device 22. This can again be done via Bluetooth.

[0095] The data processing device 22 further comprises a data processing unit 26 and a data storage unit 28.

[0096] The data storage unit 28 comprises a computer-readable medium 30.

[0097] A computer program 32 is stored on the computer-readable medium 30, which includes instructions that, when the computer program 32 is executed by the data processing unit 26, or more generally by a computer, cause it to perform a procedure for determining a blood-related, individual anaerobic threshold ANS of a subject 10.

[0098] Thus, the computer-readable medium also includes 30 instructions which, when executed by the data processing unit 26, or more generally by a computer, cause it to execute a procedure for determining a blood-related, individual anaerobic threshold ANS of a subject 10.

[0099] The data processing unit 26 and the data storage unit 28 thus represent means 34 for carrying out a procedure for determining a blood-related, individual anaerobic threshold ANS of a subject 10.

[0100] The following section explains in more detail the procedure for determining a blood-related, individual anaerobic threshold ANS.

[0101] In a first step, Si l, initial data Dl are obtained, comprising performance values ​​P describing the subject's performance, time values ​​T describing each corresponding recording time, and corresponding salivary lactate concentration values ​​KS. The initial data Dl are determined during the execution of the performance level test, with the performance values ​​P describing the subject's performance being determined by the second sensor unit 20 and the salivary lactate concentration values ​​KS being determined by the first sensor unit 18. After both the first sensor unit 18 and the second sensor unit 20 are communicatively coupled to the data processing device 22, which includes the time unit 24, a time value T describing a recording time can be assigned to each performance value P and each lactate concentration value KS.

[0102] Consequently, the first data Dl are received at the device 22 for data processing and / or are available at the device 22 for data processing.

[0103] The first data Dl are graphically represented in Figure 2, with time plotted on the horizontal axis and salivary lactate concentration KS on a first vertical axis.

[0104] Figure 2 illustrates, in this context, both some specific measured values ​​for the salivary lactate concentration KS and a curve using which the measured values ​​have been interpolated. It should be understood that both the measured values ​​and the curve are purely illustrative.

[0105] Since the power output increases by a predetermined power level, e.g. 30 watts, at predetermined time intervals, e.g. 3 minutes, a power value P can be assigned to each salivary lactate concentration value KS.

[0106] Since the performance values ​​differ for each subject 10, they are simply labeled PI to PI 2 in Figure 2. In this example, subject 10 reached their maximum performance at performance value P6. This means that the performance values ​​increase continuously from PI to P6. The performance values ​​P7 to PI 2 are comparatively low, for example, in the range of performance value PI. The performance values ​​P are also to be understood as purely illustrative.

[0107] It is noted that if subject 10 cannot maintain the highest power output P for the entire assigned time interval, the highest power output P can be determined by interpolation. For example, to determine the highest power output P, ​​the percentage of the fixed power level corresponding to the proportion of the time interval during which subject 10 was able to maintain that power output can be added to the previous power output P. Thus, if the subject was able to maintain the highest power output at a power level of 30 watts for 80% of the predefined time interval, for example, 3 minutes, the highest power output is determined by adding 80% of 30 watts to the previous power level.

[0108] It becomes particularly clear from the representation in Figure 2 that the initial data Dl exceed a maximum salivary lactate concentration value KS. Furthermore, it becomes clear that the maximum salivary lactate concentration value KS only occurs after the maximum performance value P has been reached.

[0109] For better understanding, Figure 2 also shows an exemplary curve of a blood lactate concentration (KB) using a dashed line. The blood lactate concentration (KB) is plotted along a second vertical axis, which differs from the first vertical axis. This is solely for illustrative purposes. In the illustrated embodiment, the blood lactate concentration is explicitly not measured directly, but rather determined solely based on the sensor-measured salivary lactate concentration, as will be explained in detail below.

[0110] In a second step (S 12), the saliva-related individual aerobic threshold (SAES) is determined. For this purpose, a quotient of the recorded salivary lactate concentration values ​​(KS) and the corresponding performance values ​​(P) is calculated based on the initial data (Dl) and plotted virtually or actually against the performance data. The performance value (P) at which the quotient is minimal is the saliva-related individual aerobic threshold (SAES) (see Figure 3).

[0111] In an alternative version of the second step S12, the saliva-related, individual aerobic threshold SAES can also be determined by evaluating a rate of change over time of the salivary lactate concentration values ​​KS from the first data Dl.

[0112] As mentioned previously, the aerobic threshold is defined as the lowest power output at which a lactate concentration rises above the resting lactate concentration. In other words, the saliva-related aerobic threshold (SAES) can be determined as the power output at which the gradient of the salivary lactate concentration (KS) first deviates significantly from zero. A significant deviation can be defined as the gradient deviating from zero by at least a predetermined value. For this purpose, the lactate concentration values ​​(KS) from the initial data (Dl) can be used, which can also be interpolated to determine a gradient (see Figure 2).

[0113] In a subsequent third step S13, the saliva-related, individual anaerobic threshold SANS is determined.

[0114] For this purpose, the D-max method, which is well-known in itself, is used. In this context, a straight line is created that extends from the lactate concentration KS present at the saliva-related, individual aerobic threshold SAES determined in the second step, to a maximum salivary lactate concentration KS.

[0115] A salivary lactate concentration value KS is then determined that exhibits the maximum deviation from this line. The saliva-related, individual anaerobic threshold SANS is then defined as the performance value P that corresponds to the salivary lactate concentration value KS with the greatest deviation.

[0116] This can be done graphically using Figure 2. Alternatively, this can be done by the data processing device 22.

[0117] In an alternative iteration of the third step S13, the straight line is drawn between the minimum salivary lactate concentration value KS and the maximum salivary lactate concentration value KS, instead of between the salivary lactate concentration value KS present at the aerobic threshold SAES and the maximum salivary lactate concentration value KS. It is understood that in this alternative iteration of the third step S13, it is not necessary to determine the saliva-related, individual aerobic threshold SAES in the second step S12. This can also be done graphically using Figure 2 or by the data processing device 22.

[0118] In another alternative iteration of the third step S13, the saliva-related individual anaerobic threshold (SANS) is obtained by evaluating the rate of change of salivary lactate concentration values ​​(KS) over time. As previously explained, the saliva-related individual anaerobic threshold (SANS) is defined by the point at which lactate is generated so rapidly that it can no longer be maintained in steady state by the body's own degradation processes. This means that at the anaerobic threshold (SANS), a gradient of salivary lactate concentration (KS) begins to rise from a substantially constant value. This can be determined either using a virtual or actual curve describing lactate concentrations (KS). This can also be done based on Figure 2.

[0119] After performing the third step S13, the saliva-related individual anaerobic threshold SANS of subject 10 is known. In a fourth step S14, the saliva-related individual anaerobic threshold SANS of subject 10 is converted into a blood-related individual anaerobic threshold ANS of subject 10.

[0120] For this purpose, a data model M is used, which in the present embodiment is a regression model based on linear regression. The data model M describes a relationship between the blood-related, individual anaerobic threshold ANS and a saliva-related, individual anaerobic threshold SANS in the form of a straight line, and can therefore be described by a linear equation (see Figure 4).

[0121] The present linear regression model was created by determining both the blood-related individual anaerobic threshold (ANS) and the saliva-related individual anaerobic threshold (SANS) in a number of subjects. The exercise test used here, for example, can be employed for this purpose. The determination of the blood-related individual anaerobic threshold (ANS) can be based on actual blood samples. Consequently, the relationship between the blood-related individual anaerobic threshold (ANS) and the saliva-related individual anaerobic threshold (SANS) can be determined based on the majority of subjects. This relationship can then be used in the form of the regression model to determine the blood-related individual anaerobic threshold (ANS) in further subjects, e.g., subject 10 described here, thus eliminating the need for further blood sample analysis.

[0122] The linear regression model is illustrated in Figure 4, which shows, as an example, a plurality of data points, each corresponding to one of the plurality of subjects who participated in creating the data model M. The resulting straight line from these data points, which represents the actual data model M, is also shown in Figure 4.

[0123] In the fourth step, S14, the saliva-related, individual anaerobic threshold SANS determined in the third step, S13, can be entered into the data model M. The data model M then outputs the blood-related, individual anaerobic threshold ANS.

[0124] In summary, by applying the procedure for determining a subject's blood-related, individual anaerobic threshold ANS 10, the blood-related, individual anaerobic threshold ANS is determined based on the initial data Dl and using a data model M. No blood samples or blood analyses are necessary.

[0125] The procedure for determining a subject's blood-related, individual anaerobic threshold can also be designed according to a first variant. The following discussion focuses solely on the differences compared to the previously explained basic variant of the procedure.

[0126] The key difference is that the data model M used, which again includes a regression model, describes a relationship between the blood-related individual aerobic threshold AES and the saliva-related individual aerobic threshold SAES.

[0127] The first step Si l and the second step S12 remain the same compared to the basic variant already explained, whereby the saliva-related, individual aerobic threshold SAES is determined as the result of the second step S12.

[0128] Now, using the data model M, which describes a relationship between the blood-related individual aerobic threshold AES and the saliva-related individual aerobic threshold SAES, the blood-related individual aerobic threshold AES of subject 10 is determined.

[0129] This data model M can be a linear regression model, illustrated in Figure 5, which shows an exemplary plurality of data points, each corresponding to one of the plurality of subjects who participated in creating the data model M. The resulting straight line from these data points, which represents the actual data model M, is also shown in Figure 5.

[0130] Furthermore, another data model is used, which also includes a regression model, specifically a linear regression model. This data model describes a relationship between a maximum salivary lactate concentration value KS and a maximum blood lactate concentration value KB and is illustrated in Figure 6. Here, a plurality of measurement points are again shown as examples, each corresponding to one of the plurality of subjects who participated in the creation of the data model M. The resulting straight line from these measurement points, which represents the actual data model M, is also shown in Figure 6.

[0131] Using this data model, a maximum blood lactate concentration value KB can be determined from the maximum salivary lactate concentration value KS from the first data Dl.

[0132] Now the D-max method, which is known in itself, can be applied again. This time, however, the straight line between the lactate concentration value of the blood-related, individual aerobic threshold (AES) and the maximum blood lactate concentration value (KB) is drawn.

[0133] Furthermore, to perform the D-max procedure, salivary lactate concentration values ​​(KS) that lie between the salivary-related individual aerobic threshold (SAES) and the maximum salivary lactate concentration value (KS) must be converted into blood lactate concentration values ​​(KB). The salivary lactate concentration values ​​(KS) can be converted individually into blood lactate concentration values ​​(KB). Optionally, a curve can then be fitted through the blood lactate concentration values ​​(KB), e.g., during a fitting process. Alternatively, a curve can be fitted through the salivary lactate concentration values ​​(KS), and then the curve describing the salivary lactate concentration values ​​(KS) can be converted into a curve describing the blood lactate concentration values ​​(KB). The procedure for determining a subject's blood-related individual anaerobic threshold (ANS) can also be designed according to a second variant.The following discussion will focus solely on the differences between this procedure and the variants already explained.

[0134] The key difference is that in the second variant, the data model M includes a trained data model M. This trained data model M is configured to use the initial data Dl as input and directly output the blood-related, individual anaerobic threshold ANS (see illustration in Figure 7). This means that the trained data model M can be used directly after the first step SI 1, and the steps S12, S13, and S14 described above can be omitted.

[0135] The trained data model M is, for example, implemented as an artificial neural network. This artificial neural network can be implemented, for example, as a convolutional neural network (CNN), a recurrent neural network (RNN), or a transformer.

[0136] The training of the trained data model M can be designed as follows (see also Figure 8).

[0137] To train the data model M, an annotated dataset can be used, comprising several training data elements DT. Each of the training data elements DT includes a performance value P, which describes a power output, a time value T, which describes an associated time of measurement, and a salivary lactate concentration value KS. These training data elements DT are annotated with associated blood-related, individual anaerobic thresholds ANS.

[0138] Consequently, the data model M can learn during training to determine blood-related, individual anaerobic thresholds ANS based on initial data Dl. As already explained, the initial data Dl comprises descriptive performance values ​​P of the subject, corresponding time values ​​T for each measurement, and corresponding salivary lactate concentration values ​​KS.

[0139] Figure 9 again shows the subject 10, who now, however, is not performing a performance level test using a bicycle ergometer 12, but is riding a bicycle 36 outdoors.

[0140] Subject 10 uses a System 38 to monitor his blood-related, individual anaerobic threshold ANS.

[0141] This system 38 comprises a first sensor unit 40 for detecting a salivary lactate concentration value KS. In the illustrated embodiment, the first sensor unit 40 is designed as an intraoral sensor unit. This means that the subject 10 carries the first sensor unit 18 in his mouth.

[0142] The first sensor unit 40 can be used to continuously or quasi-continuously measure a salivary lactate concentration value KS.

[0143] The first sensor unit 18, which was used in connection with the procedure for determining the blood-related individual anaerobic threshold ANS, and the first sensor unit 40 can be identically designed or even actually identical. In the latter case, the same first sensor unit 18, 40 can therefore be both a component of system 16 for determining a blood-related individual anaerobic threshold ANS and a component of system 38 for monitoring a blood-related individual anaerobic threshold ANS.

[0144] Furthermore, the system 38 includes a device 42 for data processing.

[0145] The data processing device 42, in the illustrated embodiment, is designed as a mobile electronic device attached to the handlebars of the bicycle 36. The mobile electronic device could be, for example, a mobile phone or a tablet. It is also possible for the mobile electronic device to be a smartwatch.

[0146] The first sensor unit 40 is coupled to the data processing device 42 via communication technology. In the illustrated embodiment, the communication technology coupling is a wireless coupling, for example Bluetooth. This is illustrated by a dashed line.

[0147] System 38 further comprises a time unit 44, which is configured to provide a time value T. In the illustrated embodiment, the time unit 44 is designed as a component of the data processing device 42 and is therefore communicatively coupled to all other components of the data processing device 42. Alternatively, it would also be conceivable to provide the time unit 44 as a separate unit from the data processing device 42. In this alternative, the time unit 44 must be communicatively coupled to the data processing device 42. This can again be done via Bluetooth.

[0148] The system 38 also includes an output interface 46 for outputting an indicator which describes a relationship between at least one salivary lactate concentration value KS, which was recorded at an associated recording time T, and the saliva-related individual anaerobic threshold SANS or the blood-related individual anaerobic threshold ANS.

[0149] In the illustrated embodiment, the output interface 46 is formed by a screen unit of the electronic device that constitutes the data processing device 42. The output interface 46 is thus, for example, formed as a screen unit of a mobile phone, tablet, or smartwatch.

[0150] Consequently, the relationship of at least one salivary lactate concentration value KS, which was recorded at an associated recording time T, to the saliva-related individual anaerobic threshold SANS or to the blood-related individual anaerobic threshold ANS can be graphically represented, i.e. output for the subject 10.

[0151] The output interface 46 is thus also coupled to the other components of the device 42 for data processing via communication technology. More generally speaking, the output interface 46 is coupled to the device 42 for data processing.

[0152] The device 42 for data processing further comprises a data processing unit 48 and a data storage unit 50.

[0153] The data storage unit 50 comprises a computer-readable medium 52.

[0154] A computer program 54 is stored on the computer-readable medium 52, which includes instructions that, when the computer program 54 is executed by the data processing unit 48, or more generally by a computer, cause it to perform a procedure for monitoring a blood-related, individual anaerobic threshold ANS of a subject 10.

[0155] Thus, the computer-readable medium also includes 52 instructions which, when executed by the data processing unit 48, or more generally by a computer, cause it to execute a procedure for monitoring a blood-related, individual anaerobic threshold ANS of a subject 10.

[0156] The data processing unit 48 and the data storage unit 50 thus represent means 56 for carrying out a procedure for monitoring a blood-related, individual anaerobic threshold ANS of a subject 10.

[0157] The following section explains in more detail the procedure for monitoring a blood-related, individual anaerobic threshold (ANS).

[0158] In the first step S21 of this procedure, second data D2 are obtained, comprising salivary lactate concentration values ​​KS and corresponding time values ​​T describing the respective acquisition time. This second data D2 is provided by the first sensor unit 40 and by the time unit 44.

[0159] Furthermore, in a second step S22, which can be performed before, during, or after the first step S21, the blood-related, individual anaerobic threshold ANS of subject 10 is obtained. This can be provided in the form of third-party data D3.

[0160] The blood-related individual anaerobic threshold ANS of subject 10 may have been determined using the previously explained procedure for determining the blood-related individual anaerobic threshold ANS.

[0161] Alternatively, the blood-related individual anaerobic threshold (ANS) can be determined by obtaining the rate of change of the salivary lactate concentration values ​​(KS) from the second data set (D2) and then calculating a salivary individual anaerobic threshold (SANS) based on this rate of change. Furthermore, the salivary individual anaerobic threshold (SANS) must be converted into the blood-related individual anaerobic threshold (ANS). This conversion was explained above in connection with the procedure for determining the blood-related individual anaerobic threshold (ANS).

[0162] It is emphasized that in both variants the blood-related, individual anaerobic threshold ANS can be provided without the need for blood sampling and / or blood analysis.

[0163] In a third step S23 of the procedure for monitoring the blood-related individual anaerobic threshold ANS, an indicator is provided which describes a relationship of the second data D2 to the saliva-related individual anaerobic threshold SANS or to the blood-related individual anaerobic threshold ANS.

[0164] This can be done graphically using the output interface 46, which is designed as a screen unit. With reference to the saliva-related individual anaerobic threshold SANS, the subject 10 can thus be informed via the screen unit whether the salivary lactate concentration values ​​KS contained in the second data D2 indicate performance below, above, or at the saliva-related individual anaerobic threshold SANS.

[0165] Subject 10 can use this information to control their training. Because salivary lactate concentration (KS) typically lags behind blood lactate concentration (KB) by 6 to 10 minutes, this is particularly useful when the workload, i.e., the power output (P) of the subject, fluctuates little over the duration of the workload and the workload is relatively long.

[0166] Alternatively or additionally, the indicator, which describes the relationship between the second data point D2 and the salivary lactate concentration (SANS), can be recorded over the training period and for a certain time afterward, thus documenting the training. In particular, time periods and / or time fractions are documented during which the salivary lactate concentration values ​​(KS) contained in the second data point D2 indicate performance below, above, or at the salivary lactate concentration (SANS).

[0167] With reference to the blood-related individual anaerobic threshold ANS, the subject can be shown via the screen unit whether the salivary lactate concentration values ​​KS contained in the second data D2 indicate a performance below, above or at the blood-related individual anaerobic threshold ANS.

[0168] Subject 10 can use this relationship of the second data point D2 to the saliva-based individual anaerobic threshold (SANS) or the blood-based individual anaerobic threshold (ANS) to control their training and thus make it more efficient and / or effective. Optionally, the system 38, which monitors the subject's blood-based individual anaerobic threshold (ANS), is also designed to indicate training zones defined by Subject 10's blood-based individual aerobic threshold (AES) and blood-based individual anaerobic threshold (ANS).

[0169] In this context, a first training zone Bl is characterized by a blood lactate concentration value KB below the blood-related, individual aerobic threshold AES of the subject 10 (see Figure 10).

[0170] A second training zone B2 is characterized by a blood lactate concentration value KB that lies between the blood-related, individual aerobic threshold AES of subject 10 and the blood-related, individual anaerobic threshold ANS of subject 10.

[0171] A third training zone B3 is characterized by a blood lactate concentration value KB above the blood-related, individual anaerobic threshold ANS of subject 10.

[0172] To determine these training zones Bl, B2, B3, the blood-related individual aerobic threshold AES of subject 10 and the blood-related individual anaerobic threshold ANS of subject 10 can be determined in the same way as explained above.

[0173] Alternatively, it is possible to directly determine the training zones Bl, B2, and B3 based on the second data set D2 and using a further data model, in particular a suitably trained data model. In such a case, the blood-related individual aerobic threshold (AES) and the blood-related individual anaerobic threshold (ANS) of subject 10 do not need to be explicitly determined beforehand.

[0174] The training zones can also be used for training control and / or training documentation. To display a current training zone Bl, B2, B3 and / or past training zones Bl, B2, B3 to subject 10, the output interface 46 of system 38 can again be used to monitor a blood-related, individual anaerobic threshold ANS.

[0175] It is understood that, due to the fact that the training zones Bl, B2, B3 can be determined even without explicit specification of the blood-related, individual aerobic threshold AES of subject 10 and the blood-related, individual anaerobic threshold ANS of subject 10, the training zones Bl, B2, B3 can also be determined independently of the procedure for monitoring a blood-related, individual anaerobic threshold ANS of subject 10, i.e. independently.

[0176] The following explanations describe System 16 and the procedure for determining a blood-related, individual anaerobic threshold (ANS) using a subject 10 who uses a bicycle ergometer 12. System 38 and the procedure for monitoring the blood-related, individual anaerobic threshold (ANS) were explained using a subject 10 who rides a bicycle 36. It is understood that this sport was chosen for the sake of simplicity. Of course, it is also possible to use the systems 16 and 38 and the procedures described above in connection with other sports.

[0177] Furthermore, it is understood that even though the same subject 10 was used to explain system 16 and the procedure for determining a blood-related, individual anaerobic threshold ANS and to explain system 38 and the procedure for monitoring the blood-related, individual anaerobic threshold ANS, it is of course also possible to use different subjects here, i.e., that one subject uses system 16 and the procedure for determining a blood-related, individual anaerobic threshold ANS and another subject uses system 38 and the procedure for monitoring the blood-related, individual anaerobic threshold ANS.

[0178] Furthermore, the data processing device 22 and its components were presented here as separate units from the data processing device 42 and its components. This is one possibility. However, it is also possible that a single data processing device is configured to perform both the method for determining the blood-related individual anaerobic threshold (ANS) and the method for monitoring the blood-related individual anaerobic threshold (ANS). Such a data processing device can therefore be both part of a system 16 for determining the blood-related individual anaerobic threshold (ANS) and

[0179] be part of a system 38 for monitoring the blood-related, individual anaerobic threshold ANS.

[0180] Reference symbol list

[0181] 10 test subjects

[0182] 12 exercise bikes

[0183] 14 pedals

[0184] 16 System for determining a blood-related, individual anaerobic threshold

[0185] 18 First sensor unit for detecting at least one salivary lactate concentration value

[0186] 20 second sensor unit for recording at least one performance value

[0187] 22 Device for data processing

[0188] 24 time units

[0189] 26 Data processing unit

[0190] 28 Data storage unit

[0191] 30 computer-readable media

[0192] 32 Computer program

[0193] 34 Means for carrying out the procedure for determining a blood-related, individual anaerobic threshold of a subject

[0194] 36 bicycle

[0195] 38 System for monitoring a blood-related, individual anaerobic threshold

[0196] 40 first sensor unit

[0197] 42 Device for data processing

[0198] 44 time units

[0199] 46 Output interface

[0200] 48 Data processing unit

[0201] 50 data storage units

[0202] 52 computer-readable media

[0203] 54 Computer program

[0204] 56 Means for carrying out the procedure for monitoring a subject's blood-related individual anaerobic threshold ANS blood-related individual anaerobic threshold

[0205] AES blood-related, individual aerobic threshold

[0206] B 1 first training area

[0207] B2 second training area

[0208] B3 third training area

[0209] First data

[0210] D2 second data

[0211] D3 third data

[0212] DT Training Data Element

[0213] KS salivary lactate concentration value

[0214] KB blood lactate concentration value

[0215] M data model

[0216] P performance value

[0217] SANS saliva-related individual anaerobic threshold

[0218] SAES saliva-related individual aerobic threshold

[0219] 511 First step of the procedure for determining a blood-related, individual anaerobic threshold

[0220] 512 second step of the procedure for determining a blood-related, individual anaerobic threshold

[0221] 513 third step of the procedure for determining a blood-related, individual anaerobic threshold

[0222] 514 fourth step of the procedure for determining a blood-related, individual anaerobic threshold

[0223] 521 First step of the procedure for monitoring a blood-related, individual anaerobic threshold

[0224] 522 second step of the procedure for monitoring a blood-related, individual anaerobic threshold S23 third step of the procedure for monitoring a blood-related, individual anaerobic threshold

[0225] T Time value

Claims

1. Patent claims 1. Method for determining a blood-related, individual anaerobic threshold (ANS) of a subject (10), comprising: - Obtaining initial data (Dl) which include performance values ​​(P) describing the subject's performance (10), time values ​​(T) describing a corresponding recording time, and corresponding salivary lactate concentration values ​​(KS), wherein the initial data (Dl) exceed a maximum salivary lactate concentration value (KS) (Si l), and - Obtaining the blood-related individual anaerobic threshold (ANS) based on the initial data (Dl) and using a data model (M) (S14).

2. Method according to claim 1, wherein the data model (M) comprises a trained data model (M) which uses the first data (Dl) as input data and is trained to output the blood-related individual anaerobic threshold (ANS).

3. A method according to claim 1 or 2, wherein the data model (M) comprises a regression model that describes a relationship between the blood-related individual anaerobic threshold (ANS) and a saliva-related individual anaerobic threshold (SANS), and / or wherein the data model (M) comprises a regression model that describes a relationship between the blood-related individual aerobic threshold (AES) and a saliva-related individual aerobic threshold (SAES), and / or wherein the data model (M) comprises a regression model that describes a relationship between a maximum salivary lactate concentration (KS) and a maximum blood lactate concentration (KB).

4. Method according to any of the preceding claims, wherein obtaining the blood-related individual anaerobic threshold (ANS) comprises obtaining a saliva-related individual anaerobic threshold (SANS) based on the first data (Dl) (S13).

5. The method of claim 4, wherein maintaining the saliva-related individual anaerobic threshold (SANS) comprises maintaining a saliva-related individual aerobic threshold (SAES) (S12).

6. Method according to claim 5, wherein the salivary individual aerobic threshold (SAES) is obtained as the performance value (P) at which a quotient of salivary lactate concentration value (KS) and associated performance value (P) is minimal, or wherein the salivary individual aerobic threshold (SAES) comprises an evaluation of a time rate of change of the salivary lactate concentration values ​​(KS).

7. A method according to any one of claims 4 to 6, wherein obtaining the saliva-related individual anaerobic threshold (SANS) comprises determining a maximum distance of a straight line extending from a minimum salivary lactate concentration value (KS) to the maximum salivary lactate concentration value (KS) or extending from a salivary lactate concentration value (KS) associated with an individual aerobic threshold (SAES) to the maximum salivary lactate concentration value (KS), from a curve connecting the salivary lactate concentration values ​​(KS), wherein the salivary-related individual anaerobic threshold (SANS) is obtained as a power value (P) at the point of maximum distance.

8. The method of claim 4, wherein obtaining the saliva-related individual anaerobic threshold (SANS) comprises evaluating a rate of change over time of the salivary lactate concentration values ​​(KS).

9. Method for monitoring a blood-related individual anaerobic threshold (ANS) of a subject (10), comprising: - Obtaining second data (D2) which includes salivary lactate concentration values ​​(KS) and time values ​​(T) describing the respective measurement time (S21), - Obtaining the blood-related individual anaerobic threshold (ANS) of the subject (10), wherein the blood-related individual anaerobic threshold (ANS) of the subject (10) is determined by the method according to one of the preceding claims (S22), or Obtaining a time-dependent rate of change of salivary lactate concentration (KS) values ​​and obtaining a saliva-related individual anaerobic threshold (SANS) based on the time-dependent rate of change of salivary lactate concentration (KS) values, and - Providing an indicator that describes a relationship of the second data (D2) to the saliva-based individual anaerobic threshold (SANS) or to the blood-based individual anaerobic threshold (ANS) (S23).

10. Method for training a data model (M) to determine a blood-related individual anaerobic threshold (ANS), wherein the data model (M) is trained using an annotated dataset comprising several training data elements (DT) which include performance values ​​(P) describing a performance, time values ​​(T) describing a respective recording time, and respective salivary lactate concentration values ​​(KS), wherein the training data elements (DT) are annotated with associated blood-related individual anaerobic thresholds (ANS).

11. Device (22, 42) for data processing, comprising means (34, 56) for carrying out at least one of the methods according to one of claims 1 to 10.

12. Computer program (32, 54), comprising instructions which, when the computer program (32, 54) is executed by a computer, cause it to execute at least one of the methods according to any one of claims 1 to 10.

13. Computer-readable medium (30, 52) comprising instructions which, when executed by a computer, cause it to execute at least one of the methods according to any one of claims 1 to 10.

14. System (16) for determining a blood-related, individual anaerobic threshold (ANS) of a subject (10), comprising: - a device (22) for data processing, comprising means (34) for carrying out the method according to any one of claims 1 to 8, - a first sensor unit (18) for detecting at least one salivary lactate concentration value (KS), - a second sensor unit (20) for recording at least one performance value (P) which describes a performance of the subject (10), - a time unit (24) for providing a time value (T), wherein the first sensor unit (18), the second sensor unit (20) and the time unit (24) are communicatively coupled to the device (22) for data processing or are components of the device (22) for data processing.

15. System ( 8) for monitoring a blood-related, individual anaerobic threshold (ANS) of a subject (10), comprising: - a device (42) for data processing, comprising means (56) for carrying out the method according to claim 9, - a first sensor unit (40) for detecting at least one salivary lactate concentration value (KS), - a time unit (44) to provide a time value (T), - an output interface (46) for outputting an indicator which describes a relationship of at least one salivary lactate concentration value (CS) recorded at an associated recording time (T) to the salivary individual anaerobic threshold (SANS) or to the blood individual anaerobic threshold (ANS), wherein the first sensor unit (40), the time unit (44) and the output interface (46) are coupled to the device (42) for data processing via communication technology or are components of the device (42) for data processing.

Images