Apparatus and method for determining mild cognitive impairment by using key input

Abstract

The present invention provides an apparatus for determining mild cognitive impairment, the apparatus comprising: a key input unit that generates keystroke data according to a user's key input; a main storage unit storing an MCI determination unit and an MCI DB that stores the keystroke data and reference data serving as MCI determination criteria; and a calculation unit that executes the MCI determination unit, wherein the MCI determination unit is configured to store and collect the keystroke data in the MCI DB, and when the keystroke data is collected in an amount greater than or equal to a reference capacity, compare the collected keystroke data with the reference data serving as the MCI determination criteria to determine whether MCI is present.

Description

Device and method for determining mild cognitive impairment using key input

[0001] The present invention relates to the determination of mild cognitive impairment, and more particularly, to a device and method for determining mild cognitive impairment that can determine mild cognitive impairment using key input.

[0002] Mild cognitive impairment (MCI), a preclinical stage of Alzheimer's disease (AD), causes cognitive decline, including memory and executive function, beyond what occurs with normal aging. The importance of early intervention for patients with MCI has been emphasized. Therefore, individuals with MCI should undergo ongoing monitoring to ensure they do not progress from MCI to AD.

[0003] Screening methods for MCI include sensor-based tests and neuropsychological tests, which are primarily administered in clinical settings.

[0004] Sensor-based tests use accelerometers, voice recorders, motion capture sensors, and other sensors to detect and analyze patient movement, conversation, and motion signals to determine MCI. However, sensor-based tests often rely on data derived from difficult-to-use and evaluate sensors like accelerometers, voice recorders, and motion capture sensors, which can produce results that are often out of sync with everyday life (see non-patent literature).

[0005] Neuropsychological testing is a paper-based neuropsychological test administered by physicians, therapists, clinical psychologists, and other professionals in clinical settings like hospitals and public health centers to assess cognitive function. The testee's cognitive function level is assessed by recording their responses and questions, and these are compared to normative data to determine the extent of cognitive impairment and determine mild cognitive impairment. These neuropsychological tests can identify mild cognitive impairment with high sensitivity and specificity.

[0006] However, to undergo neuropsychological testing, an individual must visit a clinical institution and undergo testing by a specialist, which is cumbersome.

[0007] Additionally, neuropsychological tests have a problem in effectively distinguishing early MCI from healthy aging.

[0008] In addition to cognitive decline, MCI has been reported to be associated with motor dysfunction in the upper limbs of MCI patients (see non-patent literature).

[0009] Accordingly, the present invention aims to provide a device and method for determining mild cognitive impairment, which can determine whether or not a person has MCI by collecting keystroke information generated by key input dynamics representing the user's upper limb motor function during daily life.

[0010] However, the problem to be solved by the present invention is not limited to the problem mentioned above, and may be expanded in various ways without departing from the spirit and scope of the present invention.

[0011] An embodiment of the present invention provides a mild cognitive impairment determination device comprising: a key input unit that generates keystroke data according to a user's key input; an MCI DB that stores the keystroke data and standard data that are MCI determination criteria, and a main storage unit that stores an MCI determination unit; and an operation unit that executes the MCI determination unit, wherein the MCI determination unit is configured to store the keystroke data in the MCI DB and collect it, and then, when the keystroke data is collected beyond a standard capacity, compare the collected keystroke data with the standard data to determine whether or not it is MCI.

[0012] The above keystroke data may be data that includes, in a time series, a hold time (HC), which is the time during which a key is held down during a series of keystrokes; and a flight time (FC), which is the time from when a key is released until the next key is pressed.

[0013] The above MCI DB may further store keyboard type data including one or more keyboard type information that enables selection of the keyboard type of the key input unit; and user information of a user performing the MCI determination.

[0014] The above keyboard type information may include keyboard type information of one or more of a software keyboard and a hardware keyboard.

[0015] The keyboard type information of the above software keyboard may be one or more of QWERTY, Cheonjiin, Cheonjiin Plus, Naratgul, Windows Touch Interface, Linux Onboard, or Google G Board.

[0016] The above MCI determination unit may be configured to include a keystroke data collection module that collects the keystroke data generated from the key input unit and stores it in the MCI DB; a user information collection module that collects user information for the MCI determination; and an MCI determination module that, when the collected keystroke data exceeds a standard capacity, compares the collected keystroke data and the user information with the standard data to determine whether or not it is MCI.

[0017] The above MCI determination unit may further include an update module that receives updated standard data transmitted from an MCI server and updates the standard data.

[0018] Another embodiment of the present invention provides an MCI determination server comprising: a server communication unit for receiving keystroke data generated and transmitted from a user's terminal according to a user's key input through a communication network; a server MCI DB for storing the keystroke data and standard data that are MCI determination criteria, and a server main storage unit for storing a server MCI determination unit; and a server operation unit for executing the server MCI determination unit, wherein the server MCI determination unit collects the keystroke data by storing it in the server MCI DB, and when the keystroke data is collected in excess of a standard capacity, compares the collected keystroke data with the standard data to determine whether the user is an MCI corresponding to the keystroke data, thereby providing an MCI determination service through a communication network.

[0019] Another embodiment of the present invention provides a method for determining mild cognitive impairment using a mild cognitive impairment determination device including a key input unit, an MCI DB storing keystroke data and standard data that are MCI determination criteria, and a main storage unit storing an MCI determination unit and a calculation unit, the method comprising: a keystroke data collection step in which the MCI determination unit collects keystroke data according to a key input by the user using the key input unit; a user information collection step in which the MCI determination unit receives user information for MCI determination from a user, collects the collected information, and stores the collected information in the MCI DB; and a mild cognitive impairment determination step in which the MCI determination unit, when the keystroke data is collected in excess of a standard capacity for MCI determination, compares the collected keystroke data with the standard data to determine whether or not MCI is present.

[0020] The above method for determining mild cognitive impairment may further include an MCI determination unit installation step of receiving and installing the MCI determination unit recorded as a code executed by the operation unit.

[0021] The above keystroke data may be data that includes, in a time series, a hold time (HC), which is the time during which a key is held down during a series of keystrokes; and a flight time (FC), which is the time from when a key is released until the next key is pressed.

[0022] An embodiment of the present invention records and analyzes key inputs of a mobile device used on a daily basis to easily determine mild cognitive impairment.

[0023] An embodiment of the present invention can easily prevent the development of Alzheimer's dementia by early detection of mild cognitive impairment and providing cognitive treatment at an appropriate time.

[0024] The effects of the present invention are not limited to those mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description below.

[0025] Figure 1 is a functional block diagram of a cognitive impairment determination device (1) of an embodiment of the present invention.

[0026] Figure 2 is a diagram showing a keystroke data structure.

[0027] Figure 3 is a functional block diagram of the mild cognitive impairment determination unit (MCI determination unit) (23).

[0028] FIG. 4 is a diagram illustrating one embodiment of a user information input interface.

[0029] Figure 5 is a flowchart showing the processing process of a method for determining mild cognitive impairment according to an embodiment of the present invention.

[0030] Figure 6 is a diagram showing an example of the output of the result of determining the degree of cognitive impairment according to an embodiment of the present invention.

[0031] Figure 7 is a configuration diagram of an MCI determination system for updating standard data and an experimental example of an embodiment of the present invention.

[0032] Figure 8 is a graph showing the ROC curves of five predictive factors.

[0033] Specific structural or functional descriptions of embodiments according to the concept of the present invention disclosed in this specification are merely illustrative for the purpose of explaining embodiments according to the concept of the present invention, and embodiments according to the concept of the present invention may be implemented in various forms and are not limited to the embodiments described in this specification.

[0034] Embodiments according to the concept of the present invention may have various modifications and take various forms, and thus, embodiments are illustrated in the drawings and described in detail in this specification. However, this is not intended to limit embodiments according to the concept of the present invention to specific disclosed forms, but rather includes modifications, equivalents, or alternatives that fall within the spirit and technical scope of the present invention.

[0035] While terms such as "first" or "second" may be used to describe various components, these components should not be limited by these terms. These terms are intended solely to distinguish one component from another. For example, a first component may be referred to as a "second component," and similarly, a second component may also be referred to as a "first component," without departing from the scope of the invention.

[0036] When a component is referred to as being "connected" or "connected" to another component, it should be understood that it may be directly connected or connected to that other component, but that there may be other components in between. Conversely, when a component is referred to as being "directly connected" or "directly connected" to another component, it should be understood that there are no other components in between. Expressions that describe relationships between components, such as "between," "immediately between," or "directly adjacent to," should be interpreted similarly.

[0037] The terminology used herein is for the purpose of describing specific embodiments only and is not intended to limit the present invention. The singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, it should be understood that the terms "comprises" or "has" are intended to specify the presence of a described feature, number, step, operation, component, part, or combination thereof, but do not preclude the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0038] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. Terms defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and will not be interpreted in an idealized or overly formal sense unless explicitly defined herein.

[0039] Hereinafter, embodiments will be described in detail with reference to the attached drawings. However, the scope of the patent application is not limited or restricted by these embodiments. The same reference numerals in each drawing represent the same components.

[0040] Fig. 1 is a functional block diagram of a mild cognitive impairment determination device (1) (referred to as MCI determination device (1)) according to an embodiment of the present invention. Fig. 2 is a diagram showing a keystroke data structure. Fig. 3 is a functional block diagram of a mild cognitive impairment determination unit (hereinafter, MCI determination unit) (23).

[0041] The above MCI determination device (1) may be a variety of electronic devices including a key input means, such as a mobile device such as a smartphone, laptop, tablet PC, or a personal PC such as a desktop.

[0042] As shown in Fig. 1, the MCI determination device (1) may be configured to include a key input unit (10), a main storage unit (20), an operation unit (30), an auxiliary storage unit (40), an output unit (50), and a communication unit (60).

[0043] The above key input unit (10) may be configured to generate keystroke data according to a user's key input. The above key input unit (10) may be at least one of a software keyboard or a hardware keyboard that enables data to be input into the MCI determination device (1).

[0044] The keyboard type of the above software keyboard may be one or more of QWERTY, Cheonjiin, Cheonjiin Plus, Naratgul, Windows Touch Interface, Linux Onboard, or Google G Board. The types of the above software keyboard are described as examples and are not limited thereto.

[0045] The above main storage unit (20) may be a storage device that stores the MCI DB (21) that stores the keystroke data and the standard data that is the MCI determination criterion, and the MCI determination unit (23). The above main storage unit (20) may include a non-volatile storage device.

[0046] The above MCI DB (21) may be a database structured to store keystroke data, MCI determination criteria standard data, keyboard type data, and user information.

[0047] As shown in Fig. 2, the keystroke data may be data that includes, in a time series, a hold time (HC), which is the time during which a key is held down during a series of keystrokes, and a flight time (FC), which is the time from when the key is released until the next key is pressed.

[0048] Referring again to Figure 1, the above standard data may be the highest Uden index preset for MCI determination.

[0049] The above keyboard type data may include keyboard type information of one or more of a software keyboard and a hardware keyboard. The keyboard type information of the software keyboard may be keyboard type information of one or more of QWERTY, Cheonjiin, Cheonjiin Plus, Naratgul, Windows Touch Interface, Linux Onboard, or Google G Board.

[0050] The above user information may be information collected by user input so that the MCI determination unit (23) can perform MCI determination. The user information may include one or more of age, gender, education level, keyboard type, and number of years of mobile device use.

[0051] The above MCI determination unit (23) can be configured to collect the keystroke data generated when a user inputs a key and compare it with standard data to determine whether it is MCI.

[0052] The above MCI determination unit (23) can be configured to perform MCI determination using the standard deviation of keystroke data or MCI determination using the Uden index.

[0053] The MCI determination using the standard deviation of the above keystroke data collects keystroke data of healthy people to construct a standard data. The mean and standard deviation can be derived from the constructed standard data, and the sum of the mean and the adjusted standard deviation can be set as the MCI determination criterion value. For example, the MCI determination criterion value can be set as (mean + yx standard deviation). In this case, the coefficient y of the standard deviation can have a value of 0.5 to 2.5. Once the MCI determination criterion value is determined, if the flight time or hold time of the user's keystroke data is greater than the MCI determination criterion value, it can be determined as MCI. The MCI determination using the Uden index collects keystroke data of healthy people and people classified as mild cognitive impairment, and then performs receiver operating characteristic curve (ROC curve) analysis. The analysis results are presented with multiple sensitivities and specificities according to the values ​​(cut-off points) that can distinguish between two points. The MCI determination unit (23) calculates the Youden index (sensitivity +1, specificity -1) using the sensitivity and specificity values. At this time, the highest Youden index becomes the point that considers the optimal sensitivity and specificity. The MCI determination unit (23) can determine that the hold time or flight time of the user's keystroke data is MCI if it is greater than the cut-off point corresponding to the highest Youden index, and determine that it is normal if it is smaller.

[0054] The above MCI determination unit (23) is described in more detail below with reference to FIG. 3.

[0055] The above-mentioned operation unit (30) may be configured to execute the above-mentioned MCI determination unit (23) and then cause the MCI determination unit (23) to perform an operation for MCI determination. The above-mentioned operation unit (30) may be a calculation device such as a central processing unit, a GPU (Graphics Processing Unit), a microcontroller, or an FPGA.

[0056] The auxiliary storage unit (40) may be a memory device into which the MCI determination unit (23) is loaded and executed. The auxiliary storage unit (40) may be a volatile memory device that temporarily stores data generated by the MCI determination unit (23).

[0057] The above output unit (50) may be configured to output the driving status and generation information of the MCI determination unit (23). The output unit (50) may include a display device, a serial or parallel data output port, a printer, an audio device for sound or auditory output, etc.

[0058] The above communication unit (60) may be a communication interface device that enables the MCI determination unit (23) to communicate with an external device. The communication unit (60) may be configured to enable the MCI determination unit (23) to transmit collected keystroke data to an external device or to receive updated standard data from an external device.

[0059] Again, referring to FIG. 3, the MCI determination unit (23) may be configured to include a keystroke data collection module (231), a user information collection module (233), an MCI determination module (235), and an update module (237) to determine whether or not MCI is present using keystroke data.

[0060] The above keystroke data collection module (231) can be configured to collect keystroke data generated from the key input unit (10) and store it in the MCI DB.

[0061] The user information collection module (233) may be configured to collect user information for MCI determination. To this end, the user information collection module (233) may be configured to output a user information input interface for user information collection.

[0062] FIG. 4 is a diagram illustrating one embodiment of a user information input interface.

[0063] As shown in Fig. 4, the user information input interface can be configured to allow the user to input age, gender, education level, keyboard type, and number of years of mobile phone use.

[0064] Again, referring to FIG. 3, the MCI determination module (235) may be configured to determine whether or not an MCI is present by comparing the collected keystroke data and the user information with the standard data when the collected keystroke data exceeds a reference capacity. The MCI determination module (235) may be an artificial intelligence MCI determination model learned by an artificial intelligence algorithm to automatically determine whether or not an MCI is present from the keystroke data and the user information by using the keystroke data, the user's MCI history data, the user information, and the standard data as learning data.

[0065] The above update module (237) can be configured to update the standard data stored in the MCI DB (21) with the standard data for update received from an external device such as the MCI server (9, see FIG. 7) by the MCI determination unit (23).

[0066] Figure 5 is a flowchart showing the processing process of a method for determining mild cognitive impairment according to an embodiment of the present invention.

[0067] As shown in FIG. 5, the method for determining mild cognitive impairment (hereinafter referred to as the MCI determination method) of the embodiment is performed by the MCI determination device (1) of FIGS. 1 to 4, and may be configured to include a user information collection step (S30), a keystroke data collection step (S40), a standard capacity satisfaction determination step (S50), and a mild cognitive impairment determination step (S60) (hereinafter referred to as the MCI determination step (S50)).

[0068] The above user information collection step (S30) may be a step in which the user information collection module (233) of the MCI determination unit (23) receives user information for MCI determination from the user, collects the information, and stores it in the MCI DB (21).

[0069] The above keystroke data collection step (S40) may be a step in which the keystroke data collection module (231) of the MCI determination unit (23) collects keystroke data according to the user's key input using the key input unit after the user information collection step (S30) is performed.

[0070] The above-mentioned standard capacity satisfaction determination step (S50) may be a step for determining whether the capacity of the keystroke data collected and stored in the keystroke data collection module (231) is greater than or equal to the standard capacity for MCI determination.

[0071] The above MCI determination step (S50) may be a step in which the MCI determination module (235) of the MCI determination unit (23) determines whether or not MCI is present by comparing the collected keystroke data and user information with the standard data when the keystroke data is collected in excess of the standard capacity for MCI determination.

[0072] Fig. 6 is a diagram showing an example of the output of the mild cognitive impairment determination result of an embodiment of the present invention. As shown in Fig. 6, the determination result of the MCI determination step (S50) can be output through the output unit (50).

[0073] This is explained again with reference to Figure 5.

[0074] The above MCI determination unit (23) may be software written in code executed by the operation unit (30). In this case, the MCI determination unit (23) may be received from an external device such as an MCI server (9, see FIG. 7) through a communication unit (60) and installed in the main storage unit (20) by the operation unit (30).

[0075] In this case, the MCI determination method may further include an MCI determination unit installation step (S10) and an MCI determination unit execution step (S20) performed before the user information collection step (S30), as shown in FIG. 5.

[0076] The above MCI determination unit installation step (S10) may be a step in which the operation unit (30) downloads the software MCI determination unit (23) and then installs it in the main storage unit (20). The MCI determination unit (23) installed in the main storage unit (20) may create an MCI DB (21) in the main storage unit (20) during the installation process.

[0077] The above MCI determination unit execution step (S20) may be a step in which the operation unit (30) loads the MCI determination unit (23) into the auxiliary storage unit (40) and executes it in response to an MCI determination unit execution command from a user after the MCI determination unit (23) is installed. The MCI determination unit (23) may output the user information input interface of FIG. 4 when initially executed. In addition, the MCI determination unit (23) may be executed in the background to collect the keystroke data.

[0078] The MCI determination unit (23) of the embodiment of the present invention may be configured to be driven by a mobile device and perform MCI determination independently, as described in FIGS. 1 to 6. Alternatively, the MCI determination unit (23) may be configured to collect keystroke data from the mobile device and then transmit it to an MCI server (9) to perform MCI determination, and then receive and output the MCI determination result from the MCI server (9).

[0079] Figure 7 is a configuration diagram of an MCI determination system for updating standard data and an experimental example of an embodiment of the present invention.

[0080] As shown in Fig. 7a, the MCI determination system may be configured such that a plurality of mobile devices, each of which is an MCI determination device (1) having an MCI determination unit (23) installed, communicate with an MCI server (9) via a communication network. The user terminal (2) including the mobile device may include, for example, a smartphone, a laptop, a personal PC, etc.

[0081] As shown in b of Fig. 7, the MCI server (9) may be configured as a server computer including a server input unit (110), a server main storage unit (120) that stores a server MCI DB (121) and a server MCI determination unit (123), a server operation unit (130), a server auxiliary storage unit (140), a server output unit (150), and a server communication unit (160).

[0082] It may be a communication port to which a keyboard or an administrator's laptop, etc., configured to receive control commands from the administrator of the above server input unit (110) is connected.

[0083] The above server main storage unit (120) may be a storage device that stores a server MCI DB (121) and a server MCI determination unit (123) that store keystroke data generated by a user's key input transmitted from a user terminal connected through a communication network and standard data that is a MCI determination criterion. The server main storage unit (120) may include a non-volatile storage device.

[0084] The above server MCI DB (121) may be a database structured to store keystroke data received through a communication network, standard data described with reference to FIGS. 1 to 6 as MCI determination criteria, keyboard type data, and user information.

[0085] The above server MCI determination unit (123) can be configured to perform MCI determination using the standard deviation of keystroke data or MCI determination using the Uden index, similarly to the MCI determination unit (23) of the mild cognitive impairment determination device (1).

[0086] The server MCI determination unit (123) may be configured to collect the keystroke data by storing it in the server MCI DB (121), and then, if the keystroke data is collected beyond a standard capacity, compare the collected keystroke data with the standard data to determine whether the user's MCI corresponds to the keystroke data. The server MCI determination unit (123) may transmit the determined MCI determination result to the user terminal (2) via the server communication network (130).

[0087] The above server operation unit (130) may be configured to execute the server MCI determination unit (123) and then perform an operation to determine whether the user is an MCI user using the received keystroke data. The server operation unit (130) may be a calculation device such as a central processing unit, a GPU (Graphics Processing Unit), a microcontroller, or an FPGA.

[0088] The server auxiliary storage unit (140) may be a memory device into which the server MCI determination unit (123) is loaded and executed. The server auxiliary storage unit (140) may be a volatile memory device that temporarily stores data generated by the server MCI determination unit (123).

[0089] The server output unit (150) may be configured to output the operating status and generation information of the server MCI determination unit (123). The server output unit (150) may include a display device, a serial or parallel data output port, a printer, etc.

[0090] The server communication unit (160) may be a communication interface device that enables the server MCI determination unit (123) to communicate with an external device. The server communication unit (160) may be configured to allow the server MCI determination unit (123) to receive user keystroke data from a user terminal and transmit the MCI determination result to the user terminal (2) or an external device. In addition, the server communication unit (160) may be configured to receive standard data to be updated from an external device. The standard data to be updated may be input through the server input unit (110).

[0091] <Experimental Example>

[0092] The performance of the MCI determination method using the MCI determination device (1) using keystroke data of the present invention was tested.

[0093] The participants were divided into two groups: 64 healthy controls (HC) and 47 MCI patients. According to a previous study [Peterson RC. Mild cognitive impairment as a diagnostic entity. J Intern Med. 2004;256(3):189-194. [doi: 10.1111 / j.1365-2796.2004.01388]], the inclusion criteria for MCI were a) subjective memory complaints, b) memory impairment compared to age- and education-matched healthy controls as determined by neuropsychological battery performance (<1.5 standard deviations), c) intact global cognitive function as determined by the Cognitive Impairment Test score, d) independent activities of daily living, and e) smartphone use for at least 3 months. The exclusion criteria were a) clinician-diagnosed dementia, b) neurological or psychiatric disorders such as stroke and depression, and c) visual or hearing impairment. These criteria are based on the original Peterson criteria, which limit MCI to memory problems only (amnestic MCI).

[0094] Ethical Considerations

[0095] All participants provided written informed consent prior to the experiment. The experiment was approved by the Institutional Review Board of Soonchunhyang University (202306-SB-070-04) and registered with the Thai Clinical Trials Registry (TCTR20220415002).

[0096] <Procedure>

[0097] All participants performed the Computerized Cognitive Behavior Test (CBT) task, a Korean version of the Montreal Cognitive Assessment (MoCA-K), prior to collecting keystroke data. They then installed the free "Neurokeys" mobile application, developed for Android and iOS, which enabled them to measure their health status through typing on their smartphones. Participants were instructed to familiarize themselves with the Neurokeys keyboard, which replaced their default keyboard, for a week. While regularly typing, keyboard interactions (keystroke dynamics) that generated keystroke data were recorded in the background, generating the keystroke data shown in Figure 2. Keystroke dynamics were stored in JSON format and indexed in a database for use by the application. The application periodically transmitted uniquely encoded keystroke dynamics information to the MCI server (9), a remote cloud server (Microsoft Azure), while the participant's device was connected to Wi-Fi and charging.

[0098] Data collection took place over a period of one month, and all participants were instructed to ensure that no one other than themselves could type on their phones. Over the course of one month, 3,530 typing sessions were collected: 2,250 sessions from 64 HC patients and 1,280 sessions from 35 MCI patients. Of the 3,530 sessions, 2,740 sessions with more than 40 keystrokes per session were ultimately used in the analysis.

[0099] In this experiment, we analyzed two variables from the key input dynamics of the saved JSON files: hold time (HT) and flight time (FT). These two variables were selected because they were found to be significant in identifying MCI among other key input dynamics. HT represents the time interval between key press and release, and FT represents the time interval between key press and the next key press (see Figure 2). For data preprocessing, HTs exceeding 700 μs and FTs exceeding 3 seconds were excluded.

[0100] <Resulting Action>

[0101] The MoCA-K was developed to differentiate MCI from normal aging and comprises visuospatial / executive function, attention, memory, language, abstraction, and orientation. Scores range from 0 to 30, with higher scores indicating better overall cognitive function. The MoCA-K cutoff score for MCI is 23, with an additional 1 point added for subjects with less than 6 years of education. In a previous study using a cutoff score of 23, the sensitivity and specificity were 94.2% and 40.5%, respectively.

[0102] Computer-based cognitive testing (CBT) tests were used to assess attention and working memory.

[0103] In the experiment, nine white squares were randomly placed on a tablet monitor. While the color of some squares changed sequentially from white to red, participants were asked to memorize the location and order of the changed squares. Participants were then instructed to point to the changed square by touching the screen. The experiment set the number of changing squares to five, and 15 trials were conducted.

[0104] <Statistical Analysis>

[0105] SPSS for Windows (version 22.0) was used for data analysis. Participants' demographic characteristics were analyzed using descriptive statistics. Independent t-tests and chi-square tests were used to compare the two groups. Receiver operating characteristic (ROC) curve analysis was used to determine sensitivity and specificity. The cutoff score for MCI patients was determined based on the highest Youden index (sensitivity + specificity - 1), which can serve as a criterion for selecting the optimal cutoff score. Predicted probabilities from logistic regression were used to perform ROC curve analysis with the combination of keystroke dynamics as a single variable. Spearman's correlation coefficient was used to examine the relationship between keystroke dynamics and overall cognitive function. Statistical significance was set at p < 0.05.

[0106] <Result>

[0107] - General and clinical characteristics of the two groups

[0108] There were no statistically significant differences in demographic characteristics.

[0109] There were no significant differences between the two groups (P>.05), except for one (P<0.001). These results indicate that HCs were age- and education-matched to subjects with MCI, except for cognitive function (see Table 1).

[0110] [Table 1]

[0111]

[0112] Table 1 presents the general clinical characteristics of the participants (N=111). The values ​​presented in Table 1 are means (standard deviations).

[0113] The terms for each are as follows:

[0114] MMSE-K: Korean version of Mini-Mental State Examination, MoCA-K: Korean version of Montreal Cognitive Assessment, CBT: Cauchy Block Test, HT: Hold Time, FT: Flight Time.

[0115] Sensitivity, specificity, and discriminatory power

[0116] [Table 2]

[0117]

[0118] Table 2 shows the sensitivity and specificity for detecting MCI (N = 111).

[0119] Each variable in Table 2 is as follows.

[0120] P<0.001. AUC: area under the curve; CBT: Corsi block test; MoCA-K: Montreal Cognitive Assessment Korean version; FT: flight time; HT: hold time

[0121] Figure 8 is a graph showing the ROC curves of five predictive factors.

[0122] When separating MCI from matched HC, both HT and FT showed higher Youden indices than the MoCA-K (HT: .847; FT: .947; MoCA-K: .469), suggesting that key input dynamics can better discriminate MCI than existing MCI screening tools (see Table 2). In particular, FT showed the highest sensitivity (97.9%) and specificity (94.7%). Interestingly, when combined with MoCA-K, HT showed a higher Youden indices, whereas FT did not (see Table 2 and Figure 8).

[0123] <Correlation between Key Input Dynamics and Cognitive Function>

[0124] [Table 3]

[0125]

[0126] Table 3 shows the correlation between global cognitive functions and keystroke dynamics.

[0127] Key input dynamics were correlated with MoCA-K (HT: r=-.468, P<0.001; FT: r=-.491, P<0.001) and CBT tests (HT: r=-.487, P<0.001; FT: r=-.492, P<0.001) (see Table 3). These results indicate that keystroke dynamics are related to overall cognitive function and working memory.

[0128] The experimental results showed that MCI patients had significantly longer keystroke latencies than the control group. In particular, the delay between keypresses had the highest sensitivity (97.9%) and specificity (96.9%). Furthermore, keystroke dynamics were significantly correlated with the MoCA-K (hold time: r= -0.468, P<0.001; flight time: r= -0.497, P<0.001).

[0129] The experimental results confirmed that the longer HT and FT times in MCI patients are more likely due to working memory deficits rather than purely motor-related issues. Therefore, the experiment suggests that smartphone typing in MCI patients may be negatively affected by working memory deficits, resulting in longer keystroke dynamics, particularly HT and FT reaction times.

[0130] Furthermore, HT and FT were able to detect individuals with MCI with Uden's indices of .847 and .947, respectively. Both HT and FT showed greater discriminatory power in MCI than the MoCA-K, confirming that key input dynamics may be more advantageous in identifying MCI.

[0131] MCI occurs more frequently in normal aging than in existing screening tools. The primary factor underlying this comparison is the metrics between keystroke dynamics and the MoCA-K. Delayed HT and FT may indicate working memory deficits, a hallmark of MCI. Indeed, HT and FT showed significant correlations with CBT tests that support working memory indices in keystroke dynamics. In contrast, the MoCA-K reflects global cognitive function, including orientation and language, which are not clearly impaired in MCI. These differences in metrics indicate the superiority of keystroke dynamics for MCI screening.

[0132] In particular, the FT showed the highest Youden index among all outcome measures, and the Youden index did not increase when combined with the MoCA-K. These results indicate that the FT is more useful than the HT in differentiating MCI. Furthermore, the delayed FT can be used alone as a good marker of MCI, suggesting that the keystroke dynamics monitor can be a cost-effective method for identifying MCI, given that it does not require additional neuropsychological testing to improve the accuracy of the test.

[0133] In addition to demonstrating superior keystroke dynamics compared to MoCA-K, which is widely used to detect MCI through experiments, FT has the following advantages:

[0134] Experimental results confirmed that keystroke dynamics, which reflect both motor and cognitive deficits, are a more clinically useful digital biomarker of MCI than neuropsychological assessment tools. These experimental results offer a novel perspective on detecting MCI through keystroke dynamics associated with smartphone use. These promising results suggest that measurements made via smartphone typing can serve as an ecologically valid digital biomarker that can replace laboratory-based neuropsychological screening tools in nonclinical settings.

[0135] The devices described above may be implemented as hardware components, software components, and / or a combination of hardware components and software components. For example, the devices and components described in the embodiments may be implemented using one or more general-purpose computers or special-purpose computers, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing instructions and responding to them. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to the execution of the software. For ease of understanding, the processing device is sometimes described as being used alone; however, one of ordinary skill in the art will recognize that the processing device may include multiple processing elements and / or multiple types of processing elements. For example, a processing unit may include multiple processors, or a processor and a controller. Other processing configurations, such as parallel processors, are also possible.

[0136] Software may include a computer program, code, instructions, or a combination of one or more of these, which may configure a processing device to perform a desired operation or may, independently or collectively, command the processing device. The software and / or data may be permanently or temporarily embodied in any type of machine, component, physical device, virtual equipment, computer storage medium or device, or transmitted signal wave, for interpretation by the processing device or for providing instructions or data to the processing device. The software may also be distributed over networked computer systems and stored or executed in a distributed manner. The software and data may be stored on one or more computer-readable recording media.

[0137] The method according to the embodiment may be implemented in the form of program commands that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program commands, data files, data structures, etc., alone or in combination. The program commands recorded on the medium may be those specially designed and configured for the embodiment or may be those known and available to those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, and hardware devices specially configured to store and execute program commands, such as ROMs, RAMs, and flash memories. Examples of the program commands include not only machine language codes generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter, etc. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiment, and vice versa.

[0138] Although the embodiments described above have been described with limited drawings, those skilled in the art will recognize that various modifications and variations can be made based on the above description. For example, appropriate results can still be achieved even if the described techniques are performed in a different order than described, and / or components of the described systems, structures, devices, circuits, etc. are combined or combined in a different manner than described, or are replaced or substituted with other components or equivalents.

[0139] Therefore, other implementations, other embodiments, and equivalents to the claims also fall within the scope of the claims described below.

Claims

1. A key input unit that generates keystroke data according to the user's keystroke; A main storage unit storing the MCI DB and the MCI determination unit that store the above keystroke data and the standard data that is the MCI determination criterion; and Includes a computation unit that executes the above MCI determination unit, The above MCI determination unit, After collecting the above keystroke data by storing it in the MCI DB, if the above keystroke data is collected more than the standard capacity, it is configured to compare the collected keystroke data with the standard data to determine whether it is MCI. Mild cognitive impairment screening device.

2. In paragraph 1, The above keystroke data is, During continuous keystroke input, Hold time (HC), which is the time the key is pressed; and Flight time (FC), which is the time from when a key is released until the next key is pressed, is data that contains the time series. Mild cognitive impairment screening device.

3. In paragraph 1, The above MCI DB is Keyboard type data including one or more keyboard type information that enables selection of the keyboard type of the above key input unit; and Further storing user information of the user performing the above MCI determination Mild cognitive impairment screening device.

4. In paragraph 3, The above keyboard type information is Contains information about one or more keyboard types, either software keyboard or hardware keyboard. Mild cognitive impairment screening device.

5. In paragraph 4, Keyboard type information of the above software keyboard is: Information on one or more keyboard types: QWERTY, Cheonjiin, Cheonjiin Plus, Naratgul, Windows Touch Interface, Linux Onboard, or Google Gboard Mild cognitive impairment screening device.

6. In paragraph 1, The above MCI determination unit, A keystroke data collection module that collects the keystroke data generated in the above key input unit and stores it in the MCI DB; A user information collection module that collects user information for determining the above MCI; and If the collected keystroke data exceeds the standard capacity, the collected keystroke data and the user information are compared with the standard data to determine whether or not it is MCI, and the MCI determination module is included. Mild cognitive impairment screening device.

7. In paragraph 6, The above MCI determination unit, It further comprises an update module that receives updated standard data transmitted from the MCI server and updates the standard data. Mild cognitive impairment screening device.

8. A server communication unit that receives keystroke data generated and transmitted from the user's terminal according to the user's keystroke input through a communication network; A server MCI DB that stores the above keystroke data and the standard data that is the MCI determination criterion, and a server main storage unit that stores the server MCI determination unit; and Includes a server operation unit that executes the above server MCI determination unit, The above server MCI determination unit After collecting the above keystroke data by storing it in the server MCI DB, if the above keystroke data is collected more than the standard capacity, the collected keystroke data is compared with the standard data to determine whether the user corresponding to the keystroke data is MCI, and an MCI determination service is provided through a communication network. MCI Disclosure Server.

9. In a method for determining mild cognitive impairment using a mild cognitive impairment determination device including a main storage unit and a calculation unit storing a key input unit, an MCI DB storing keystroke data and standard data that are MCI determination criteria, and an MCI determination unit, A keystroke data collection step in which the above MCI determination unit collects keystroke data according to keystroke input by the user using the above key input unit; A user information collection step in which the MCI determination unit receives user information for MCI determination from the user, collects the information, and stores it in the MCI DB; and The above MCI determination unit includes a mild cognitive impairment determination step of determining whether or not MCI is present by comparing the collected keystroke data with the standard data when the keystroke data is collected in excess of the standard capacity for MCI determination. How to determine mild cognitive impairment.

10. In paragraph 9, The above keystroke data is, During continuous keystroke input, Hold time (HC), which is the time the key is pressed; and Flight time (FC), which is the time from when a key is released until the next key is pressed, is data that contains the time series. How to determine mild cognitive impairment.

Images