CN116344042B - Cognitive reserve intervention lifting method and system based on multi-modal analysis - Google Patents
Cognitive reserve intervention lifting method and system based on multi-modal analysis Download PDFInfo
- Publication number
- CN116344042B CN116344042B CN202310634247.8A CN202310634247A CN116344042B CN 116344042 B CN116344042 B CN 116344042B CN 202310634247 A CN202310634247 A CN 202310634247A CN 116344042 B CN116344042 B CN 116344042B
- Authority
- CN
- China
- Prior art keywords
- cognitive
- cognitive reserve
- user
- reserve
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000001149 cognitive effect Effects 0.000 title claims abstract description 238
- 238000000034 method Methods 0.000 title claims abstract description 47
- 238000004458 analytical method Methods 0.000 title claims abstract description 24
- 238000003745 diagnosis Methods 0.000 claims abstract description 84
- 238000012549 training Methods 0.000 claims abstract description 77
- 230000009467 reduction Effects 0.000 claims abstract description 8
- 238000011156 evaluation Methods 0.000 claims description 26
- 238000009826 distribution Methods 0.000 claims description 24
- 238000012545 processing Methods 0.000 claims description 19
- 238000012552 review Methods 0.000 claims description 18
- 239000011159 matrix material Substances 0.000 claims description 10
- 230000003925 brain function Effects 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 6
- 238000006243 chemical reaction Methods 0.000 claims description 4
- 238000000354 decomposition reaction Methods 0.000 claims description 3
- 230000006735 deficit Effects 0.000 claims description 3
- 230000000284 resting effect Effects 0.000 claims description 3
- 230000006870 function Effects 0.000 description 13
- 230000008569 process Effects 0.000 description 11
- 230000003920 cognitive function Effects 0.000 description 8
- 206010012289 Dementia Diseases 0.000 description 6
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 6
- 239000008280 blood Substances 0.000 description 6
- 210000004369 blood Anatomy 0.000 description 6
- 210000004556 brain Anatomy 0.000 description 6
- 230000004927 fusion Effects 0.000 description 6
- 238000005259 measurement Methods 0.000 description 6
- 229910052760 oxygen Inorganic materials 0.000 description 6
- 239000001301 oxygen Substances 0.000 description 6
- 238000010276 construction Methods 0.000 description 5
- 230000029058 respiratory gaseous exchange Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000002372 labelling Methods 0.000 description 4
- 238000005481 NMR spectroscopy Methods 0.000 description 3
- 239000000090 biomarker Substances 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000001537 neural effect Effects 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 208000024827 Alzheimer disease Diseases 0.000 description 2
- 208000028698 Cognitive impairment Diseases 0.000 description 2
- 230000009471 action Effects 0.000 description 2
- 208000010877 cognitive disease Diseases 0.000 description 2
- 238000001816 cooling Methods 0.000 description 2
- 230000006378 damage Effects 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 208000015122 neurodegenerative disease Diseases 0.000 description 2
- 230000007171 neuropathology Effects 0.000 description 2
- 230000035484 reaction time Effects 0.000 description 2
- 206010003694 Atrophy Diseases 0.000 description 1
- 208000024806 Brain atrophy Diseases 0.000 description 1
- 238000000342 Monte Carlo simulation Methods 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000008485 antagonism Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000037444 atrophy Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 208000029028 brain injury Diseases 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000002490 cerebral effect Effects 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 230000019771 cognition Effects 0.000 description 1
- 230000003931 cognitive performance Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000002599 functional magnetic resonance imaging Methods 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000003340 mental effect Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000004770 neurodegeneration Effects 0.000 description 1
- 238000013421 nuclear magnetic resonance imaging Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000002792 vascular Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0475—Generative networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/094—Adversarial learning
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Molecular Biology (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Epidemiology (AREA)
- Pathology (AREA)
- Primary Health Care (AREA)
- Databases & Information Systems (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses a cognitive reserve intervention lifting method and system based on multi-modal analysis. The method comprises the following steps: performing cognitive reserve diagnosis on a user based on a preset cognitive reserve diagnosis model; pushing a cognitive reserve training scheme to the user according to the cognitive reserve diagnosis result of the user; after a preset period, carrying out cognitive reserve re-diagnosis on the user; wherein, the cognitive reserve diagnostic model is constructed by the steps of: acquiring multi-mode data of a user; carrying out noise reduction treatment on the multi-mode data; data enhancement is performed on the multi-modal data to learn the original data set against the network according to the generation type and generate a new data set; constructing a preliminary cognitive reserve diagnostic model according to the new data set and the priori knowledge; and iteratively updating the preliminary cognitive reserve diagnostic model through back propagation learning to generate an optimal cognitive reserve diagnostic model.
Description
Technical Field
The invention relates to a cognitive reserve intervention and lifting method based on multi-modal analysis, and also relates to a corresponding cognitive reserve intervention and lifting system, belonging to the field of medical care informatics.
Background
Alzheimer's disease is associated with brain atrophy, progressive cognitive impairment and final dementia. However, the onset and level of cognitive impairment associated with the extent of atrophy varies greatly from patient to patient. The concept of cognitive stores (Cognition Reserve, abbreviated CR) can explain this inter-individual heterogeneity, which describes the ability to maintain cognitive function in the presence of neuropathology. However, to date, no directly measured scale or other technical means has been able to directly reflect differences in cognitive reserve in individuals. In existing researches or applications, indexes such as education degree, pre-onset IQ, professional achievement and the like are often used for reflecting differences of cognitive stores of individuals.
Educational level is one of the most widely used cognitive reserve predictors. Research shows that education can reduce the influence of neurodegenerative diseases such as dementia on the cognitive function of individuals by forming an effective neuropathology foundation. Studies have also shown that education improves cognitive function by creating new cognitive strategies, thereby reducing the risk of dementia.
Pre-onset IQ is also one of the indicators that effectively predicts cognitive reserves. The mental stimulus may cope with the influence of cerebral neurodegenerative disorders and cerebrovascular damage on cognitive function by some function compensation mechanism, and thereby delay the occurrence of Alzheimer's disease.
Professional achievement is of great importance in predicting cognitive reserve in an individual and analyzing the individual's risk of dementia. Research shows that high-level professions such as academic experts, enterprise management layers and the like can protect cognitive functions and reduce the probability of individuals suffering from dementia.
At present, the existing measurement mode of the cognitive reserve is relatively single, objective physiological indexes are lacking, and development of brain electricity and nuclear magnetic function imaging technology enables measurement and calculation of the cognitive reserve to be included in the biological marker indexes for analysis, so that measurement granularity of the cognitive reserve is further refined.
Disclosure of Invention
The primary technical problem to be solved by the invention is to provide a cognitive reserve intervention and lifting method based on multi-modal analysis.
The invention aims to provide a cognitive reserve intervention and lifting system based on multi-modal analysis.
In order to achieve the technical purpose, the invention adopts the following technical scheme:
according to a first aspect of an embodiment of the present invention, there is provided a cognitive reserve intervention promotion method based on multimodal analysis, including the steps of:
performing cognitive reserve diagnosis on a user based on a preset cognitive reserve diagnosis model to obtain a cognitive reserve diagnosis result of the user;
pushing a cognitive reserve training scheme to the user according to the cognitive reserve diagnosis result of the user;
after the cognitive reserve training of a preset period, performing cognitive reserve re-diagnosis on the user;
wherein the cognitive reserve diagnostic model is constructed by:
acquiring multi-mode data of a user; wherein the multi-modal data includes at least: demographic data, IQ and nuclear magnetic functional image data, and cognitive assessment data;
carrying out noise reduction treatment on the multi-mode data;
performing data enhancement on the multi-modal data to learn the original data set against the network according to a generation type and generate a new data set;
constructing a preliminary cognitive reserve diagnosis model according to the new data set and the priori knowledge, and outputting a cognitive reserve predicted value of a user;
and iteratively updating the cognitive reserve diagnostic model through back propagation learning to generate an optimal cognitive reserve diagnostic model.
Wherein preferably, the cognitive reserve intervention and promotion method further comprises the following steps:
acquiring diagnosis and treatment times, training frequency and training interval duration of a user;
if the diagnosis and treatment times are 1, pushing a cognitive reserve diagnosis scheme to the user so as to perform cognitive reserve diagnosis on the user based on a preset cognitive reserve diagnosis model;
if the diagnosis and treatment times are greater than 1 and the training frequency and the training interval duration meet a first preset condition, pushing a cognitive reserve training scheme to the user;
and if the diagnosis and treatment times are greater than 1 and the training frequency and the training interval duration meet the second preset condition, pushing a cognitive reserve review scheme to the user.
Preferably, the noise reduction processing is performed on the multi-mode data, which specifically includes:
decomposing the multi-modal data into two components, a coarse scale and a fine scale; wherein,
on a rough scale, a low-rank and sparse component decomposition method is adopted, coefficient constraint is applied to the multi-modal data, and the overall structural information of the image is mined by utilizing the internal association relation of the multi-modal data according to the local correlation of the multi-modal data so as to realize information growth and noise removal of the multi-modal data;
at a fine scale, data details are recovered from the resting brain function image and the data of the functional brain function image impairment.
Preferably, the multi-modal data is subjected to data enhancement so as to learn the original data set and generate a new data set according to the generation type antagonism network, which specifically comprises the following steps:
learning, by a generator, a potential law or distribution of the multimodal data to generate new data;
judging the true or false of the new data through a discriminator;
if the judging result of the judging device is false, adjusting the parameters of the generator to regenerate new data until the judging result of the judging device is true;
wherein the generator and the arbiter together form the generative countermeasure network.
Wherein preferably, the cognitive reserve diagnostic model is constructed by:
forming a first input by sampling the prior distribution for potential variables; wherein the a priori distribution is formed based on the a priori knowledge;
forming a second input by sampling the posterior distribution for the latent variable; wherein the posterior distribution is formed based on the new dataset;
inputting the first input and the second input into a generator network together to output a predicted value CR of the cognitive reserve, thereby constructing a preliminary cognitive reserve diagnostic model;
where cr=pω (y|x, z), ω is a parameter set of the generator network, P (y|x) is a model conditional probability distribution, z is a latent variable representing the inherent uncertainty of the salient region.
Wherein preferably, the initial training level of the cognitive reserve training regimen is obtained by:
acquiring education degrees, professional attributes, cognitive assessment results and cognitive reserve diagnosis results of a user;
acquiring a first weight wid of educational degrees and professional attributes of the user on a cognitive reserve CR;
acquiring a second weight wuk of the cognitive assessment result of the user on the cognitive reserve CR;
calculating an initial training grade S of the cognitive reserve training scheme based on initial difficulty CR100 of a default cognitive reserve task, a first weight wid and a second weight wuk of the system; where s=cr100×wid Wuk.
Wherein preferably, the first weight wid is obtained by:
constructing a demographic attribute weight matrix E multiplied by O multiplied by CR based on the education level ei, the professional attribute oi and the cognitive reserve CR of the user;
carrying out standardization processing on the demographic attribute weight matrix;
acquiring a first weight wid of educational degrees and professional attributes of the user on a cognitive reserve CR based on a standardized demographic attribute weight matrix; wherein, wid E (e×o×cr).
Wherein preferably, the second weight wuk is obtained by:
the user sequentially completes the evaluation of n preset comprehensive evaluation scales and generates original scores X1, X2 and X3..
Performing standardized conversion on the original scores in the comprehensive evaluation table according to the comprehensive normal modulus parameters of the healthy user, and generating a second weight wuk of the cognitive evaluation result of the user on the cognitive reserve CR, wherein:
Wuk＝Wmn*{100－10*（Xi－i）/σi}
wherein, the value range of i is 1-n; xi represents the original score in the comprehensive evaluation scale; i represents the average value of the original scores of comprehensive evaluation scales of healthy people matched with the ages, sexes, professions and education degrees of the users; σi represents the standard deviation of the original score of the comprehensive evaluation scale of healthy people matched with the age, sex, occupation and education degree of the user; wmn represents the weight of the scale lm e L to the capacity kn e K; l represents a set of scales; k represents a set of weights.
According to a second aspect of embodiments of the present invention, there is provided a cognitive reserve intervention boost system based on multimodal analysis, comprising:
the data acquisition equipment at least comprises a cognitive reserve diagnosis acquisition unit, a cognitive reserve training acquisition unit and a cognitive reserve review acquisition unit, so as to acquire multi-mode data of a user in a cognitive reserve diagnosis scene, a cognitive reserve training scene and a cognitive reserve review scene respectively;
the central processing unit is connected with the data acquisition equipment and is provided with the cognitive reserve diagnosis model in advance, and the central processing unit is used for carrying out cognitive reserve diagnosis on a user in a cognitive reserve diagnosis scene;
and the display is connected with the central processing unit and used for displaying the cognitive reserve diagnosis result or the cognitive reserve re-diagnosis result.
Compared with the prior art, the invention has the following technical characteristics:
1. through gathering multimode data such as educational level, IQ and nuclear magnetic function image, cognitive training action to can carry out more accurate, objectified measurement to user's cognitive reserve.
2. The cognitive reserve diagnostic model is constructed by using a knowledge and data hybrid driven model, and a model system with priori features and posterior optimization iteration is constructed according to the essential features of the cognitive reserve. The system comprises a data noise reduction, data enhancement and fusion model construction process, and comprehensively applies a multi-mode fusion model to realize the construction of a diagnosis model. In addition, the data of different modes often contain different characteristics, when the multi-mode data processing is performed, the characteristics and knowledge of the data can be considered to be introduced besides the multi-mode data, a knowledge and data mixed driving model is built, and the performance and the interpretability of the model are enhanced.
3. Three different diagnosis and treatment scenes are constructed in advance, and according to diagnosis and treatment times, training frequency and interval duration information of a user using a system, a use logic is constructed, and multi-scene division is carried out on cognitive reserve diagnosis and treatment of the user, so that a closed-loop cognitive reserve intervention scheme is formed. And when the cognitive reserve training scheme is pushed, the initial training level of the user is determined by integrating the demographic attributes and the cognitive level evaluation data of the user.
4. The corresponding cognitive reserve intervention lifting system is a comprehensive diagnosis and treatment equipment system integrating software and hardware. In terms of software, complete implementation logic is constructed, and the system is ensured to be capable of realizing effective diagnosis and accurate intervention of cognitive reserve. In terms of hardware, different neurophysiologic index acquisition frames are built according to the needs of different scenes, and the effective cognitive reserve level feedback of an individual at an effective time node is ensured by taking scientific, efficient, economical and convenient principles as a general principle.
Drawings
FIG. 1 is a flowchart of a cognitive reserve intervention promotion method based on multi-modal analysis according to a first embodiment of the present invention;
FIG. 2 is a diagram showing the composition of multi-modal data in a first embodiment of the present invention;
FIG. 3 is a schematic diagram of a generated countermeasure network according to a first embodiment of the present invention;
FIG. 4 is a schematic diagram of a preliminary cognitive reserve diagnostic model in a first embodiment of the invention;
FIG. 5 is a flowchart of a cognitive reserve intervention boost method based on multi-modal analysis according to a second embodiment of the present invention;
FIG. 6 is a block diagram of a cognitive reserve intervention boost system based on multi-modal analysis according to a third embodiment of the present invention;
fig. 7 is a flowchart of a cognitive reserve intervention boost system based on multi-modal analysis according to a third embodiment of the present invention.
Detailed Description
The technical contents of the present invention will be described in detail with reference to the accompanying drawings and specific examples.
First embodiment:
as shown in fig. 1, the cognitive reserve intervention and promotion method based on multi-modal analysis provided by the first embodiment of the present invention specifically includes steps S1 to S5:
s1: and acquiring multi-mode data of the user.
It can be appreciated that the efficient definition of the connotation and the accurate acquisition of the corresponding data of cognitive reserves as a composite evaluation index for indirect observation are important items. The cognitive stores are defined in this example as follows, based on the results of the comprehensive diversity analysis:
cognitive reserve is the ability of a user to achieve cognitive compensation for brain injury using a brain adaptive neural network. Specifically, the cognitive reserve is a comprehensive representation of multidimensional information such as individual physiological structures, cognitive functions, performance and the like, and the cognitive reserve maintains better cognitive functions and normal performance by affecting the level of brain biomarkers and regulating brain activation patterns.
Thus, based on the above definition, the present embodiment further specifies the multi-modal data types required for the cognitive reserve measurement, including but not limited to the following categories:
(1) diagnostic data variables, such as: clinical diagnosis of dementia lineages and vascular lineages;
(2) biomarker physical data variables, such as: neurophysiologic data such as respiration, blood oxygen, nuclear magnetic resonance fMRI, and the like;
(3) behavior characteristic data variables, such as: cognitive function data such as attention span, situational memory capacity and the like of a user;
(4) demographic information data variables, such as: educational level, professional attributes, etc. of the user.
Thus, referring to fig. 2, on the basis of the explicit data type, the acquisition professionals and implementation feasibility are comprehensively evaluated, and the embodiment explicitly acquires the multi-modal data dimension.
S2: and carrying out noise reduction treatment on the multi-mode data.
In the process of model construction, quality damage caused by noise pollution to any party of data can influence the final result, so that the processing of the noisy multi-source data is an essential link before modeling. In this embodiment, a noise reduction method of a low rank matrix is used to perform noise processing on multi-mode data.
Specifically, the method comprises the steps S21 to S23:
s21: decomposing the multi-modal data into two components, a coarse scale and a fine scale;
s22: on a rough scale, a low-rank and sparse component decomposition method is adopted, coefficient constraint is applied to the multi-modal data, and the overall structural information of the image is mined by utilizing the internal association relation of the multi-modal data according to the local correlation of the multi-modal data so as to realize information growth and noise removal of the multi-modal data;
s23: at a fine scale, data details are recovered from the resting brain function image and the data of the functional brain function image impairment.
S3: and carrying out data enhancement on the multi-mode data.
It will be appreciated that a sample set of sufficient volume may not always be collected due to the scarcity or excessive acquisition costs of some samples. In the case of a sufficient number of samples, the model performance may be affected due to poor data quality, resulting in a model being in an overfitted state, and thus a significant decrease in generalization ability of the model.
The data enhancement is a method for increasing the abundance of data, can effectively improve the generalization capability of a model and the capability of solving class imbalance, and aims to expand the number of the existing limited data and generate more diversified data.
In this embodiment, a generative countermeasure network (generative adversarial network, abbreviated GAN) is used to automatically learn the representation of the original dataset and generate a new "real" dataset to improve the number and quality of the original dataset, thereby improving the generalization performance of the model and reducing the overfitting in the training process.
Referring to fig. 3, the generated countermeasure network is mainly composed of a generator and a discriminator, and the generator learns the potential rules or distributions of the multi-modal data to generate new data. Then, the discriminator judges whether the new data is true or false. If the judging result of the judging device is false, the parameters of the generator are adjusted to regenerate new data until the judging result of the judging device is true. Thus, the generator and the arbiter conduct resistance training in one zero and game, and the parameters are continuously adjusted by the mutual resistance, so that the two networks are promoted to improve performance during training. In an ideal case, the dynamic process will eventually reach an equilibrium state, and the generator will generate new data by learning the potential distribution of the target data, and the discriminator cannot determine the authenticity of the new data, thereby achieving the effect of spurious.
S4: and (3) carrying out multi-mode data fusion analysis, and constructing a primary cognitive reserve diagnosis model.
Referring to fig. 4, the method specifically includes steps S41 to S43:
s41: forming a first input by sampling the prior distribution for potential variables; wherein the prior distribution is formed based on prior knowledge;
in this embodiment, the latent variables refer to: variable factors that can affect cognitive reserve and cannot be expressed directly with existing knowledge, such as: personality traits, nationality, professional attributes, and so forth.
S42: forming a second input by sampling the posterior distribution for the latent variable; wherein the posterior distribution is formed based on the new dataset;
s43: inputting the first input and the second input into a generator network together to output a predicted value CR of the cognitive reserve, thereby constructing a preliminary cognitive reserve diagnostic model;
where cr=pω (y|x, z), ω is a parameter set of the generator network, P (y|x) is a model conditional probability distribution, z is a latent variable representing the inherent uncertainty of the salient region.
It can be appreciated that in this embodiment, the fusion analysis of a priori knowledge (e.g., expert knowledge) and data is implemented by using an active learning method. The execution of the sample learning and weak supervision learning related algorithm flow does not need to interact with human, the automatic labeling of the unlabeled sample is realized by utilizing a network with a certain classification performance obtained by training the existing labeling data, and the unlabeled data is utilized based on the self so as to improve the generalization capability of the model. While active learning is focused on how to use as few samples as possible to maintain performance comparable to classifiers trained with a large number of samples. More specifically, the goal of active learning is to select the sample from the unlabeled sample set that has the greatest value in enhancing the performance of the model and give it to the relevant expert for labeling, to minimize sample labeling costs while maintaining model performance as much as possible.
The method comprises the steps of acquiring 'difficult' sample data which is helpful for improving the model performance but is prone to error division of a classifier by means of a correlation algorithm of machine learning or deep learning, giving related experts to review, confirm or mark again, then training the valuable data marked by the experts obtained in the previous step again by adopting a supervised or semi-supervised learning algorithm, and gradually improving the model performance by repeating the process, so that the experience of the experts is integrated into the learning of the model.
S5: and (5) iteratively updating to generate an optimal cognitive reserve diagnosis model.
Specifically, the method comprises the steps S51 to S54:
s51: constructing an inference model, and mapping the continuous latent variable z into a prediction function Y of the multivariable input X; wherein the inference model is defined as:
z ∼ P（z）＝N（0，I），Y＝fω（X，z）＋ε，ε∼ N（0，diag（σ）2）
wherein X represents a multimodal variable; y represents a preliminary model predictive value based on multivariate X; z is a latent variable generated by model prediction; omega represents a parameter vector generated by the model; p (z) is the a priori distribution of z; the conditional distribution of Y for a given X is pω (y|x) = R P (z) pω (y|x, z) dz and integrates the latent variable z; pθ (z|x) represents a priori distribution of the cognitive reserve model; pω (z|x, Y) ABP represents posterior distribution of cognitive reserve model; n () represents a data distribution; n (0, diag (σ) 2) represents a normal distribution with diag (σ) 2 as the variance.
S52: evaluating model parameters by adopting a loss function;
wherein the loss function is LABP= = ≡
Where i represents a specific sample and n represents the total number of samples.
S53: performing iterative optimization on the preliminary cognitive reserve diagnostic model by adopting a gradient descent method to obtain parameters for minimizing a loss function;
wherein the desired term EP ω (z|x, Y) is approximated as a sample taken from P ω (z|x, Y) and a monte carlo average is calculated; θ represents likelihood function estimation parameters; θω represents the prediction likelihood function estimation parameter.
S54: iteration is carried out on the potential variable z by adopting a gradient-based Monte Carlo method so as to generate an optimal cognitive reserve diagnosis model;
z t＋1 ＝z t ＋ ＋sN（0，I d ）。
wherein z is t＋1 Representing the new potential variable; z t Representing the existing latent variable; s represents a generator predicted value; i d Representing training dataset variances.
Thus, in this embodiment, the preliminary cognitive reserve diagnostic model is iteratively updated by means of back propagation learning, so as to generate an optimal cognitive reserve diagnostic model.
Second embodiment:
referring to fig. 5, a second embodiment of the present invention provides a cognitive reserve intervention improving method based on multi-modal analysis based on the first embodiment. In this embodiment, the diagnosis and treatment times ci, training frequency fi and training interval time ti (in month units in this embodiment) of the user are analyzed to improve the cognitive reserve of the user. Specifically, the method comprises the steps S10 to S20:
s10: and obtaining a cognitive reserve diagnosis result of the user.
Specifically, if the diagnosis and treatment frequency ci is 1 (i.e., the user does not go through the diagnosis and treatment of cognitive reserve), pushing a cognitive reserve diagnosis scheme to the user so as to perform the diagnosis of cognitive reserve on the user based on the cognitive reserve diagnosis model, thereby obtaining the result of cognitive reserve of the user. Wherein the cognitive reserve diagnostic model is constructed by the method of the first embodiment described above.
S20: and pushing a cognitive reserve training scheme to the user according to the cognitive reserve result of the user.
If the diagnosis and treatment times ci are greater than 1, and the training frequency fi and the training interval time ti meet a first preset condition, a cognitive reserve training scheme is pushed to a user.
Specifically, in the present embodiment, when ci > 1, w=cimod 30 is calculated;
if w is more than 0, pushing the cognitive training scene scheme;
if w=0 & fi is less than or equal to 30 & ti is less than or equal to 3, pushing the cognitive training scene scheme.
It can be understood that the first preset condition in this embodiment is: the user training times ci is not a multiple of 30; alternatively, the number of user exercises ci is a multiple of 30, but the frequency of exercises is no more than 30 and the duration of the exercises is no more than 3 months.
In addition, in the step S20, when the cognitive reserve training scheme needs to be pushed to the user, an initial training level of the cognitive reserve training scheme needs to be determined first, and specifically includes steps S201 to S204:
s201: and obtaining education degrees, professional attributes, cognitive assessment results and cognitive reserve diagnosis results of the user.
S202: a first weight wid of the user's educational level and professional attributes on the cognitive reserve CR is obtained.
Specifically, a demographic attribute weight matrix e×o×cr is constructed based on the educational level ei, professional attribute oi, and cognitive reserve CR of the user. The demographic attribute weight matrix is then normalized. Finally, acquiring a first weight wid of educational degree and professional attribute of the user on the cognitive reserve CR based on the standardized demographic attribute weight matrix; wherein, wid E (e×o×cr).
S203: and acquiring a second weight wuk of the cognitive assessment result of the user on the cognitive reserve CR.
First, the user sequentially completes evaluation of n comprehensive evaluation amounts set in advance, and generates original scores X1, X2, X3.. Then, the original scores in the comprehensive evaluation scale are subjected to standardized conversion according to the comprehensive normal modulus parameters of the healthy user, and a second weight wuk of the cognitive evaluation result of the user to the cognitive reserve CR is generated, wherein the formula is as follows:
Wuk＝Wmn*{100－10*（Xi－i）/σi}
wherein, the value range of i is 1-n; xi represents the original score in the comprehensive evaluation scale; i represents the average value of the original scores of comprehensive evaluation scales of healthy people matched with the ages, sexes, professions and education degrees of the users; σi represents the standard deviation of the original score of the comprehensive evaluation scale of healthy people matched with the age, sex, occupation and education degree of the user; wmn represents the weight of the scale lm e L to the capacity kn e K; l represents a set of scales; k represents a set of weights.
S204: calculating an initial training grade S of a cognitive reserve training scheme based on initial difficulty CR100 of a cognitive reserve task defaulted by a system, a first weight wid and a second weight wuk; where s=cr100×wid Wuk.
S30: after the cognitive reserve training in a preset period, a cognitive reserve review scheme is pushed to a user.
And if the diagnosis and treatment times are greater than 1 and the training frequency and the training interval duration meet a second preset condition, pushing a cognitive reserve review scheme to the user.
When ci > 1, w=cimod 30 is calculated;
if w=0 & fi >30 & ti >3, pushing the cognitive reserve review judgment scene scheme.
It can be understood that the second preset condition in this embodiment is: the number of user training ci is a multiple of 30, and the training frequency is higher than 30 times and the training interval time is longer than 3 months.
Third embodiment:
referring to fig. 6, in addition to the second embodiment, a cognitive reserve intervention promoting system based on multi-modal analysis is further provided in a third embodiment of the present invention, which includes a data acquisition device 10, a central processing unit 20 and a display 30. The data acquisition device 10 is used for acquiring multi-mode data, and the central processing unit 20 is used for carrying out data processing on the multi-mode data so as to output a specific result of an intervention scheme; the display 30 is used for visual display to display specific results of the intervention program.
Specifically, the data acquisition device 10 includes at least a cognitive reserve diagnosis acquisition unit 101, a cognitive reserve training acquisition unit 102, and a cognitive reserve review acquisition unit 103. The cognitive reserve diagnosis acquisition unit 101 may include a nuclear magnetic resonance imager, a respiration acquisition instrument, a blood oxygen acquisition instrument, and the like, and is configured to acquire multi-modal data of a user in a cognitive reserve diagnosis scene. The cognitive reserve training acquisition unit 102 may include a breath acquisition instrument, a blood oxygen acquisition instrument, a response time recorder, etc. for acquiring multi-modal data of a user in a cognitive reserve review scenario. The cognitive reserve review acquisition unit 103 may include a breath acquisition instrument, a blood oxygen acquisition instrument, a video acquisition device, a voice acquisition device, and the like, for acquiring multi-modal data of a user in a cognitive reserve training scenario.
It will be appreciated that the data acquisition device 10 includes a plurality of acquisition instruments therein, and that in different scenarios, the corresponding acquisition instruments participate in data acquisition tasks, such as: the nuclear magnetic resonance imaging instrument only collects brain image data of the user in a cognitive reserve diagnosis scene, and the respiration collecting instrument collects respiration data of the user in three scenes.
In one embodiment of the present invention, a nuclear magnetic resonance imager generally includes a main magnet, a gradient system, a radio frequency system, a spectrometer system, a computer, and auxiliary equipment. The main magnet uses a permanent magnet device with a field strength of 1T or below; the gradient system comprises a gradient coil, a gradient controller, a digital-to-analog converter, a gradient amplifier, a gradient cooling system and the like; the radio frequency system comprises a radio frequency coil, a radio frequency generator and a receiver; the spectrometer system realizes the functions of the emission time sequence of the radio frequency, the matched application time sequence of the gradient and the like; the computer and auxiliary facilities include main control computer, image display, inspection bed and radio frequency shielding, magnetic shielding, UPS power supply, cooling system, etc. and the functions of the computer and auxiliary facilities are to ensure that the process from the start of inspection to the acquisition of MR image can be orderly and accurately performed.
The respiration acquisition instrument is a piezoresistor sensor. The blood oxygen collector uses blood oxygen finger clamps, one end of which is connected to the device, and the other end is clamped on the fingers of a person. The video capture device uses a notebook/desktop self-contained camera, requiring a capture resolution of no less than 1024 x 768 pixels.
The voice acquisition device is provided with a microphone of a compatible sound card, can acquire data of an audio frequency band of not lower than 20 HZ-100 HZ, and generates files in a common audio format such as WAV.
The reaction time recorder has the functions of recording the total time, reaction time, movement completion time, total clicking times and the like required by the project, and can record millisecond time of 0.001-9.999 seconds in the reaction
The central processing unit 20 is connected with the data acquisition device 10 and preset with a cognitive reserve diagnostic model constructed by the method of the first embodiment for performing a cognitive reserve diagnosis on a user in a cognitive reserve diagnostic scenario. Moreover, the central processor 20 can also push a cognitive reserve training scheme to the user in a training scenario based on the cognitive reserve intervention and lifting method described in the second embodiment, so as to form a diagnosis and treatment analysis report; and on the other hand, carrying out cognitive reserve review on the user in a review scene to form effect evaluation feedback or review push suggestions.
The display 30 is connected to the central processing unit 20, and is used for displaying the cognitive reserve diagnosis result or the cognitive reserve re-diagnosis result.
As shown in fig. 7, the specific working process of the above cognitive reserve intervention and lifting system is as follows:
s100: after the user logs in, the system judges whether the user logs in for the first time. If the user does not log in for the first time, directly calling the last record report; if the user logs in for the first time, demographic information is filled in.
S200: the user wears the wearable device according to the system prompt.
S300: the system judges whether the user needs to perform cognitive reserve diagnosis, if so, the system enters a multi-mode data acquisition process to complete data acquisition processes of brain image data, cognitive evaluation data and the like; after diagnosing the user based on a cognitive reserve diagnosis model preset by the system, further judging whether the user needs intervention, namely: whether cognitive reserve training is required.
If the user is a historical user and does not need to perform cognitive reserve diagnosis, the system directly judges whether the user needs intervention according to the last record report.
S400: if the system judges that the user does not need to intervene, the system directly ends, if the system judges that the user needs to intervene, the cognitive reserve training scheme is pushed to the user, and after the user finishes the cognitive reserve training scheme, the system generates a report to display a result to the user, so that the cognitive reserve intervention process is finished.
S500: after the user is subjected to the cognitive reserve training of a preset period, the system can be used for carrying out re-diagnosis on the user so as to feed back the cognitive reserve training effect to the user and push re-diagnosis suggestions to the user.
In summary, the cognitive reserve intervention lifting method and system based on multi-modal analysis provided by the embodiment of the invention have the following beneficial effects:
1. through gathering multimode data such as educational level, IQ and nuclear magnetic function image, cognitive training action to can carry out more accurate, objectified measurement to user's cognitive reserve.
2. The cognitive reserve diagnostic model is constructed by using a knowledge and data hybrid driven model, and a model system with priori features and posterior optimization iteration is constructed according to the essential features of the cognitive reserve. The system comprises a data noise reduction, data enhancement and fusion model construction process, and comprehensively applies a multi-mode fusion model to realize the construction of a diagnosis model. In addition, the data of different modes often contain different characteristics, when the multi-mode data processing is performed, the characteristics and knowledge of the data can be considered to be introduced besides the multi-mode data, a knowledge and data mixed driving model is built, and the performance and the interpretability of the model are enhanced.
3. Three different diagnosis and treatment scenes are constructed in advance, and according to diagnosis and treatment times, training frequency and interval duration information of a user using a system, a use logic is constructed, and multi-scene division is carried out on cognitive reserve diagnosis and treatment of the user, so that a closed-loop cognitive reserve intervention scheme is formed. And when the cognitive reserve training scheme is pushed, the initial training level of the user is determined by integrating the demographic attributes and the cognitive level evaluation data of the user.
4. The corresponding cognitive reserve intervention lifting system is a comprehensive diagnosis and treatment equipment system integrating software and hardware. In terms of software, complete implementation logic is constructed, and the system is ensured to be capable of realizing effective diagnosis and accurate intervention of cognitive reserve. In terms of hardware, different neurophysiologic index acquisition frames are built according to the needs of different scenes, and the effective cognitive reserve level feedback of an individual at an effective time node is ensured by taking scientific, efficient, economical and convenient principles as a general principle.
The cognitive reserve intervention lifting method and the system based on the multi-modal analysis provided by the invention are described in detail. Any obvious modifications to the present invention, without departing from the spirit thereof, would constitute an infringement of the patent rights of the invention and would take on corresponding legal liabilities.
Claims (9)
1. A cognitive reserve intervention lifting method based on multi-modal analysis is characterized by comprising the following steps:
performing cognitive reserve diagnosis on a user based on a preset cognitive reserve diagnosis model to obtain a cognitive reserve diagnosis result of the user;
pushing a cognitive reserve training scheme to the user according to the cognitive reserve diagnosis result of the user;
after the cognitive reserve training of a preset period, performing cognitive reserve re-diagnosis on the user;
the preset cognitive reserve diagnostic model is constructed through the following steps:
acquiring multi-mode data of a user; wherein the multi-modal data includes at least: demographic data, IQ and nuclear magnetic functional image data, and cognitive assessment data;
carrying out noise reduction treatment on the multi-mode data;
performing data enhancement on the multi-modal data to learn the original data set against the network according to a generation type and generate a new data set;
constructing a preliminary cognitive reserve diagnosis model according to the new data set and the priori knowledge, and outputting a cognitive reserve predicted value of a user; the method specifically comprises the following steps: forming a first input by sampling the prior distribution for potential variables; wherein the a priori distribution is formed based on the a priori knowledge; forming a second input by sampling the posterior distribution for the latent variable; wherein the posterior distribution is formed based on the new dataset; inputting the first input and the second input into a generator network together to output a predicted value CR of the cognitive reserve, thereby constructing a preliminary cognitive reserve diagnostic model; where cr=pω (y|x, z), ω being a parameter set of the generator network, z representing a latent variable;
and iteratively updating the preliminary cognitive reserve diagnostic model through back propagation learning to generate an optimal cognitive reserve diagnostic model.
2. The cognitive reserve intervention boost method of claim 1, further comprising:
acquiring diagnosis and treatment times, training frequency and training interval duration of a user;
if the diagnosis and treatment times are 1, pushing a cognitive reserve diagnosis scheme to the user so as to perform cognitive reserve diagnosis on the user based on a preset cognitive reserve diagnosis model;
if the diagnosis and treatment times are greater than 1 and the training frequency and the training interval duration meet a first preset condition, pushing a cognitive reserve training scheme to the user;
and if the diagnosis and treatment times are greater than 1 and the training frequency and the training interval duration meet the second preset condition, pushing a cognitive reserve review scheme to the user.
3. The cognitive reserve intervention boost method of claim 1, wherein the multi-modal data is noise reduced, comprising:
decomposing the multi-modal data into two components, a coarse scale and a fine scale; wherein,
on a rough scale, a low-rank and sparse component decomposition method is adopted, coefficient constraint is applied to the multi-modal data, and the overall structural information of the image is mined by utilizing the internal association relation of the multi-modal data according to the local correlation of the multi-modal data so as to realize information growth and noise removal of the multi-modal data;
at a fine scale, data details are recovered from the resting brain function image and the data of the functional brain function image impairment.
4. The cognitive reserve intervention boost method of claim 1, wherein the multi-modal data is data enhanced to learn the original data set against the network and to generate a new data set according to a generative challenge network, comprising:
learning, by a generator, a potential law or distribution of the multimodal data to generate new data;
judging the true or false of the new data through a discriminator;
if the judging result of the judging device is false, adjusting the parameters of the generator to regenerate new data until the judging result of the judging device is true;
wherein the generator and the arbiter together form the generative countermeasure network.
5. The cognitive reserve intervention boost method of claim 1, wherein the initial training level of the cognitive reserve training regimen is obtained by:
acquiring education degrees, professional attributes, cognitive assessment results and cognitive reserve diagnosis results of a user;
acquiring a first weight wid of educational degrees and professional attributes of the user on cognitive reserves;
acquiring a second weight wuk of the cognitive assessment result of the user on the cognitive reserve;
calculating an initial training grade S of the cognitive reserve training scheme based on initial difficulty CR100 of a default cognitive reserve task, a first weight wid and a second weight wuk of the system; wherein,
。
6. the cognitive reserve intervention boost method of claim 5, wherein the first weight wid is obtained by:
constructing a demographic attribute weight matrix E multiplied by O multiplied by CR based on the education level E, the professional attribute O and the predicted value CR of the cognitive reserve of the user;
carrying out standardization processing on the demographic attribute weight matrix;
acquiring a first weight wid of educational degrees and professional attributes of the user on cognitive reserves based on the standardized demographic attribute weight matrix; wherein wid ε (E X O X CR).
7. The cognitive reserve intervention boost method of claim 5, wherein the second weight wuk is obtained by:
the user sequentially completes the evaluation of n preset comprehensive evaluation scales and generates original scores X1, X2 and X3..
Performing standardized conversion on the original scores in the comprehensive evaluation scale according to the comprehensive normal mode parameters of the healthy user to generate a second weight wuk of the cognitive evaluation result of the user on the cognitive reserve; wherein,
wherein Xi represents the original score in the comprehensive evaluation scale, and the value range of i in Xi is 1-n; j represents the average value of the original scores of the comprehensive evaluation scale of healthy people matched with the age, sex, occupation and education degree of the user; σi represents the standard deviation of the original score of the comprehensive evaluation scale of healthy people matched with the age, sex, occupation and education degree of the user; wmn represents the weight of the scale lm to the capacity kn, where lm ε L, kn ε K, L represents the set of scales; k represents a set of weights.
8. A cognitive reserve intervention boost system based on multimodal analysis, comprising:
the data acquisition equipment at least comprises a cognitive reserve diagnosis acquisition unit, a cognitive reserve training acquisition unit and a cognitive reserve review acquisition unit, so as to acquire multi-mode data of a user in a cognitive reserve diagnosis scene, a cognitive reserve training scene and a cognitive reserve review scene respectively;
the central processing unit is connected with the data acquisition equipment and preset with the optimal cognitive reserve diagnosis model according to any one of claims 1-4, and is used for performing cognitive reserve diagnosis on a user in a cognitive reserve diagnosis scene;
and the display is connected with the central processing unit and used for displaying the cognitive reserve diagnosis result or the cognitive reserve re-diagnosis result.
9. The cognitive reserve intervention boost system of claim 8, wherein:
in a cognitive reserve training scene, the central processing unit pushes a cognitive reserve training scheme to a user; and in the cognitive reserve review scene, the central processing unit carries out cognitive reserve review on the user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310634247.8A CN116344042B (en) | 2023-05-31 | 2023-05-31 | Cognitive reserve intervention lifting method and system based on multi-modal analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310634247.8A CN116344042B (en) | 2023-05-31 | 2023-05-31 | Cognitive reserve intervention lifting method and system based on multi-modal analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116344042A CN116344042A (en) | 2023-06-27 |
| CN116344042B true CN116344042B (en) | 2023-12-01 |
Family
ID=86879125
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310634247.8A Active CN116344042B (en) | 2023-05-31 | 2023-05-31 | Cognitive reserve intervention lifting method and system based on multi-modal analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116344042B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112353381A (en) * | 2020-11-24 | 2021-02-12 | 杭州冉曼智能科技有限公司 | Alzheimer's disease comprehensive diagnosis system based on multi-modal brain images |
| CN113298830A (en) * | 2021-06-22 | 2021-08-24 | 西南大学 | Acute intracranial ICH region image segmentation method based on self-supervision |
| CN113724880A (en) * | 2021-11-03 | 2021-11-30 | 深圳先进技术研究院 | Abnormal brain connection prediction system, method and device and readable storage medium |
| CN113902129A (en) * | 2021-10-28 | 2022-01-07 | 华中师范大学 | Multi-mode unified intelligent learning diagnosis modeling method, system, medium and terminal |
| CN114188013A (en) * | 2021-09-01 | 2022-03-15 | 北京智精灵科技有限公司 | Cognitive and brain image data integration evaluation method for Alzheimer's disease |
| CN115153452A (en) * | 2022-03-02 | 2022-10-11 | 国家康复辅具研究中心 | Cognitive regulation and training system |
| CN116092673A (en) * | 2023-04-10 | 2023-05-09 | 华南理工大学 | Portable multi-information fusion analysis and intervention evaluation system and method thereof |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP4160523A1 (en) * | 2021-09-29 | 2023-04-05 | Siemens Healthcare GmbH | Computer-implemented method for identifying a conspicuous structure |
-
2023
- 2023-05-31 CN CN202310634247.8A patent/CN116344042B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112353381A (en) * | 2020-11-24 | 2021-02-12 | 杭州冉曼智能科技有限公司 | Alzheimer's disease comprehensive diagnosis system based on multi-modal brain images |
| CN113298830A (en) * | 2021-06-22 | 2021-08-24 | 西南大学 | Acute intracranial ICH region image segmentation method based on self-supervision |
| CN114188013A (en) * | 2021-09-01 | 2022-03-15 | 北京智精灵科技有限公司 | Cognitive and brain image data integration evaluation method for Alzheimer's disease |
| CN113902129A (en) * | 2021-10-28 | 2022-01-07 | 华中师范大学 | Multi-mode unified intelligent learning diagnosis modeling method, system, medium and terminal |
| CN113724880A (en) * | 2021-11-03 | 2021-11-30 | 深圳先进技术研究院 | Abnormal brain connection prediction system, method and device and readable storage medium |
| CN115153452A (en) * | 2022-03-02 | 2022-10-11 | 国家康复辅具研究中心 | Cognitive regulation and training system |
| CN116092673A (en) * | 2023-04-10 | 2023-05-09 | 华南理工大学 | Portable multi-information fusion analysis and intervention evaluation system and method thereof |
Non-Patent Citations (2)
| Title |
|---|
| Wanyun Lin 等.Bidirectional Mapping of Brain MRI and PET With 3D Reversible GAN for the Diagnosis of Alzheimer's Disease.《Frontiers in Neuroscience》.2021,第15卷第1-13页. * |
| 融合多维特征的认知诊断模型与方法研究;杨华利;《中国博士学位论文全文数据库 社会科学Ⅱ辑》(第2期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116344042A (en) | 2023-06-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7293050B2 (en) | Mild Cognitive Impairment Judgment System | |
| KR102197112B1 (en) | Computer program and method for artificial neural network model learning based on time series bio-signals | |
| JP7299427B2 (en) | A method for predicting mental health and providing a mental health solution by learning psychological index data and physical index data based on machine learning, and a mental health evaluation device using the same | |
| US20220093215A1 (en) | Discovering genomes to use in machine learning techniques | |
| Bach et al. | Case representation and similarity assessment in the self back decision support system | |
| US20180211727A1 (en) | Automated Evidence Based Identification of Medical Conditions and Evaluation of Health and Financial Benefits Of Health Management Intervention Programs | |
| KR101841222B1 (en) | Method for generating prediction results for early prediction of fatal symptoms of a subject and apparatus using the same | |
| JP7191443B2 (en) | Target object attribute prediction method based on machine learning, related equipment and computer program | |
| JP2020532025A (en) | A method for generating a prediction result for predicting the occurrence of a fatal symptom of a subject at an early stage, and a device using the prediction result. | |
| Hiesh et al. | Classification of schizophrenia using genetic algorithm-support vector machine (ga-svm) | |
| JP7173482B2 (en) | Health care data analysis system, health care data analysis method and health care data analysis program | |
| Khazaal et al. | Predicting Coronary Artery Disease Utilizing Support Vector Machines: Optimizing Predictive Model | |
| Pingle | Evaluation of mental stress using predictive analysis | |
| CN116344042B (en) | Cognitive reserve intervention lifting method and system based on multi-modal analysis | |
| Yin et al. | A hybrid intelligent diagnosis approach for quick screening of Alzheimer’s disease based on multiple neuropsychological rating scales | |
| US11547346B2 (en) | Method, server, and computer program for classifying severe cognitive impairment patients by analyzing EEG data | |
| US20240008785A1 (en) | Information processing system, information processing device, information processing method, and information processing program | |
| Juliet | Investigations on Machine Learning Models for Mental Health Analysis and Prediction | |
| Cheng et al. | Mining discriminative patterns to predict health status for cardiopulmonary patients | |
| Obayya et al. | A novel automated Parkinson’s disease identification approach using deep learning and EEG | |
| Strobel et al. | Healthcare in the Era of Digital twins: towards a Domain-Specific Taxonomy. | |
| KR20170098414A (en) | Systems and algorithms for self differential diagnosis of diseases | |
| Grzywalski et al. | Interactive Lungs Auscultation with Reinforcement Learning Agent | |
| Dixit et al. | Managing the Scarcity of Monitoring Data Through Machine Learning for Human Behavior in Mental Health Care | |
| Dong et al. | Readmission prediction of diabetic patients based on AdaBoost-RandomForest mixed model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |