JP2019162288A - Determination value calculation device, method and program, and micro error occurrence determination device, method and program - Google Patents

Abstract

An object of the present invention is to automatically determine occurrence or non-occurrence of a micro error. A timing at which it is determined that a micro error has occurred while a subject performs the task of the instrumental daily life activity is associated with data of the speed of a finger performing the task. From the time-series data of the speed of the finger performing the task, the time-series data of the speed of the segment including the timing at which it was determined that the micro-error occurred was determined, and the non-stationary state of the finger for determining that the micro-error occurred. Is calculated. [Selection diagram] FIG.

Description

  The present invention relates to a determination value calculation device, method, and program, and a micro error occurrence determination device, method, and program.

  Conventionally, methods and systems for distinguishing between patients with mild cognitive impairment (MCI) and healthy individuals without impaired brain function have focused on cognitive function testing. In addition, a memory test using a questionnaire (Patent Document 5) has been proposed. Related technologies include the following patent documents 1 to 12.

JP 2017-156402 A JP 2017-104289 A Japanese Patent Laying-Open No. 2015-173695 JP 2014-018341 A JP 2007-282929 A Special table 2011-511762 gazette Special table 2011-502564 gazette Special table 2009-528103 gazette Special table 2009-515568 Special table 2008-510580 gazette Special table 2007-532219 gazette Table 2014/013999

  However, patients with MCI have a decline in function in IADL (Instrumental Activities of Daily Living), which is a complex and higher-order operation group than ADL (Activities of Daily Living). I know. As a new screening index for specifying MCI, the necessity of paying attention to an action index has been shown.

  In addition, ADL (Daily Life Activity) refers to actions and actions that are normally performed in daily life. Specifically, basic actions such as meals, excretion, preparation, movement, bathing, etc. Point. On the other hand, IADL (mean daily life activities) is a more complex and higher-order operation than ADL, specifically, housework such as shopping, washing or cleaning, money management, medication management, going out and vehicle An action such as riding a car.

  One of the behavior indicators is a human error called micro-error. A micro error is a stagnation of an operation that is distinguished from an erroneous operation assumed when performing ADL or IADL. It is known that the frequency of occurrence of micro errors in patients with MCI is significantly different from that of healthy individuals, and micro errors are attracting attention as an effective index for early detection of MCI. There are various types of micro errors, but typically there are the following two types. First, there is a reach and contact (reach touch) type in which an undesired object is extended and touched, but at the moment of touching, the object is released and the hand is extended to a different object. For example, while making coffee, reach out and touch an unnecessary jelly bottle, and change the action plan at the moment of touching. Secondly, there is a reach-and-contact type (reach no-touch) in which a hand is reached and touched to an unnecessary object, and the action is changed to an action of reaching a different object on the verge. For example, while making coffee, you reach out to an unnecessary jelly bottle, but do not touch it, and reach out to a different object just before it.

  However, at present, micro errors are identified using a moving image observation method by a trained observer, and a technique for automatically detecting the occurrence or non-occurrence of micro errors has not been established.

  The technology of the present disclosure automatically determines whether or not a micro error has occurred by automatically determining whether or not a micro error has occurred. An object of the present invention is to provide a micro error occurrence determination apparatus, method, and program capable of performing the above.

  In order to achieve the above object, a first aspect of the technology of the present disclosure includes a display unit that displays an environment for a subject to perform a predetermined activity, a measurement unit that measures the position of the physical part of the subject, Time-series data of the velocity of the physical part from the data of the position measured in time series by the measurement unit of the physical part of the subject who tried to perform the predetermined action in the environment displayed by the display unit The speed calculation unit for calculating the time and the time of the speed within a predetermined range including a time when a micro error, which is an incorrect action of the subject, is distinguished from an erroneous action assumed when performing the predetermined activity A determination value calculation device that calculates a determination value for determining that the micro error has occurred from series data, wherein the determination value is a non-physical part of the subject. Stationary Is a value representing a tree.

  The display unit displays an environment for the subject to perform a predetermined activity, and the measurement unit measures the position of the physical part of the subject. From the data of the position measured in time series by the measurement unit of the physical part of the subject who tried to perform the predetermined action in the environment displayed by the display unit, the speed calculation unit of the physical part Calculate time-series data of speed. Time-series data of the speed within a predetermined range including a time when a micro error, which is an erroneous action of the subject, is distinguished from an erroneous action assumed when the judgment value calculation unit performs the predetermined activity Then, a determination value for determining that the micro error has occurred is calculated.

  The determination value is a value representing an unsteady movement of the physical part of the subject.

  The second aspect of the technology of the present disclosure is displayed by the display unit that displays an environment for the subject to perform a predetermined activity, the measurement unit that measures the position of the physical part of the subject, and the display unit. A speed calculation unit that calculates time-series data of the speed of the physical part from data of positions measured in time series by the measurement unit of the physical part of the subject who has attempted the predetermined action in the environment; The micro calculated in advance from the time-series data of the speed in a predetermined range including the time when a micro error, which is a grudge of the subject's motion that is distinguished from an erroneous motion assumed when performing a predetermined activity, is generated From the determination value for determining that an error has occurred and the time-series data of the speed of the physical part calculated by the speed calculation unit, the physical part of the subject who tried to perform the predetermined action A determination unit that determines whether or not the micro error has occurred, and the determination value is a value that represents an unsteady movement of the physical part of the subject. .

  The display unit displays an environment for the subject to perform a predetermined activity, and the measurement unit measures the position of the physical part of the subject. From the data of the position measured in time series by the measurement unit of the physical part of the subject who tried to perform the predetermined action in the environment displayed by the display unit, the speed calculation unit of the physical part Calculate time-series data of speed. The determination unit is calculated in advance from time-series data of the speed within a predetermined range including a time when a micro error, which is a grudge of the subject's motion that is distinguished from an erroneous motion assumed when performing a predetermined activity, is generated. From the determination value for determining that the micro error has occurred, and the time-series data of the speed of the physical part calculated by the speed calculation unit, the physical movement of the subject who tried to perform the predetermined action It is determined whether or not the micro error has occurred at the site.

  The determination value is a value representing an unsteady movement of the physical part of the subject.

  A third aspect of the technology of the present disclosure is a determination value calculation program that causes a computer to function as the speed calculation unit and the determination value calculation unit of the first aspect.

  A third aspect of the technology of the present disclosure is a micro error occurrence determination program that causes a computer to function as the speed calculation unit and the determination unit of the second aspect.

  The technology of the present disclosure can enable to automatically determine whether a micro error has occurred or not.

  The technology of the present disclosure can automatically determine whether a micro error has occurred or not.

1 is a diagram illustrating a data processing device 100. FIG. (A) shows a first image representing the first environment of the instrumental daily life activity in a virtual space, and (B) shows a first image of the instrumental daily life activity different from the first environment. It is a figure which shows the 2nd image which expressed 2 environment to virtual space. 2 is a block diagram of an electrical system configuration of the data processing apparatus 100. FIG. 4 is a block diagram of functions of a CPU 20 of the data processing apparatus 100. FIG. It is a flowchart which shows a data processing program. It is a flowchart of the program of Data collection processing. It is a flowchart of the program of a Data cleaning process. It is a flowchart of the program of Segmentation processing. It is a flowchart of the program of the 1st process of Feature extraction process. It is a flowchart of the program of the 2nd process of Feature extraction process. It is a flowchart of the program of the 3rd process of Feature extraction process. It is a figure which shows a mode that a test subject moves a finger | toe when a test subject performs the task which is predetermined | prescribed operation | movement of a means daily life activity. It is a graph which shows the time series data of the speed of a finger. It is a figure which shows a mode that the speed | velocity | rate of the finger at the time of performing one operation | movement of a task gradually increases from speed 0, reaches a peak, and after reaching a peak, it gradually decreases and becomes speed 0. It is a figure which shows a mode that the time series data of the speed of the finger | toe were divided | segmented into the group (segment) of each operation | movement. It is a figure which shows dividing | segmenting the time series data of the speed of the finger | toe of each segment according to whether the micro error has generate | occur | produced. It is a figure which shows the skewness of the speed data in each segment in the time series data of speed. It is a figure which shows the value regarding the autoregressive model of the data of the speed of each segment in the time series data of speed. It is a figure which shows the value obtained by carrying out the specific spectrum conversion of the speed data of each segment in the time series data of speed. It is a figure which shows that the activity in a segment is moving a target from a start point (Start) to a target point (Target), and a finger moves from a start point (Start) to a target point (Target). (A) shows time-series data of the speed of each segment, (B) shows the difficulty calculated based on the Fitts model, and (C) shows the segment related to the occurrence of ME. FIG. (A) shows the result of ME detection of the Reach No-touch type when the task is prepared for breakfast, (B) shows the result of ME detection of the Reach touch type when the task is prepared for breakfast, and (C) shows FIG. 4D is a diagram showing a result of ME detection of the Reach No-touch type when the task is prepared for lunch, and FIG. 4D is a diagram showing a result of ME detection of the Reach touch type when the task is prepared for lunch.

  FIG. 1 shows a data processing apparatus 100. FIG. 2A shows a first image in which a first environment of instrumental daily activities of daily living (IADL) is expressed in a virtual space. FIG. 2B shows a second image in which a second environment different from the first environment of the instrumental daily life activity is represented in a virtual space. FIG. 3 shows a block diagram of the data processing apparatus 100.

  As shown in FIG. 3, the data processing apparatus 100 is an image (IADL environment) that expresses an environment (IADL environment) in which a subject conducts a means of daily life activity using a virtual reality (VR) technique (virtual reality) technology in a virtual space. A display 12 (see also FIG. 1) for displaying FIG. 2A or FIG. 2B is provided.

  Further, the data processing apparatus 100 is attached to a column from a pan head on a tripod, and includes a three-dimensional motion sensor 16 (see also FIG. 1) that measures the three-dimensional position of the subject's finger placed on the screen of the display 12. I have.

  Furthermore, the data processing apparatus 100 calculates a determination value for determining that a micro error has occurred from the time-series data of the three-dimensional position of the subject's finger who has attempted a predetermined action in the environment displayed on the display 12. A computer 10 is provided. The predetermined operation is, for example, preparation of breakfast or lunch in means of daily life activities. The computer 10 further determines whether or not a micro error has occurred from the calculated action determination value and the three-dimensional position of the subject's finger.

  As will be described in detail later, in the operation of the data processing apparatus 100, a first stage of calculating a judgment value for judging that a micro error has occurred and a micro error has occurred using the judgment value. There is a second stage for determining whether or not. The data processing device 100 functions as a determination value calculation device in the first stage, and functions as a micro error occurrence determination device in the second stage.

  The computer 10 includes a CPU 20, a ROM 22, a RAM 24, and an input / output interface (I / O) 26. The CPU 20, ROM 22, RAM 24, and input / output interface (I / O) 26 are connected to each other via a bus 28. A display 12 and a three-dimensional motion sensor 16 are connected to an input / output interface (I / O) 26. The computer 10 is configured by, for example, a personal computer (PC).

  The ROM 22 stores a data processing program (FIG. 4) described later. The data processing program (FIG. 4) is read from the ROM 22, expanded in the RAM 24, and executed by the CPU 20.

  Next, various functions realized by the CPU 20 of the data processing apparatus 100 executing a data processing program will be described with reference to FIG. The data processing program has a Data collection function, a Data cleaning function, a Segmentation function, a Feature extraction function, and a Classification function. When the CPU 20 executes the data processing program having these functions, the CPU 20 is configured as a data collection unit 32, a data cleaning unit 34, a segmentation unit 36, a feature extraction unit 38, and a classification unit 40 as shown in FIG. Function.

  Next, data processing by the data processing apparatus 100 will be described in detail with reference to FIG. The CPU 20 of the data processing apparatus 100 executes the data processing program (determination value calculation program, micro error occurrence determination program), thereby realizing the data processing shown in the flowchart of FIG.

  The display 12 displays an image (see FIG. 2) expressing the environment of instrumental daily life activities (IADL) in a virtual space. As shown in FIG. 12, while the subject performs a task (for example, preparation of a lunch box) which is a predetermined operation of the daily life activity by moving his / her finger, the three-dimensional motion sensor 16 performs a predetermined time. The three-dimensional position of the subject's finger is measured every time, and the measured three-dimensional position data is transmitted to the computer 10 every predetermined time.

  In step 102 of FIG. 5, the data collection unit 32 executes data collection (data collection) processing, and in step 104, the data cleaning unit 34 executes data cleaning (data cleansing) processing. In step 106, the segmentation unit 36 executes a segmentation process, and in step 108, the feature extraction unit 38 executes a feature extraction (feature extraction) process. In step 110, the classification unit 40 executes a classification process. Details of each process will be described later.

  FIG. 6 shows a flowchart of the data collection processing program.

  The data collection unit 32 acquires the result of detecting the three-dimensional position of the subject's finger by the three-dimensional motion sensor 16 in step 202. In step 204, the data collection unit 32 transfers the measurement data (time series data of the three-dimensional position of the finger) to the memory (RAM 24) of the PC.

  In step 206, the data collection unit 32 displays measurement data on the display 12 of the PC (computer 10). In step 208, the data collection unit 32 determines whether or not the measurement is finished by determining whether or not data is continuously transmitted from the three-dimensional motion sensor 16 for a certain period of time. Note that data is not transmitted from the three-dimensional motion sensor 16 when the subject determines that the predetermined activity has been completed and removes the finger from the measurement region of the three-dimensional motion sensor 16.

  If it is not determined that the measurement has been completed, the data collection process returns to step 202. If it is determined that the measurement has ended, the Data collection process ends.

  FIG. 7 shows a flowchart of a data cleaning process program.

  In step 212, the data cleaning unit 34 acquires the measured information (data of the three-dimensional position of the finger) from the memory (RAM 24) of the PC (computer 10).

  In step 214, the data cleaning unit 34 calculates the finger speed (in this application, “speed” refers to the absolute value of the speed and does not include the direction) from the acquired data. In step 216, the data cleaning unit 34 creates a probability distribution from the absolute value of the calculated speed. Specifically, the data cleaning unit 34 creates a histogram of absolute values of speeds, and creates a probability distribution by dividing the frequency of each speed by the number of data of the entire speed. In step 216, the data cleaning unit 34 further displays the created probability distribution on the display 12.

  In step 218, the data cleaning unit 34 determines whether there is data having a speed exceeding a value twice the standard deviation. If the determination in step 218 is affirmative, the data cleaning unit 34 removes the outlier in step 220. Thereafter, the data cleaning process returns to step 212.

  If the determination in step 218 is negative, the data cleaning unit 34 performs an interpolation process in step 222 so that the calculated speed data is arranged at predetermined time intervals. Interpolation processing is roughly divided into first interpolation processing and second interpolation processing. The first interpolation process is a process of interpolating the speed data at the time corresponding to the removed outlier in step 220 with the speed data before and after the time. By the way, as described above, the three-dimensional motion sensor 16 transmits the data of the three-dimensional position of the subject's finger to the computer 10 at predetermined time intervals. It may be off from. The second interpolation process is a process of interpolating missing data or data deviating from a predetermined time interval with data of speeds before and after each time.

  After step 222, the data cleaning process ends. As shown in FIG. 13, time series data of finger speed is obtained by the data cleaning process.

  FIG. 8 shows a flowchart of the segmentation processing program. In step 230, the segmentation unit 36 acquires speed data from the memory (RAM 24) of the PC (computer 10). In step 232, the segmentation unit 36 second-order differentiates the speed data. In step 234, the segmentation unit 36 detects the zero crossing point of the second-order differentiated data. In step 236, the segmentation unit 36 determines whether or not the changing point from the negative region to the positive region is a zero crossing point. If the determination in step 236 is affirmative, the segmentation unit 36 stores the segment point in step 238. Thereafter, the segmentation process proceeds to step 240. If the determination in step 236 is negative, the segmentation process proceeds to step 240. In step 240, the segmentation unit 36 determines whether the processing has been completed up to the end of the data point. If the determination in step 240 is negative, the segmentation process returns to step 230. If step 240 is positive, the segmentation process ends.

  As described above, in order to execute the task of the instrumental daily life activity, the finger is moved as shown in FIG. 12, but the task of the instrumental daily life activity is composed of a plurality of actions for moving the finger in this way, Composed. As shown in FIG. 14, the speed of the finger in each operation gradually increases from speed 0, reaches a peak, and after reaching the peak, gradually decreases and reaches speed 0. By the segmentation process shown in FIG. 8, the time-series data of the finger speed shown in FIG. 13 is divided into a group of each operation as shown in FIG.

  FIG. 9 shows a flowchart of the first process program of the feature extraction process.

  By the way, whether or not a micro error has occurred when multiple researchers who have been trained in micro error coding observe the movement of the subject's fingers while the subject performs the above-mentioned task of instrumental daily life activities. Judging. When it is determined that a micro error has occurred, the plurality of researchers input a micro error occurrence identification signal to the PC (computer 10) via an input device (not shown). The computer 10 associates the timing at which the micro error occurrence identification signal is input with time-series data of finger speeds as shown in FIG. In the example shown in FIG. 16, it is shown that the micro error has occurred twice.

  When the program of the first process of the feature extraction process in FIG. 9 starts, the feature extraction unit 38 extracts the velocity data of one segmentation section in step 250. In step 252, the feature extraction unit 38 determines whether a micro error (ME) has occurred by determining whether a micro error occurrence identification signal is associated with the corresponding section.

  If step 252 is affirmative, the feature extraction unit 38 stores the speed data of the section in which the ME has occurred in the ME generation section storage area 254M of the memory (RAM 24) in step 254. When the determination at step 252 is negative, the feature extraction unit 38 stores the speed data of the section where the ME is not generated in the ME non-occurrence section storage area 256M of the memory (RAM 24) at step 256. After steps 254 and 256, the feature extraction unit 38 determines in step 258 whether or not the processing has been completed up to the terminal segmentation section. If step 258 is negative, the feature extraction process returns to step 250. If step 258 is positive, the feature extraction process ends.

  FIG. 10 shows a flowchart of the second processing program of the feature extraction processing. In step 262, the feature extraction unit 38 accesses the speed data of the segment section stored in the ME non-occurrence section storage area 256M in the speed data of the section where no ME has occurred. In step 264, the feature extraction unit 38 calculates a feature amount described later in detail. In step 266, it is determined whether the processing has been completed up to the end segmentation section. If step 266 is negative, the feature extraction process returns to step 262. If step 266 is affirmative, the feature extraction process ends.

  FIG. 11 shows a flowchart of the third process program of the feature extraction process.

  In step 272, the feature extraction unit 38 accesses the speed data of the segment section in the speed data of the section where the ME has occurred, which is stored in the ME generation section storage area 254M. In step 274, the feature extraction unit 38 calculates a feature amount (determination value). In step 276, it is determined whether or not the processing has been completed up to the end segmentation section. If step 276 is negative, the feature extraction process returns to step 272. If step 276 is positive, the feature extraction process ends.

  Next, feature amounts (step 264 (FIG. 10), step 274 (FIG. 11)) in the present embodiment will be described.

  The feature amount (determination value) in step 274 (FIG. 11) is a determination value for determining that a micro error has occurred, and is a value that represents an unsteady movement of the subject's finger (physical part).

  Specifically, the determination value includes the skewness of the velocity data in the segment, the value related to the autoregressive model of the velocity data, the value obtained by performing singular spectrum conversion on the velocity data, the Fitt's raw model ( The value of the segment's activity determined using Fitz's Law)

  With reference to FIG. 17, the skewness of the speed data in each segment in the speed time-series data will be described. The skewness S is expressed as follows.

  Here, N is the total number of velocity data in the segment. xi is the i-th speed data in the segment.

Is the average value of the velocity data in the segment. i is a variable that identifies the velocity data in the segment.

  Next, with reference to FIG. 18, the value regarding the autoregressive model of the velocity data of each segment in the velocity time-series data will be described. As shown in FIG. 18, the time-series data of the speed in each segment is further upsampled, and the speed data at each time of upsampling in each segment is obtained by interpolation. The autoregressive model of velocity data is expressed as follows:

y _t is the predicted value of the velocity at time t after upsampling in each segment. y _t−1 is speed data at time ti in each segment. a _i is a coefficient of the autoregressive model, and e _i is a difference (residual) between the speed data based on the polynomial fitted by the autoregressive model and the actual speed data.

The determination value may be the coefficient a _{i of the} autoregressive model or the difference (residual) e _i .

  Next, with reference to FIG. 19, values obtained by performing singular spectrum conversion on the velocity data of each segment in the velocity time-series data will be described. As shown in FIG. 19, the following processing is performed for each segment. That is, the time series data of the speed in each segment is further upsampled, and the speed data at each time of upsampling in each segment is obtained by interpolation. For the speed data at each time t after upsampling in each segment, a Hankel matrix is created from the speed data before and after the time t, and the Hankel matrix is subjected to singular value decomposition to obtain a singular vector U, Get Q. Then, the score (Score) is obtained by calculating the inner product of the singular vector U and the singular vector Q. By the way, the speed data actually changes as indicated by a solid line (change in speed) in the lowermost graph of FIG. A plurality of scores (Score) obtained by calculation as described above are obtained as shown by a dotted line (change point Score). For example, if M pieces of velocity data are obtained for the segment by upsampling, M scores are obtained for each velocity data. By dividing the sum of the values of each of the M total scores (Score) by the total number M, the average of the scores (Score) is obtained as the feature amount (determination value) of the segment. The feature amount (determination value), which is a value obtained by performing specific spectrum conversion on the velocity data, is not limited to the average value of the score, and may be the peak time of the score.

Next, with reference to FIGS. 20 and 21, values related to activities in segments determined using the Fitts model will be described. The value related to the activity in the segment determined using the Fitts model includes the time required for the activity and the difficulty of the activity. As shown in FIG. 20, the activity in the segment is to move the object from the start point (Start) to the target point (Target), and the finger moves from the start point (Start) to the target point (Target). The time T _{subtask (i)} required for the activity in the segment determined using the Fitts model is as follows.

T _{subtask (i)} = a + blog ₂ ((D / S) +1)

  D is the distance between the start point (Start) and the target point (Target). S is the range of the target point (Target). a and b are coefficients. i is a variable for identifying each segment.

Further, the difficulty level of activity in the segment determined using the Fitts model is log ₂ ((D / S) +1).

The coefficients a and b are obtained as follows. For all segments, log ₂ ((D / S) +1) and time T _{subtask (i)} are obtained, the vertical axis is time T _{subtask (i)} , and the horizontal axis is log ₂ ((D / S) +1). Plot each value. _Obtain a linear expression of time T _{subtask (i)} for log ₂ ((D / S) +1). a is an intercept of the linear expression, b is a slope of the linear expression, and these are obtained from the linear expression.

When the time T _{subtask (i)} is obtained from the time-series data of the speed of each segment in FIG. 21A, the time T _{subtask (i)} is as shown in FIG. 21B, and FIG. As shown, the time T _{subtask (i)} of the segment associated with the occurrence of ME is the maximum value of the time T _{subtask (i)} . Therefore, the segment time T _{subtask (i)} associated with the occurrence of ME is a determination value. The _{reason why} the time T _{subtask (i)} becomes longer is that the difficulty level log ₂ ((D / S) +1) is high. Therefore, the difficulty level log ₂ ((D / S) +1) of the segment associated with the occurrence of ME is also a determination value.

  The above is the feature amount (determination value) in step 274 (FIG. 11), but the feature amount in step 264 (FIG. 10) is the same as the method for calculating the feature amount (determination value) in step 274 (FIG. 11). Get in the way.

  The feature amount (determination value) in step 274 (FIG. 11) and the feature amount in step 264 (FIG. 10) are normalized.

  The classification process in step 110 in FIG. 5 is a process for obtaining the ME detection rate in the IADL task using the feature amount (determination value). The classification unit 40 determines whether each task is based on an execution result (time-series data of finger speeds) and the above feature amount (judgment value) when the task is prepared for breakfast and when the task is prepared for lunch. It is determined whether ME has occurred in the operation. Then, the classification unit 40 obtains the ME detection rate in each IADL task from the determination result and the actual ME detection result. The results are shown in FIGS. 22 (A) to 22 (D). FIG. 22A shows the result of ME detection of the Reach No-touch type when the task is breakfast preparation. FIG. 22B shows the result of ME detection of the reach touch type when the task is breakfast preparation. FIG. 22C shows a result of ME detection of the Reach No-touch type when the task is lunch preparation. FIG. 22D shows the result of ME detection of the reach touch type when the task is lunch preparation. In each of FIGS. 22A to 22D, the horizontal axis represents a false positive rate, and the vertical axis represents a positive positive rate. 22A to 22D show an ROC (Receiver Operating Characteristic) curve, an AUC based on the ROC curve (area under the curve), a current classifier (micro error occurrence determination). The performance of the device is shown in dots. As shown in FIGS. 22A to 22D, the micro error occurrence determination device of the present embodiment can accurately determine the occurrence of a micro error.

  As described above, in this embodiment, VR (Virtual Reality) technology is used to represent the IADL environment in virtual space, the three-dimensional position of the subject's finger is measured, and the subject's finger during operation is measured. Velocity data is obtained from the three-dimensional position measurement data, and a feature amount (determination value) for determining that a micro error has occurred is calculated. Therefore, a micro error can be identified, and an MCI patient and a healthy person can be discriminated.

  By the way, the conventional MCI screening system requires that the subject go to a medical institution and take a cognitive function test for an average of 1 to 2 hours under the appropriate guidance of the care worker. This is a time-consuming method for workers. Moreover, it was difficult to detect MCI with high sensitivity and specificity only by evaluating cognitive function. In this embodiment, screening is possible in a few minutes, the time cost can be greatly reduced, and since the IADL behavior index is evaluated, it is higher than the conventional evaluation of only the cognitive function. MCI can be detected with sensitivity and specificity.

In this embodiment, it is possible to automatically determine the occurrence or non-occurrence of a micro error, and to accurately determine the occurrence or non-occurrence of a micro error.
Since the judgment value is an index focusing on behavior, it can be introduced into various tasks other than IADL and has a wide range of applications.

(Modification)
In the embodiment described above, the absolute value of the speed is used as the speed, but it may not be the absolute value of the speed. For example, the moving direction of the finger may be taken into account.
In the embodiment described above, the occurrence location of ME is further visualized as information. Thereby, it becomes easy to create an action profile by an expert.
In each embodiment described above, the task is a task of means daily activities. The technology of the present disclosure is not limited to this. The task may be a daily life activity task that is less complex and less sophisticated than the routine daily life activity. The physical part may be a toe, a head, a shoulder or the like according to each activity in addition to the finger of the hand.

  In each of the embodiments described above, the display 12 displays an image that represents an environment in which a subject performs a mean daily life activity in a virtual space using the virtual reality technology. The technology of the present disclosure is not limited to this. The display 12 may display an image that two-dimensionally represents an environment for the subject to perform a daily life activity or a daily life activity.

  In each of the embodiments described above, the ROM 22 stores a data processing program (FIG. 4). The technology of the present disclosure is not limited to this. For example, a data processing program (FIG. 4) is stored in an arbitrary portable storage medium such as a CD-ROM (Compact Disk Read Only Memory), an SSD (Solid State Drive), or a USB (Universal Serial Bus) memory. May be. In this case, the data processing program (FIG. 4) in the storage medium is installed in the data processing apparatus 100, and the installed program is executed by the CPU 20.

  Further, a data processing program (FIG. 4) is stored in a storage unit such as another computer or server device connected to the data processing apparatus 100 via a communication network (not shown). May be downloaded in response to a request from the data processing apparatus 100. In this case, the downloaded data processing program (FIG. 4) is executed by the CPU 20.

  Further, the data processing described in the above embodiments is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, and the processing order may be changed within a range not departing from the spirit. Data processing may be realized only by a hardware configuration such as FPGA or ASIC, or may be realized by a combination of a software configuration using a computer and a hardware configuration.

  The above embodiment has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and includes design and the like within a scope that does not depart from the gist of the technology of the present disclosure. .

  All documents, patent applications and technical standards mentioned in this specification are to the same extent as if each individual document, patent application and technical standard were specifically and individually stated to be incorporated by reference. Incorporated by reference in the book.

12 Display 16 3D motion sensor 20 CPU
32 Data collection unit 34 Data cleaning unit 36 Segmentation unit 38 Feature extraction unit 40 Classification unit

Claims (10)

A display unit for displaying an environment for the subject to perform a predetermined activity;
A measurement unit for measuring the position of the physical part of the subject;
Time-series data of the velocity of the physical part from the data of the position measured in time series by the measurement unit of the physical part of the subject who tried to perform the predetermined action in the environment displayed by the display unit A speed calculation unit for calculating
From the time-series data of the speed within a predetermined range including the time when the micro error, which is the erroneous operation of the subject, is distinguished from the erroneous operation assumed when performing the predetermined activity, the micro error is A determination value calculation unit for calculating a determination value for determining that it has occurred;
A determination value calculation device comprising:
The judgment value is
A value representing the unsteady movement of the subject's physical part,
Determination value calculation device. The display unit virtually displays a three-dimensional environment for the subject to perform a predetermined activity.
The determination value calculation apparatus according to claim 1. The judgment value is
Skewness of the speed data within the predetermined range;
A value for the autoregressive model of the velocity data,
A value obtained by performing singular spectrum conversion on the time series data of the velocity, or a value relating to the activity in the predetermined range determined using the Fitts model.
The determination value calculation apparatus according to claim 1 or 2. A display unit for displaying an environment for the subject to perform a predetermined activity;
A measurement unit for measuring the position of the physical part of the subject;
Time-series data of the velocity of the physical part from the data of the position measured in time series by the measurement unit of the physical part of the subject who tried to perform the predetermined action in the environment displayed by the display unit A speed calculation unit for calculating
The micro error calculated in advance from the time-series data of the speed in a predetermined range including the time when a micro error, which is an erroneous operation of the subject, is distinguished from an erroneous operation assumed when performing a predetermined activity From the determination value for determining that has occurred and the time-series data of the speed of the physical part calculated by the speed calculation unit, the physical part of the subject who is going to perform the predetermined action, A determination unit for determining whether or not a micro error has occurred;
A micro error occurrence determination device comprising:
The judgment value is
A value representing the unsteady movement of the subject's physical part,
Micro error occurrence determination device. The display unit virtually displays a three-dimensional environment for the subject to perform a predetermined activity.
The micro error occurrence determination device according to claim 4. The judgment value is
Skewness of the speed data within the predetermined range;
A value for the autoregressive model of the velocity data,
A value obtained by performing singular spectrum conversion on the time series data of the velocity, or a value relating to the activity in the predetermined range determined using the Fitts model.
The micro error occurrence determination device according to claim 4 or 5. A display unit displaying an environment for the subject to perform a predetermined activity;
A measurement unit measuring the position of the physical part of the subject;
From the data of the position measured in time series by the measurement unit of the physical part of the subject who tried to perform the predetermined action in the environment displayed by the display unit, the speed calculation unit of the physical part Calculating time-series data of speed;
Time-series data of the speed within a predetermined range including a time when a micro error, which is an erroneous action of the subject, is distinguished from an erroneous action assumed when the judgment value calculation unit performs the predetermined activity From the step of calculating a determination value for determining that the micro error has occurred,
A determination value calculation method comprising:
The judgment value is
A value representing the unsteady movement of the subject's physical part,
Determination value calculation method. A display unit displaying an environment for the subject to perform a predetermined activity;
A measurement unit measuring the position of the physical part of the subject;
From the data of the position measured in time series by the measurement unit of the physical part of the subject who tried to perform the predetermined action in the environment displayed by the display unit, the speed calculation unit of the physical part Calculating time-series data of speed;
The determination unit is calculated in advance from time-series data of the speed within a predetermined range including a time when a micro error, which is an erroneous operation of the subject, which is distinguished from an erroneous operation assumed when performing a predetermined activity. From the determination value for determining that the micro error has occurred, and the time-series data of the speed of the physical part calculated by the speed calculation unit, the physical movement of the subject who tried to perform the predetermined action Determining whether or not the micro error has occurred in a site;
A micro error occurrence determination method comprising
The judgment value is
A value representing the unsteady movement of the subject's physical part,
Micro error occurrence determination method.   The determination value calculation program which makes a computer function as the said speed calculation part of any one of Claims 1-3, and the said determination value calculation part.   A micro error occurrence determination program that causes a computer to function as the speed calculation unit and the determination unit according to any one of claims 4 to 6.