Cognitive function assessment system, cognitive function assessment method, and cognitive function assessment program

Abstract

A cognitive function evaluation system capable of evaluating cognitive function decline with high accuracy is provided. [Solution] A cognitive function assessment system 1 of the present invention includes a display device 10 that displays a sign S, a measurement device 20 that measures the three-dimensional coordinates of key points on the body of a subject U who touches the sign S, and an information processing device 30 that evaluates the cognitive function of the subject U. The information processing device 30 includes a data acquisition unit 32 that acquires three-dimensional coordinate data for the number of touches divided by the time points at which the sign S was touched, an exercise data calculation unit 33 that calculates exercise data for the number of touches based on the three-dimensional coordinate data, a smoothness value calculation unit 34 that calculates a smoothness value representing the smoothness of the movement for the number of touches from the exercise data, a feature calculation unit 35 that calculates feature amounts of the movement from the smoothness values ​​for the number of touches, and an evaluation unit 36 ​​that evaluates the cognitive function of the subject based on boundary lines and feature amounts constructed using a cognitive function assessment model.

Description

[Technical Field]

[0001] The present invention relates to a cognitive function assessment system, a cognitive function assessment method, and a cognitive function assessment program for assessing cognitive function. [Background technology]

[0002] In today's aging society, there is a need to detect and prevent age-related cognitive decline at an early stage. Systems for testing or evaluating this cognitive decline have been proposed, for example, as disclosed in Patent Documents 1 and 2.

[0003] In the cognitive function testing system of Patent Document 1, a touch-sensitive display device such as a touchscreen is placed in front of a user, and cognitive function is tested by accurately measuring the speed and / or acuity of touch responses to visual marks displayed on the display device. That is, the system obtains the elapsed time from when a mark is displayed at a first position at a first time until the visual mark displayed at a second position is touched, as well as the distance between the first and second positions, and calculates a score based on the elapsed time, distance, and appropriateness of the response. Cognitive ability is tested based on this calculated score.

[0004] The system for assessing cognitive function in Patent Document 2 is equipped with a detector that detects movement information related to the subject's lower back, and evaluates cognitive function based on the time it takes for the subject to stand up from a chair, walk around a target, and sit back down in the chair, as well as the results of a test that evaluates at least one of the subject's attention function, memory function, and language function. The test to evaluate attention function uses Part B of the Trail Making Test, which is commonly used to evaluate cognitive function, and involves the subject selecting 13 numbers from "1" to "13" and 12 hiragana characters from "a" to "shi" randomly displayed on the screen, alternating between selecting numbers and hiragana characters starting with "1," and assessing attention function based on the time it takes to select the final number, "13." [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Laid-Open No. 2010-527642 [Patent Document 2] Japanese Patent Application Publication No. 2018-29706 Summary of the Invention [Problem to be solved by the invention]

[0006] However, the system of Patent Document 1 tests cognitive function based on the action of touching a visual mark once, and a single action can result in a good or bad result by chance, making the test inaccurate. Furthermore, to test accurately, the same test must be repeated multiple times, placing a heavy burden on the subject. In the system of Patent Document 2, when a test to evaluate attention function is used, the evaluation is based on the action of selecting multiple displays, making the system less susceptible to the problems of Patent Document 1.

[0007] On the other hand, it is thought that as cognitive function declines, subtle movements of the upper limbs increase due to hesitation when selecting a number or hiragana character before recognizing which number or hiragana character to select next. However, the system of Patent Document 2 does not take upper limb movements into consideration, and evaluates cognitive function solely based on the time required to select all the numbers and hiragana characters. In this case, even if a user takes a long time to select due to reasons that do not directly affect cognitive function, such as hand disability or visual impairment, the system will determine that cognitive function is declining simply because it takes a long time, resulting in a problem of being unable to perform a highly accurate evaluation.

[0008] The present invention has been made to solve the above-mentioned problems, and aims to provide a cognitive function assessment system, a cognitive function assessment method, and a cognitive function assessment program that can assess cognitive function decline with high accuracy, while excluding, as much as possible, assessments that indicate decline in cognitive function due to factors that do not directly affect cognitive function. [Means for solving the problem]

[0009] Item 1. A display device that displays multiple types of signs that can be selected in sequence randomly or in an orderly manner on a screen; a measuring device that measures three-dimensional coordinates or three-axis acceleration of key points on the body of a subject who touches the mark displayed on the display device; an information processing device that evaluates the cognitive function of the subject based on the three-dimensional coordinates or the three-axis acceleration of key points on the body of the subject measured by the measurement device; The information processing device includes: a sign display unit that displays a plurality of types of signs that can be selected in order randomly or in an orderly manner on the screen of the display device; a data acquisition unit that acquires three-dimensional coordinate data or three-axis acceleration data measured by the measuring device for the number of touches divided by the divisions, using the time points when the subject touches the multiple types of signs displayed by the sign display unit as divisions; a motion data calculation unit that calculates motion data of key points of the body for the number of touches based on the three-dimensional coordinate data or the three-axis acceleration data for the number of touches acquired by the data acquisition unit; and a smoothness value calculation unit that calculates smoothness values ​​representing smoothness of movements of key points of the subject's body for the number of touches from the motion data for the number of touches calculated by the motion data calculation unit; a feature value calculation unit that calculates feature values ​​of the motion of key points of the subject's body from the smoothness values ​​for the number of touches calculated by the smoothness value calculation unit; A cognitive function assessment system comprising: a boundary line for identifying whether or not cognitive function is declining, constructed using a cognitive function assessment model that has previously undergone machine learning to determine the correspondence between features and cognitive function; and an assessment unit that evaluates the cognitive function of a subject based on the features calculated by the feature calculation unit.

[0010] Item 2. The cognitive function assessment system according to Item 1, wherein the measuring device is a depth camera provided above the display device or an acceleration sensor attached to the wrist of the subject.

[0011] Item 3. The movement data is three-axis velocity data for the number of touches obtained by differentiating the three-dimensional coordinate data or integrating the three-axis acceleration data, the smoothness value is an average frequency and a maximum frequency in the three-axis directions for the number of touches obtained by frequency analysis of the three-axis velocity data, the feature amount is an average value, a median value, or a maximum value of the average frequency in the three axial directions for the number of touches, and an average value, a median value, or a maximum value of the maximum frequency in the three axial directions for the number of touches, Item 3. The cognitive function assessment system according to Item 1 or 2, wherein the assessment unit assesses that cognitive function is impaired when the relationship between the average value, median value, or maximum value of any of the average frequencies in the three axial directions and the average value, median value, or maximum value of any of the maximum frequencies in the three axial directions is below the boundary line.

[0012] Item 4. The movement data is three-axis velocity data for the number of touches obtained by differentiating the three-dimensional coordinate data or integrating the three-axis acceleration data, the smoothness value is the number of sign changes in the three-axis directions for the number of touches obtained by analyzing sign changes in the three-axis velocity data, the feature amount is an average value, a median value, or a maximum value of the number of sign changes in the three axis directions for the number of touches; Item 3. The cognitive function assessment system according to item 1 or 2, wherein the assessment unit assesses that cognitive function is impaired when the average, median, or maximum value of any of the number of sign changes in the three axis directions is above the boundary line.

[0013] Item 5. The movement data is three-axis jerk data for the number of touches obtained by differentiating the three-dimensional coordinate data or the three-axis acceleration data, the smoothness value is an average absolute jerk and a maximum absolute jerk for the number of touches obtained by performing jerk analysis on any one of the axis components or norms of the three-axis jerk data, the feature amount is an average value, a median value, or a maximum value of the mean absolute jerk for the number of touches, and an average value, a median value, or a maximum value of the maximum absolute jerk for the number of touches, 3. The cognitive function assessment system according to item 1 or 2, wherein the assessment unit assesses that cognitive function is impaired when a relationship between an average value, a median value, or a maximum value of the mean absolute jerk and an average value, a median value, or a maximum value of the maximum absolute jerk exceeds the boundary line.

[0014] Item 6. The cognitive function evaluation system according to any one of Items 1 to 5, wherein the signs are of two types: numbers and hiragana.

[0015] Item 7. The cognitive function evaluation system according to Item 6, wherein the signs are the numbers "1" to "12" and the hiragana characters "a" to "shi."

[0016] Item 8. The cognitive function assessment system according to Item 7, characterized in that the signs of the numbers "1" to "12" and the hiragana characters "a" to "shi" are arranged in an orderly fashion in three vertical and four horizontal rows and are displayed in two separate displays.

[0017] Item 9. A cognitive function evaluation method for evaluating the cognitive function of a subject, which is executed by an information processing device having a sign display unit, a data acquisition unit, a motion data calculation unit, a smoothness value calculation unit, a feature calculation unit, and an evaluation unit, a sign display step in which the sign display unit randomly or orderly displays a plurality of types of signs that can be selected in order on a screen of a display device; a data acquisition step in which the data acquisition unit acquires three-dimensional coordinates or three-axis accelerations of key points on the body of the subject who taps the marker displayed on the display device, measured by a measurement device, as three-dimensional coordinate data or three-axis acceleration data for the number of touches divided by the divisions, with the time points at which the subject touches the marker being used as divisions; a motion data calculation step in which the motion data calculation unit calculates motion data of key points of the body for the number of touches based on the three-dimensional coordinate data or the three-axis acceleration data for the number of touches acquired in the data acquisition step; a smoothness value calculation step in which the smoothness value calculation unit calculates smoothness values ​​representing smoothness of movements of key points of the subject's body for the number of touches from the motion data for the number of touches calculated in the motion data calculation step; a feature calculation step in which the feature calculation unit calculates feature values ​​of movements of key points of the subject's body from the smoothness values ​​for the number of touches calculated in the smoothness value calculation step; The cognitive function assessment method is characterized in that the assessment unit comprises an assessment step of assessing the cognitive function of the subject based on a boundary line for identifying whether or not cognitive function has declined, which boundary line is constructed using a cognitive function assessment model that has previously undergone machine learning to determine the correspondence between features and cognitive function, and the features calculated in the feature calculation step.

[0018] Item 10. A cognitive function assessment program for evaluating the cognitive function of a subject, a sign display step of randomly or systematically displaying a plurality of types of signs that can be selected in order on a screen of a display device; a data acquisition step of acquiring three-dimensional coordinates or three-axis accelerations of key points on the body of the subject who taps the mark displayed on the display device, measured by a measurement device, as three-dimensional coordinate data or three-axis acceleration data for the number of touches divided by the divisions, with the time points at which the subject touches the mark being used as divisions; a motion data calculation step of calculating motion data of key points of the body for the number of touches based on the three-dimensional coordinate data or the three-axis acceleration data for the number of touches acquired in the data acquisition step; a smoothness value calculation step of calculating smoothness values ​​representing smoothness of movements of key points of the subject's body for the number of touches from the motion data for the number of touches calculated in the motion data calculation step; a feature value calculation step of calculating feature values ​​of the motion of key points of the subject's body from the smoothness values ​​for the number of touches calculated in the smoothness value calculation step; A cognitive function assessment program that causes a computer to execute processing including: a boundary line that identifies whether or not cognitive function is declining, constructed using a cognitive function assessment model that has previously undergone machine learning to determine the correspondence between features and cognitive function; and an assessment step that evaluates the cognitive function of a subject based on the features calculated in the feature calculation step. [Effects of the Invention]

[0019] According to the present invention, it is possible to perform a highly accurate evaluation of cognitive function decline, excluding, as much as possible, evaluations that indicate cognitive function decline due to factors that do not directly affect cognitive function. [Brief explanation of the drawings]

[0020] [Figure 1] 1 is a block diagram showing the configuration of a cognitive function assessment system according to one embodiment of the present invention. [Figure 2] 2 is a front view showing an example of a display device and a measurement device of the cognitive function assessment system of FIG. 1. FIG. [Figure 3] 2 is a front view showing an example of measurement positions of three-dimensional coordinate data measured by a measurement device of the cognitive function evaluation system of FIG. 1. FIG. [Figure 4] 1 is a flowchart showing a cognitive function assessment method according to one embodiment of the present invention. [Figure 5] FIG. 1 is a front view showing an example of a test for evaluating cognitive function used in the present invention. [Figure 6] 1 is a graph showing the evaluation results of frequency analysis in Example 1. [Figure 7] 10 is a graph showing the evaluation results of the sign change analysis of Example 2. [Figure 8] 10 is a graph showing the evaluation results of the jerk analysis of Example 3. DETAILED DESCRIPTION OF THE INVENTION

[0021] Hereinafter, an embodiment of the present invention will be described with reference to the accompanying drawings. The cognitive function assessment system of the present invention is a system for assessing the cognitive function of a subject using a test similar to TMT part B (Trail Making Test Part B), which is generally used for assessing cognitive function.

[0022] A cognitive function assessment system 1 according to this embodiment will be described with reference to FIGS. 1 to 3. FIG. 1 is a block diagram showing the configuration of the cognitive function assessment system 1 according to this embodiment. FIG. 2 is a front view showing an example of the display device 10 and the measurement device 20 of the cognitive function assessment system 1 according to this embodiment. FIG. 3 is a front view showing an example of the measurement position of three-dimensional coordinate data measured by the measurement device 20 of the cognitive function assessment system 1 according to this embodiment. In this specification, the front-rear direction, left-right direction, and up-down direction are defined based on the subject U whose cognitive function is assessed in the cognitive function assessment system 1 according to this embodiment. That is, the front side of the subject U is defined as the front direction, the right hand side of the subject U is defined as the right direction, and the left hand side of the subject U is defined as the left direction.

[0023] 1, the cognitive function assessment system 1 includes a display device 10, a measurement device 20, and an information processing device 30. The display device 10, the measurement device 20, and the information processing device 30 are connected via a computer network such as the Internet or a WAN (Wide Area Network).

[0024] The display device 10 is installed in front of the subject U and displays a test for assessing cognitive function. Specifically, the display device 10 randomly or systematically displays multiple types of signs S, which can be selected in sequence, on a screen 11. The display device 10 may also display the results of cognitive function assessment by the cognitive function assessment system 1. The display device 10 is provided with a touch panel on the screen 11 that can detect the subject U's touch on the screen 11 when the subject U touches the screen 11. The display device 10 is, for example, a screen for displaying images, such as a television monitor or display, that has a touch panel.

[0025] The subject U takes a test to evaluate cognitive function while standing or sitting. At this time, the screen 11 of the display device 10 is preferably positioned at eye level with the subject U. Furthermore, the screen 11 of the display device 10 is preferably large enough to allow the subject U to move his / her arms to some extent when touching the screen 11, and large enough that the subject U has to use both hands to touch the left and right edges of the screen 11. By making the screen 11 of the display device 10 such a size, the subject U takes the test using both hands, and by observing the movements of the subject U's hands, the decline in cognitive function can be accurately evaluated.

[0026] In this embodiment, the signs S displayed on the display device 10 are composed of two types: numbers and hiragana. Specifically, the signs S are the numbers "1" to "12" and the hiragana "a" to "shi." On the screen 11 of the display device 10, the signs S of the numbers "1" to "12" and the hiragana "a" to "shi" are displayed in an orderly fashion in three columns and four rows, twice. The signs S can be in any combination as long as there is a general order, such as numbers and alphabets, numbers and katakana, or hiragana and alphabets, and there are about 12 or more of each type.

[0027] The measuring device 20 measures the three-dimensional coordinates of key points on the body of the subject U who touches the marker S displayed on the display device 10. Specifically, the measuring device 20 measures the three-dimensional coordinates of key points (key points) for understanding the movement of the subject U when touching the marker S, such as joint points such as the wrist joint, finger joint, elbow joint, and shoulder joint, as indicated by black circles in FIG. 3 . Because hesitation in movement due to a decline in cognitive function manifests itself in the movement of the subject U's upper body, particularly from the upper arm to the fingertips, it is preferable that the key points be located near the upper body, particularly the joints of the fingertips. Here, the three-dimensional coordinates refer to the coordinates of the subject U in the vertical direction, the horizontal direction, and the front-to-back direction of each key point on the subject U's body. In this embodiment, as shown in FIG. 2 , the measuring device 20 is configured, for example, with a depth camera and is installed above the display device 10, which is installed in front of the subject U whose cognitive function is being evaluated. The measuring device 20 measures the three-dimensional coordinates of the subject U at predetermined sampling intervals.

[0028] The information processing device 30 is a device that evaluates the cognitive function of the subject U based on the three-dimensional coordinates of key points on the body of the subject U measured by the measurement device 20. The information processing device 30 includes a sign display unit 31, a data acquisition unit 32, a motion data calculation unit 33, a smoothness value calculation unit 34, a feature calculation unit 35, an evaluation unit 36, a storage unit 37, and a control unit 38. The information processing device 30 is, for example, a computer such as a smartphone, a tablet, a microcomputer, a personal computer, or a server computer. The information processing device 30 may be provided integrally with the display device 10 or may be provided in a location separate from the display device 10.

[0029] The sign display unit 31 displays a plurality of types of signs S, which can be selected in order, on the screen 11 of the display device 10 in a random or orderly manner.

[0030] The data acquisition unit 32 acquires 3D coordinate data measured by the measurement device 20 for the number of touches divided by the divisions, using the time points when the subject U touches the multiple types of signs S displayed by the sign display unit 31 as divisions. Specifically, in this embodiment, 12 numbers from "1" to "12" and 12 hiragana characters from "a" to "shi" are used as the signs S. As will be described in detail later, the sign S is touched 24 times in the following order: number "1," hiragana character "a," number "2," hiragana character "i," number "11," hiragana character "sa," number "12," and hiragana character "shi." At this time, the time points when each number or each hiragana character is touched are divisions. In other words, since there are 24 touches, the total number of touches is 24, and the data is divided into 24 divisions. The data acquisition unit 32 acquires 3D coordinate data for the number of touches made by the subject U from the time he starts the touching action until he touches the first number "1," which is counted as one touch, and from the time he touches the number "1" until he touches the next hiragana character "a" which is counted as two touches, up to the time he touches the last hiragana character "shi," in other words, 24 touches.

[0031] The measuring device 20 measures the three-dimensional coordinates of the subject U at each predetermined sampling time. For example, there are multiple pieces of three-dimensional coordinate data for each sampling time between touching the digit "1" and touching the hiragana character "a." The data acquiring unit 32 determines the time when the screen 11 was touched from the three-dimensional coordinate data and acquires multiple pieces of three-dimensional coordinate data as time-series data for each section separated by the time of touch. At this time, the data acquiring unit 32 determines whether the subject U touched the screen 11 with the right or left hand based on the three-dimensional coordinate data of key points, such as near the joint at the tip of the right hand, and the three-dimensional coordinate data of key points, such as near the joint at the tip of the left hand, when the subject U touched the sign S. Then, the data acquiring unit 32 acquires three-dimensional coordinate data of one key point on the hand that touched the screen 11 where the subject U's movement is prominent. That is, for example, if the subject U touches the screen 11 with his or her right hand and a prominent movement is observed at the wrist joint, the data acquiring unit 32 acquires three-dimensional coordinate data of the wrist joint of the right hand.

[0032] The motion data calculation unit 33 calculates motion data of the body key points for the number of touches based on the three-dimensional coordinate data for the number of touches acquired by the data acquisition unit 32. Here, the motion data is velocity data in three axial directions, acceleration data in three axial directions, jerk data in three axial directions, or the like, of the key points on the body of the subject U.

[0033] The smoothness value calculation unit 34 calculates smoothness values, which represent the smoothness of the movements of the key points of the body of the subject U, for the number of touches from the movement data for the number of touches calculated by the movement data calculation unit 33. Here, the smoothness value is a numerical value of an index that quantitatively indicates the smoothness of the movement. The smoothness value is, for example, a frequency component obtained by frequency analysis, a number of sign changes obtained by sign change analysis, a jerk index obtained by jerk analysis, etc.

[0034] The feature calculation unit 35 calculates feature amounts of the movements of the key points of the body of the subject U from the smoothness values ​​for the number of touches calculated by the smoothness value calculation unit 34. Here, the feature amount is a numerical value that indicates the features of the movements of the subject U, calculated based on the smoothness value, which is an index showing the smoothness of the movements. The feature amount is the average, median, or maximum value of the frequency components, the average, median, or maximum value of the number of sign changes, the average, median, or maximum value of the jerk index, etc.

[0035] The evaluation unit 36 ​​evaluates the cognitive function of the subject U based on a boundary line for identifying whether or not cognitive function has declined, which is constructed using the cognitive function evaluation model, and the feature values ​​calculated by the feature value calculation unit 35. The cognitive function evaluation model is a trained model in which the correspondence between feature values ​​and cognitive function has been previously machine-learned. Specifically, the cognitive function evaluation model is a model capable of binary classification, which is created using known machine learning techniques such as logistic regression, support vector machine (SVM), decision tree, and gradient boosting. The created cognitive function evaluation model constructs a boundary line for identifying whether or not cognitive function has declined. The evaluation unit 36 ​​evaluates the cognitive function of the subject U by determining whether the feature values ​​calculated by the feature value calculation unit 35 are above or below the boundary line.

[0036] The configurations of the motion data calculation unit 33, smoothness value calculation unit 34, feature calculation unit 35, and evaluation unit 36 ​​will be described in more detail below for each method of quantifying the smoothness of movement. In the field of rehabilitation, frequency analysis, sign change analysis, and jerk analysis are mainly used as methods for quantifying the smoothness of movement.

[0037] First, we will explain the case where frequency analysis is used. The movement data calculation unit 33 calculates, as movement data, three-axis velocity data for the number of touches obtained by first-order differentiation of the three-dimensional coordinate data with respect to the sampling time measured by the measurement device 20. The smoothness value calculation unit 34 calculates, as smoothness values, the average frequency and maximum frequency in each of the three axial directions for the number of touches, obtained by performing a fast Fourier transform on the three-axis velocity data for the number of touches. The feature calculation unit 35 calculates, as feature values, the average, median, or maximum value of the average frequency in each of the three axial directions for the number of touches, and the average, median, or maximum value of the maximum frequency in each of the three axial directions for the number of touches.

[0038] When moving a hand toward a sign S to touch it, if the person becomes confused while searching for the sign S to touch, the hand movement slows and the hand shakes more. In this case, the maximum frequency tends to be low because the hand movement starts slowly, and the average frequency tends to be high because the hand shakes more. On the other hand, if the person does not become confused while searching for the sign S, the hand movement becomes faster and the hand moves in a straight line toward the sign S, resulting in less hand shaking. In this case, the maximum frequency tends to be high because the hand movement is fast, and the average frequency tends to be low because the hand shakes less. Based on the above trends, the evaluation unit 36 ​​evaluates that cognitive function is impaired if the relationship between the average, median, or maximum value of the average frequency and the average, median, or maximum value of the maximum frequency is below the boundary line. The feature used in this evaluation is the same component value as the feature used for training the cognitive function evaluation model. That is, for example, when constructing a cognitive function evaluation model, if the correspondence between the median of the average frequency on an axis perpendicular to the screen 11 (front-back direction) and the median of the maximum frequency in the front-back direction and the cognitive function is learned by machine learning, the evaluation unit 36 ​​performs evaluation using the median of the average frequency in the front-back direction and the median of the maximum frequency in the front-back direction calculated by the feature calculation unit 35. Note that the feature used for learning the cognitive function evaluation model and for evaluating cognitive function in the evaluation unit 36 ​​may be the average, median, or maximum of the average frequency in the up-down direction or the left-right direction, and the average, median, or maximum of the maximum frequency in the up-down direction or the left-right direction.

[0039] Second, a case where sign change analysis is used will be described. The movement data calculation unit 33 calculates, as movement data, three-axis velocity data for the number of touches, obtained by first-order differentiation of the three-dimensional coordinate data with respect to the sampling time measured by the measurement device 20. The smoothness value calculation unit 34 calculates, as a smoothness value, the number of sign changes in the three-axis directions for the number of touches, obtained by performing sign change analysis on the three-axis velocity data for the number of touches. Here, the sign change analysis in this embodiment is to calculate the number of changes in the sign (positive or negative) of the slope of the velocity waveform for each sampling time of the three-axis velocity data calculated as time-series data. The number of sign changes is calculated as the number of sign changes. The feature calculation unit 35 calculates, as a feature, the average, median, or maximum of the number of sign changes in each of the three-axis directions for the number of touches.

[0040] When moving a hand toward a sign S to touch it, if the person becomes confused while searching for the sign S to touch, the hand tends to shake more, resulting in more changes in the sign of the slope of the velocity waveform. On the other hand, if the person does not become confused while searching for the sign S, the hand tends to shake less because the hand moves in a straight line toward the sign S, resulting in fewer changes in the sign of the slope of the velocity waveform. Based on this tendency, the evaluation unit 36 ​​evaluates that cognitive function has declined if the average, median, or maximum number of sign changes exceeds the boundary line. The feature used for this evaluation is the value of the same component as the feature used for training the cognitive function evaluation model. That is, for example, if machine learning is performed to determine the correspondence between the average number of sign changes on an axis perpendicular to the screen 11 (front-to-back direction) and cognitive function when constructing the cognitive function evaluation model, the evaluation unit 36 ​​performs evaluation using the average number of sign changes in the front-to-back direction calculated by the feature calculation unit 35. Note that the feature used for training the cognitive function evaluation model and for evaluating cognitive function in the evaluation unit 36 ​​may be the average, median, or maximum number of sign changes in the up-down or left-to-right direction.

[0041] Third, a case where jerk analysis is used will be described. The motion data calculation unit 33 calculates three-axis jerk data for the number of touches as motion data by third-order differentiating the three-dimensional coordinate data with respect to the sampling time measured by the measurement device 20. The smoothness value calculation unit 34 calculates a norm from the three-axis jerk data for the number of touches and performs jerk analysis on the calculated norm to calculate the average absolute jerk and maximum absolute jerk for the number of touches as smoothness values. Here, the jerk analysis in this embodiment involves analyzing the jerk component from the time-series norm calculated from the three-axis jerk data, which is time-series data. Specifically, the absolute values ​​for each sampling time of the norm of the time series for each segment separated at the time when the subject U touches the marker S are averaged to calculate the average absolute jerk. Furthermore, the maximum absolute jerk is calculated by taking the square root of the absolute value for each sampling time of the time-series norm. The feature amount calculation unit 35 calculates, as feature amounts, the average value, median value, or maximum value of the mean absolute jerk for the number of touches, and the average value, median value, or maximum value of the maximum absolute jerk for the number of touches. Note that the smoothness value calculation unit 34 may directly calculate the average absolute jerk for each of the three axial directions and the maximum absolute jerk for each of the three axial directions without calculating a norm from the three-axis jerk data. In this case, the feature amount calculation unit 35 calculates, as feature amounts, the average value, median value, or maximum value of the mean absolute jerk for each of the three axial directions for the number of touches, and the average value, median value, or maximum value of the maximum absolute jerk for each of the three axial directions for the number of touches.

[0042] When moving a hand toward a sign S to touch it, if the person becomes confused while searching for the sign S to touch, the hand will shake more and sudden braking will occur more frequently due to sudden changes in direction, etc. At this time, the mean absolute jerk tends to be higher due to the increased hand shaking, and the maximum absolute jerk tends to be higher due to the increased number of sudden brakings. On the other hand, if the person does not become confused while searching for the sign S, the person moves the hand in a straight line toward the sign S, resulting in less hand shaking and fewer sudden brakings. At this time, the mean absolute jerk tends to be lower due to the reduced hand shaking, and the maximum absolute jerk tends to be lower due to the reduced number of sudden brakings. Based on the above trends, the evaluation unit 36 ​​evaluates that cognitive function has deteriorated if the relationship between the average value, median, or maximum value of the mean absolute jerk and the average value, median, or maximum value of the maximum absolute jerk exceeds a boundary line. The feature used for this evaluation is the same component value as the feature used for training the cognitive function evaluation model. That is, for example, if the correspondence between the average value of the mean absolute jerk and the average value of the maximum absolute jerk and the cognitive function is learned by machine learning when constructing a cognitive function assessment model, evaluation unit 36 ​​performs evaluation using the average value of the mean absolute jerk and the average value of the maximum absolute jerk calculated by feature calculation unit 35. Note that the feature used for learning the cognitive function assessment model and for evaluating cognitive function in evaluation unit 36 ​​may be the average value, median value, or maximum value of the mean absolute jerk in the forward / backward, upward, or left / right direction, and the average value, median value, or maximum value of the maximum absolute jerk in the forward / backward, upward, or left / right direction.

[0043] The storage unit 37 stores various information, data, programs, etc. The storage unit 37 is configured with a ROM (Read Only Memory) and a RAM (Random Access Memory). The storage unit 37 stores a cognitive function evaluation program for evaluating the cognitive function of the subject U.

[0044] The control unit 38 is configured by a processor such as a CPU (Central Processing Unit). The control unit 38 executes a program to control the operations of the sign display unit 31, the data acquisition unit 32, the motion data calculation unit 33, the smoothness value calculation unit 34, the feature calculation unit 35, the evaluation unit 36, and the storage unit 37.

[0045] Next, with reference to FIGS. 4 and 5, a cognitive function assessment method for assessing the cognitive function of a subject U, which is executed by the cognitive function assessment system 1, will be described. FIG. 4 is a flowchart showing the cognitive function assessment method according to this embodiment. FIG. 5 is a front view showing an example of a test for assessing cognitive function used in the present invention. FIG. 5(a) shows a first test screen of the test for assessing cognitive function, and FIG. 5(b) shows a second test screen of the test for assessing cognitive function. Each process of the cognitive function assessment method is executed by a cognitive function assessment program stored in the storage unit 37 of the information processing device 30. In the following, the smoothness of a movement is quantified using frequency analysis, and the median of the average frequency and the median of the maximum frequency on an axis perpendicular to the screen 11 (front-back direction) are used as feature quantities. When the smoothness of a movement is quantified using sign change analysis or jerk analysis, the same process is performed when the average frequency and maximum frequency in the up-down direction or left-right direction are used as feature quantities.

[0046] A subject U taking a test to evaluate cognitive function faces the display device 10 while standing or sitting. As shown in FIG. 4, when the test to evaluate cognitive function starts, the sign display unit 31 executes a sign display step of randomly or orderly displaying multiple types of signs S, which can be selected in order, on the screen 11 of the display device 10 (S10). In this embodiment, the signs S are a first type of twelve numbers from "1" to "12," which can be selected in ascending order, and a second type of twelve hiragana characters from "a" to "shi," which can be selected in alphabetical order. As shown in FIG. 5(a), the sign display unit 31 first displays the first six numbers from "1" to "6" and the first six hiragana characters from "a" to "ka" on the screen 11 as a first test screen.

[0047] When signs S are displayed on the screen 11, the subject U touches the left half of the screen 11 (to the left of the dashed line in FIG. 5(a)) with his / her left hand and the right half of the screen 11 (to the right of the dashed line in FIG. 5(a)) with his / her right hand, alternately touching and selecting signs S according to type. In this embodiment, numbers and hiragana are selected alternately, numbers in ascending order, and hiragana in alphabetical order. That is, the order of selection is number "1", hiragana "a", number "2", hiragana "i", number "5", hiragana "o", number "6", hiragana "ka". Correctly selected signs S disappear in order at the timing of their selection.

[0048] During this time, if an incorrect sign S is selected, the control unit 38 issues a warning to inform the subject U that the incorrect sign S has been selected. Selecting an incorrect sign S refers, for example, to selecting the hiragana character "u" instead of the number "3," or selecting a sign S displayed on the left half of the screen 11 with the right hand instead of the left hand. The sign S includes information indicating whether it is displayed on the right or left half of the screen, and the correct hand is determined by comparing it with the three-dimensional coordinate data of the key points of the left and right hands of the subject U selecting the sign S. The warning may be displayed by the selected sign S not disappearing, or by displaying a message such as "Wrong" on the screen 11 of the display device 10, or by generating a warning sound using a sound device (not shown) such as a speaker in the cognitive function assessment system 1. In this case, even if the subject selects the next correct sign S by skipping the incorrect sign S, the selection will be invalid until the correct sign S is selected.

[0049] Once all the signs S on the first test screen for the test to evaluate cognitive function have been selected, as shown in FIG. 5(b), the sign display unit 31 then displays the last six numbers from "7" to "12" and the last six hiragana characters from "ki" to "shi" on the screen 11 as a second test screen. As with the first test screen, the subject U uses both hands to alternately touch and select signs S by type in order. That is, the subject selects the signs in the following order: number "7," hiragana character "ki," number "8," hiragana character "ku," number "11," hiragana character "sa," number "12," and hiragana character "shi." At this time, too, if an incorrect sign S is selected, a warning is issued.

[0050] While the subject U is selecting the marker S, the measurement device 20 measures the three-dimensional coordinates of key points on the body of the subject U tapping the marker S displayed on the display device 10 at predetermined sampling times (S12). When the last marker S is selected and the test for assessing the cognitive function of the subject U is completed, the measurement device 20 ends the measurement.

[0051] When the test for evaluating the cognitive function of the subject U is completed, the data acquisition unit 32 executes a data acquisition step of acquiring the three-dimensional coordinates measured by the measurement device 20 as three-dimensional coordinate data for the number of touches, divided by the time points when the subject U touched the marker S (S14). The time points of the touches are determined from the front-to-back positions of the three-dimensional coordinates. In this embodiment, the three-dimensional coordinate data for the number of touches is acquired, with one touch from the start of the test until the number "1" is touched being counted as one touch, two touches from the touch of the number "1" until the next hiragana character "a" being touched being counted as two touches, and 24 touches from the touch of the number "12" until the last hiragana character "shi" being touched being counted as 24 touches. The acquired three-dimensional coordinate data for the number of touches is transmitted to the exercise data calculation unit 33.

[0052] When the motion data calculation unit 33 receives the three-dimensional coordinate data for the number of touches from the data acquisition unit 32, it executes a motion data calculation step of calculating motion data of key points of the body in three axial directions for the number of touches based on the three-dimensional coordinate data for the number of touches acquired in the data acquisition step (S16). In this embodiment, in order to quantify the smoothness of the movement using frequency analysis, the motion data calculation unit 33 calculates three-axial velocity data for the number of touches by first-order differentiating the three-dimensional coordinate data in each direction with respect to the sampling time measured by the measurement device 20. The calculated motion data in the three axial directions for the number of touches is sent to the smoothness value calculation unit 34.

[0053] When the smoothness value calculation unit 34 receives the motion data in the three axial directions for the number of touches from the motion data calculation unit 33, it executes a smoothness value calculation step of calculating smoothness values ​​in the three axial directions, representing the smoothness of the movements of key points of the body of the subject U, for the number of touches, from the motion data in the three axial directions for the number of touches calculated in the motion data calculation step (S18). In this embodiment, in order to quantify the smoothness of the movements using frequency analysis, the smoothness value calculation unit 34 performs frequency analysis on the three axial velocity data for the number of touches to calculate the average frequency and maximum frequency for each of the three axial directions for the number of touches. The calculated smoothness values ​​in the three axial directions for the number of touches are transmitted to the feature calculation unit 35.

[0054] When the feature calculation unit 35 receives the smoothness values ​​in the three axial directions for the number of touches from the smoothness value calculation unit 34, it executes a feature calculation step of calculating feature values ​​of the movements of the key points of the body of the subject U in the three axial directions from the smoothness values ​​for the number of touches calculated in the smoothness value calculation step (S20). In this embodiment, the smoothness of the movements is quantified using frequency analysis, and the median of the average frequency in the front-rear direction for the number of touches and the median of the maximum frequency in the front-rear direction for the number of touches are used, so that the median of the average frequency in the front-rear direction for the number of touches and the median of the maximum frequency in the front-rear direction for the number of touches are calculated. The calculated feature values ​​are transmitted to the evaluation unit 36.

[0055] Upon receiving the features from the feature calculation unit 35, the evaluation unit 36 ​​executes an evaluation step of evaluating the cognitive function of the subject U based on the boundary line for identifying whether or not cognitive function has declined, which is constructed using the cognitive function evaluation model, and the feature line calculated in the feature calculation step (S22). In this embodiment, frequency analysis is used to quantify the smoothness of movement. As cognitive function declines, the mean frequency tends to increase and the maximum frequency tends to decrease. Therefore, the evaluation unit 36 ​​evaluates that cognitive function has declined when the relationship between the median of the mean frequency and the median of the maximum frequency is below the boundary line, based on the boundary line constructed by the cognitive function evaluation model. The evaluation result is transmitted to the control unit 38.

[0056] When the control unit 38 receives the evaluation results, it displays the evaluation results on the screen 11 of the display device 10 (S24). If there are no problems with cognitive function, the evaluation results display a mark indicating that cognitive function is good and a message such as "No changes were observed in cognitive function. A good condition has been maintained." If there are problems with cognitive function, a mark indicating that cognitive function has declined and a message such as "Mild changes were observed in cognitive function" are displayed. Then, each process in the cognitive function evaluation method ends. [Example]

[0057] Next, the cognitive function assessment system 1 according to the above embodiment will be specifically described using examples.

[0058] Examples will be described with reference to Fig. 6 to Fig. 8. Fig. 6 is a graph showing the evaluation results by frequency analysis in Example 1. Fig. 7 is a graph showing the evaluation results by sign change analysis in Example 2. Fig. 8 is a graph showing the evaluation results by jerk analysis in Example 3. In Fig. 6 to Fig. 8, the independent elderly group is plotted with white circles (○) and the cognitive decline group is plotted with black circles (●).

[0059] In this example, cognitive function is evaluated using the cognitive function evaluation system 1 of the present invention for a subject U with normal cognitive function (hereinafter referred to as an "independent elderly person") and a subject U with cognitive decline (hereinafter referred to as a "person with cognitive decline"), and the effectiveness of the present invention is verified. Whether the cognitive function of subject U is normal or declined is determined using the Mini Mental State Examination (MMSE), which has long been known as a screening test for assessing the state of cognitive function. Specifically, subjects U who score 28 or more on the MMSE, which indicates a healthy cognitive function, are considered independent elderly people, and subjects U who score 27 or less on the MMSE, which indicates a suspected decline in cognitive function, are considered cognitively declined.

[0060] In this example, 10 independent elderly people and 10 people with cognitive impairment were asked to take a test for evaluating cognitive function according to the above embodiment, and the results were evaluated by the evaluation unit 36 ​​of the above embodiment. In evaluating cognitive function, frequency analysis, sign change analysis, and jerk analysis, which are mainly used in the field of rehabilitation, are adopted as methods for quantifying the smoothness of movement. In Example 1, frequency analysis was used as a method for quantifying the smoothness of movement. In Example 2, sign change analysis was used as a method for quantifying the smoothness of movement. In Example 3, jerk analysis was used as a method for quantifying the smoothness of movement.

[0061] In Example 1, the feature amount is calculated using the median of the average frequency in the anterior-posterior direction and the median of the maximum frequency in the anterior-posterior direction. As shown in FIG. 6, boundary line B is shown as a linear function in which the median of the average frequency and the median of the maximum frequency are proportional to each other. FIG. 6 shows the results of cognitive function assessment of 10 independent elderly people and 10 people with cognitive decline using the cognitive function assessment system 1 of the above embodiment. As can be seen from FIG. 6, the independent elderly people are plotted above boundary line B, and the people with cognitive decline are plotted below boundary line B. As described above, in Example 1, cognitive function is assessed as declined when the score is below boundary line B, and therefore it can be seen that cognitive function can be accurately assessed by the cognitive function assessment system 1 of the present invention.

[0062] In Example 2, the feature amount is calculated as the average value of the number of sign changes in the forward and backward directions. Boundary line B is shown as a threshold value of the number of sign changes, as shown in Figure 7. Figure 7 shows the results of cognitive function assessment of 10 independent elderly people and 10 people with cognitive decline using the cognitive function assessment system 1 of the above embodiment. As can be seen from Figure 7, the independent elderly people are plotted below boundary line B, and the people with cognitive decline are plotted above boundary line B. As mentioned above, in Example 2, cognitive function is assessed as declining when it is above boundary line B, so it can be seen that cognitive function can be accurately assessed by the cognitive function assessment system 1 of the present invention.

[0063] In Example 3, the feature amount is calculated by calculating a norm from the three-axis jerk data and using the average value of the mean absolute jerk of the norm and the average value of the maximum absolute jerk of the norm. As shown in FIG. 8, boundary line B is shown as a linear function in which the average value of the mean absolute jerk and the average value of the maximum absolute jerk are proportional to each other. FIG. 8 shows the results of cognitive function assessment of 10 independent elderly people and 10 people with cognitive decline using the cognitive function assessment system 1 of the above embodiment. As can be seen from FIG. 8, the independent elderly people are plotted below boundary line B, and the people with cognitive decline are plotted above boundary line B. As described above, in Example 3, cognitive function is assessed as declined when the value exceeds boundary line B, and therefore it can be seen that cognitive function can be accurately assessed by the cognitive function assessment system 1 of the present invention.

[0064] As described above, the cognitive function assessment system 1 of the present invention measures the three-dimensional coordinates of key points on the body when testing cognitive function. Then, based on the three-dimensional coordinate data, a smoothness value and feature values ​​that quantify the smoothness of the movements are calculated, and cognitive function is assessed using the calculated feature values. That is, cognitive function is assessed by taking into account the movements of key points on the body of the subject U. A conventional test for assessing cognitive function, known as TMTpartB, simply measures the length of time it takes to select all of the signs S. Assessing cognitive function solely by the overall length of time can overlook localized hesitation or delayed reactions during the test. While localized hesitation and delayed reactions are signs of cognitive function decline, simply looking at the overall length of time can miss these signs. Generally, when people experience hesitation or delayed reactions, their hand and arm movements become awkward. In the present invention, the smoothness of the subject U's hand and arm movements is quantified for assessment, allowing for assessment of cognitive function without overlooking localized hesitation and delayed reactions, or the associated signs of cognitive function decline. As a result, cognitive function decline can be assessed with high accuracy.

[0065] Although the embodiment of the present invention has been described above, the present invention is not limited to the above embodiment, and various modifications are possible without departing from the spirit of the present invention.

[0066] For example, in the above embodiment, the measuring device 20 is configured as a depth sensor attached to the display device 10 and measures the three-dimensional coordinates of key points on the body of the subject U, but this configuration is not necessarily required. For example, the measuring device 20 may be configured as an acceleration sensor attached to a joint point such as the wrist of the subject U, and may be a device that measures the three-axis acceleration of a key point on the body of the subject U who touches a sign S displayed on the display device 10. In this case, the measuring device 20 measures the acceleration in the up-down direction, the acceleration in the left-right direction, and the acceleration in the front-back direction of the subject U at each key point on the body to which the measuring device 20 is attached.

[0067] When the measurement device 20 is configured as an acceleration sensor, the information processing device 30 evaluates the cognitive function of the subject U based on the three-axis acceleration of key points on the body of the subject U measured by the measurement device 20. Specifically, the data acquisition unit 32 acquires three-axis acceleration data for the number of touches divided by segments as time-series data. Furthermore, the motion data calculation unit 33 calculates the motion data of the body key points for the number of touches based on the three-axis acceleration data for the number of touches. That is, when frequency analysis and sign change analysis are used, three-axis velocity data for the number of touches obtained by first-order integration of the three-axis acceleration data with respect to the sampling time is calculated as the motion data of the body key points. When jerk analysis is used, three-axis jerk data for the number of touches obtained by first-order differentiation of the three-axis acceleration data with respect to the sampling time is calculated as the motion data of the body key points. The configurations of the other units, such as the marker display unit 31, smoothness value calculation unit 34, feature calculation unit 35, and evaluation unit 36, and the cognitive function evaluation method are the same as those of the above-described embodiment.

[0068] Even when the measuring device 20 is configured as an acceleration sensor, a smoothness value and feature values ​​that quantify the smoothness of movement are calculated based on the triaxial acceleration data, and cognitive function is evaluated using the calculated feature values. That is, cognitive function is evaluated taking into account the movements of key points on the body of the subject U. This achieves the above-described effect of the present invention, that is, cognitive function can be evaluated without overlooking localized hesitation, delayed reactions, and accompanying signs of cognitive function decline, and cognitive function decline can be evaluated with high accuracy.

[0069] Furthermore, in the above embodiment, the signs S are of two types, 12 numbers and 12 hiragana characters, and are displayed in an orderly fashion in three columns and four rows on the screen 11, but this configuration is not necessarily required. For example, the total number of signs S is not limited to 24, and can be any number. Furthermore, the combination of signs S is not limited to two types, such as numbers and hiragana characters, but can be three or more types, such as numbers, hiragana characters, and alphabets. Furthermore, when the signs S are displayed in an orderly fashion on the screen 11 of the display device 10, the arrangement is not limited to three columns and four rows, and the signs S can be arranged in any number of columns, or they can be displayed randomly rather than in an orderly fashion. [Industrial Applicability]

[0070] The cognitive function assessment system, cognitive function assessment method, and cognitive function assessment program of the present invention are used to assess cognitive function in elderly care facilities, etc., and are used to detect decline in cognitive function early, and therefore have industrial applicability. [Explanation of symbols]

[0071] 1. Cognitive function assessment system 10 Display device 11 screens 20 Measuring Equipment 30 Information processing equipment 31 Sign display section 32 Data Acquisition Section 33 Exercise data calculation unit 34 Smoothness value calculation section 35 Feature calculation unit 36 Evaluation Department S sign U Subject

Claims

1. A display device that displays a plurality of types of signs that can be selected in sequence on a screen randomly or in an orderly manner; a measuring device that measures three-dimensional coordinates or three-axis acceleration of key points on the body of a subject who touches the mark displayed on the display device; an information processing device that evaluates the cognitive function of the subject based on the three-dimensional coordinates or the three-axis acceleration of key points on the body of the subject measured by the measurement device, The information processing device includes: a sign display unit that displays a plurality of types of signs that can be selected in order randomly or in an orderly manner on the screen of the display device; a data acquisition unit that acquires three-dimensional coordinate data or three-axis acceleration data measured by the measuring device for the number of touches divided by the divisions, using the time points when the subject touches the multiple types of signs displayed by the sign display unit as divisions; a motion data calculation unit that calculates motion data of key points of the body for the number of touches based on the three-dimensional coordinate data or the three-axis acceleration data for the number of touches acquired by the data acquisition unit; and a smoothness value calculation unit that calculates smoothness values ​​representing smoothness of movements of key points of the subject's body for the number of touches from the motion data for the number of touches calculated by the motion data calculation unit; a feature value calculation unit that calculates feature values ​​of the motion of key points of the subject's body from the smoothness values ​​for the number of touches calculated by the smoothness value calculation unit; A cognitive function assessment system comprising: a boundary line for identifying whether or not cognitive function is declining, constructed using a cognitive function assessment model that has previously undergone machine learning to determine the correspondence between features and cognitive function; and an assessment unit that evaluates the cognitive function of a subject based on the features calculated by the feature calculation unit.

2. 2. The cognitive function assessment system according to claim 1, wherein the measurement device is a depth camera provided above the display device or an acceleration sensor attached to the wrist of the subject.

3. the movement data is three-axis velocity data for the number of touches obtained by differentiating the three-dimensional coordinate data or integrating the three-axis acceleration data, the smoothness value is an average frequency and a maximum frequency in the three-axis directions for the number of touches obtained by frequency analysis of the three-axis velocity data, the feature amount is an average value, a median value, or a maximum value of the average frequency in the three axial directions for the number of touches, and an average value, a median value, or a maximum value of the maximum frequency in the three axial directions for the number of touches, The cognitive function assessment system according to claim 1 or 2, characterized in that the evaluation unit evaluates that cognitive function is impaired when the relationship between the average value, median value, or maximum value of any of the average frequencies in the three axial directions and the average value, median value, or maximum value of any of the maximum frequencies in the three axial directions is below the boundary line.

4. the movement data is three-axis velocity data for the number of touches obtained by differentiating the three-dimensional coordinate data or integrating the three-axis acceleration data, the smoothness value is the number of sign changes in the three-axis directions for the number of touches obtained by analyzing sign changes in the three-axis velocity data, the feature amount is an average value, a median value, or a maximum value of the number of sign changes in the three axis directions for the number of touches, The cognitive function assessment system described in claim 1 or 2, characterized in that the evaluation unit evaluates that cognitive function is declining if the average, median, or maximum value of any of the sign change numbers in the three axial directions exceeds the boundary line.

5. the motion data is three-axis jerk data for the number of touches obtained by differentiating the three-dimensional coordinate data or the three-axis acceleration data, the smoothness value is an average absolute jerk and a maximum absolute jerk for the number of touches obtained by performing jerk analysis on any one of the axis components or norms of the three-axis jerk data, the feature amount is an average value, a median value, or a maximum value of the mean absolute jerk for the number of touches, and an average value, a median value, or a maximum value of the maximum absolute jerk for the number of touches, 3. The cognitive function assessment system according to claim 1, wherein the assessment unit assesses that cognitive function is impaired when a relationship between an average value, a median value, or a maximum value of the mean absolute jerk and an average value, a median value, or a maximum value of the maximum absolute jerk exceeds the boundary line.

6. 2. The cognitive function evaluation system according to claim 1, wherein the signs are of two types: numbers and hiragana.

7. 7. The cognitive function evaluation system according to claim 6, wherein the signs are the numbers "1" to "12" and the hiragana characters "a" to "shi."

8. The cognitive function assessment system of claim 7, characterized in that the signs for the numbers "1" to "12" and the hiragana characters "a" to "shi" are arranged in an orderly fashion in three vertical and four horizontal rows and are displayed in two separate displays.

9. A cognitive function evaluation method for evaluating a cognitive function of a subject, which is executed by an information processing device having a sign display unit, a data acquisition unit, a motion data calculation unit, a smoothness value calculation unit, a feature calculation unit, and an evaluation unit, a sign display step in which the sign display unit randomly or orderly displays a plurality of types of signs that can be selected in order on a screen of a display device; a data acquisition step in which the data acquisition unit acquires three-dimensional coordinates or three-axis accelerations of key points on the body of the subject who taps the marker displayed on the display device, measured by a measurement device, as three-dimensional coordinate data or three-axis acceleration data for the number of touches divided by the divisions, with the time points at which the subject touches the marker being used as divisions; a motion data calculation step in which the motion data calculation unit calculates motion data of key points of the body for the number of touches based on the three-dimensional coordinate data or the three-axis acceleration data for the number of touches acquired in the data acquisition step; a smoothness value calculation step in which the smoothness value calculation unit calculates smoothness values ​​representing smoothness of movements of key points of the subject's body for the number of touches from the motion data for the number of touches calculated in the motion data calculation step; a feature calculation step in which the feature calculation unit calculates feature values ​​of the motion of key points on the body of the subject from the smoothness values ​​for the number of touches calculated in the smoothness value calculation step; The cognitive function assessment method is characterized in that the assessment unit comprises an assessment step of assessing the cognitive function of the subject based on a boundary line for identifying whether or not cognitive function has declined, which boundary line is constructed using a cognitive function assessment model that has previously undergone machine learning to determine the correspondence between features and cognitive function, and the features calculated in the feature calculation step.

10. A cognitive function assessment program for evaluating the cognitive function of a subject, a sign display step of randomly or systematically displaying a plurality of types of signs that can be selected in order on a screen of a display device; a data acquisition step of acquiring three-dimensional coordinates or three-axis accelerations of key points on the body of the subject who taps the mark displayed on the display device, measured by a measurement device, as three-dimensional coordinate data or three-axis acceleration data for the number of touches divided by the divisions, with the time points at which the subject touches the mark; a motion data calculation step of calculating motion data of key points of the body for the number of touches based on the three-dimensional coordinate data or the three-axis acceleration data for the number of touches acquired in the data acquisition step; a smoothness value calculation step of calculating smoothness values ​​representing smoothness of movements of key points of the subject's body for the number of touches from the motion data for the number of touches calculated in the motion data calculation step; a feature value calculation step of calculating feature values ​​of the motion of key points of the subject's body from the smoothness values ​​for the number of touches calculated in the smoothness value calculation step; A cognitive function assessment program that causes a computer to execute processing including: a boundary line that identifies whether or not cognitive function is declining, constructed using a cognitive function assessment model that has previously undergone machine learning to determine the correspondence between features and cognitive function; and an assessment step that evaluates the cognitive function of a subject based on the features calculated in the feature calculation step.

Images