TWI791323B - Method and computer program product for predicting Mini-Mental Examination (MMSE) scores using human-computer interactive games - Google Patents

Abstract

A method for predicting scores of the Mini-Mental Examination (MMSE) with a human-computer interactive game, the method is executed by a computer device; the method includes: providing a game test program by a human-computer interactive game module, the game test program is based on the response The four intelligent items of power, attention, coordination, and memory produce several questions respectively, and each item game test has a limited time; the human-computer interaction game module calculates and records a testee's execution of each item game The number of correct answers, correct rate, and correct answer reaction time performance parameters produced by the test; a simple mental scale score prediction module extracts the number of correct answers of the reaction force and the number of correct answers of the memory as parameters for calculation, and obtains the Mini-Mental Scale Prediction Scores for Subjects. The invention can effectively predict the cognitive function and risk of cognitive impairment of the elderly.

Description

Method and computer program product for predicting Mini-Mental Examination (MMSE) scores using human-computer interactive games

The present invention is related to the detection of cognitive function, more specifically, it is a method and computer program for predicting the score of mini-mental state examination (MMSE) through serious game (serious game) or functional game (applied game) product.

Global aging has become an irreversible trend. The age structure of Taiwan's elderly population is also growing rapidly, and the incidence of dementia is also gradually increasing. Therefore, how to easily and objectively assess the cognitive function of the elderly has become an urgent need.

TW201424692 discloses a system for analyzing mental function indicators, including at least one user end, which includes a set of wireless sensing interaction devices, an interactive display module and a body identification module, the body identification module is located in the On the interactive display module, user identity identification can be carried out. The wireless sensing interactive device has a device connection module and a meter input module, which can provide users with different operation methods, and the interactive display The module informs the operation content in the form of voice and image; and a system terminal, which is connected to the user terminal, which includes a user database and a mental function analysis module. The user database stores at least A user account information and a personalized information, the personalized information at least includes the user's gender, age, education level, frontal assessment battery (FAB), Mini-Mental State Examination Scale (Mini-Mental State Examination, MMSE) and Montreal cognitive assessment (MoCA) and other mental degradation index data, the mental function The analysis module first connects to the user database to retrieve the user account information and the personalized information represented by the user identification code from the user identification code sent by the identity recognition module, The mental function analysis module can analyze the user's mental function status through the operation of the wireless sensing interaction device, which includes a normalized processing module, a hierarchical classification module and an output module, all of which are personalized The mental degradation index data will be converted according to the standard value of the index in one of the normalized computing units in the normalization processing module, normalized to the same interval for comparison between the different mental degradation index data, and the normalized data will be imported The evaluation of the degradation status in the hierarchical classification module can classify the user into three categories: dementia, mild dementia and normal through a classification algorithm, and output the result to the user terminal through the output module for display .

TW201926233A discloses a method for providing corresponding training based on cognitive diagnosis results, the method at least includes the following steps: defining a plurality of cognitive testing parameters; establishing a cognitive test question bank based on these cognitive testing parameters; A cognitive ability test; judging whether the subject's cognitive function is abnormal according to the test process and test results of the subject; and when the cognitive function is abnormal, providing the subject with intensive training corresponding to the cognitive function .

TWI494084B discloses a cognitive function detection device, including: a sensing module that senses a user's movement to generate at least one movement signal; a step analysis module that performs step analysis based on the movement signal to generate multiple gaits parameters; and a cognitive function analysis module that performs a cognitive function analysis based on the gait parameters to generate a cognitive function analysis result, wherein the cognitive function analysis result includes at least one of the gait parameters and a The mapping relationship of cognitive function scores, and the cognitive function analysis module generates a disease assessment result based on the cognitive function analysis result and a cognitive function test information, the cognitive function test information includes A cognitive function actual score.

TWI605862 discloses a personalized game system that can detect cognitive functions, which is built in an electronic device, and the electronic device is connected to an image detector. The detector automatically captures the user's posture and gestures, which are used to detect the user's posture, movement and reaction when completing a checkpoint, and calculate the user's position in the virtual 3D space; an identity recognition and game recording module, the system Directly capture the on-site photos through the image detector, and combine the position of the limbs positioned by the image detector to capture the face photos in the photos, and compare the obtained face photos in the database through the face recognition method Take out the correct user's identity information and game records, and update the history of the game records; a game level score module, which includes multiple game levels, and records the user's answering speed and body movements in each game level Parameters of response, correct rate, and answering time, wherein the game level is used to evaluate at least two or more cognitive abilities; a cognitive function analysis module is used to analyze the correlation between the user's performance in each level and cognitive ability sex, analyzing the correlation between the user's physical response parameters and cognitive ability, and analyzing the game user's cognitive ability, where the cognitive ability refers to the user's attention, memory, oral comprehension behavior ability, construction ability, computing ability and Orientation.

TWI634528 discloses a computer program product for evaluating cognitive ability, which is loaded into a computer to perform the following steps: Step S1: Enter a setting interface, in which at least one test control parameter is selectively set and/or Or at least one physical environment parameter; step S2: enter a test interface, make a test subject accept a test program, the test program is configured to test a short-term memory accuracy rate of the test subject, the computer according to the at least one test The control parameter and the at least one physical environment parameter control a peripheral device to adjust a physical factor of a physical environment with a control signal; step S3: enter a scoring mode, and calculate the subject with a preset scoring criterion Accepting a score obtained by the test procedure; and step S4: entering an analysis mode, generating an analysis report according to the score, the at least one test control parameter and the at least one physical environment parameter; wherein, the at least one test control parameter is used as controlling the variable, and using the at least one physical environment parameter as the operating variable to generate the analysis report; or using the at least one physical environment parameter as the control variable and using the at least one test control parameter as the operating variable, to generate the analysis report.

TWM606750 discloses a game system for detecting the cognitive function of the elderly, including: a master control device, the master control device drives the data storage unit through the processing unit for corresponding linkage according to the received signal, and the data storage unit includes: A body identification module, including identity creation sub-module and identity recognition sub-module; a game level module, including a sub-module of personalized life experience material, a cognitive function game level sub-module, and a game response parameter recording sub-module and a sub-module for recording game scores; and a cognitive function analysis module, including a questionnaire module, a sub-module for analyzing the correlation between game scores and questionnaires, and a sub-module for tracking cognitive function processes; The control device includes a large-area interactive display plane and a touch module, and transmits touch signals received by the touch module to the processing unit. The game levels set by using personalized life experience materials enable the elderly to complete the assessment of cognitive function during the game process, and through long-term tracking to know the changes in the cognitive function of the elderly, and then provide a reference for clinical dementia prevention and treatment.

The development of a standard and objective cognitive function assessment model combined with serious game elements has become a feasible strategy for screening the risk of cognitive impairment in the elderly. The present invention designs a human-computer interactive game capable of predicting the scores of the Mini-Mental Examination (MMSE) and generates relevant parameters, and uses the parameters to calculate the predicted scores of the Mini-Mental Examination (MMSE) to evaluate the risk of cognitive impairment of the elderly.

The technical feature of the present invention: a computer program product for predicting scores of the Simple Mental Scale (MMSE) with a human-computer interactive game, which is loaded into a computer device; including: a human-computer interactive game module, the human-computer interactive game module provides A subject conducts a game test program; the game test is based on four intelligent items such as reaction, attention, coordination and memory to produce several questions, and each item game test has a limited time; the man-machine The interactive game module calculates and records the performance parameters such as the number of correct answers, the correct rate, and the reaction time of the correct answer produced by the subject's execution of the game test for each item; a simple mental scale score prediction module receives the All the performance parameters transmitted by the human-computer interaction game module, the number of correct answers to the reaction force and the number of correct answers to the memory are extracted as calculation parameters to obtain the predictive score of the simple mental scale of the subject.

Efficacy of the present invention:

The present invention does not use traditional questionnaires to perform MMSE evaluation, but uses the game test of the present invention to predict MMSE scores through program calculations. The predicted MMSE scores have high reliability, validity and explanatory power through statistical analysis verification, and can effectively predict the cognitive function of the elderly and risk of cognitive impairment. The invention can be used as a feasible tool for evaluating the cognitive function of the elderly.

1: Human-computer interaction game module

a: red light

b: blue light

c: green light

〔Fig. 1〕The present invention reaction force game test topic schematic diagram.

[Fig. 2] Schematic diagram of the attention game test item of the present invention.

[Fig. 3] Schematic diagram of the subject of the coordination game test of the present invention.

[Fig. 4] Schematic diagram of the memory game test subject of the present invention.

The computer program product of the present invention to predict the scores of the Mini-Mental Examination (MMSE) is loaded into in a computer device. The computer devices include but are not limited to personal computers, tablet computers, smart phones, or similar devices.

The computer program product of the present invention includes a personal information module, a man-machine interactive game module, and a simple mental scale score prediction module.

The personal information module receives and records a subject's name, gender, age, education level, the testee's performance parameters produced by operating the human-computer interaction game module, and the score prediction of the simple mental scale The MMSE prediction score calculated by the module based on the game score. .

The man-machine interactive game module is a game program module built in the computer program product of the present invention, or a physical game device linked with the computer program product of the present invention. The man-machine interactive game module is used for the subject to conduct a game test. The game test is based on the four intelligent items of reaction, attention, coordination and memory, and several questions are produced respectively. Each game test has a limited time. Subjects complete within this limited time. The program of the game test and its limited time are 2 minutes of reaction, 2 minutes of attention, 2 minutes of coordination and 4 minutes of memory, and the interval time of each game test is 30 seconds. The human-computer interactive game module calculates and records the performance parameters such as the number of correct answers, the correct rate, and the reaction time of correct answers produced by the subject performing each game test, and transmits the performance parameters to the simple mind Scale score prediction module. The number of correct answers is defined as the number of correct answers within the limited time. The correct rate is defined as the number of correct answers divided by the total number of answers within the limited time, and the unit is a percentage. The reaction time of correct answer is the time required to answer the question correctly within the limited time, and the unit is millisecond.

The game program module or physical game device mainly provides a palace-style game, which includes several grids arranged in a matrix, preferably nine grids or sixteen grids, and each grid can be Produces a light. If the palace-style game is provided for the game program module, the palace-style game is displayed on the touch screen of the computer device, and the user plays the game by touching the light on the grid through the touch screen. If the palace-style game is provided for the physical game device, the user can touch, press, and tap the lighted physical objects in the palace grid with fingers or palms to perform game actions. As shown in Fig. 1 to Fig. 4, the human-computer interaction game module 1 of the present invention adopts the sixteen-square format game, and a light is generated in each square.

As shown in Figure 1, the subject of the reaction game test is to turn on a red light a at random, and the subject presses the red light a to answer, so as to test the subject's ability to respond quickly and correctly after seeing the light. The human-computer interaction game module 1 calculates and records the testee's performance parameters such as the number of correct answers in the reaction game test, the correct rate, and the reaction time for correct answers, and transmits them to the personal information module and the simple mind Scale score prediction module.

As shown in Figure 2, the question of the attention game test is to randomly turn on one to several red lights a, blue lights b, and green lights c, and the testee presses the red light a to answer. The ability to judge correctly. The human-computer interaction game module 1 calculates and records the number of correct answers, correct rate, and reaction time of correct answers of the subject in the attention game test, and transmits them to the personal information module and the simple mental quantity Table score prediction module.

As shown in Figure 3, the topic of the coordination game test is to randomly turn on one to a symmetrical red light a, and the testee answers by pressing the symmetrical red light with both hands. The ability to respond correctly. The human-computer interaction game module 1 calculates and records the testee's performance parameters such as the number of correct answers in the coordination game test, the correct rate, and the reaction time for correct answers, and transmits them to the personal information module and the simple mind Scale score prediction module.

As shown in Figure 4, the topic of the memory game test is to turn on one to several red lights a in sequence, and the testees are Ask to memorize the order and position of these red lights a and press the red light to answer. It is used to test the ability of the subjects to quickly answer the correct order and position after seeing the light signals. The man-machine interactive game module 1 calculates and records the testee's performance parameters such as the number of correct answers in the memory game test, the correct rate, and the reaction time for correct answers, and transmits them to the personal information module and the simple mental quantity Table score prediction module.

The score prediction module of the Simple Mental Scale receives all the achievement parameters transmitted by the human-computer interaction game module, and extracts the number of correct answers of the reaction force (RT-CN) and the number of correct answers of the memory power (ME-CN), as follows: Calculations are performed to obtain the predicted score of the mini-mental state examination (MMSE) of the subject, hereinafter referred to as the predicted score of MMSE. The MMSE prediction score is stored in the personal information module.

MMSE=-4.432+0.613 * RT-CN+0.388 * ME-CN

The Simple Mental Scale Score Prediction Module can extract the subject’s gender (Gender, female is number 0, male is number 1) from the personal information module, and the following formula is calculated to obtain a more predictive MMSE prediction Fraction.

MMSE=-4.432+0.613 * RT-CN+0.388 * ME-CN+0.197 * Gender

The full score of the MMSE scale is 30 points. According to the education level of the testees, those with a junior high school education or above are defined as cognitive dysfunction with a score of 24 points or less, those with a primary school education are defined as cognitive dysfunction with a score of 21 points or less, and those with no education are defined as a cognitive dysfunction with a score of 16 points. The following is defined as cognitive impairment.

The correlation between the number of correct answers of reaction force (RT-CN), the number of correct answers of memory (ME-CN) and the predicted score of MMSE is verified by the following experimental and statistical analysis methods.

The human-computer interaction game module in this experiment is the Lohas intelligent training machine (SODA, idea bus) developed by Dianjing Technology Company. The test method is a 12-minute human-computer interactive game with a 16-grid keyboard device. The sequence of the test is 2-minute reaction force, 2-minute attention, and 2-minute coordination Power, and four-minute memory, rest 30 seconds in each test. During the test, the testee adopts a sitting posture for testing, and the testee can only perform finger movements during the test, so as to avoid excessive body movements from affecting the test results. The human-computer interaction game module calculates and records the performance parameters such as the number of correct answers, the correct rate, and the reaction time for correct answers produced by the subject when performing each game test. There were 51 subjects in this experiment (17 males and 34 females), with an average age of 79.92±8.68 years old and an average score of 23.33±6.61 on the MMSE scale.

The data collected in this study were firstly deleted with missing values and filtered out extreme values with twice the standard deviation, and then analyzed with IBM SPSS Statistics 23 statistical software. The basic data of the subjects were analyzed with descriptive statistics, and the reliability analysis of the human-computer interaction game was conducted with the dependent sample t-test. The correlation between the testee's game performance parameters and MMSE predicted scores was analyzed by Pearson product difference correlation. Multiple regression analysis was used to analyze the predictability of the testee's game performance parameters and MMSE prediction scores. Risk prediction of cognitive impairment by most game performance parameters analyzed using two logistic regression analyzes. The statistical significance level was set at α=.05.

Descriptive statistical analysis shows that the game performance parameters of reaction, attention, and coordination have similar performances, which are relatively easy items for the testees, while the correct answer rate of memory is low, and the reaction time is relatively low. The variation is also large, indicating that it is a more difficult project for the subjects. The correct answer rate of the four indicators of reaction, attention, coordination and memory has a significant positive correlation with the MMSE predicted score, while the correct answer reaction time of the four indicators is significantly negatively correlated with the MMSE predicted score. Between moderate and high correlation.

In order to avoid collinearity problems from affecting the regression analysis results of this experiment, this experiment adopts a two-stage method for analysis. First, the three parameters of the four indicators (number of correct answers, correct rate, and response time to correct answers) are used as predictive variables, and the MMSE prediction score is used as a dependent variable to find out the parameters with predictive power in each indicator. The results showed that the parameter with predictive power was responsiveness The number of questions, the number of correct attention questions, the reaction time of coordination, the correct rate of coordination, the number of correct answers of memory, and the reaction time of correct answers of memory.

These six parameters are used as predictive variables, and considering that the number of genders in this experiment is not equal, in order to avoid affecting the prediction results, gender is included in the predictive variables for stepwise regression. The results showed that the number of correct answers in the first response (RT-CN) and the number of correct answers in memory (ME-CN) had significant positive predictive power on the MMSE prediction score, and the number of correct answers in response (RT-CN) had a greater impact , and the number of correct answers for memory (ME-CN) is next, and the fitting equation of the regression model is the equation shown above.

Claims (2)

A computer program product for predicting scores of the Mini-Mental Examination (MMSE) by man-machine interactive games, loaded into a computer device; including: a man-machine interactive game module, the man-machine interactive game module provides a game in a palace format, the The palace-style game includes several grids arranged in a matrix, and a light can be generated in each grid; the human-computer interaction game module is used for a subject to perform a game test; the game test is based on reaction force, attention The four intelligent items of power, coordination and memory produce several questions respectively, and each item of the game test has a limited time; The number of correct answers, correct rate, and correct answer reaction time performance parameters; it is characterized in that: the computer program product further includes a simple mental scale score prediction module, which receives all the scores transmitted by the human-computer interaction game module Parameters, extract the number of correct answers for the reaction force and the number of correct answers for the memory, and the subject’s gender as parameters to perform the calculation of the following formula to obtain the predicted score of the subject’s Simple Mental Scale; MMSE=-4.432 +0.613 * RT-CN+0.388 * ME-CN+0.197 * Gender Among them, RT-CN is the number of correct answers in the reaction game test; ME-CN is the number of correct answers in the memory game test; MMSE is the simple mental Scale prediction score; Gender is the gender of the subject, female is the number 0, and male is the number 1. As described in claim item 1, the computer program product predicting the Minimal Mental Scale (MMSE) score with a human-computer interaction game, wherein the limited time of the reaction game test is two minutes, and the limited time of the attention game test is 2 minutes. Two minutes, the limited time of the coordination game test is two minutes, the limited time of the memory game test is four minutes, and the interval time of the various game tests is 30 seconds.

Images