Method and device for determining cognitive impairment

Abstract

The present invention discloses a model for determining whether or not a cognitive disorder is present based on gait characteristic values ​​measured from an accelerometer attached to a subject, and a configuration for determining whether or not a cognitive disorder is present by inputting personal information data and gait data and using a model including a mathematical leading regression equation including walking speed and walking time variability as predictor variables. The present invention provides a method for determining cognitive disorder that can easily and quickly select cognitive disorders such as dementia regardless of the underlying cause and accurately determine even early-stage cognitive disorders.

Description

[Technical field]

[0001] The present disclosure provides a method, apparatus, and computer program product for determining cognitive impairment. [Background technology]

[0002] Human gait is associated with the peripheral and central nervous systems, and gait speed, cadence, step length, step time, variability, and asymmetry change with age. Human gait also changes in cognitive disorders (CD), such as mild cognitive impairment and dementia. Slower gait speed and higher gait variability are associated with the prevalence and incidence of cognitive disorders. Previous studies have shown that slower gait speed is associated with brain regions related to information processing speed, while temporal gait variability is associated with brain regions related to cortical sensorimotor control. These findings suggest that slower gait speed is associated with the risk of vascular dementia or Parkinson's dementia, while temporal gait variability is associated with the risk of cognitive decline unrelated to cerebrovascular and / or Parkinson's disease. The motor cognitive risk syndrome, characterized by cognitive decline and slow gait speed, is a new high-risk condition for dementia, which has also been reported to be associated with a higher risk of vascular dementia and Parkinson's disease dementia than Alzheimer's disease. Temporal gait variability has also been associated with cognitive decline, but no previous studies have used temporal gait variability instead of gait speed, or used gait speed and temporal gait variability simultaneously to predict or discriminate cognitive impairment.

[0003] On the other hand, the lack of such studies may be due to the problem of a lack of reliable and approachable methods to measure temporal gait variability. Although gait can be analyzed in various ways, gait analysis using lab-based 3D motion capture systems and instrumented walkways can only measure a limited number of steps due to spatial constraints, and therefore has the disadvantage of being unreliable and expensive as a tool for screening cognitive disorders.

[0004] This has created a need for developing algorithms to measure gait using wearable acceleration sensors that are less spatially constrained and less expensive, and to use the data obtained through gait measurement to identify cognitive impairment.

[0005] The above-mentioned background art is technical information possessed by the inventor for the purpose of deriving the present invention or acquired in the process of deriving the present invention, and is not necessarily publicly known art that was disclosed to the general public prior to the filing of the application of the present invention. Summary of the Invention [Problem to be solved by the invention]

[0006] The present invention discloses various embodiments relating to a cognitive impairment determination method, an apparatus, and a computer program product. The technical problem to be achieved by the present embodiments is not limited to the above-mentioned technical problem, and other technical problems can be inferred from the following embodiments. [Means for solving the problem]

[0007] As a technical means for achieving the above-mentioned technical problem, a first aspect of the present disclosure discloses a cognitive impairment diagnosis method in which each step is executed by a processor, the method including the steps of: determining N subjects to be diagnosed with cognitive impairment and acquiring personal information data of the subjects; acquiring clinical result data for the subjects; acquiring gait data for the subjects based on gait measurements measured using an inertia measurement unit attached to the subjects' center of body mass; and developing a cognitive impairment diagnosis model including a plurality of predictor variables based on the acquired data set, wherein the cognitive impairment diagnosis model includes walking speed and step time variability included in the gait data as part of the predictor variables.

[0008] Here, obtaining the personal information data means having the subject enter the age, sex, years of education, height, weight, whether or not the subject has lower limb arthritis, and the results of a Mini Mental State Examination (MMSE).

[0009] Here, obtaining the clinical result data means inputting the results of diagnosing cognitive impairment, including mild cognitive impairment and dementia, according to pre-established diagnostic criteria, and classifying subjects who have been diagnosed with a cognitive impairment into a first group and subjects who have not been diagnosed with a cognitive impairment into a second group.

[0010] Obtaining the walking data includes the steps of acquiring signals measured while the subject is walking using the inertial measurement unit attached to the subject's center of gravity, and acquiring the number of steps per minute, walking time, walking speed, stride length and walking time variability based on the acquired signals.

[0011] Here, the step of developing the cognitive impairment discrimination model includes the steps of: dividing the dataset into a development dataset and a validation dataset; developing a cognitive impairment discrimination model including the walking speed and walking time variability as some of the predictor variables by using binary logistic regression analysis on the development dataset with a portion of the walking data and a portion of the personal information data as independent variables; developing a comparison model including a simple mental status test score and age as predictor variables by using binary logistic regression analysis on the development dataset with a portion of the personal information data as independent variables; and generating Receiver Operating Characteristics (ROC) curves for the developed cognitive impairment discrimination model and the comparison model against the development dataset and the validation dataset, and comparing the areas under the curve (AUC) to evaluate the discrimination performance of the cognitive impairment discrimination model.

[0012] The method further includes a step of determining whether or not a subject has a cognitive impairment using the cognitive impairment determination model developed herein.

[0013] Here, the method for determining whether or not the subject has a cognitive impairment includes obtaining the number of years of education of the subject, receiving a tri-axial acceleration signal and a tri-axial angular velocity signal measured from an inertial measurement unit attached to the center of gravity of the subject, obtaining walking speed and walking time variability based on the signals, and substituting the obtained values ​​into the cognitive impairment determination model to determine whether or not the subject has a cognitive impairment.

[0014] A second aspect of the present disclosure provides a computer program product including one or more computer-readable recording media having stored thereon a program for executing a method including the steps of determining N subjects to be diagnosed with cognitive impairment and acquiring personal information data of the subjects, acquiring clinical outcome data for the subjects, acquiring gait data for the subjects based on gait measurements measured using an inertial measurement unit attached to the subjects' center of gravity, developing a cognitive impairment discrimination model including a plurality of predictive variables based on the acquired data set, and determining whether or not the subject has cognitive impairment using the developed cognitive impairment discrimination model, wherein the cognitive impairment discrimination model includes walking speed and walking time variability included in the gait data as part of the predictive variables.

[0015] A third aspect of the present disclosure provides an apparatus including: an input / output unit that receives personal information data and clinical outcome data of a person to be measured and a person to be discriminated, and outputs the presence or absence of cognitive impairment of the person to be discriminated; a network unit that receives gait measurement values ​​measured using an inertial measurement unit attached to the center of gravity of the person to be measured and the person to be discriminated; a processor that controls the input / output unit to acquire personal information data and clinical outcome data of the person to be measured and the person to be discriminated who is to be discriminated for cognitive impairment, acquires gait data based on the received gait measurement values, develops a cognitive impairment discrimination model including a plurality of predictive variables based on the acquired data set, and discriminates the presence or absence of cognitive impairment of the person to be discriminated via the developed cognitive impairment discrimination model; and a memory unit that stores the acquired personal information data, clinical outcome data, and gait data, wherein the cognitive impairment discrimination model includes walking speed and walking time variability included in the walking data as part of the predictive variables.

[0016] Other aspects, features and advantages in addition to those described above will become apparent from the following drawings, claims and detailed description of the invention. Effect of the Invention

[0017] According to the problem-solving means of the present disclosure described above, a method for determining whether or not a cognitive impairment is present is provided, which can easily and quickly select for cognitive impairments such as dementia regardless of the underlying cause by determining the presence or absence of cognitive impairment based on gait characteristic values ​​measured by an accelerometer, and can accurately determine even early-stage cognitive impairment.

[0018] In addition, according to the problem-solving means of the present disclosure, a cognitive impairment diagnosis method is provided that allows a subject to easily and repeatedly perform cognitive impairment diagnosis in their daily lives, and derives accurate cognitive impairment diagnosis results that are not affected by learning effects or social education. [Brief description of the drawings]

[0019] [Figure 1] FIG. 1 is a block diagram of a cognitive impairment detection system according to one embodiment. [Diagram 2] 1 is a flowchart showing a method for developing a cognitive impairment determination model included in a cognitive impairment determination method according to one embodiment. [Diagram 3] 3 is a detailed flowchart showing step 140 of FIG. 2. [Figure 4] 3 is a detailed flow chart showing step 150 of FIG. 2. [Diagram 5] 13 is a table showing the characteristics of a data set of subjects who participated in the development of the cognitive impairment discrimination device 20. [Figure 6] 1 is a table showing characteristics of a cognitive impairment discrimination model according to the present invention. [Figure 7A-7C] The results of performance verification of the cognitive impairment discrimination model are shown. [Figure 8] 1 is a table showing the correlation of weighted composite scores of the developed models. [Figure 9] FIG. 1 is a block diagram showing a cognitive impairment detection device 20 according to one embodiment. [Figure 10] FIG. 13 is a block diagram showing the configuration of a cognitive impairment detection device 20 according to another embodiment. [Figure 11] 1 is a flowchart illustrating a method for determining cognitive impairment according to one embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0020] The present invention discloses various embodiments of a cognitive impairment diagnosis method, an apparatus, and a computer program product, and discloses a cognitive impairment diagnosis method in which each step is executed by a processor, the method including the steps of: determining N subjects to be diagnosed with cognitive impairment and acquiring personal information data of the subjects; acquiring clinical result data for the subjects; acquiring gait data for the subjects based on gait measurements measured using an inertia measurement unit attached to the subject's center of body mass; and developing a cognitive impairment diagnosis model including a plurality of predictor variables based on the acquired data set, the cognitive impairment diagnosis model including walking speed and step time variability included in the gait data as part of the predictor variables.

[0021] Hereinafter, the embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention. However, the present invention may be embodied in various different forms and is not limited to the embodiments described herein. In addition, in the drawings, parts that are not related to the description are omitted in order to clearly explain the present invention, and similar parts are designated by similar reference numerals throughout the specification.

[0022] In the following embodiments, the terms "first", "second" and the like are not used in a limiting sense but are used for the purpose of distinguishing one component from another component.

[0023] In the following embodiments, singular expressions include plural expressions unless the context clearly indicates otherwise.

[0024] In the following embodiments, terms such as "include" or "have" mean the presence of features or components described in the specification, and do not preclude the possibility that one or more other features or components may be added.

[0025] In the following embodiments, when a part such as a film, region, or component is said to be on or above another part, this includes not only the case where it is directly on top of the other part, but also the case where another film, region, component, etc. is interposed between them.

[0026] In the drawings, the size of components may be exaggerated or reduced for convenience of explanation. For example, the shape, size, and thickness of each component shown in the drawings are arbitrarily shown for convenience of explanation, and the present invention is not necessarily limited to those shown.

[0027] In the following embodiments, the x-axis, y-axis, and z-axis are not limited to the three axes on a Cartesian coordinate system, but may be interpreted in a broad sense including them. For example, the x-axis, y-axis, and z-axis may be mutually orthogonal, but may also point in different directions that are not mutually orthogonal.

[0028] When an embodiment can be embodied in other ways, certain process or steps may be performed out of the order described. For example, two processes or steps described in succession may be performed substantially simultaneously or may be performed in the reverse order from that described.

[0029] The present disclosure will now be described in detail with reference to the accompanying drawings.

[0030] FIG. 1 is a block diagram of a cognitive impairment determination system according to one embodiment.

[0031] The cognitive impairment determination system according to one embodiment may include an inertial measurement unit 10 , a cognitive impairment determination device 20 , and a user terminal 30 .

[0032] The inertial measurement unit 10 is a device that measures acceleration and angular velocity, and may be a device that is attached to a subject and measures three-axis acceleration, three-axis angular velocity, etc., during the movement of the subject. That is, the inertial measurement unit 10 may be a device for measuring walking characteristics in space and time.

[0033] The inertial measurement unit 10 can be attached to the center of gravity of the subject. When the inertial measurement unit 10 is attached to the center of gravity, the most accurate and reliable gait characteristics can be measured. The inertial measurement unit 10 can be a three-axis accelerometer attached to the person's waist using a certain adhesive material. The inertial measurement unit 10 can transmit the measured value in the form of a signal to the cognitive impairment discrimination device 20 via a network.

[0034] The cognitive impairment detection device 20 can receive signals from the inertial measurement unit 10, receive data from the user terminal 30, or receive data directly from the user, perform cognitive impairment detection based on the received values, and then provide the result to the user. The cognitive impairment detection device can be composed of a computer device or multiple computer devices that communicate via a network and provide instructions, codes, files, content, services, etc.

[0035] According to an alternative embodiment, the cognitive impairment detection device 20 may be in the form of a program or application downloaded to a user terminal 30 or the like, or may be in the form of a server that provides a cognitive impairment detection method via the web. Hereinafter, devices, servers, terminals, applications, programs, and recording media storing applications / programs that execute the cognitive impairment detection method of the present invention will all be referred to as the cognitive impairment detection device 20.

[0036] The user terminal 30 may be a mobile or non-mobile computing device such as a smartphone, a tablet, a PC, a wearable smart device, or a smart device dedicated to healthcare. The user terminal 30 may include any kind of device that communicates with other devices via a network. The user terminal 30 may receive a result from the cognitive impairment determination device 20 and may also transmit data of the subject to the cognitive impairment determination device 20.

[0037] According to an alternative embodiment of the present invention, the user terminal 30 and the cognitive impairment assessment device 20 may be configured as an integrated unit. In this case, the cognitive impairment assessment device 20 may be installed in the user terminal 30 in the form of an application.

[0038] According to another optional embodiment, the user terminal 30, the cognitive impairment detection device 20, and the inertial measurement unit 10 may be configured as an integrated unit. In this case, the inertial measurement unit 10 is built into the user terminal 30, and the cognitive impairment detection device 20 may be installed in the user terminal 30 in the form of an application.

[0039] According to another alternative embodiment, the user terminal 30 and the inertial measurement unit 10 are configured as an integrated unit, and the cognitive impairment determination device 20 can receive values ​​measured by the user terminal 30.

[0040] According to another alternative embodiment, the system of Fig. 1 may further include a medical service providing server (not shown). The medical service providing server may be a computing device that communicates with other devices via a network. The medical service providing server may receive results from the cognitive impairment assessment device 20 and may transmit data of the subject to the cognitive impairment assessment device 20.

[0041] The network includes a local area network (LAN), a wide area network (WAN), a value added network (VAN), a mobile radio communication network, a satellite communication network, and a combination thereof, and is a comprehensive data communication network that enables the subjects of each network configuration shown in Fig. 1 to communicate with each other smoothly, and may include wired Internet, wireless Internet, and mobile wireless communication networks. In addition, wireless communication may include, but is not limited to, wireless LAN (Wi-Fi), Bluetooth (registered trademark), Bluetooth (registered trademark) low energy, ZigBee (registered trademark), WFD (Wi-Fi Direct), UWB (ultra-wideband), infrared communication (IrDA, infrared Data Association), NFC (Near Field Communication), etc.

[0042] 2 is a flowchart showing a method for developing a cognitive impairment detection model included in a cognitive impairment detection method according to an embodiment. Each step shown in FIG. 2 can be executed by a cognitive impairment detection device, and more specifically, can be executed by a processor included in the cognitive impairment detection device.

[0043] First, the cognitive impairment assessment device 20 determines the names of N subjects. Here, the names of N subjects can be determined according to criteria stored in the cognitive impairment assessment device 20.

[0044] Next, the cognitive impairment determination device 20 acquires personal information data of the determined subject. Here, the personal information data may be directly input to the cognitive impairment determination device 20, or may be received and acquired from the user terminal 30 or a medical service providing server. Here, the personal information data refers to demographic and anthropometric data of the subject. According to one embodiment, the personal information data may include the subject's age, sex, years of education, height, weight, presence or absence of lower limb arthritis, and simple mental state test results (scores). Here, the lower limb arthritis may be, for example, osteoarthritis, rheumatoid arthritis, or post-traumatic arthritis. Meanwhile, the personal information data may further include foot length.

[0045] Next, the cognitive impairment discrimination device 20 acquires clinical result data of the subject. The clinical data result may be directly input to the cognitive impairment discrimination device 20, or may be received and acquired from the user terminal 30 or the medical service providing server. Here, the clinical result data may indicate whether or not the subject has been diagnosed with a cognitive impairment. In the case where a CD (cognitive disorder) group is defined as a group including subjects diagnosed with a cognitive impairment, and a CN (cognitively normal) group is defined as a group including subjects not diagnosed with a cognitive impairment, after diagnosing the cognitive impairment of the subject according to a preset diagnostic criterion, the subject diagnosed with a cognitive impairment can be placed in the CD group, and the subject not diagnosed with a cognitive impairment can be placed in the CN group. As for the clinical result data, the name of the group diagnosed for the subject can be input as the clinical result data.

[0046] Next, the cognitive impairment discrimination device 20 acquires walking data of the subject. The cognitive impairment discrimination device receives the three-axis acceleration signal and the three-axis angular velocity signal measured from the inertial measurement unit 10, and processes them to acquire walking data. The walking data may include the number of steps per minute, walking time, walking speed, stride length, walking time variability, and walking time asymmetry.

[0047] Next, the cognitive impairment discrimination device 20 develops a cognitive impairment discrimination model based on the acquired data set. The cognitive impairment discrimination device 20 develops the cognitive impairment discrimination model through logistic regression analysis, and can evaluate the diagnostic performance of the developed model through comparative models and clinical outcome data.

[0048] Meanwhile, in addition to FIG. 2, the cognitive impairment discrimination device 20 can discriminate the presence or absence of a cognitive impairment in a person to be discriminated using the developed cognitive impairment discrimination model, and then provide the result to a user.

[0049] Each step is explained in more detail below.

[0050] In step 110, N subjects are determined from a population containing a total of M participants.

[0051] First, the device determines the population according to the first criterion. Here, the first criterion may be based on a first study. In particular, a total of M participants in the first study may be set as the population, and N subjects may be determined from the population. Here, the first study may be a nationwide population-based cohort study on elderly Koreans. According to one embodiment, the first study may be a Korean Longitudinal Study on Cognitive Aging and Dementia (KLOSCAD), specifically, the first study may be "Overview of the Korean longitudinal study on cognitive aging and dementia" (JWHan, TH Kim, KP Kwak, K. Kim, BJ Kim, SG Kim, JL Kim, TH Kim, SW Moon, and JY Park, Psychiatry investigation, vol. 15, no. 8, pp. 767, 2018.), but the present invention is not limited thereto, and subjects may be changed according to other studies or criteria. For example, the device can form a population of 6,818 community-dwelling Koreans aged 60 years or older randomly selected from residential registers in 30 regions across the country based on the first study.

[0052] Next, the device determines N subjects from the determined population according to a second criterion. Here, the second criterion can be a criterion for excluding people who have a medical history that may affect their walking characteristics. Specifically, among the participants in the population, participants who have a medical history that may affect their walking characteristics based on their medical history can be determined as N subjects. Specifically, of the M participants in the population, N people who are community-dwelling Koreans aged 60 years or older can be determined as subjects, excluding participants who have a history of stroke, Parkinson's disease or movement disorder, a history of unilateral or bilateral knee or hip arthroplasty, a history of knee, hip or ankle joint surgery, or a history of spinal disease such as spinal stenosis, scoliosis, spondylolysis and spondylolisthesis.

[0053] In step 120, personal information data of the subject is obtained.

[0054] The apparatus can input demographic and anthropometric data of the subjects as personal information data. For example, a skilled researcher can collect the above-mentioned demographic and anthropometric data from each subject and input it as personal information data to the apparatus. Here, the personal information data is the demographic and anthropometric data of the subjects, and can include the age, sex, years of education, height, weight, and presence or absence of lower limb arthritis of each subject. Here, the lower limb arthritis can be, for example, osteoarthritis, rheumatoid arthritis, or post-traumatic arthritis. Furthermore, the personal information data can further include the length of the foot.

[0055] Furthermore, the device can input various neuropsychological test results of the subjects as personal information data. For example, a neuropsychologist or an experienced research nurse can administer a Korean version of a neurocognitive test (Consortium to Establish a Registry for Alzheimer's Disease Assessment Packet, CERAD-K), a neuropsychological assessment battery, a digit span test, a frontal assessment battery, and a simple mental status test to each subject, and then input the results or scores into the device as personal information data.

[0056] In step 130, the device acquires the subject's clinical outcome data.

[0057] The device can input the results of a cognitive impairment diagnosis of the subject based on preset diagnostic criteria as clinical result data.

[0058] For example, a geriatric psychiatrist or geriatrician meets with the subject, obtains a detailed medical history, and performs laboratory tests, physical tests, and neurological tests. A panel of geriatric psychiatrists then diagnoses the subject according to pre-established diagnostic criteria. Here, the diagnostic criteria are as follows: dementia is diagnosed according to the first criterion, mild cognitive impairment is diagnosed according to the second criterion, Alzheimer's disease is diagnosed according to the third criterion, and dementia with Lewy bodies is diagnosed according to the fourth criterion. Here, the first criterion may be the criterion in the Diagnostic and Statistical Manual of Mental Disorders, 4th edition, the second criterion may be the consensus criteria of the International Practice Group on Mild Cognitive Impairment, the third criterion may be the criteria of the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer's Disease and Related Disorders Association, and the fourth criterion may be the consensus criteria proposed in McKeith IG, Dickson D, Lowe J, et al. Diagnosis and management of dementia with Lewy bodies third report of the DLB consortium. Neurology 2005;65:1863-1872. However, the first criterion may be the criterion in the Diagnostic and Statistical Manual of Mental Disorders, 4th edition, and the second criterion may be the consensus criteria proposed in McKeith IG, Dickson D, Lowe J, et al. Diagnosis and management of dementia with Lewy bodies third report of the DLB consortium. Neurology 2005;65:1863-1872. However, the first criterion may be the consensus criteria proposed in McKeith IG, Dickson D, Lowe J, et al. Diagnosis and management of dementia with Lewy bodies third report of the DLB consortium. Neurology 2005;65:1863-1872.

[0059] Next, subjects diagnosed with dementia and mild cognitive impairment can be classified into the CD group, and subjects who function independently in the community and do not show any symptoms of cognitive impairment in an objective neuropsychological evaluation can be classified into the CN group. In this case, obtaining clinical outcome data can mean inputting the results of the subjects being classified into the CD group or the CN group according to the above-mentioned criteria.

[0060] In step 140, the device acquires gait data of the subject.

[0061] The device receives or inputs gait measurements from an inertial measurement unit attached to the subject, and generates gait data based on the input gait measurements.

[0062] Gait data can be generated based on signals measured from an inertial measurement unit and can include steps per minute, walking time, walking speed, stride length, walking time variability, and walking time asymmetry.

[0063] Figure 3 is a detailed flow chart of step 140 of figure 2. Step 140 will now be described in more detail with reference to figure 3.

[0064] Referring to FIG. 3, in step 141, the subject wears the inertial measurement unit 10 and performs gait assessment.

[0065] According to one embodiment, the inertial measurement unit 10 may include a digital 3-axis accelerometer sensor and a gyroscope. For example, the inertial measurement unit may be a FITMETER® from FitLife Inc., which is a hexahedron with soft edges measuring 3.5×3.5×1.3 cm and weighing 14 g, or an ActiGraph® from SMD solution, which is a hexahedron with soft edges measuring 3.0×4.0×1.0 cm and weighing 17 g. This inertial measurement unit may measure 3-axis accelerations of up to ±8 g (resolution 0.004 g / 0.00024 g) and 3-axis angular velocities of up to ±1,000 / s (resolution 0.03 / s) at a sample rate of 250 Hz. However, the present invention is not limited thereto, and a smartphone or smart medical device incorporating a digital accelerometer sensor and a gyroscope may be used as the inertial measurement unit 10.

[0066] With the inertial measurement unit 10 attached to the subject's center of gravity, a gait evaluation of the subject is carried out.

[0067] The inertial measurement unit 10 can be fixed to the third or fourth lumbar vertebrae of the back corresponding to the center of gravity of the subject using a certain adhesive or an elastic band.

[0068] The gait assessment is carried out by having the subject wearing the inertial measurement unit 10 walk in a walking section three times in a relaxed state at a speed selected by the subject. Here, the walking section is a 14m flat straight walkway, and the movement line can be designed to turn after passing the 14m line. However, the gait assessment method and the walking section can be variously modified without being limited to the above.

[0069] In step 142 , the cognitive impairment discrimination device 20 receives the signals measured using the inertial measurement unit 10 .

[0070] The inertial measurement unit 10 transmits signals measured while the subject is walking through the walking section to the cognitive impairment detection device, where the signals include various gait measurements, such as three-axis acceleration and three-axis angular velocity.

[0071] According to one embodiment, the cognitive impairment determination device 20 can pre-process the received signal. Specifically, the cognitive impairment determination device 20 can remove signals corresponding to a walking section of a preset length before and after the acquired signal in order to minimize the acceleration effect of the subject and measure the walking in a normal state. For example, the cognitive impairment determination device 20 can remove signals corresponding to a walking section of 2 m length before the reference point and before the subject turns, and treat only the remaining signals corresponding to the intermediate 10 m walking section as valid signals.

[0072] In step 143, the cognitive impairment detection device 20 generates gait data based on the acquired signals.

[0073] The cognitive impairment discrimination device 20 processes the signal and identifies each step. The cognitive impairment discrimination device 20 acquires gait data based on the signal acquired from the inertial measurement unit 10 and the previously acquired personal information data of the subject. Here, the gait data may include the number of steps per minute, walking time, walking speed, stride length, walking time variability, and walking time asymmetry.

[0074] The number of steps per minute means the number of steps per minute, that is, the walking rate, and may be expressed in units of steps / min. The number of steps per minute may be calculated by calculating how many steps are included in one minute when the section between peaks of the acceleration signal is defined as walking after the device receives the acceleration signal of the subject for a certain period of time.

[0075] Walking time is in seconds and is defined as the duration of each step from first foot contact on one side to first foot contact on the other side. Walking time is estimated as the inverse credence of steps per minute multiplied by 60.

[0076] Walking speed is defined as the distance traveled by the body on flat ground in cm / s, and is defined as cm per second. Walking speed can be derived from a mathematical linear regression equation including roll angle, yaw angle, and weight as walking factors. For example, from Equation 1, it can be seen that the linear regression equation of the first model includes age, steps per minute, vertical height displacement (VHD), foot length, weight, roll angle, and yaw angle as walking factors for estimating walking speed, and each walking factor can have a corresponding coefficient. Here, the vertical height displacement is the average difference between the maximum and minimum vertical heights of the center of gravity within one walking cycle, and can be calculated from data of an inertial measurement unit. The foot length is the average length of both feet, and can be obtained from personal information data. The roll angle and yaw angle of the center of gravity can reflect the angular motion of the trunk and limbs. The roll angle and yaw angle are calculated from data of an inertial measurement unit, and the values ​​can be re-oriented to Cartesian coordinates.

[0077]

number

[0078] However, the present invention is not limited to deriving the walking speed using Equation 1, and the walking speed can also be derived using various models described in Patent Application No. 10-2021-0194313 filed by the applicant of the present invention.

[0079] The stride length is expressed in cm and can be defined as the distance between the heel of the previous foot and the heel of the current foot. The stride length can be calculated by multiplying the walking speed by 60 and then dividing it by the number of steps per minute, as shown in Equation 2 below.

[0080]

number

[0081] The unit of the walking time variability is %, and in this specification, the walking time variability may be expressed as the coefficient of variance step time (CV step time). The walking time variability may be defined as an index showing how constant the walking cycles of the left and right legs are maintained. The coefficient of variance step time may be derived as shown in the following Equation 3. The coefficient of variance step time may be derived using a model described in Patent Application No. 10-2016-0124602 (Method of detecting walking factors and walking events based on a three-axis accelerometer) filed by the applicant of the present invention. In detail, the coefficient of variance step time for left leg walking is calculated using the standard deviation of the walking time and the mean of the walking time for left leg walking, and the coefficient of variance step time for right leg walking is calculated using the standard deviation of the walking time and the mean of the walking time for right leg walking, and the sum of the coefficients of variance step time for left and right leg walking is divided by 2 to derive the final coefficient of variance step time.

[0082]

number

[0083] Meanwhile, in this specification, the gait time variability may be the same value as the temporal gait variability. In general, gait variability can include temporal gait variability and spatial gait variability. Among them, temporal gait variability is more sensitive to early abnormal gait variability and more reliable than spatial gait variability. Therefore, a reliable cognitive impairment discrimination model can be developed by using the gait time variability corresponding to temporal gait variability as an independent variable of the cognitive impairment discrimination model.

[0084] Gait time asymmetry can be defined as the absolute difference, in seconds (sec) or milliseconds (ms), between the mean gait time for the left leg step and the mean gait time for the right leg step.

[0085] Returning again to FIG. 2, in step 150, a cognitive impairment discrimination model is developed based on the obtained data set.

[0086] The cognitive impairment discrimination model can be a gait-based algorithm and can include multiple predictor variables. Specifically, the cognitive impairment discrimination model can include walking speed and walking time variability as part of the predictor variables.

[0087] Figure 4 is a detailed flow chart of step 150 of Figure 2. In the following, the method for developing a cognitive impairment discrimination model will be described in more detail with reference to Figure 4.

[0088] Referring to FIG. 4, in step 151, the cognitive impairment detection device 20 divides all acquired data sets into a development data set and a validation data set.

[0089] Here, the data set refers to the entire data including the personal information data, clinical result data, and gait data of each subject.

[0090] The cognitive impairment discrimination device 20 can randomly divide the data set. However, it is not limited to this, and the data set of subjects in the CD group and the data set of subjects in the CN group can be divided so that the same proportions of the data set of subjects in the CD group and the data set of subjects in the CN group are distributed in both sets. For example, the data set of subjects in the CD group is about 17% of the data set of subjects in both the development data set and the validation data set. Here, the CD group and the CN group are obtained by classifying subjects who are owners of the data set, and who have been diagnosed with dementia, mild cognitive impairment, Alzheimer's disease, or dementia due to Lewy bodies according to diagnostic criteria into the CD group, and subjects who function independently in the community and have not shown any symptoms of cognitive impairment in an objective neuropsychological evaluation into the CN group. Information on the CD group and the CN group can be obtained from clinical result data acquired in advance.

[0091] On the other hand, the development dataset can be split to include a larger amount of data than the validation dataset, for example 80% of the entire dataset for the development dataset and 20% for the validation dataset.

[0092] FIG. 5 is a table showing the characteristics of the data set of subjects who participated in the development of the cognitive impairment discrimination device 20.

[0093] 5, CN is a dataset of subjects included in the CN group, CD is a dataset of subjects included in the CD group, All is the entire dataset, Development dataset is the development dataset, and Validation dataset is the validation dataset.

[0094] Among the values ​​in the table in Figure 5, continuous variables are expressed as averages (standard deviations) and categorical variables are expressed as percentages. Meanwhile, continuous variables were compared using Student's t-test, and categorical variables were compared using chi-squared test to obtain p-value (p*).

[0095] 5, of the total 595 subjects, 101 (17%) were included in the CD group, and the remaining 494 were included in the CN group. In detail, according to the obtained clinical outcome data, of the 101 subjects in the CD group, 94 were diagnosed with mild cognitive impairment, 6 with dementia due to Alzheimer's disease, and 1 with dementia due to Lewy bodies.

[0096] The personal information data of subjects in the CD group shows that they are older, have fewer years of education, are shorter, and have lower scores on the simple mental status test compared to the CN group. Also, the walking data of subjects in the CD group shows that they have lower steps per minute, longer walking time, greater walking time variability, greater walking time asymmetry, slower walking speed, and shorter stride length compared to the CN group.

[0097] In step 152, the cognitive impairment discrimination device 20 develops a cognitive impairment discrimination model using logistic regression analysis on the development data set.

[0098] The cognitive impairment discrimination model can be a regression model for discriminating the presence or absence of cognitive impairment in a subject using the measured values ​​of the inertial measurement unit 10. Here, discriminating the presence or absence of cognitive impairment means estimating cases in which the subject is likely to be diagnosed with dementia and mild cognitive impairment according to the above-mentioned diagnostic criteria.

[0099] According to one embodiment, a binary logistic regression model with forward selection is used to develop a regression-based model for discriminating cognitive impairment.

[0100] The binomial logistic regression analysis model is a model that inputs multiple independent variables and determines which of the input independent variables is superior in discriminating the dependent variable. The forward selection method is a method in which the input independent variable with the best explanatory power is added, and if the input independent variable meets the statistical significance level, the model sequentially adopts that independent variable.

[0101] Here, the dependent variable refers to the presence or absence of cognitive impairment, and a binary result of "with or without cognitive impairment" is derived.

[0102] Here, the independent variables used are walking time variability, walking speed, age, sex, years of education, height, and the presence or absence of lower limb arthritis. The cognitive impairment discrimination device 20 stores personal information data and walking data of the subject. The personal information data includes age, sex, years of education, height, weight, the presence or absence of lower limb arthritis, and the results of a simple mental state test, and the walking data includes the number of steps per minute, walking time, walking speed, stride length, walking time variability, and walking time asymmetry.

[0103] The cognitive impairment detection device 20 can use walking time variability, walking speed, age, sex, years of education, height, and presence or absence of lower limb arthritis selected according to a criterion from among these variables as independent variables. The criterion for selecting the independent variables can be determined by determining whether the variables have a large influence on whether or not cognitive impairment is detected.

[0104] For example, age, gender, and years of education have been confirmed as variables that affect the diagnosis of cognitive impairment in various existing studies.

[0105] Gait time variability and walking speed are also important factors in distinguishing cognitive impairment. For more information, Blumen HM, Brown LL, Habeck C, et al. Gray matter volume covariance patterns associated with gait speed in older adults: a multi-cohort MRI study. Brain imaging and behavior 2019;13:446-460. According to existing neuroimaging studies, walking speed is related to the volume of the brainstem, cerebellum, primary and supplementary motor areas, and frontal cortical areas involved in information processing speed and executive function. On the other hand, existing neuroimaging studies such as Tian Q, Chastan N, Bair WN, Resnick SM, Ferrucci L, Studenski SA. The brain map of gait variability in aging, cognitive impairment, and dementia-a systematic review. Neuroscience & Biobehavioral Reviews 2017;74:149-162. have shown that gait variability is associated with the volume of the primary sensorimotor cortex, hippocampus, anterior cingulate cortex, basal ganglia, and posterior thalamic radiation, which are involved in overall gait, memory, and executive function.

[0106] Height is highly correlated with walking time variability and walking speed, and the presence or absence of lower limb arthritis is also highly correlated with walking time variability and walking speed.

[0107] The apparatus inputs the independent variables selected according to these criteria into a binary logistic regression model to which a forward selection method is applied, and then selects the independent variables that have a large influence on the derivation of the dependent variable as predictor variables, and develops a cognitive impairment discrimination model including the selected predictor variables. According to one embodiment of the present invention, the developed cognitive impairment discrimination model may include walking speed, walking time variability, and years of education as significant predictors. In other words, the cognitive impairment discrimination model according to one embodiment of the present invention is a model that includes walking measurements as predictor variables when determining whether or not a subject has cognitive impairment.

[0108] Equation 4 is the mathematical linear regression equation of the developed cognitive impairment discrimination model. Gait speed is the walking speed, which is extracted from the walking data, step time variability is the walking time variability, which is also extracted from the walking data. Education is the number of years of education, which is extracted from the personal information data.

[0109]

number

[0110] FIG. 6 is a table showing characteristics of the cognitive impairment discrimination model according to the present invention.

[0111] Referring to Equation 4 and FIG. 6, the cognitive impairment discrimination model includes predictor variables and regression coefficients (B).

[0112] The predictor variables are walking speed, walking time variability (shown as gait variability in Table 6) and years of education, and each predictor variable may have its own regression coefficient (B in FIG. 6). For example, the regression coefficient of walking speed is -0.040, the regression coefficient of walking time variability is 0.535, and the regression coefficient of years of education is -0.115. Here, each regression coefficient means a relative measure of discriminatory power for the predictor variable, and can represent the power of how accurately the predictor variable included in the model discriminates the dependent variable. The larger the regression coefficient, the greater the influence the corresponding predictor variable has on the discrimination of the dependent variable, and among the predictor variables included in the cognitive impairment discrimination model of the present invention, walking time variability corresponds to the variable with the greatest explanatory power. Meanwhile, 2.286 in Equation 4 corresponds to the regression constant (constant in FIG. 6).

[0113] As shown in the p-values ​​(p) in Figure 6, walking speed, walking time variability, and years of education have low p-values ​​(p<0.05), which confirms that they are statistically significant.

[0114] The OR in Figure 6 means Odds ratio, and indicates the likelihood of the dependent variable changing when the independent variables entered into the model change. For example, in the case of walking speed, a decrease of 1 cm / s increases the risk of being diagnosed with cognitive impairment by 0.961 times. 95% CI means confidence interval, and indicates the confidence interval of the OR.

[0115] In the cognitive impairment discrimination model expressed by Equation 4, an optimal cutoff for discriminating cognitive impairment can be derived at >-1.56. Here, the optimal cutoff corresponds to a standard score for discriminating cognitive impairment. That is, when the walking speed, walking time variability, and years of education of a subject are inputted into the cognitive impairment discrimination model, a real value is derived, and if the real value is greater than -1.56, the subject can be discriminated as having a cognitive impairment. If the real value derived from the model is equal to or less than -1.56, the subject can be discriminated as not having a cognitive impairment.

[0116] For reference, the optimal cut points can be derived through bootstrap sampling, showing a sensitivity of 72.0±6.6% and a specificity of 74.4±5.9%.

[0117] In step 153, the cognitive impairment discrimination device 20 develops a comparison model using logistic regression analysis on the development data set.

[0118] The comparison model may be a regression model for determining cognitive impairment of a subject using a simple mental status test score instead of the measurement value of the inertial measurement unit. The comparison model may be derived to verify the cognitive impairment determination model. A method for deriving the comparison model is similar to the method for deriving the cognitive impairment determination model described above, and therefore a duplicated description will be omitted.

[0119] The SME score and age can be used as independent variables for deriving the comparison model. The comparison model derived in this way can include the SME score and age as important predictor variables. In other words, the comparison model is a model that includes the SME score as a predictor variable when determining whether or not a subject has a cognitive disorder.

[0120] Equation 5 is the mathematical linear regression equation of the comparison model. MMSE score is the Mini-Mental Status Examination score, which is extracted from the personal information data, and age is the age, which is also extracted from the personal information data. For reference, the optimal cut point for the comparison model to discriminate cognitive impairment was derived to be >-1.16, with a sensitivity of 56.7±3.2% and a specificity of 88.3±1.4%.

[0121]

number

[0122] In step 154, the cognitive impairment discrimination device 20 validates the developed cognitive impairment discrimination model.

[0123] The cognitive impairment discrimination model can be validated based on statistical criteria, can be validated by a comparison model, or can be validated through a validation dataset.

[0124] First, the cognitive impairment discriminator 20 can test the goodness of fit of the model based on statistical criteria.

[0125] Specifically, the cognitive impairment discrimination device 20 can evaluate the statistical goodness of fit of the model using the Hosmer-Lemeshow test with p>0.05 and the algorithm explanatory power (Nagelkerke R2 index). The regression coefficients can be interpreted as the relative weights of the independent variables to the dependent variable. Then, a weighted composite score can be calculated using a logit score to which the weights of the regression coefficients of the logistic regression model are applied.

[0126] In addition, the cognitive impairment discrimination device 20 can verify the performance of the cognitive impairment discrimination model through AUC. Here, performance refers to cognitive impairment discrimination performance or diagnostic performance, and refers to how accurately the presence or absence of a cognitive impairment can be discriminated.

[0127] As an embodiment, the cognitive impairment discrimination device 20 can verify the performance by deriving and comparing the AUC of the cognitive impairment discrimination model and the comparison model. Specifically, the cognitive impairment discrimination device can verify the performance of the cognitive impairment discrimination model by deriving and comparing the ROC curves of both models for the development dataset and the validation dataset. More specifically, the cognitive impairment discrimination device can generate ROC curves for each of the two models and evaluate the discrimination performance or diagnostic performance of the cognitive impairment discrimination model by comparing the AUCs of both models using the method proposed in Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983;148:839-843.

[0128] In one embodiment, the cognitive impairment discrimination device 20 can verify the performance by deriving and comparing the AUC of both models for the development dataset. First, the cognitive impairment discrimination device can estimate the optimal cut point, sensitivity, and specificity of the weighted composite score using a bootstrap sampling estimation method. In this case, 80% of the development dataset can be randomly selected in each run. The cognitive impairment discrimination device can obtain bootstrap estimates of the AUC, the optimal cut point, the sensitivity, and the specificity by combining the results of 100 runs. Next, validation can be performed by deriving and comparing the bootstrap estimated average AUC between the cognitive impairment discrimination model and the comparison model.

[0129] In another embodiment, the cognitive impairment discrimination device 20 can compare the AUC for the validation data set with the AUC for the development data set. Specifically, the cognitive impairment discrimination device can compare the AUC of the cognitive impairment discrimination model obtained from the validation data set with the AUC of the cognitive impairment discrimination model obtained from the development data set to check the possibility of overfitting.

[0130] Furthermore, the cognitive impairment discrimination device 20 can perform validation by comparing the AUCs of both models obtained from the validation data set.

[0131] 7A to 7C show the performance verification results of the cognitive impairment discrimination model. In the following, the performance verification results of the cognitive impairment discrimination model will be reviewed with reference to FIGS. 7A to 7C. The performance verification results described below are for reference only, and the specific result values ​​may vary depending on conditions such as the composition of the data set and the size of the variables. Meanwhile, the statistical significance used in this specification may be set to p-value p<0.05 in a two-sided test.

[0132] First, the Gait-based model, which includes the algorithm of the gait database, shows a good statistical fit. In detail, the results of the Hosmer-Lemeshow test, χ 2 = 4.16, df = 8, p = 0.843, and algorithm explanatory power = 0.243, showing appropriate statistical fitness. In addition, the algorithm classification accuracy (%) was calculated to be 83.8%, which was also highly accurate.

[0133] For reference, the comparison model including the Mini-Mental Status Examination score-based algorithm (MMSE-based model) also shows a good statistical fit. In detail, the results of the Hosmer-LeMeshow test χ 2 = 12.99, df = 8, p = 0.112, and algorithm explanatory power = 0.271, which show a good statistical fit. In addition, the algorithm discrimination accuracy was calculated to be 86.1%, which is also highly accurate.

[0134] On the other hand, as shown in Figures 7A to 7C, the cognitive impairment discrimination model shows a high AUC value, and it can be confirmed that the discrimination performance or diagnostic performance is excellent (good). AUC is a value that evaluates the performance of a model, and an AUC value of 0.7-0.8 is judged as fair, 0.8-0.9 is good, and 0.9-1.0 is judged as excellent.

[0135] Referring to FIG. 7A, the results of calculating the AUC of the cognitive impairment discrimination model in the development dataset and the AUC of the comparative model are shown. Specifically, the cognitive impairment discrimination model has an AUC of 0.788 in the development dataset, with a standard error [SE] of 0.027, a 95% CI of 0.748-0.823, and a p-value of p<0.001. The comparative model has an AUC of 0.762 in the development dataset, with a standard error [SE] of 0.033, a 95% CI of 0.721-0.799, and a p-value of p<0.001. As such, since the AUC of the cognitive impairment discrimination model (Gait-based model) including the gait database algorithm is larger than the AUC of the comparative model (MMSE-based model) including the simplified mental status examination score-based algorithm, it can be confirmed that the cognitive impairment discrimination model has superior performance.

[0136] Next, 100 samples were extracted from the development dataset by bootstrap sampling, and the AUCs of the two models were compared. The average bootstrap estimated AUC of the cognitive impairment discrimination model was 0.787 ± 0.015, and the average bootstrap estimated AUC of the comparison model was 0.760 ± 0.016. In this case, too, it can be confirmed that the AUC of the cognitive impairment discrimination model is larger than that of the comparison model.

[0137] Furthermore, when the cognitive impairment discrimination model is applied to an independent validation data set, the discrimination or diagnostic performance is evaluated as good.Referring to Figure 7B, in this case, the AUC of the cognitive impairment model is 0.811, the SE is 0.046, the 95% CI is 0.729-0.877, and the p-value is p<0.001.It can be confirmed that the AUC of the cognitive impairment discrimination model is greater than the AUC of the comparison model.

[0138] In addition, a high AUC can be confirmed when the cognitive impairment discrimination model is applied to both the validation data set and the development data set. Referring to FIG. 7C, when the cognitive impairment discrimination model is applied to the development data set, the AUC is confirmed to be 0.788, and when the cognitive impairment discrimination model is applied to the validation data set, the AUC is confirmed to be 0.811. That is, it can be confirmed that the cognitive impairment discrimination model shows similarly high discrimination or diagnostic performance in both data sets.

[0139] Meanwhile, according to an alternative embodiment, the cognitive impairment identification device can use Pearson's correlation analysis to confirm that the models are correlated with each other and have similar performance in identifying cognitive impairment, although there are slight differences. The correlation analysis can be implemented by analyzing the correlation between the cognitive impairment identification model, the comparison model, and even the weighted composite score of the simple mental status test score using Pearson's correlation analysis on various data sets.

[0140] FIG. 8 is a table showing the correlations of the weighted composite scores of the developed models.

[0141] 8, it can be seen that the correlation between the weighted composite scores for the CD group and the CN group for all subjects in the cognitive impairment discrimination model and the comparison model is very significant. It can also be seen that the correlation between the weighted composite scores for the CD group and the CN group for subjects in the cognitive impairment discrimination model and the simple mental status test score is very significant.

[0142] In summary, the cognitive impairment identification device according to an embodiment includes a model for identifying subjects exhibiting cognitive impairment. The model can be developed and verified through gait characteristics obtained based on measurements taken via a wearable inertial sensor attached to the center of gravity of the subject. As can be seen from the verification results, it can be confirmed that the cognitive impairment identification device exhibits high levels of identification or diagnostic performance. In particular, when comparing the cognitive impairment identification model with a comparison model using a simplified mental status test score, which is widely used in clinical and research to identify cognitive impairment, it was confirmed that the cognitive impairment identification model exhibited similar or superior performance.

[0143] According to one embodiment, the cognitive impairment discrimination model includes walking speed and walking time variability as main predictor variables, which is effective in discriminating cognitive impairment of any type or cause.

[0144] Specifically, according to existing neuroimaging studies, walking speed is related to the volume of the brainstem, cerebellum, primary and supplementary motor areas, and frontal cortical regions involved in information processing speed and executive function. In contrast, gait variability is related to the volume of the primary somatosensory cortex, hippocampus, anterior cingulate cortex, basal ganglia, and posterior thalamic radiation, which are involved in overall gait, memory, and executive function. Therefore, the cognitive impairment discrimination model of the present invention is characterized by being able to successfully discriminate cognitive impairment of all causes by including walking speed and walking time variability as major variables. Alzheimer's disease is the most common type of dementia, accounting for 60-80% of dementia cases, followed by vascular dementia and Lewy body dementia / Parkinson's disease dementia, which account for 20%. Meanwhile, 46% of Alzheimer's disease patients have vascular pathology. Therefore, the cognitive impairment discrimination method according to an embodiment of the present invention can be usefully used to screen for dementia regardless of the underlying cause.

[0145] The RMS test is widely used in clinical and research to identify cognitive disorders, but it has several drawbacks. The RMS test cannot be administered by the individual to be assessed, and there is a learning effect when administered repeatedly, which can easily affect social education. In addition, the RMS test is not suitable for screening for early-stage cognitive disorders.

[0146] However, the cognitive impairment determination method according to an embodiment can be implemented by being installed in a user's terminal as an application, and has an advantage that it can be easily and repeatedly used by individuals in daily life because the person to be determined wears a wearable inertial sensor and analyzes the gait through a signal measured by the person to be determined. Furthermore, the cognitive impairment determination method is not affected by learning effects and is free from social and educational influences. The cognitive impairment determination method is characterized in that it can be used not only to determine cognitive impairment but also to continuously monitor the person to be determined by being used in the real life of the person to be determined.

[0147] FIG. 9 is a block diagram showing a cognitive impairment detection device 20 according to one embodiment.

[0148] The cognitive impairment detection device 20 may include a network unit 21, a storage unit 22, a processor 23, and an input / output unit 24. Fig. 9 shows only components related to one embodiment, and the cognitive impairment detection device may further include other general-purpose components in addition to the components shown in Fig. 9.

[0149] The network unit 21 may include one or more components that enable wired / wireless communication with other nodes. For example, the network unit 21 may include at least one of a short-range communication unit (not shown), a mobile communication unit (not shown), and a broadcast receiving unit (not shown).

[0150] The network unit 21 can communicate with the inertial measurement unit 10 to receive gait measurement values, communicate with the user terminal 30 to receive personal information data, and communicate with a medical service providing server to receive determination information and clinical result data for N subjects.

[0151] The memory unit 22 is hardware that stores various data processed within the cognitive impairment discrimination device, and can store programs for processing and control of the processor 23.

[0152] The storage unit 22 can store the personal information data acquired in step 120 of FIG. 2, the clinical result data acquired in step 130, and the walking data acquired in step 140.

[0153] The storage unit 22 can also store various programs for developing the cognitive impairment discrimination device in step 150 of Fig. 2 and various programs used to verify the developed cognitive impairment discrimination device. For example, the programs can be executed using R version 4.2.0 (Foundation for Statistical Computing, Vienna, Austria), the Statistical Package for the Social Sciences for Windows version 20 (International Business Machines Corporation, Armonk, NY), and MedCalc for Windows version 18.11.3 (MedCalc Software, Mariakerke, Belgium), and the like, but the present invention is not limited thereto.

[0154] The memory unit 22 may include random access memory (RAM), such as dynamic random access memory (DRAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), CD-ROM, Blu-ray or other optical disk storage, hard disk drive (HDD), solid state drive (SSD), or flash memory.

[0155] The input / output unit 24 may include an input unit including a keypad, a touch screen, etc., and an output unit including a display, a speaker, etc., and the input / output unit 24 may be embodied as a touch display.

[0156] The input / output unit 24 can directly input the personal information data in step 120 and the clinical result data in step 130 in Fig. 2 from the user. In addition, the input / output unit 24 can provide the user with the cognitive impairment discrimination result by the cognitive impairment discrimination device 20 or the discrimination result that is a dependent variable of the cognitive impairment discrimination model.

[0157] The processor 23 controls the overall operation of the cognitive impairment discrimination device 20. For example, the processor 23 executes a program stored in the memory unit 22, thereby generally controlling the input / output unit 24, the network unit 21, the memory unit 22, etc.

[0158] The processor 23 can execute each step of Figures 2, 3 and 4, and in detail can control the network unit 21 to receive measurement values, information and data, generate walking data from the measurement values, and control the memory unit 22 to store the acquired data set, and can control the configuration unit to execute a program stored in the memory unit 22 to develop a cognitive impairment discrimination model using the data set, verify this, and use the developed cognitive impairment discrimination model to determine the presence or absence of cognitive impairment and provide the result to the user.

[0159] The processor 23 may be implemented using at least one of ASICs (application specific integrated circuits), DSPs (digital signal processors), DSPDs (digital signal processing devices), PLDs (programmable logic devices), FPGAs (field programmable gate arrays), controllers, micro-controllers, microprocessors, and other electrical units for performing functions.

[0160] Fig. 10 is a block diagram showing a configuration of a cognitive impairment detection device 20 according to another embodiment. Fig. 11 is a flowchart showing a cognitive impairment detection method according to one embodiment. The cognitive impairment detection method using the cognitive impairment detection device 20 will be specifically described with reference to Figs. 10 and 11.

[0161] The cognitive impairment discrimination device 20 in FIG. 10 may include a network unit 21, a personal information data storage unit 221, a walking data storage and generation unit 231, a cognitive impairment discrimination unit 232, a discrimination result storage unit 2220, and an input / output unit 24.

[0162] In step 201, the cognitive impairment discrimination device 20 asks for the input of personal information data of the person to be discriminated. The person to be discriminated can input the personal information data via the input / output unit 24, and the input personal information data can be stored in the personal information data storage unit 221. Here, the personal information data includes years of education, and can further include age, weight, leg length, etc.

[0163] In step 202, the cognitive impairment discrimination device 20 receives a signal from the inertial measurement unit 10 attached to the person to be discriminated. The person to be discriminated attaches the inertial measurement unit 10 to the center of gravity and performs a gait evaluation. The inertial measurement unit 10 transmits a signal measured during walking to the cognitive impairment discrimination device 20. The cognitive impairment discrimination device 20 stores the signal received via the network unit 21 in the walking data storage and generation unit 231. Here, the signal may be a three-axis acceleration or three-axis angular velocity measurement value measured during walking.

[0164] In step 203, the cognitive impairment detection device 20 generates walking data from the stored signals. Here, the walking data can include walking speed and walking time variability. Specifically, the walking data storage and generation unit 231 can calculate walking speed based on the signals and the stored personal information data, and calculate walking time variability based on the stored signals.

[0165] In step 204, the cognitive impairment discrimination device 20 can discriminate the presence or absence of cognitive impairment by substituting the personal information data and the walking data into the cognitive impairment discrimination model. The cognitive impairment discrimination unit 232 includes the previously developed cognitive impairment discrimination model. The cognitive impairment discrimination model is developed by a binomial logistic regression method to which a forward selection method is applied, and has the form of a linear regression equation, and includes walking speed, walking time variability, and years of education as predictor variables, and each predictor variable has a preset regression constant, and also includes a preset regression constant. The cognitive impairment discrimination model is a linear regression equation and has the following Equation 6. The optimal cutoff point at which the linear regression equation of Equation 6 becomes a standard score for discriminating the presence or absence of cognitive impairment may be >-1.56.

[0166]

number

[0167] The cognitive impairment discrimination unit 232 inputs the walking speed, walking time variability, and years of education of the person to be discriminated into the model as inputs, and obtains a score expressed in real numbers as an output. If the obtained score is greater than the optimal cutoff point, the cognitive impairment discrimination unit 232 can conclude that the person to be discriminated has a cognitive impairment. If the obtained score is equal to or less than the optimal cutoff point, the cognitive impairment discrimination unit 232 can conclude that the person to be discriminated does not have a cognitive impairment.

[0168] In step 205, the cognitive impairment discrimination device 20 provides the discrimination result to the user. Here, the user may be the person to be discriminated, or may be a medical staff member or a guardian. The cognitive impairment discrimination device 20 can directly provide the discrimination result to the user via the input / output unit 24. However, the present invention is not limited to this, and the discrimination result can also be provided to the guardian or medical staff member by transmitting the discrimination result to another user terminal or a medical service providing server via the network unit 21.

[0169] Meanwhile, according to another embodiment of the present invention, the steps performed by the processor may be embodied as software (e.g., a program) including one or more instructions stored in a storage medium readable by a machine.

[0170] For example, a processor of a device can call up at least one instruction word from the storage medium and execute it, thereby enabling the device to operate to perform at least one function according to the at least one instruction word called up. The at least one instruction word may include a code generated by a compiler or a code that can be executed by an interpreter. The device-readable storage medium may be provided in the form of a non-transitory storage medium. Here, "non-transitory" only means that the storage medium is a tangible device and does not include a signal (e.g., electromagnetic wave), and this term does not distinguish between a case where data is semi-permanently stored in the storage medium and a case where data is temporarily stored in the storage medium.

[0171] Meanwhile, according to another embodiment of the present invention, each step executed by the processor may be provided in a computer program product.

[0172] The computer program product can be traded between a seller and a buyer as a commodity. The computer program product can be distributed in the form of a device-readable storage medium (e.g., compact disc read only memory (CD-ROM)) or distributed online (e.g., downloaded or uploaded) via an application store (e.g., Play Store™) or directly between two user devices. In the case of online distribution, at least a part of the computer program product can be at least temporarily stored or temporarily generated in a device-readable storage medium such as the memory of a manufacturer's server, an application store's server, or an intermediary server.

[0173] Furthermore, in this specification, a "unit" may be a hardware component, such as a processor or a circuit, and / or a software component executed by a hardware component, such as a processor.

[0174] The scope of the present embodiments is indicated by the claims below rather than the above detailed description, and should be interpreted to include all modifications and variations derived from the meaning and scope of the claims and their equivalent concepts.

Claims

1. A method for determining cognitive impairment, each step of which is executed by a processor, comprising: A step of determining N subjects to be assessed for cognitive impairment and acquiring personal information data of the subjects; acquiring clinical outcome data for the subject; acquiring gait data based on gait measurements measured by an inertial measurement unit attached to the subject's center of gravity; developing a cognitive impairment discrimination model including a plurality of predictor variables based on the acquired data set; Including, A method in which the cognitive impairment discrimination model includes walking speed and walking time variability contained in the walking data as part of predictive variables.

2. The acquisition of said personal information data is The method according to claim 1, wherein the subject's age, sex, years of education, height, weight, presence or absence of lower limb arthritis, and simple mental state test results are input and acquired.

3. Obtaining the clinical outcome data includes: The method according to claim 1, comprising having a subject enter the results of diagnosing cognitive impairment, including mild cognitive impairment and dementia, according to preset diagnostic criteria, classifying subjects diagnosed with cognitive impairment into a first group and subjects not diagnosed with cognitive impairment into a second group.

4. The acquiring of the gait data includes: acquiring a signal measured while the subject is walking using the inertial measurement unit attached to the subject's center of gravity; obtaining steps per minute, walking time, walking speed, stride length, and walking time variability based on the obtained signals; The method of claim 1 , comprising:

5. The step of developing the cognitive impairment discrimination model includes: splitting the dataset into a development dataset and a validation dataset; A step of developing a cognitive impairment discrimination model including the walking speed and walking time variability as part of predictive variables by using a binary logistic regression analysis on the development dataset with a part of the walking data and a part of the personal information data as independent variables; A step of developing a comparison model including a simple mental status test score and age as predictor variables by performing a binary logistic regression analysis on the development dataset using a part of the personal information data as an independent variable; A step of generating ROC curves for the developed cognitive impairment discrimination model and the comparison model against the development dataset and the validation dataset, and comparing AUCs to evaluate the discrimination performance of the cognitive impairment discrimination model; The method of claim 1 , comprising:

6. A step of determining whether or not a subject has a cognitive impairment using the developed cognitive impairment determination model. The method of claim 1 further comprising:

7. The method for determining whether or not a subject has a cognitive impairment comprises: Obtain the number of years of education of the person to be discriminated; receiving a three-axis acceleration signal and a three-axis angular velocity signal measured by an inertial measurement device attached to the center of gravity of the person to be discriminated, and acquiring a walking speed and a walking time variability based on the three-axis acceleration signal and a three-axis angular velocity signal; The method according to claim 6 , further comprising substituting the acquired value into the cognitive impairment discrimination model to discriminate whether or not a cognitive impairment is present.

8. A step of determining N subjects to be assessed for cognitive impairment and acquiring personal information data of the subjects; acquiring clinical outcome data for the subject; acquiring gait data based on gait measurements measured by an inertial measurement unit attached to the subject's center of gravity; developing a cognitive impairment discrimination model including a plurality of predictor variables based on the acquired data set; A step of determining whether or not a person to be determined has a cognitive impairment using the developed cognitive impairment determination model; Including, A computer program product comprising one or more computer-readable recording media having stored thereon a program for executing a method in which the cognitive impairment discrimination model includes walking speed and walking time variability contained in the walking data as part of predictive variables.

9. An input / output unit that receives personal information data and clinical result data of the subject and the discrimination target, and outputs whether or not the subject has a cognitive impairment as determined by the input / output unit; A network unit that receives gait measurement values ​​measured using an inertial measurement unit attached to the center of gravity of the subject and the discrimination target; a processor that controls an input / output unit to acquire personal information data and clinical result data of subjects who are to be diagnosed with cognitive impairment and subjects to be diagnosed, acquires gait data based on the received gait measurement values, develops a cognitive impairment discrimination model including a plurality of predictive variables based on the acquired data set, and discriminates whether or not the subjects to be diagnosed have a cognitive impairment via the developed cognitive impairment discrimination model; A storage unit that stores the acquired personal information data, clinical result data, and walking data; Including, The apparatus, wherein the cognitive impairment discrimination model includes walking speed and walking time variability contained in the walking data as part of predictive variables.

Images