Information processing apparatus and method of creating functions

Abstract

To provide more detailed information about whether a subject is in a specific disease state on the basis of a biological signal waveform acquired from the subject.SOLUTION: An input interface 11 receives waveform data WD corresponding to a biological signal waveform of the subject 20. A processor 12 calculates a probability that the subject 20 is in a specific pathological condition based on the waveform data WD, calculates a prevalence rate as a probability that the subject 20 is actually in the pathological condition based on the probability, and causes a visualization device 32 to visualize the prevalence rate.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to an information processing device that causes a visualization device to visualize a prevalence rate as a probability that a subject actually has a specific pathological condition based on the subject's biological signal waveform. The present disclosure also relates to a method for creating a function for calculating the prevalence rate in the information processing device. [Background technology]

[0002] Patent Document 1 discloses a Holter electrocardiograph that acquires an electrocardiogram waveform, which is an example of a biological signal waveform, from a subject. Data corresponding to the electrocardiogram waveform is received by a remotely located information processing device. The information processing device is configured to analyze the data to determine whether the subject has a heart disease, which is an example of a specific disease. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2014-054391 Summary of the Invention [Problem to be solved by the invention]

[0004] There is a demand for providing more detailed information on whether a subject is suffering from a particular pathological condition based on the biological signal waveforms acquired from the subject. [Means for solving the problem]

[0005] One example aspect provided by the present disclosure is an information processing device, an interface that accepts waveform data corresponding to a biological signal waveform of the subject; a processor that calculates a probability that the subject has a specific pathological condition based on the waveform data, calculates a prevalence rate as a probability that the subject actually has the pathological condition based on the calculated probability, and causes a visualization device to visualize the prevalence rate; It is equipped with:

[0006] The calculated probability is merely a predicted value. For example, even if the probability that a biological signal waveform acquired from a certain subject contains a waveform portion suspected of a specific disease is calculated to be 100%, this does not necessarily mean that the subject actually has the disease. The above configuration provides a new evaluation axis: "If the probability that a subject has a specific pathological condition is inferred to be X, then the probability that the subject actually has the condition (prevalence) is Y." Visualizing the calculated prevalence in this way can provide more detailed information on whether the subject has a specific pathological condition, thereby assisting medical professionals in making diagnoses.

[0007] A method for creating a function for calculating the prevalence rate in the information processing device may also be one of the exemplary aspects provided by the present disclosure. A step A of calculating the probability by inputting waveform data corresponding to a biological signal waveform obtained from the subject into the processor; Step B: identifying which of a plurality of predefined numerical ranges the calculated probability falls within and confirming whether the subject actually has the specific pathological condition; Step C: repeating steps A and B for a plurality of biological signal waveforms acquired at different times to obtain a distribution of the prevalence rate for each of the plurality of numerical ranges; a step D of determining said function as a line or curve that approximates said distribution; Contains:

[0008] One example aspect provided by the present disclosure is an information processing device, an interface that accepts waveform data corresponding to a biological signal waveform of the subject; a processor that calculates the probability that the subject has each of N types of pathological conditions (N is an integer of 2 or more) based on the waveform data; It is equipped with The processor causes a visualization device to visualize, in a viewable manner, the probabilities of each of M pathological conditions (M is an integer equal to or less than N) selected from the N pathological conditions.

[0009] According to the above configuration, the probability that a subject has each of a plurality of pathological conditions is calculated based on a common biological signal waveform acquired from the subject, and the calculated probability is visualized in a visible manner, thereby providing an environment in which a more multifaceted and comprehensive pathological condition evaluation of the subject can be performed. [Brief explanation of the drawings]

[0010] [Figure 1] 2 illustrates an example of a functional configuration of an information processing device according to an embodiment. [Figure 2] 2 shows an example of a flow of processing executed by the processor of FIG. 1. [Figure 3] 3 illustrates an example of an image visualized by a visualization device through the processing of FIG. 2. [Figure 4] 10 shows another example of the flow of processing executed by the processor of FIG. [Figure 5] 5 shows an example of an image visualized by the visualization device through the processing of FIG. 4. [Figure 6] 5 shows another example of an image visualized by the visualization device through the process of FIG. 4. [Figure 7] 5 shows another example of an image visualized by the visualization device through the process of FIG. 4. [Figure 8] 5 shows another example of an image visualized by the visualization device through the process of FIG. 4. [Figure 9] 5 shows another example of an image visualized by the visualization device through the process of FIG. 4. [Figure 10] 2 illustrates an example of a configuration of an information processing system including the information processing device of FIG. [Figure 11]2 shows another example of the configuration of an information processing system including the information processing device of FIG. DETAILED DESCRIPTION OF THE INVENTION

[0011] Exemplary embodiments will now be described in detail with reference to the accompanying drawings.

[0012] 1 illustrates an example of a functional configuration of an information processing device 10 according to an embodiment. The information processing device 10 is configured to perform a process of calculating the probability that an electrocardiogram waveform acquired from a subject 20 includes a waveform portion suspected of atrial fibrillation.

[0013] An electrocardiogram waveform is an example of a biological signal waveform. Arrhythmia, including atrial fibrillation, is an example of a cardiac disease. Cardiac disease is an example of a pathological condition. The probability that an electrocardiogram waveform acquired from subject 20 contains a waveform portion suspected of atrial fibrillation is an example of the probability that the subject has a pathological condition.

[0014] The information processing device 10 includes an input interface 11. The input interface 11 is configured as a hardware interface that receives waveform data WD corresponding to an electrocardiogram waveform of a subject 20 acquired via an electrocardiograph or the like.

[0015] The waveform data WD may be digital data or analog data depending on the specifications of the electrocardiograph, etc. If the waveform data WD is analog data, the input interface 11 is provided with an appropriate conversion circuit including an A / D converter. This description also applies to other signals and data that can be accepted by the input interface 11, which will be described later.

[0016] The information processing device 10 includes a processor 12. The processor 12 is configured as a computing element that executes the processing exemplified in FIG.

[0017] First, a process is executed to calculate the probability that the electrocardiogram waveform corresponding to the waveform data WD contains a waveform portion suspected of atrial fibrillation (STEP 11). Specifically, the processor 12 inputs data corresponding to each partial electrocardiogram waveform into an inference model 31 illustrated in Fig. 1. The inference model 31 may be provided as a part of the information processing device 10, or may be provided as a part of a separate device capable of communicating with the information processing device 10.

[0018] The inference model 31 is an inference algorithm generated through machine learning, which will be described later. The inference model 31 is configured to receive waveform data WD as input and output the probability that the electrocardiogram waveform corresponding to the waveform data WD contains a waveform portion suspected of being atrial fibrillation as an inference result (risk value). The risk value ranges from 0 to 1. A risk value of 0 corresponds to a probability of 0%. A risk value of 1 corresponds to a probability of 100%.

[0019] The inference model 31 can be generated by machine learning using training data. The training data is provided as a combination of electrocardiogram waveforms acquired under different measurement environments and conditions of the subject, and annotations indicating whether the measured waveforms contain waveform segments suspected of atrial fibrillation. Examples of algorithms used for machine learning include neural networks, decision trees, random forests, support vector machines, Bayesian estimation, and Gaussian processes. Examples of statistical methods include methods using the coefficient of variation. The inference model 31 may also be generated through machine learning without using training data or through appropriate statistical methods.

[0020] Next, a process is executed to calculate a prevalence rate, which is the probability that the subject 20 actually develops atrial fibrillation, based on the probability calculated as described above (STEP 12 in FIG. 2). As illustrated in FIG. 1, the information processing device 10 includes a storage 13. The processor 12 calculates the prevalence rate by referring to relationship data RD that defines the relationship between the risk value and the prevalence rate stored in the storage 13. The storage 13 is a storage device that can be realized by a semiconductor memory, a hard disk drive, a magnetic tape drive, or the like.

[0021] FIG. 3 shows an example of the relational data RD. The relational data RD includes a function f with the risk value as the explanatory variable and the prevalence rate as the objective variable. In this example, the function f is defined as an approximate straight line of the distribution d of a plurality of samples measured in advance. The specific method for creating the function f is as follows. A) Waveform data WD corresponding to an electrocardiogram waveform obtained from a subject is input into the inference model 31, and the probability that the electrocardiogram waveform contains a waveform portion suspected of being atrial fibrillation is calculated as a risk value. B) Identify which of several pre-defined numerical ranges the calculated risk value falls within, and confirm whether the subject from whom the above electrocardiogram waveform was obtained actually has atrial fibrillation. C) By repeating the above A) and B) for multiple electrocardiogram waveforms acquired at different times, a distribution d of the probability that the subject actually develops atrial fibrillation when the inference model 31 outputs a risk value that falls within a specific numerical range can be obtained. The multiple electrocardiogram waveforms may be acquired from the same subject, or may be acquired from multiple different subjects. D) Define the function f as an approximate straight line of the distribution d. Note that if the distribution d can be approximated, the function f may be expressed as a curve.

[0022] The processor 12 can uniquely calculate the prevalence value by substituting the risk value output by the inference model 31 into the function f.

[0023] Next, a process is executed in which the prevalence rate is visualized by the visualization device 32 illustrated in Fig. 1 (STEP 13 in Fig. 2). Examples of the visualization device 32 include a display device and a printing device.

[0024] 1, the information processing device 10 includes an output interface 14. The processor 12 is configured to output, from the output interface 14, a display control signal DC that causes the visualization device 32 to visualize the prevalence rate.

[0025] The output interface 14 is configured as a hardware interface. The display control signal DC may be a digital signal or an analog signal depending on the specifications of the visualization device 32. If the display control signal DC is an analog signal, the output interface 14 is provided with an appropriate conversion circuit including a D / A converter. This description also applies to other signals and data that can be output by the output interface 14, which will be described later.

[0026] 3 shows an example of an image 40 visualized by the visualization device 32. The image 40 includes an area 41 in which the calculated prevalence is visualized.

[0027] The risk value output by the inference model 31 is no more than a predicted value based on the machine learning performance of the inference model 31. For example, even if the probability that an electrocardiogram waveform acquired from a certain subject contains a waveform portion suspected of atrial fibrillation is calculated to be 100%, this does not necessarily mean that the subject actually has atrial fibrillation. In this embodiment, the prevalence is calculated based on relational data RD that statistically indicates whether the subject actually has atrial fibrillation when the probability of atrial fibrillation takes a specific value. Therefore, a new evaluation axis can be provided, such as "when the probability of atrial fibrillation is inferred to be X, the probability (prevalence) of actual atrial fibrillation is Y."

[0028] The prevalence calculated in this way can be visualized, providing more detailed information on whether a subject has atrial fibrillation and assisting medical professionals in making a diagnosis.

[0029] 3, an area 41 of an image 40 visualized by the visualization device 32 includes information indicating the risk of the subject 20 identified based on the prevalence calculated as described above. In this embodiment, a prevalence of 30% or more is determined to be a "high risk," a prevalence of 10% or more but less than 30% is determined to be a "medium risk," and a prevalence of less than 10% is determined to be a "low risk."

[0030] The risk is represented by an index 41a. In this embodiment, the index 41a includes the characters "high," "medium," and "low" and colors associated with the characters.

[0031] According to this configuration, it is possible to provide additional information that allows the subject 20 to more intuitively grasp the risk state related to the onset of atrial fibrillation.

[0032] The number of categories representing the presence or absence, or the level of risk, can be determined as appropriate. The form of the indicator 41a can also be determined as appropriate using numbers, symbols, patterns, images, etc. in addition to or instead of the above-mentioned letters and colors.

[0033] 3, in image 40 visualized by visualization device 32, the risk values ​​used in calculating the prevalence may also be visualized in area 41. In addition, image 40 may include area 42 that visualizes the relationship data RD used in calculating the prevalence.

[0034] According to this configuration, the basis for how the visualized prevalence rate was calculated can be provided to the user, thereby preventing a decrease in the user's satisfaction with the calculation results.

[0035] 3, an area 41 of the image 40 may visualize information suggesting actions to take depending on the identified risk of developing atrial fibrillation. In this example, a message recommending long-term Holter ECG testing is visualized based on the situation where the risk of developing atrial fibrillation is determined to be high.

[0036] According to this configuration, it is possible to prompt the user to take appropriate action according to the risk state of the identified subject 20.

[0037] When creating the function f in the relational data RD described above, it is preferable to select the waveform data WD according to the subject's age before inputting it into the inference model 31. Age is an example of a subject's attribute. That is, different functions f can be created based on electrocardiogram waveforms acquired from subjects belonging to one age range and electrocardiogram waveforms acquired from subjects belonging to another age range.

[0038] In this case, as illustrated in Fig. 1, the input interface 11 of the information processing device 10 may be configured to receive age data AG indicating the age of the subject 20 from a user interface 33. The user interface 33 may be a part of the information processing device 10. The age data AG is an example of attribute data. The processor 12 may be configured to calculate the prevalence rate using a function f of the relationship data RD created to match the age of the subject 20 indicated by the age data AG.

[0039] The prevalence of atrial fibrillation may vary depending on the age group of the subject. With the above configuration, the accuracy of calculating the prevalence can be optimized depending on the age of the subject 20.

[0040] When creating the function f in the relational data RD described above, it is preferable that the waveform data WD be selected according to the medical institution where the information processing device 10 is installed before inputting it into the inference model 31. The selection may be performed on a department-by-department basis within the medical institution. That is, different functions f may be created based on an electrocardiogram waveform acquired from a subject who visited one medical institution and an electrocardiogram waveform acquired from a subject who visited another medical institution.

[0041] With this configuration, it is possible to reflect in the function f the tendency of prevalence rates that may differ depending on the medical institution, thereby improving the clinical relevance of the calculated prevalence rates. In addition, even if the information processing device 10 is installed in a medical institution other than the medical institution where the waveform data WD was acquired, the created function f can be made versatile for use in medical institutions of similar size and facilities.

[0042] Other examples of the subject's attributes that are referred to when selecting the waveform data WD include gender, height, weight, medical history, residential area, nationality, race, etc. In this case, the user interface 33 can be configured to allow input of attribute data indicating the attributes used when creating the relationship data RD.

[0043] 4 shows another example of the flow of processing executed by the processor 12 of the information processing device 10. In this example, the processor 12 is configured to calculate the probability that the subject 20 has each of N types of pathological conditions (N is an integer equal to or greater than 2) based on the waveform data WD received by the input interface 11 (STEP 21). In other words, the inference model 31 of this example is configured to, in response to input of specific waveform data WD, output as an inference result (risk value) the probability that the electrocardiogram waveform corresponding to the waveform data WD contains a waveform portion suspected of each of the N types of pathological conditions.

[0044] The inference model 31 can be generated by machine learning using training data. The training data is provided as a combination of electrocardiogram waveforms acquired under different measurement environments and subject conditions, and annotations indicating whether the measured waveforms contain waveform regions suspected of each of the N pathological conditions. Examples of algorithms used for machine learning include neural networks, decision trees, random forests, support vector machines, Bayesian estimation, and Gaussian processes. Examples of statistical methods include a method using the coefficient of variation. The inference model 31 may also be generated through machine learning without using training data or an appropriate statistical method.

[0045] 5 shows an example of an image 50 visualized by the visualization device 32 based on the processing of this example. The image 50 includes an area 51 listing N pathologies for which risk values ​​can be calculated by the inference model 31 based on input of an electrocardiogram waveform. In this example, three pathologies (diagnoses) are listed: paroxysmal atrial fibrillation, acute coronary syndrome, and mitral regurgitation.

[0046] Next, a process is executed in which M types of pathologies (M is an integer equal to or less than N) are selected from the N types of pathologies (STEP 22 in FIG. 4).

[0047] In the image 50 shown in FIG. 5, an "x" mark is displayed next to each of the N types of diagnosis names listed in the area 51. When the mark is clicked or touched, the display of the diagnosis name is canceled. In this example, the display is not canceled and three pathologies are selected (M=N). For example, when the "x" mark next to "Mitral Valve Regurgitation" is operated, "Mitral Valve Regurgitation" is deleted from the area 51 (M <N)。

[0048] Next, a process is executed in which the visualization device 32 visualizes a risk value for each of the selected M pathological conditions (STEP 23 in FIG. 4). The risk value for each of the M pathological conditions is visualized in a viewable manner.

[0049] As shown in FIG. 5, the image 50 includes an area 52 in which the risk values ​​of each of the selected M pathologies are visualized in a visible manner. In this example, each risk value is visualized in the form of a bar graph. As long as the visibility is ensured, the display format of each risk value can be determined appropriately. Examples of other display formats include numerical values, radar charts, balloon charts, etc.

[0050] According to the above configuration, the probability that the subject 20 has each of a plurality of pathological conditions is calculated based on a common electrocardiogram waveform acquired from the subject 20, and the calculated probability is visualized in a visible manner. This provides an environment in which a more multifaceted and comprehensive pathological evaluation of the subject 20 can be performed.

[0051] 6 shows another example of a method for selecting M pathologies from N pathologies for which risk values ​​have been calculated. In addition to an area 51 listing the N pathologies, an image 50 may include an area 53 for selecting the M pathologies.

[0052] In this example, the user selects the pathology to be visualized by moving at least one of the icon images representing N types of pathology displayed in area 51 to area 53 using a pointer operation or a touch operation.

[0053] 7, the N pathologies displayed in the area 51 may be divided into at least one group containing M pathologies. The group division may be determined based on the strength of mutual correlation, etc.

[0054] In this case, when an icon image representing one pathology belonging to a specific group is moved to area 53 by a user's pointer operation or touch operation, other pathologies belonging to the same group are also moved to area 53. This configuration can improve the efficiency of the operation to select pathologies to be visualized.

[0055] 8 also shows another example of a method for selecting M pathologies from N pathologies for which risk values ​​have been calculated. In this example, when one of the icon images representing the N pathologies displayed in area 51 is moved to area 53 by the user's pointer operation or touch operation, a process is performed to highlight the icon image of the pathology recommended for selection. The recommendation for selection can be made based on the degree of relevance to the pathology associated with the icon image moved to area 53, the calculated risk value, etc.

[0056] In this example, the icon image of the pathology to be recommended is left in its initial state, while the remaining icon images are displayed in a grayed-out state. The user may select the icon image related to the recommended pathology, or may select the grayed-out icon image. This configuration can efficiently support the user's decision-making in selecting the pathology to be visualized.

[0057] It should be noted that the system may be configured such that when the user selects one of the N pathologies displayed in the area 51, at least one of the M pathologies to be visualized is automatically determined.

[0058] Selection recommendations may be made through text-based messages in addition to or instead of being based on icon images as described above.

[0059] As illustrated in FIG. 5, the image 50 visualized by the visualization device 32 may include an area 54 for visualizing an indicator representing the risk that the subject 20 has, as identified based on the risk value for each of the selected M types of pathological conditions.

[0060] In this example, a risk value of 0.3 or greater is judged to be a "high risk," a risk value of 0.1% or greater but less than 0.3 is judged to be a "medium risk," and a risk value of less than 0.1 is judged to be a "low risk." In this example, symbols of different shapes are assigned to each risk.

[0061] According to this configuration, it is possible to provide additional information that allows the risk state of subject 20 related to each of the multiple visualized pathologies to be more intuitively understood.

[0062] The number of categories representing the presence or absence, or the high or low level of risk, can be determined appropriately. The form of the indicator can also be determined appropriately using letters, colors, numbers, patterns, images, etc. in addition to or instead of the above symbols.

[0063] The processor 12 of the information processing device 10 may be configured to associate the risk value calculated for each of the N types of pathological conditions as described above with the time point at which the electrocardiogram waveform corresponding to the waveform data WD was acquired, and store the risk value as history data in the storage 13. In this case, the processor 12 is configured to cause the visualization device 32 to visualize, in a visible manner, the change over time in the risk value for each of the selected M types of pathological conditions, based on the history data.

[0064] 9 illustrates an image 60 showing the change over time in the risk value visualized by the visualization device 32. Conditions such as the period of time to be used for visualization and the number of electrocardiogram tests to be used for visualization can be set appropriately via the user interface 33.

[0065] This configuration allows the time-dependent changes in risk values ​​relating to multiple pathological conditions to be viewed at a glance, providing an environment in which changes in the condition of the subject 20, the effects of treatment, etc. can be evaluated from multiple angles and in a comprehensive manner.

[0066] The information processing device 10, the inference model 31, the visualization device 32, and the user interface 33 described above can constitute an information processing system 70 that performs processing to calculate the probability that an electrocardiogram waveform W acquired from a subject 20 includes a waveform portion suspected of being atrial fibrillation. The information processing system 70 can be configured in various ways.

[0067] 10 shows an example of the configuration of an information processing system 70. The information processing system 70 includes an electrocardiograph 71. The electrocardiograph 71 is a device that acquires an electrocardiogram waveform W from the subject 20 and generates corresponding waveform data WD.

[0068] The information processing device 10 and the inference model 31 are mounted on an electrocardiograph 71. The visualization device 32 and the user interface 33 are each connected to the information processing device 10 via a communication network N so as to be able to communicate with each other.

[0069] The information processing system 70 may include a server device 72. The server device 72 is communicatively connected to the information processing device 10 via a communication network N. The inference model 31 may be mounted on the server device 72. In this case, the processor 12 of the information processing device 10 transmits a portion of the waveform data WD from the output interface 14 to the server device 72, and receives data corresponding to the probability calculated by the inference model 31 via the input interface 11.

[0070] The visualization device 32 may be mounted on the electrocardiograph 71. The user interface 33 may be mounted on the visualization device 32 or the electrocardiograph 71.

[0071] 11 shows another example of the configuration of an information processing system 70. The information processing device 10 according to this example is an independent device communicatively connected to an electrocardiograph 71. Communication between the electrocardiograph 71 and the information processing device 10 may be performed via a communication network N.

[0072] In this example, the inference model 31 is installed in the information processing device 10. However, similar to the example described with reference to FIG.

[0073] The visualization device 32 and the user interface 33 are each communicably connected to the information processing device 10 via a communication network N. However, the visualization device 32 may be mounted on the electrocardiograph 71 or on the information processing device 10. The user interface 33 may be mounted on the visualization device 32, on the electrocardiograph 71, or on the information processing device 10.

[0074] When the inference model 31 is installed on a server device 72, the inference model 31 can be shared by multiple information processing devices 10. In this case, the ease of centralized management of the inference model 31 becomes more pronounced as the number of information processing devices 10 used for sharing increases.

[0075] The processor 12 of the information processing device 10 having the various functions described above may be realized by a general-purpose microprocessor operating in cooperation with general-purpose memory. Examples of general-purpose microprocessors include a CPU, an MPU, and a GPU. Examples of general-purpose memory include a ROM and a RAM. In this case, a computer program that realizes the various functions described above may be stored in the ROM. The ROM is an example of a non-transitory computer-readable medium that stores a computer program. The general-purpose microprocessor specifies at least a portion of the program stored in the ROM, deploys it on the RAM, and executes the above-described processing in cooperation with the RAM. The computer program may be pre-installed in the general-purpose memory, or may be downloaded from an external server device via a communication network N and then installed in the general-purpose memory. In this case, the external server device is an example of a non-transitory computer-readable medium that stores a computer program. Note that the general-purpose memory may be used as the storage 13.

[0076] The processor 12 of the information processing device 10 having the various functions described above may be realized by a dedicated integrated circuit, such as a microcontroller, ASIC, or FPGA, capable of executing the computer program. In this case, the computer program is pre-installed in a memory element included in the dedicated integrated circuit. The memory element is an example of a computer-readable medium storing a computer program. The processor 12 of the information processing device 10 having the various functions described above may also be realized by a combination of a general-purpose microprocessor and a dedicated integrated circuit.

[0077] The various configurations described above are merely examples for facilitating understanding of the present disclosure. Each configuration example can be appropriately modified or combined with other configurations within the scope of the present disclosure.

[0078] The cardiac disease used to calculate the prevalence rate by the information processing device 10 is not limited to atrial fibrillation. Other examples of cardiac diseases include atrial premature contractions, paroxysmal supraventricular tachycardia, ventricular premature contractions, ventricular tachycardia, ventricular fibrillation, myocardial infarction, angina pectoris, aortic dissection, heart failure, valvular disease, and chronic kidney disease.

[0079] The combination of the biological signal waveform acquired from the subject 20 and the pathology used to calculate the prevalence by the information processing device 10 is not limited to an electrocardiogram waveform and a cardiac disease. By appropriately setting the type of biological signal waveform associated with the waveform data WD input to the inference model 31, the pathology associated with the probability calculated by the inference model 31, and the pathology associated with the prevalence, various combinations of biological signal waveforms and pathologies used to calculate the prevalence can be realized. For example, the prevalence of diabetes may be calculated based on an electrocardiogram waveform, or the prevalence of epileptic seizures may be calculated based on an electroencephalogram. Instead of the waveform data WD, a combination of data obtained from the subject's 20 medical record, such as age, weight, and medical history, or the subject's electrocardiogram, electroencephalogram, pulse wave, blood oxygen saturation, and other vital data obtainable from the subject may be used.

[0080] The expression "having a particular pathological condition" as used herein also means "having a risk factor that leads to a particular pathological condition." For example, having risk factors such as a smoking history or high blood pressure is also an example of "having a particular pathological condition." Therefore, the probability of having a risk factor such as a smoking history or high blood pressure can also be calculated based on the electrocardiogram waveform. This probability is also included in the meaning of the term "prevalence" as used herein.

[0081] The configurations listed below also form part of this disclosure. Item 1: an interface that accepts waveform data corresponding to a biological signal waveform of the subject; a processor that calculates a probability that the subject has a specific pathological condition based on the waveform data, calculates a prevalence rate as a probability that the subject actually has the pathological condition based on the calculated probability, and causes a visualization device to visualize the prevalence rate; Equipped with Information processing device. Item 2: the processor causes the visualization device to visualize an index representing a risk that the subject has, which is identified based on the prevalence rate; Item 1. An information processing device according to item 1. Item 3: the interface accepts attribute data indicating attributes of the subject; The processor calculates the prevalence rate by referring to the attribute. Item 3. The information processing device according to item 1 or 2. Item 4: The probability is calculated by inputting the waveform data into an inference model generated through machine learning using a plurality of biological signal waveforms determined to include waveform portions suspected of indicating the specific pathological condition. 4. The information processing device according to any one of items 1 to 3. Item 5: the biological signal waveform is an electrocardiogram waveform, The specific pathology is heart disease. 5. The information processing device according to any one of items 1 to 4. Item 6: 6. A method for creating a function for calculating the prevalence rate in the information processing device according to any one of items 1 to 5, A step A of calculating the probability by inputting waveform data corresponding to a biological signal waveform obtained from the subject into the processor; Step B: identifying which of a plurality of predefined numerical ranges the calculated probability falls within and confirming whether the subject actually has the specific pathological condition; Step C: repeating steps A and B for a plurality of biological signal waveforms acquired at different times to obtain a distribution of the prevalence rate for each of the plurality of numerical ranges; a step D of determining said function as a line or curve that approximates said distribution; Contains, How to create a function. Item 7: selecting the waveform data according to attributes of the subject; How to create a function as described in item 6. Item 8: a step of selecting the waveform data according to a medical institution in which the information processing device is installed, How to create a function as described in item 6 or 7. Item 9: an interface that accepts waveform data corresponding to a biological signal waveform of the subject; a processor that calculates the probability that the subject has each of N types of pathological conditions (N is an integer of 2 or more) based on the waveform data; It is equipped with The processor visualizes the probability of each of M pathological conditions (M is an integer equal to or less than N) selected from the N pathological conditions in a visible manner on a visualization device. Information processing device. Item 10: The M type pathology can be selected by the user. Item 10. The information processing device according to item 9. Item 11: When one of the N pathologies is designated by a user, the processor suggests other pathologies to be selected based on the designation. Item 11. The information processing device according to item 9 or 10. Item 12: The processor causes the visualization device to visualize an index representing a risk that the subject has, which is identified based on the probability, for each of the M types of pathological conditions. 12. The information processing device according to any one of items 9 to 11. Item 13: a storage device that stores the calculated probability for each of the N pathological conditions as history data in association with the time point at which the biological signal waveform was acquired; the processor causes the visualization device to visualize, based on the history data, the change in the probability over time for each of the M types of pathological conditions in a viewable manner; 13. The information processing device according to any one of items 9 to 12. Item 14: the biological signal waveform is an electrocardiogram waveform, One of the N pathological conditions is heart disease. 14. The information processing device according to any one of items 9 to 13. [Explanation of symbols]

[0082] 10: Information processing device, 11: Input interface, 12: Processor, 13: Storage, 20: Subject, 31: Inference model, 32: Visualization device, 41a: Index, d: Distribution, f: Function, AG: Age data, W: Electrocardiogram waveform, WD: Waveform data

Claims

1. an interface that accepts waveform data corresponding to a biological signal waveform of the subject; a processor that calculates a probability that the subject has a specific pathological condition based on the waveform data, calculates a prevalence rate as a probability that the subject actually has the pathological condition based on the calculated probability, and causes a visualization device to visualize the prevalence rate; Equipped with Information processing device.

2. the processor causes the visualization device to visualize an index representing a risk that the subject has, which is identified based on the prevalence rate; The information processing device according to claim 1 .

3. the interface accepts attribute data indicating attributes of the subject; The processor calculates the prevalence rate by referring to the attribute. The information processing device according to claim 1 .

4. The probability is calculated by inputting the waveform data into an inference model generated through machine learning using a plurality of biological signal waveforms determined to include waveform portions suspected of indicating the specific pathological condition. The information processing device according to claim 1 .

5. the biological signal waveform is an electrocardiogram waveform, The specific pathology is heart disease. The information processing device according to claim 1 .

6. A method for creating a function for calculating the prevalence rate in the information processing device according to claim 1, comprising: A step A of calculating the probability by inputting waveform data corresponding to a biological signal waveform acquired from the subject into the processor; Step B: identifying which of a plurality of predefined numerical ranges the calculated probability falls within and confirming whether the subject actually has the specific pathological condition; Step C: repeating steps A and B for a plurality of biological signal waveforms acquired at different times to obtain a distribution of the prevalence rate for each of the plurality of numerical ranges; a step D of determining the function as a line or curve that approximates the distribution; Contains, How to create a function.

7. selecting the waveform data according to attributes of the subject; The method for creating a function according to claim 6.

8. a step of selecting the waveform data according to a medical institution in which the information processing device is installed, The method for creating a function according to claim 6.

9. an interface that accepts waveform data corresponding to a biological signal waveform of the subject; a processor that calculates the probability that the subject has each of N types of pathological conditions (N is an integer of 2 or more) based on the waveform data; It is equipped with the processor visualizes, on a visualization device, the probability of each of M pathological conditions (M is an integer equal to or less than N) selected from the N pathological conditions in a viewable manner; Information processing device.

10. The M type pathology can be selected by the user. The information processing device according to claim 9 .

11. When one of the N pathologies is designated by a user, the processor recommends other pathologies to be selected based on the designation. The information processing device according to claim 9 .

12. The processor causes the visualization device to visualize an index representing a risk that the subject has, which is identified based on the probability, for each of the M types of pathological conditions. The information processing device according to claim 9 .

13. a storage device that stores the calculated probability for each of the N pathological conditions as history data in association with a time point at which the biological signal waveform was acquired; the processor causes the visualization device to visualize, based on the history data, the change in the probability over time for each of the M types of pathological conditions in a viewable manner; The information processing device according to claim 9 .

14. the biological signal waveform is an electrocardiogram waveform, One of the N pathological conditions is heart disease. The information processing device according to claim 9 .

Images