Analyte concentration data-based processing device and related method

Abstract

Provided in the present disclosure are an analyte concentration data-based processing device and a related method. The processing device comprises an acquisition module and a processing module. The acquisition module is configured to receive first analyte data related to a glucose concentration and second analyte data related to a ketone body concentration obtained by continuously monitoring at least two analytes of a target subject using an analyte sensor. The processing module is configured to acquire, on the basis of the first analyte data and the second analyte data, analyte concentration data comprising glucose concentration data and ketone body concentration data, acquire, on the basis of the analyte concentration data within a time period, at least one feature set, and determine, on the basis of the at least one feature set, target information at least comprising energy information. The energy information is the relationship between estimated energy intake and energy expenditure of the target subject, where the energy intake is the energy derived from ingested carbohydrates, and the energy expenditure is the energy that is consumed. Therefore, the convenience of evaluating the relationship between carbohydrate intake and consumption can be improved.

Description

Analyte concentration data processing device and related methods Technical Field

[0001] This disclosure relates to health-related information systems, and in particular to a processing device and related method based on analyte concentration data. Background Technology

[0002] With the fast pace of modern life, people are paying increasing attention to health management. As a primary energy source for the human body, managing carbohydrate intake is crucial for maintaining good health. For example, properly controlling carbohydrate intake plays an important role in preventing and managing chronic diseases such as diabetes and obesity.

[0003] Currently, existing methods generally estimate an individual's total carbohydrate intake by recording the nutritional components and intake of food, and then estimate energy expenditure by calculating physical activity and basal metabolic rate, thereby determining the individual's carbohydrate intake and expenditure.

[0004] However, existing methods rely on individuals to keep detailed records and calculate their food intake, which requires individuals to have a high degree of self-monitoring ability and recording habits. In addition, in terms of estimating energy consumption, additional equipment is usually needed to monitor physical activity levels, which undoubtedly increases the difficulty of monitoring. Summary of the Invention

[0005] This disclosure is made in view of the above-mentioned situation, and its purpose is to provide an apparatus and related method for processing analyte concentration data that can improve the convenience of assessing the relationship between carbohydrate intake and consumption.

[0006] To this end, a first aspect of this disclosure provides a processing apparatus based on analyte concentration data, comprising an acquisition module and a processing module. The acquisition module is configured to receive first analyte data related to glucose concentration and second analyte data related to ketone body concentration obtained by continuous monitoring of at least two analytes of a target object by an analyte sensor. The processing module is configured to acquire analyte concentration data including glucose concentration data and ketone body concentration data based on the first analyte data and the second analyte data, acquire at least one feature set based on the analyte concentration data over a time period, and determine target information including at least energy information based on the at least one feature set. The energy information is an estimated relationship between the target object's energy intake and energy expenditure, wherein the energy intake is the energy from carbohydrates ingested by the target object, and the energy expenditure is the energy consumed by the target object. In this case, acquiring the relationship between the target object's energy intake and energy expenditure through glucose concentration data and ketone body concentration data improves the convenience of assessing the relationship between carbohydrate intake and expenditure compared to existing methods that rely on individual habits or require additional equipment. Furthermore, estimating the relationship between the target object's energy intake and energy expenditure facilitates quantitative assessment of energy metabolism. Additionally, combining glucose concentration data allows for quantitative measurement of energy intake, thereby improving the accuracy of energy information.

[0007] Additionally, in the processing apparatus according to the first aspect of this disclosure, optionally, the at least one feature set includes: a first feature set comprising raw data including the analyte concentration data, and / or a second feature set related to a scatter plot determined by the analyte concentration data, wherein the positions of the data points in the scatter plot are determined by the glucose concentration and ketone body concentration at the same time. In this case, the raw data contains all possible information, reducing the risk of losing important information and facilitating the observer to comprehensively identify useful information from the raw data. Furthermore, the scatter plot allows for a visual identification of the relationship between changes in glucose concentration and changes in ketone body concentration.

[0008] Additionally, in the processing apparatus according to the first aspect of this disclosure, optionally, the at least one feature set is further selected from at least one of the following options: a third feature set related to the analyte fluctuation curve determined by the analyte concentration data; a fourth feature set related to the analyte concentration data during the nighttime period within the time period; and a fifth feature set including statistical values ​​of the raw data of the analyte concentration data, wherein the statistical values ​​include a range. In this case, the third feature set is advantageous in highlighting the fluctuation characteristics of the concentration data, while the fluctuation characteristics of glucose concentration help identify energy intake, and the fluctuation characteristics of ketone body concentration help identify energy expenditure, thereby improving the accuracy of energy information. Furthermore, the fourth feature set reflects energy metabolism, which is beneficial for determining energy information. Additionally, the fifth feature set allows for a preliminary assessment of the relationship between energy intake and energy expenditure.

[0009] Furthermore, in the processing apparatus according to the first aspect of this disclosure, optionally, the second feature set includes at least one of the outline of the scatter plot, a first span, and a second span, and / or the scatter plot itself, wherein the first span is the span of glucose concentration, and the second span is the span of ketone body concentration. In this case, the scatter plot is more intuitive than the original data, which is beneficial for highlighting information related to energy metabolism. Additionally, the outline of the scatter plot helps the observer to roughly judge the relationship between changes in glucose concentration and changes in ketone body concentration. Furthermore, the span of the scatter plot helps reduce the influence of noise in the concentration data on the span, thereby obtaining more accurate first and second spans, which can be used to preliminarily determine the relationship between energy intake and energy expenditure.

[0010] Furthermore, in the processing apparatus according to the first aspect of this disclosure, optionally, the fourth feature set includes at least one of the average glucose concentration, the range of ketone body concentration, and proportionality, wherein the proportionality is the ratio between the average glucose concentration and the average ketone body concentration. In this case, the average glucose concentration during the nighttime period is affected by the energy intake during the day, and combining the average glucose concentration can improve the accuracy of energy information. Additionally, regarding the range of ketone body concentration, ketone body concentration fluctuates relatively regularly at night and can reflect energy metabolism; combining the range of ketone body concentration can improve the accuracy of energy information. Furthermore, regarding proportionality, the inventors have found an inverse relationship between the average glucose concentration and the average ketone body concentration at night, and that the average glucose concentration is related to energy intake; combining the proportionality can improve the accuracy of energy information.

[0011] Furthermore, in the processing apparatus according to the first aspect of this disclosure, the scatter plot may optionally be in the form of a heatmap. In this case, the concentration data has high repeatability, and the heatmap not only shows the distribution of data points but also the concentration of the distribution, providing more information and helping observers focus on important information in high-heat areas, thereby improving the accuracy of energy information.

[0012] Additionally, in the processing apparatus according to the first aspect of this disclosure, the time period may optionally be one day. In this case, the concentration data includes at least data on the energy intake process and the energy expenditure process, facilitating the estimation of the relationship between the energy intake and energy expenditure of the target object.

[0013] Furthermore, in the processing apparatus according to the first aspect of this disclosure, the energy information may optionally be determined by a target model, wherein the target model is a regression model. In this case, since the regression model predicts continuous values, the energy information can be continuous, which helps to quantitatively assess energy metabolism.

[0014] Furthermore, the processing apparatus according to the first aspect of this disclosure may optionally include a training module, which is used to train the target model based on historical concentration data of the target object. During the training of the target model, if the amount of historical concentration data is less than a preset amount, features are extracted from the historical concentration data, and the extracted features are used to train the target model; otherwise, the original historical concentration data is used to train the target model. This helps the target model learn corresponding feature representations with limited data and fully utilize the original data to learn more complex feature representations with large amounts of data.

[0015] A second aspect of this disclosure provides a processing method based on analyte concentration data, comprising: receiving first analyte data related to glucose concentration and second analyte data related to ketone body concentration obtained by continuous monitoring of at least two analytes of a target object by an analyte sensor; and acquiring analyte concentration data including glucose concentration data and ketone body concentration data based on the first analyte data and the second analyte data, acquiring at least one feature set based on the analyte concentration data over a time period, and determining target information including at least energy information based on the at least one feature set, wherein the energy information is an estimated relationship between the target object's energy intake and energy expenditure, the energy intake being the energy from carbohydrates ingested by the target object, and the energy expenditure being the energy consumed by the target object. In this case, acquiring the relationship between the target object's energy intake and energy expenditure through glucose concentration data and ketone body concentration data improves the convenience of assessing the relationship between carbohydrate intake and expenditure compared to existing methods that rely on individual habits or require additional equipment. Furthermore, estimating the relationship between the target object's energy intake and energy expenditure facilitates a quantitative assessment of energy metabolism. Additionally, combining glucose concentration data allows for a quantitative measurement of energy intake, thereby improving the accuracy of energy information.

[0016] According to this disclosure, an apparatus and related method for processing analyte concentration data are provided, which can improve the convenience of assessing the relationship between carbohydrate intake and consumption. Attached Figure Description

[0017] This disclosure will now be explained in further detail by way of example only with reference to the accompanying drawings.

[0018] Figure 1 is a schematic diagram illustrating a monitoring environment for simultaneous monitoring of glucose and ketone body concentrations as described in this disclosure example.

[0019] Figure 2 is a schematic diagram illustrating the control device involved in the example of this disclosure.

[0020] Figure 3A is an exemplary block diagram illustrating a first embodiment of the processing apparatus according to the present disclosure.

[0021] Figure 3B is an exemplary block diagram illustrating a second embodiment of the processing apparatus according to the present disclosure.

[0022] Figure 4 is an exemplary flowchart illustrating the execution process of the processing module involved in this disclosure example.

[0023] Figure 5A is a schematic diagram illustrating a scatter plot involved in the examples of this disclosure.

[0024] Figure 5B is a schematic diagram illustrating the glucose fluctuation curve involved in the example of this disclosure.

[0025] Figure 6 is an exemplary flowchart illustrating the processing method involved in the example of this disclosure. Detailed Implementation

[0026] Hereinafter, preferred embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. In the following description, the same reference numerals are used for the same components, and repeated descriptions are omitted. Furthermore, the drawings are merely schematic diagrams, and the proportions of the components or the shapes of the components may differ from actual figures. It should be noted that the terms "comprising" and "having," and any variations thereof, in this disclosure, do not necessarily limit the process, method, system, product, or apparatus to the explicitly listed steps or units, but may include or have other steps or units not explicitly listed or inherent to these processes, methods, products, or apparatuses.

[0027] First, let me introduce the relevant terminology used in this disclosure.

[0028] An "observer" can refer to any object that needs to obtain information from information related to the data being analyzed. For example, an observer could be a guardian, medical personnel, the target object, or the target model.

[0029] "Sample object" can refer to a user who provides data related to the training samples (such as analyte data) during the training phase of the model, or a user who provides data for analysis and statistics to determine some thresholds (such as empirical values).

[0030] As mentioned above, existing methods require individuals to have a high degree of self-monitoring ability and recording habits, and are also quite difficult to monitor. The inventors discovered through research that combining glucose concentration data and ketone body concentration data can serve as an indicator of the relationship (e.g., whether it is balanced) between the energy from ingested carbohydrates (i.e., the calories from ingested carbohydrates) and the energy expended. Hereinafter, the energy from ingested carbohydrates will be referred to as ingested energy, the energy expended as expended energy, and the relationship between ingested energy and expended energy as energy metabolism.

[0031] Specifically, carbohydrate intake can store glycogen, and glycogen depletion leads to gluconeogenesis. Increased gluconeogenesis results in elevated ketone body concentrations. Therefore, combining glucose and ketone body concentration data can identify features related to energy metabolism to determine its status. For example, regarding the range of glucose and ketone body concentrations: if the glucose concentration range is greater than a certain value (i.e., not particularly small) and the ketone body concentration range is less than a certain value (i.e., not particularly large), it can be estimated that energy intake essentially offsets energy expenditure. Conversely, if the glucose concentration range is less than a certain value (i.e., particularly small) and the ketone body concentration range is greater than a certain value (i.e., particularly large), it can be estimated that energy intake cannot offset energy expenditure. It should be noted that the specific range of glucose and ketone body concentration ranges is not specifically limited in this disclosure; appropriate thresholds can be set by statistically analyzing analyte data from samples with known energy metabolism status.

[0032] Therefore, the inventors have provided several solutions, and corresponding embodiments that at least partially address the aforementioned problems, thereby improving the ease of assessing the relationship between carbohydrate intake and consumption. Examples of this disclosure will be described in detail below.

[0033] The analyte monitoring system disclosed herein can be configured to continuously monitor analyte levels. In some examples, the analyte monitoring system may include a sensor control device (hereinafter referred to as the control device). The control device may be configured to acquire signals related to the analyte level (hereinafter referred to as analyte signals), process the analyte signals, and / or transmit information related to the analyte signals (hereinafter referred to as analyte information).

[0034] In some examples, the control device may include an analyte sensor and electronic components. The analyte sensor may be configured to generate an analyte signal, and the electronic components may be configured to receive the analyte signal, process the analyte signal, and / or transmit analyte information to a receiving device. In some examples, the analyte monitoring system may also include a receiving device that receives data from the control device.

[0035] In some examples, analyte information can be an analyte signal and / or an analyte level. The analyte signal can be any signal that corresponds to an analyte level, and the analyte signal can be converted into an analyte level. In some examples, when the analyte sensor is based on electrochemical measurement, the analyte signal can be an electrical signal, from which the corresponding analyte level can be obtained.

[0036] In some examples, for the analyte being glucose, the analyte sensor can be a glucose sensor, and the analyte level can be the glucose concentration. In some examples, for the analyte being ketone bodies, the analyte sensor can be a ketone body sensor, and the analyte level can be the ketone body concentration.

[0037] In some examples, the analyte sensor can be a multi-analyte sensor. Specifically, the analyte sensor can be a sensor that simultaneously monitors glucose concentration and ketone body concentration. In this case, the time synchronization of glucose concentration data and ketone body concentration data can be improved, and compared to using separate devices to collect glucose concentration data and ketone body concentration data separately, the usage process can be simplified, the purchase cost can be reduced, and the wearing space can be saved.

[0038] Additionally, analyte data can be data related to analyte levels. Analyte data may include analyte information at at least one time point. Furthermore, analyte concentration data may include analyte levels at at least one time point (hereinafter referred to as concentration data).

[0039] For ease of description, data related to glucose concentration will be referred to as first analyte data, and data related to ketone body concentration will be referred to as second analyte data. Furthermore, some examples illustrate simultaneous monitoring of glucose and ketone body concentrations; however, this is not intended to limit this disclosure. Unless contradictory, the descriptions also apply to any other monitoring method capable of obtaining continuous glucose and ketone body concentrations.

[0040] Examples of this disclosure will now be described in detail with reference to the accompanying drawings. Figure 1 is a schematic diagram illustrating a monitoring environment for simultaneously monitoring glucose and ketone body concentrations as described in an example of this disclosure.

[0041] Referring to Figure 1, the monitoring environment may include a control device 7. The control device 7 can be configured to acquire analyte data (i.e., first analyte data and second analyte data) of the target object 8. Specifically, a portion of the control device 7 can be located on the surface of the target object 8, and another portion can be located subcutaneously, thereby enabling monitoring of analyte levels. It should be noted that although the application scenario of the control device 7 described herein is exemplarily based on monitoring analyte levels by placing a portion of the control device 7 subcutaneously on the target object 8, this does not constitute a limitation of this disclosure.

[0042] As described above, in some examples, the analyte monitoring system may also include a receiving device 9 that receives data from the control device 7. In some examples, the monitoring environment may also include the receiving device 9, to which the control device 7 may be coupled. In some examples, the control device 7 may be directly or indirectly communicatively coupled to the receiving device 9. The control device 7 may be communicatively coupled to the receiving device 9 via one or more communication links. For example, the communication link may include at least one of a proprietary wireless protocol, a wired communication link (e.g., serial communication), and a wireless communication link (e.g., Bluetooth).

[0043] In some examples, the receiving device 9 can be a device with a display unit. This facilitates the provision of rich information related to the analyte data. In some examples, the receiving device 9 can be independent of the control device 7. Specifically, the receiving device 9 can be an electronic device arranged outside of and communicatively connected to the control device 7. For example, the receiving device 9 can be a mobile device. Alternatively, the receiving device 9 can also be a smartphone, tablet computer, or wearable device.

[0044] In some examples, the monitoring environment may also include a processing device 1, which can be used to determine target information based on concentration data. Alternatively, the concentration data can be determined from analyte data. In some examples, the processing device 1 can receive analyte data from a control device 7 or a receiving device 9.

[0045] In some examples, the processing device 1 may be independent of the control device 7 or the receiving device 9. Specifically, the processing device 1 may be a device located outside of and communicatively connected to the control device 7 or the receiving device 9. For example, the processing device 1 may be a server or a dedicated device.

[0046] In some examples, the processing device 1 can be integrated into the control device 7 or the receiving device 9. That is, the control device 7 or the receiving device 9 can be used to determine target information based on concentration data. In other words, the operation of determining target information based on concentration data can be performed in the control device 7 or the receiving device 9.

[0047] Figure 2 is a schematic diagram illustrating the control device 7 involved in the example of this disclosure.

[0048] Before describing the processing apparatus 1 in detail, let's first introduce the control apparatus 7 involved in this disclosure. Referring to FIG2, in some examples, the control apparatus 7 may include an analyte sensor 71 and an electronic component 72. The analyte sensor 71 may be configured to generate a signal related to glucose concentration (hereinafter referred to as a first signal) and a signal related to ketone body concentration (hereinafter referred to as a second signal), respectively, and the electronic component 72 may be configured to receive the first signal and the second signal.

[0049] In some examples, the analyte sensor 71 can be configured to be at least partially implanted under the skin of the target object 8. This allows at least a portion of the analyte sensor 71 to come into contact with the subcutaneous fluid of the target object 8 to generate an analyte signal. In other examples, the analyte sensor 71 may not come into contact with the subcutaneous fluid of the target object 8. For example, when the analyte sensor 71 is based on optical signal measurement, the analyte signal can be obtained without contact with the subcutaneous fluid of the target object 8 via optical signal.

[0050] In some examples, the analyte sensor 71 may include a working electrode. In some examples, there may be multiple working electrodes, and each working electrode may be used to detect different analytes. Thus, it is possible to monitor multiple analytes. In some examples, the working electrode may include a first working electrode and a second working electrode, the first working electrode being used to detect glucose and the second working electrode being used to detect ketone bodies. In some examples, glucose and ketone bodies can be monitored separately by placing enzyme layers of two analytes on the two working electrodes.

[0051] Furthermore, ketone bodies can be a collective term for acetoacetic acid, β-hydroxybutyric acid, and acetone, intermediate products of fatty acid oxidation in the liver. In the examples of this disclosure, ketone bodies can be monitored by detecting any one or more of acetoacetic acid, β-hydroxybutyric acid, and acetone. That is, ketone bodies can be at least one of acetoacetic acid, β-hydroxybutyric acid, and acetone. Preferably, parameters related to ketone bodies can be obtained by monitoring β-hydroxybutyric acid. In this case, β-hydroxybutyric acid is more sensitive and representative in reflecting ketone body levels, which can improve the reliability of measuring ketone body concentration.

[0052] Referring again to Figure 2, in some examples, the electronic component 72 can be connected to the analyte sensor 71 to receive analyte signals (i.e., the first signal and the second signal) from the analyte sensor 71. In some examples, the electronic component 72 can be electrically connected to the analyte sensor 71. In some examples, the electronic component 72 can be configured to adhere to the surface of the target object 8 (i.e., the skin surface). That is, in some states (e.g., when using the control device 7), the electronic component 72 can adhere to the surface of the target object 8. This facilitates cooperation with the analyte sensor 71 to receive analyte signals.

[0053] FIG3A is an exemplary block diagram illustrating a first embodiment of the processing apparatus 1 according to the present disclosure. FIG3B is an exemplary block diagram illustrating a second embodiment of the processing apparatus 1 according to the present disclosure.

[0054] Referring to Figure 3A, in some examples, the processing device 1 may include an acquisition module 11 and a processing module 12. The acquisition module 11 may be used to receive analyte data (i.e., first analyte data and second analyte data), and the processing module 12 may be used to determine target information (e.g., energy information described later) based on the analyte data from the acquisition module 11.

[0055] Referring to Figure 3B, in some examples, the processing device 1 may further include a training module 13. The training module 13 can be used to train a target model for determining energy information within the target information based on historical concentration data of the target object 8. This allows for the acquisition of a target model more suitable for the target object 8.

[0056] Referring to Figure 3B, in some examples, the processing device 1 may also include a display module 14. The display module 14 may be used to display target information and / or information related to the analyte data.

[0057] Referring back to Figure 3A, the acquisition module 11 can acquire analyte data from any location. In some examples, the acquisition module 11 can acquire analyte data from the analyte sensor 71.

[0058] In some examples, analyte data can be obtained by continuously monitoring at least two analytes of the target object 8 using an analyte sensor 71. This allows for the acquisition of continuous analyte information. For example, the analyte sensor 71 can acquire analyte signals at preset intervals, and the analyte signals and / or analyte levels determined based on the analyte signals can be received by the acquisition module 11. Furthermore, the preset interval can be set according to monitoring requirements. In some examples, the preset interval can be 1 minute, 3 minutes, 4 minutes, 5 minutes, 8 minutes, 10 minutes, or 15 minutes.

[0059] Figure 4 is an exemplary flowchart illustrating the execution process of the processing module 12 involved in this disclosure example.

[0060] As described above, processing module 12 can be used to determine target information based on analyte data from source acquisition module 11. Referring to FIG4, in some examples, the execution process of processing module 12 may include acquiring concentration data including glucose concentration data and ketone body concentration data (step S101), acquiring at least one feature set based on the concentration data within a time period (step S102), and determining target information including at least energy information based on at least one feature set (step S103). Furthermore, the energy information can be the estimated relationship between energy intake and energy expenditure of the target subject 8. In this case, acquiring the relationship between energy intake and energy expenditure of the target subject 8 through glucose concentration data and ketone body concentration data (i.e., by combining glucose concentration data and ketone body concentration data, an estimate of the energy expenditure balance can be made) improves the convenience of assessing the relationship between carbohydrate intake and expenditure compared to existing methods that rely on individual habits or require additional equipment. Furthermore, by estimating the relationship between energy intake and energy expenditure of the target subject 8, a quantitative assessment of energy metabolism can be achieved. Additionally, combining glucose concentration data allows for the quantitative measurement of energy intake, thereby improving the accuracy of energy information.

[0061] It should be noted that in some other examples, the target information may not include energy information. That is, unless there is a contradiction, the execution process of processing module 12 can also be applied to other information.

[0062] Referring to Figure 4, in some examples, in step S101, concentration data is obtained based on the analyte data. In some examples, if the analyte information in the analyte data is the analyte level, the analyte data can be directly used as concentration data. For example, if the analyte information is glucose concentration and ketone body concentration, the glucose concentration data can include multiple consecutive glucose concentrations, and the ketone body concentration data can include multiple consecutive ketone body concentrations. In some examples, if the analyte information in the analyte data is the analyte signal, the analyte signal can be converted into the analyte level to obtain concentration data.

[0063] Referring to Figure 4, in some examples, in step S102, data from a single time period can be selected from the received concentration data using the time period as a search criterion to obtain at least one feature set.

[0064] Furthermore, the time period can be a period that involves at least both energy intake and energy expenditure processes. In some examples, the time period can be one day. That is, a feature set can be obtained based on daily concentration data. In this case, the concentration data includes at least data on both energy intake and energy expenditure processes, which facilitates the estimation of the relationship between energy intake and energy expenditure for target object 8. In addition, the glucose concentration data for one day can also reflect the energy intake throughout the day, making the energy intake information carried in the glucose concentration data more accurate.

[0065] In some examples, the feature set may include at least one of the features of the raw concentration data and features obtained by feature extraction from the concentration data. The inventors have found that scatter plots are helpful in assessing energy metabolism. In some examples, at least one feature set may include a first feature set associated with the raw data and / or a second feature set associated with the scatter plot. In some examples, at least one feature set may also be selected from at least one of the following options: a third feature set, a fourth feature set, and a fifth feature set. Each feature set is described in detail below.

[0066] In some examples, the first feature set may include the raw data. In this case, the raw data contains all possible information, reducing the risk of missing important information and facilitating the observer (e.g., a user or the target model described later) to comprehensively identify useful information from the raw data. In some examples, the first feature set may include time-series data of glucose concentration and time-series data of ketone body concentration. Additionally, the time-series data may be analyte levels arranged chronologically.

[0067] The inventors discovered that scatter plots determined by glucose and ketone body concentrations help identify the relationship between changes in glucose and ketone body concentrations to assess energy metabolism, and also make comparisons between glucose and ketone body concentrations more intuitive. For example, the span of the scatter plot in the horizontal and vertical directions can identify the corresponding changes in ketone body concentration over a time period as glucose concentration changes, and these changes can provide a preliminary assessment of energy metabolism.

[0068] An example is as follows: assuming that the horizontal span represents the span of glucose concentration and the vertical span represents the span of ketone body concentration, if the scatter plot is narrow in the horizontal direction and long in the vertical direction, that is, the scatter plot is roughly a narrow strip with the narrow side representing glucose concentration, then it can be preliminarily concluded that the target subject 8 has very little energy intake, which is insufficient to cover energy consumption.

[0069] Based on the above findings, the inventors introduced a second feature set. This second feature set can be correlated with a scatter plot determined by concentration data. That is, a scatter plot can be generated based on concentration data, and the second feature set can be obtained from the scatter plot. This allows for a direct identification of the relationship between changes in glucose concentration and changes in ketone body concentration.

[0070] Additionally, a scatter plot can include multiple data points, the locations of which can be determined by glucose and ketone body concentrations at the same time. Furthermore, "the same time" can be any time that allows for correlation between glucose and ketone body concentrations collected at similar or identical times. In some examples, "the same time" can refer to the same instant or the same time period.

[0071] Figure 5A is a schematic diagram illustrating a scatter plot related to the examples of this disclosure. Figure 5B is a schematic diagram illustrating a glucose fluctuation curve related to the examples of this disclosure.

[0072] In some examples, the scale of the first axis of the scatter plot can be a scale for glucose concentration, and the scale of the second axis can be a scale for ketone body concentration. As an example, Figure 5A shows a scatter plot for a single time period, where the scale of the horizontal axis is a scale for glucose concentration and the scale of the vertical axis is a scale for ketone body concentration, with the units being mmol / L (millimoles per liter).

[0073] In some examples, the second feature set may include the outline of the scatter plot, at least one of a first span and a second span, and / or the scatter plot itself. That is, the second feature set may be the scatter plot itself, features extracted from the scatter plot, or a combination of both. Additionally, the first span may be a span of glucose concentration, and the second span may be a span of ketone body concentration. As an example, line L1 in Figure 5A represents the first span, and line L2 represents the second span.

[0074] In this context, scatter plots are more intuitive than raw data and are helpful in highlighting information related to energy metabolism. Furthermore, when displaying scatter plots to observers, it helps them identify the lifestyle patterns of the target subject (e.g., whether exercise is strenuous, carbohydrate intake is high, or carbohydrate intake is controlled) reflected in the concentration data, as well as noise in the concentration data.

[0075] Furthermore, the outline of a scatter plot helps observers to roughly determine the relationship between changes in glucose concentration and ketone body concentration. For example, the height or width of the outline can provide a general indication of energy metabolism. Also, if there is little noise in the concentration data, and energy metabolism is relatively balanced, the outline of the scatter plot may exhibit a hysteretic distribution. That is, there may be no data points in the middle of the scatter plot. Specifically, when ketone body concentration is high, if carbohydrate intake begins, glucose concentration starts to rise. At this point, gluconeogenesis decreases, and ketone body concentration begins to slowly decrease. When glucose concentration begins to decrease, gluconeogenesis has not yet occurred due to the presence of glycogen reserves, and ketone body concentration remains essentially unchanged. When glucose concentration decreases to a relatively stable level, ketone body concentration begins to rise. These processes are reflected in the scatter plot as a hysteretic distribution.

[0076] In addition, the span of the scatter plot helps to reduce the impact of noise in the concentration data on the span, thereby obtaining a more accurate first span and second span. The first span and second span can be used to make a preliminary judgment on the relationship between energy intake and energy consumption.

[0077] It should be noted that the span of the scatter plot can be the span of the corresponding analyte level on the scatter plot, the span of the subsequent heat area can be the span of the corresponding analyte level within the heat area, and the span involved in the subsequent fifth feature set can be the span of the corresponding analyte level in the original data.

[0078] In some examples, the second feature set may also include the density of the region of interest (GIO) in the scatter plot (e.g., the region corresponding to the target range). In this case, it reflects whether the data points fall within the GIO, facilitating the identification of the correlation between the data point distribution in the GIO and energy metabolism. Additionally, the target range can refer to a relatively safe interval for the analyte level.

[0079] In some examples, the scatter plot can take the form of a heatmap (see Figure 5A). In this case, the concentration data has high repetition, and the heatmap not only shows the distribution of data points but also the concentration of the distribution, providing more information and helping observers focus on important information in high-heat areas, thereby improving the accuracy of energy information. In some examples, the density of data points can be represented by color in the heatmap. For example, referring to Figure 5A, the brighter the color, the greater the density.

[0080] In some examples, for scatter plots in the form of heatmaps, heat regions with a density greater than a preset density can be identified based on a color threshold, and a second feature set can be obtained based on these heat regions. That is, regions with higher density are selected from the scatter plot to obtain the second feature set. In this case, it is beneficial to identify analyte data that is noisy, making the heat regions representative and less affected by noise, thus improving the accuracy of the second feature set.

[0081] Additionally, the color threshold can be an empirical value or determined based on the accuracy of the energy information. In some examples, one or more color thresholds can be determined based on the color intensity on the heatmap.

[0082] In some examples, the second feature set may include at least one of the contour, first span, second span, and density of the heat region. In some examples, the second feature set may also include the heat region itself. That is, the second feature set may include at least one of the contour, first span, second span, and density of the heat region, and / or the heat region.

[0083] Additionally, the third feature set can be correlated with the analyte fluctuation curve determined from the concentration data. As an example, Figure 5B shows a schematic of the glucose fluctuation curve determined from glucose concentration data over a single time period; a similar ketone body fluctuation curve is shown (not shown here).

[0084] In some examples, the third feature set may include glucose fluctuation curves and ketone body fluctuation curves, or glucose and ketone body fluctuation curves. Alternatively, the glucose and ketone body fluctuation curves can be curves where the glucose and ketone body fluctuation curves are relative to the same coordinate system. That is, the glucose and ketone body fluctuation curves can be used as two features, or they can be generated in the same coordinate system and used as a single feature. In this case, it is beneficial to highlight the fluctuation characteristics of the concentration data; the fluctuation characteristics of glucose concentration help identify energy intake, while the fluctuation characteristics of ketone body concentration help identify energy expenditure, thereby improving the accuracy of energy information.

[0085] Furthermore, the fourth feature set can be correlated with the concentration data during the nighttime period within the time cycle. Specifically, the fourth feature set can be obtained by extracting features from the concentration data during the nighttime period (hereinafter referred to as nighttime data). In this case, nighttime data is generally less susceptible to external interference, and its fluctuations are more regular compared to daytime concentration data (for example, daytime ketone body concentration fluctuations are generally not very regular), and some of the target object's daytime behaviors will also be reflected in the nighttime data. For example, the inventors found that under normal dietary conditions, ketone body concentration is relatively stable at night, while under low-carb diets, ketone body concentration gradually increases at night, and the glucose concentration at night remains stable within a relatively narrow range compared to normal dietary conditions. That is, nighttime data can reflect energy metabolism and is helpful in determining energy information. Accordingly, the inventors considered the range of average glucose and ketone body concentrations when extracting features from the nighttime data.

[0086] Furthermore, by fitting the average glucose concentration (hereinafter referred to as average glucose) and the average ketone body concentration (hereinafter referred to as average ketone body) of nighttime data, the inventors found that when the average glucose increases, the average ketone body decreases, and when the average glucose decreases, the average ketone body increases; that is, there is an inverse relationship between the average glucose and the average ketone body. In addition, changes in the average glucose are related to energy intake (for example, the average glucose from a normal diet is higher than the average glucose from a low-carbohydrate diet).

[0087] In some examples, the fourth feature set may include at least one of average glucose, the span of ketone body concentration, and proportionality, where proportionality can be the ratio between average glucose and average ketone body concentration. In this case, average glucose during the nighttime period is affected by daytime energy intake, and combining it with average glucose can improve the accuracy of energy information. Furthermore, regarding the span of ketone body concentration, as mentioned above, ketone body concentration fluctuates relatively regularly at night and can reflect energy metabolism; combining it with the span of ketone body concentration can improve the accuracy of energy information. Additionally, regarding proportionality, as mentioned above, the inventors have found an inverse relationship between average glucose and average ketone body concentration at night, and that average glucose is related to energy intake; combining it with proportionality can improve the accuracy of energy information.

[0088] In addition, the nighttime hours can be set according to the daily routine of target object 8, and this disclosure does not specifically limit them. For example, the nighttime hours can be from 10 pm to 6 am.

[0089] Additionally, the fifth feature set can include statistical values ​​of the raw concentration data. In some examples, these statistical values ​​can include ranges. Furthermore, the range (i.e., the span) can be the span of glucose concentration and the span of ketone body concentration. Thus, the relationship between energy intake and energy expenditure can be preliminarily determined by the span of glucose concentration and the span of ketone body concentration.

[0090] For ease of description, the features in at least one feature set will be referred to as a feature list. In some examples, the feature list includes at least features that reflect energy intake and expenditure over the entire time period. In some examples, the feature list may also be determined by combining the characteristics of data from specific time periods within the time period or the amount of concentration data. Therefore, this disclosure provides some examples of feature lists, which are not intended to limit this disclosure, as follows:

[0091] 1. The feature list consists of features from the first feature set. For example, time-series data of glucose concentration and ketone body concentration can be directly input into the target model, described later, to obtain energy information.

[0092] 2. The feature list is a scatter plot or heat map of the second feature set. For example, the scatter plot or heat map can be directly input into the target model to obtain energy information.

[0093] 3. The feature list is a third feature set. For example, glucose fluctuation curves and ketone body fluctuation curves, or glucose-ketone body fluctuation curves, can be directly input into the target model to obtain energy information.

[0094] 4. The feature list is a combination of the first feature list, the second feature list, or the third feature list with features from the second feature set excluding scatter plots or heat regions, features from the fourth feature set, and / or features from the fifth feature set.

[0095] 5. The feature list is a combination of at least two lists from the first feature list, the second feature list, or the third feature list with features from the second feature set excluding scatter plots or heat regions, features from the fourth feature set, and / or features from the fifth feature set.

[0096] Referring back to Figure 4, in some examples, energy information can be determined through a target model in step S103, where the target model can be a machine learning model. In this case, the relationship between features in the feature set and energy information can be automatically learned to obtain energy information. In other examples, association rules between features in the feature set and energy information can also be established, and energy information can be determined through these association rules.

[0097] In some examples, at least one feature set can be input into the target model to output energy information. In other examples, personal baseline information and at least one feature set can be input into the target model to output energy information. For example, personal baseline information may include age, gender, height, or weight.

[0098] In some examples, the target model can be a regression model. In this case, because the regression model predicts continuous values, it allows energy information to be continuous, which helps in the quantitative assessment of energy metabolism. In other examples, the target model can also be a non-regression model, such as a classification model.

[0099] In some examples, the regression model may include at least one of multinomial regression, decision tree regression, random forest regression, support vector machine regression (SVR), and neural network regression.

[0100] In some examples, a target model can be trained based on different feature combinations from at least one feature set to obtain target models for each feature combination. The final energy information is obtained by combining the energy information of multiple target models. For example, the final energy information can be the mean, median, mode, minimum, or other values ​​of multiple energy information, depending on the specific application scenario of the energy information.

[0101] As mentioned above, target information can include energy information. In some examples, energy information can be used to guide diet in the next time period. For instance, energy information can be used to estimate how much food target 8 should eat the next day if they do not exercise or how much exercise they do.

[0102] As mentioned above, energy information can be used to estimate the relationship between energy intake and energy expenditure for target object 8. In some examples, the relationship can be the difference between energy intake and energy expenditure. This allows for a quantitative assessment of energy metabolism. In some examples, the relationship can be whether energy intake and energy expenditure are balanced. In some examples, the relationship can be whether energy intake is insufficient.

[0103] In some examples, the difference between energy intake and energy expenditure, and whether they are balanced or insufficient, can be directly obtained from the target model. Specifically, this can be achieved by adjusting the labels of the training samples during target model training. In other examples, the difference between energy intake and energy expenditure can also be directly obtained from the target model, and then the balance or insufficiency of energy intake can be determined from this difference. For example, when the difference is close to 0, energy intake and energy expenditure can be considered balanced.

[0104] In some examples, the target information may also include at least one of exercise volume information, carbohydrate control information, carbohydrate intake information, weight control recommendations, and ketogenic recommendations. This allows for the estimation of the lifestyle of the target subject 8. It should be noted that this disclosure has described the information and principles used to determine various target information, without specifically limiting the specific process. Those skilled in the art, guided by this disclosure, can determine the target information by setting appropriate thresholds (e.g., particularly small, particularly large, relatively obvious, etc.) or rules based on observed data (e.g., the range of glucose concentration and the range of ketone body concentration).

[0105] Additionally, exercise intensity information can be derived from the range of glucose and ketone concentrations. In some examples, exercise intensity information can be generated from the range of glucose and ketone concentrations. For instance, if the range of glucose concentration for subject 8 is particularly large, while the range of ketone concentration is not particularly small, it could indicate that subject 8 has consumed a significant amount of energy and engaged in vigorous exercise, thereby largely expending the ingested energy and causing a noticeable change in ketone concentration (e.g., ketone concentration during the nighttime).

[0106] Additionally, carbohydrate control information can indicate whether energy intake is being controlled. For example, significant changes in glucose concentration before and after meals suggest that target subject 8 does not intentionally control carbohydrate intake.

[0107] In addition, carbohydrate intake information can be information about energy intake. For example, carbohydrate intake information can include high energy intake, low energy intake, and very low energy intake. In some examples, carbohydrate intake information can be generated from changes in glucose concentration before and after meals, as well as the range of ketone body concentration. For example, if target subject 8's glucose concentration before and after meals does not change significantly, but the range of ketone body concentration is too large, it can be identified that target subject 8's energy intake is very low.

[0108] Furthermore, weight control recommendations can be related to gaining or losing weight. In some examples, weight control recommendations can be generated based on energy information. Specifically, the relationship between energy intake and energy expenditure can be identified through energy information, and the direction of adjustment for energy intake and expenditure can be determined based on the target subject's weight goal, thereby outputting corresponding recommendations based on the adjustment direction.

[0109] Additionally, ketogenic recommendations can be related to the production of ketone bodies. In some examples, ketogenic recommendations can be generated based on energy information. Specifically, energy information can be used to identify the relationship between energy intake and energy expenditure to determine whether adjustments to energy intake and expenditure are needed to induce ketosis, and then corresponding recommendations can be generated.

[0110] Referring back to Figure 3B, in some examples, training module 13 can be used to train a target model for determining energy information based on historical concentration data of target object 8. In some examples, before starting to determine energy information, training module 13 can be used to train the target model using historical concentration data of target object 8 over a preset number of time periods. This allows for the acquisition of a target model more suitable for target object 8. Furthermore, the historical concentration data can be historical data of concentration data.

[0111] In some examples, at least one feature set can be obtained based on the historical concentration data of target object 8, and the target model can be trained based on at least one feature set. A description of at least one feature set is provided in the relevant description of processing module 12, and will not be repeated here.

[0112] In some examples, scatter plots or heatmaps can be directly input into the target model. In other examples, for scatter plots or heatmaps, image features can first be extracted from them using a feature extraction network, and then combined with other features from at least one feature set to train the target model.

[0113] In some examples, training module 13 trains the target model online. This allows the target model to quickly adapt to real-time changes in the data, thereby improving the accuracy of energy information.

[0114] In some examples, for the target model used to determine energy information, the labels of training samples can be the relationship between energy intake and energy expenditure. In some examples, the labels of training samples can be estimated based on physiological indicators related to the physical activity level of target object 8. In some examples, energy expenditure can be estimated using heart rate to determine the labels of training samples. For example, using the difference as a relation, the energy expenditure of target object 8 can be estimated using heart rate, and then the difference can be obtained by subtracting the energy intake of target object 8. In some examples, the labels of training samples can be determined through a user interface. That is, the labels of training samples can be input by the user (e.g., target object 8).

[0115] In some examples, the target model can be trained based on the amount of historical concentration data for target object 8. In other examples, the amount of historical concentration data for target object 8 can be used to determine whether to extract features from the raw historical concentration data for training the target model. This helps the target model learn appropriate feature representations with limited data and fully utilize the raw data to learn more complex feature representations with large amounts of data.

[0116] Specifically, in training the target model, if the amount of historical concentration data is less than a preset amount, features are extracted from the historical concentration data, and the extracted features are used to train the target model; otherwise, the original data (e.g., time series data of glucose concentration and time series data of ketone body concentration) are used to train the target model. Furthermore, the preset amount can be set as needed. For example, it can be set based on the performance of the target model. Alternatively, the preset amount can also be set manually by the user.

[0117] Furthermore, features extracted from historical concentration data can come from the second, third, fourth, and / or fifth feature sets mentioned above. Preferably, features extracted from historical concentration data can come from the second and / or fourth feature sets. In this case, the target model can learn the energy metabolism-related information reflected in the scatter plot and / or nighttime data, thereby improving the accuracy of energy information.

[0118] In some examples, the target model trained using extracted features has a simpler structure than the target model trained using the original data. In such cases, a simpler target model requires fewer parameters to be tuned, allowing for better performance with less data. Conversely, when dealing with large datasets, using a more complex target model allows it to automatically learn more sophisticated feature representations from the original data.

[0119] In some examples, a base model can be trained using historical concentration data from multiple sample objects, and this base model can then be used as the initial model for the target model. In this case, using the base model as a starting point reduces training time and the amount of training data required. Furthermore, the amount of historical concentration data from the sample objects can be larger than that from the target object. This helps the base model learn more general feature representations.

[0120] In some examples, the base model can be trained on a remote device, and the trained base model can be deployed in training module 13. This reduces the resource consumption of training module 13.

[0121] In some examples, for the base model, the label of the training samples can be the relationship between energy intake and energy expenditure. In some examples, the label of the training samples can be estimated based on physiological indicators related to the physical activity level of the sample subjects. In some examples, heart rate can be used to estimate energy expenditure to determine the label of the training samples. In some examples, a metabolic bin can be used to determine energy expenditure to determine the label of the training samples.

[0122] Referring again to Figure 3B, in some examples, the display module 14 can be used to receive and display target information from the processing module 12. In some examples, the display module 14 can also be used to display a first view including a scatter plot. In some examples, the first view may include target information. This facilitates the observer's analysis of energy metabolism by combining the scatter plot and target information. In some examples, the display module 14 can be a monitor, display screen, or electronic paper.

[0123] Figure 6 is an exemplary flowchart illustrating the processing method involved in the example of this disclosure.

[0124] Furthermore, the examples in this disclosure also provide a processing method based on analyte concentration data, which may also be referred to as a processing method or estimation method. The processing method can be implemented by at least a portion of the components in the computing device. For example, the processing method can be implemented by the processor of processing device 1 or receiving device 9. It should be noted that, unless contradictory, the above description regarding processing device 1 also applies to the processing method.

[0125] Referring to Figure 6, in some examples, the processing method may include receiving analyte data (i.e., first analyte data and second analyte data) (step S201) and determining target information based on the analyte data (step S202).

[0126] In some examples, analyte data can be acquired from the analyte sensor 71 in step S201. In some examples, the analyte data can be obtained by the analyte sensor 71 continuously monitoring at least two analytes of the target object 8. See the relevant description of the acquisition module 11 for details.

[0127] In some examples, in step S202, concentration data including glucose concentration data and ketone body concentration data can be acquired. In some examples, concentration data can be acquired based on analyte data. See the relevant description of processing module 12 for details.

[0128] In some examples, in step S202, at least one feature set can be obtained based on concentration data within a time period. See the relevant description of processing module 12 for details.

[0129] In some examples, in step S202, target information, including at least energy information, can be determined based on at least one feature set. Additionally, the energy information can be the estimated relationship between the energy intake and energy expenditure of the target object 8. See the relevant description of processing module 12 for details.

[0130] Examples of this disclosure also disclose a computer-readable storage medium that can store at least one instruction, which, when executed by a processor, implements one or more steps of the processing method described above. The computer-readable storage medium may include, but is not limited to, any type of disk, including floppy disks, optical disks, DVDs, CD-ROMs, microdrives, and magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic cards or optical cards, nanosystems (including molecular memory ICs), or any type of medium or device suitable for storing instructions and / or data.

[0131] While the present disclosure has been specifically described above in conjunction with the accompanying drawings and examples, it is to be understood that the foregoing description does not limit the present disclosure in any way. Those skilled in the art can make modifications and variations to the present disclosure as needed without departing from its essential spirit and scope, and all such modifications and variations shall fall within the scope of the present disclosure.

Claims

1. A processing device based on analyte concentration data, characterized in that, The system includes an acquisition module and a processing module. The acquisition module receives first analyte data related to glucose concentration and second analyte data related to ketone body concentration obtained by continuous monitoring of at least two analytes of the target object by an analyte sensor. The processing module acquires analyte concentration data including glucose concentration data and ketone body concentration data based on the first analyte data and the second analyte data, acquires at least one feature set based on the analyte concentration data within a time period, and determines target information including at least energy information based on the at least one feature set. The energy information is an estimated relationship between the target object's energy intake and energy expenditure, where the energy intake is the energy from carbohydrates ingested by the target object, and the energy expenditure is the energy consumed by the target object.

2. The processing apparatus according to claim 1, characterized in that, The at least one feature set includes: a first feature set comprising raw data of the analyte concentration data, and / or a second feature set relating to a scatter plot determined by the analyte concentration data, wherein the positions of the data points in the scatter plot are determined by glucose concentration and ketone body concentration at the same time.

3. The processing apparatus according to claim 2, characterized in that, The at least one feature set is further selected from at least one of the following options: a third feature set related to the analyte fluctuation curve determined by the analyte concentration data; a fourth feature set related to the analyte concentration data during the nighttime period within the time period; and a fifth feature set including statistical values ​​of the raw data of the analyte concentration data, wherein the statistical values ​​include a range.

4. The processing apparatus according to claim 2, characterized in that, The second feature set includes at least one of the outline of the scatter plot, a first span and a second span, and / or the scatter plot, wherein the first span is the span of glucose concentration and the second span is the span of ketone body concentration.

5. The processing apparatus according to claim 3, characterized in that, The fourth feature set includes at least one of the average glucose concentration, the span of ketone body concentration, and proportionality, wherein the proportionality is the ratio between the average glucose concentration and the average ketone body concentration.

6. The processing apparatus according to claim 2, characterized in that, The scatter plot is in the form of a heatmap.

7. The processing apparatus according to any one of claims 1 to 6, characterized in that, The time period is one day.

8. The processing apparatus according to any one of claims 1 to 6, characterized in that, The energy information is determined by a target model, which is a regression model.

9. The processing apparatus according to claim 8, characterized in that, It also includes a training module, which is used to train the target model based on the historical concentration data of the target object. In training the target model, if the amount of historical concentration data is less than a preset amount, features are extracted from the historical concentration data and the extracted features are used to train the target model; otherwise, the original data of the historical concentration data is used to train the target model.

10. A method for processing analyte concentration data, characterized in that, include: The system receives first analyte data related to glucose concentration and second analyte data related to ketone body concentration obtained by continuously monitoring at least two analytes of a target object using an analyte sensor; and acquires analyte concentration data including glucose concentration data and ketone body concentration data based on the first analyte data and the second analyte data, acquires at least one feature set based on the analyte concentration data within a time period, and determines target information including at least energy information based on the at least one feature set, wherein the energy information is an estimated relationship between the target object's energy intake and energy expenditure, the energy intake being the energy from carbohydrates ingested by the target object, and the energy expenditure being the energy consumed by the target object.

Images