Method, apparatus and electronic device for identifying ultra-early cognitive decline state

Abstract

The present invention provides a method, apparatus and electronic device for identifying a state of ultra-early cognitive decline. The method for identifying a state of ultra-early cognitive decline includes: obtaining basic information and scanned images of a subject, where the basic information includes gender, age and APOE gene information, and the scanned images include Aβ images and Tau images; obtaining feature data based on the scanned images; and inputting the basic information and the feature data into a pre-trained ultra-early cognitive decline state recognition model to identify whether the subject will develop a state of cognitive decline within a predetermined time period. The method for identifying a state of ultra-early cognitive decline according to the present invention can screen out subjects in a state of ultra-early cognitive decline.

Description

Technical Field

[0001] The present invention relates to the field of medicine, and particularly to a method, device and electronic device for identifying ultra-early cognitive decline status. Background Art

[0002] Alzheimer disease (AD), as a central nervous system degenerative disease, is the most common type of dementia. The incidence of AD has been increasing year by year, bringing a heavy burden to families, society and the medical system. The pathogenesis of AD is complex, and currently there are mainly hypotheses such as Aβ (β-amyloid protein) deposition hypothesis and abnormal phosphorylation hypothesis of Tau protein. With the in-depth understanding of the pathological mechanism of AD and the progress of technology, after 2010, there have been major breakthroughs in the drug research and development for AD, especially since 2019, it has entered a rapid development stage, and currently there are multiple ongoing Phase III clinical trials.

[0003] According to the clinical trial results, currently approved drugs can delay the cognitive decline in early AD. For example, both donanemab and lecanemab can significantly delay the cognitive decline level of patients for 18 months. Currently, drugs cannot reverse the disease, and the subjects participating in the trials are all patients with early status or high imaging indicators. Therefore, preclinical stage (ultra-early) intervention and Aβ / Tau combined targeting have become new research directions.

[0004] How to determine suitable subjects is an important part of clinical trials and scientific research. How to screen out subjects who currently have no status but may have cognitive decline status in the next 3 to 5 years is a problem to be solved. Such subjects may be ideal candidates for ultra-early Aβ / Tau combined targeted drug intervention trials. Summary of the Invention

[0005] Aiming at the above problems of the prior art, the purpose of the present invention is to provide a method, device and electronic device for identifying ultra-early cognitive decline status, which can screen out subjects with ultra-early cognitive decline status.

[0006] To solve the above problems, on the one hand, the present invention provides a method for identifying ultra-early cognitive decline status, and the method for identifying ultra-early cognitive decline status includes:

[0007] Obtain the basic information and scan images of the subject, where the basic information includes gender, age and APOE gene information, and the scan images include Aβ images and Tau images;

[0008] Based on the scan images, obtain feature data;

[0009] Input the basic information and the feature data into a pre-trained ultra-early cognitive decline state recognition model to identify whether the subject will develop a cognitive decline state within a predetermined time period.

[0010] Further, the feature data includes the Centiloid value and the CenTauR value.

[0011] Based on the scanned image, the obtained feature data includes:

[0012] Based on the Aβ image, obtain the Centiloid value;

[0013] Based on the Tau image, obtain the CenTauR value.

[0014] Further, based on the Aβ image, the Centiloid value is obtained according to the following formula (1):

[0015] (1)

[0016] Where , is the SUVr of the subject, and are the intercept and slope obtained by linear regression, respectively.

[0017] Further, the feature data also includes the first z-score of the Aβ image and the second z-score of the Tau image.

[0018] The recognition method further includes:

[0019] When the Centiloid value is lower than a predetermined Centiloid threshold, determine that the subject is an Aβ-negative individual.

[0020] For the subject being a negative individual, calculate the first z-score of the Aβ image and the second z-score of the Tau image.

[0021] Further, the first z-score and the second z-score are obtained according to the following formula (2).

[0022] (2)

[0023] Where is the of the subject, is the mean of the Aβ-negative group, is the standard deviation of the Aβ-negative group;

[0024] The first z-scores include the posterior cingulate gyrus z-score, the anterior cingulate gyrus z-score, the lateral prefrontal cortex z-score, the orbital frontal cortex z-score, the precuneus z-score, the lateral parietal lobe z-score, the lateral occipital lobe, and the lateral temporal lobe z-score. The second z-scores include the medial temporal lobe z-score, the lateral temporal lobe z-score, the temporoparietal lobe z-score, and the frontal lobe z-score.

[0025] Furthermore, the scanned image further includes an MR3DT1w image, and the feature data further includes the hippocampal volume.

[0026] The calculation of the hippocampal volume includes:

[0027] Obtaining the initial hippocampal volume and the total intracranial volume of the subject;

[0028] Correcting the initial hippocampal volume based on the following formula (3):

[0029] (3)

[0030] where is the initial hippocampal volume of the subject, is the total intracranial volume of the subject, is the mean of the total intracranial volume, and b is the slope obtained by performing a linear regression on the hippocampal volume and the total intracranial volume.

[0031] Furthermore, the training of the ultra-early cognitive decline state recognition model includes:

[0032] Constructing a logistic regression model with the basic information and the feature data as independent variables and whether cognitive decline occurs as the dependent variable.

[0033] The logistic regression model is shown as the following formula (4):

[0034] (4)

[0035] where is the feature vector composed of the basic information and the feature data, is the parameter, is the bias, is the transpose, is whether the cognitive decline state appears;

[0036] Training the logistic regression model based on the sample data with labels indicating whether the cognitive decline state appears to obtain the ultra-early cognitive decline state recognition model.

[0037] Furthermore, the sample data includes a first sample and a second sample;

[0038] Among them, the first sample is a sample in which the score of the initial Mini-Mental State Examination reaches a first predetermined score, and after a predetermined time period, the decreased score of the preclinical AD cognitive composite scale compared to the initial is higher than a second predetermined score;

[0039] The second sample is a sample in which the score of the initial Mini-Mental State Examination reaches a first predetermined score, and after a predetermined time period, the decreased score of the preclinical AD cognitive composite scale compared to the initial is lower than or equal to the second predetermined score.

[0040] Both the first sample and the second sample are more than 200.

[0041] On the other hand, the present invention provides an identification device for a subject in a state of ultra-early cognitive decline. The identification device for a subject in a state of ultra-early cognitive decline includes:

[0042] An acquisition module for acquiring basic information and a scanned image of a subject, where the basic information includes gender, age, and APOE gene information, and the scanned image includes an Aβ image and a Tau image;

[0043] A feature data calculation module for obtaining feature data based on the scanned image;

[0044] An identification module for inputting the basic information and the feature data into a pre-trained ultra-early cognitive decline state identification model to identify whether the subject will have a cognitive decline state within a predetermined time period.

[0045] On the other hand, the present invention provides an electronic device. The electronic device includes a processor and a memory. At least one instruction or at least one program segment is stored in the memory, and the at least one instruction or the at least one program segment is loaded and executed by the processor to implement the ultra-early cognitive decline state identification method as described in any one of the above.

[0046] Due to the above technical solutions, the present invention has the following beneficial effects:

[0047] The method for identifying the ultra-early cognitive decline state according to the present invention collects basic information of the subject and scans the subject to obtain scanned images including Aβ images and Tau images. Feature data is obtained based on the scanned images. The basic information of the subject and the feature data are input into a pre-trained ultra-early cognitive state decline recognition model, so as to identify whether the subject will have a cognitive decline state within a predetermined time period (for example, within 3 to 5 years). This helps to screen out the subjects who will benefit from the ultra-early Aβ / Tau combined targeted drug intervention, improve the efficacy detection ability of such trials, and reduce the trial cost. If the ultra-early Aβ / Tau combined targeted drug intervention is effective, it can also be used to screen the drug administration population. The ultra-early cognitive decline recognition model integrates basic information (demographics, genetic information) and feature data (imaging information), has high robustness, and has good recognition effects. By combining the manifestations of images of two AD-related proteins (Aβ images and Tau images), as well as global and regional information, possible ultra-early changes are captured. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the present invention, the accompanying drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other accompanying drawings according to these drawings without creative efforts.

[0049] Figure 1 is a flowchart of a method for identifying the ultra-early cognitive decline state according to an embodiment of the present invention;

[0050] Figure 2 is a schematic diagram of an apparatus for identifying the ultra-early cognitive decline state according to an embodiment of the present invention;

[0051] Figure 3 is a schematic diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0052] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0053] It should be noted that the terms "first", "second", etc. in the description, claims and above-mentioned drawings of the present invention are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that such data used can be interchanged under appropriate circumstances, so that the embodiments of the present invention described herein can be implemented in an order different from those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion.

[0054] Next, a method for identifying the ultra-early cognitive decline state of the embodiments of the present invention will be described.

[0055] As Figure 1 shown, the method for identifying the ultra-early cognitive decline state of the embodiments of the present invention includes:

[0056] Step S1, obtaining the basic information and scan images of the subject, the basic information includes gender, age and APOE gene information, and the scan images include Aβ images and Tau images. Among them, the APOE (Apolipoprotein E) gene information is important genetic risk factor information for AD.

[0057] The subject is scanned to obtain scan images including Aβ images and Tau images.

[0058] Step S2, obtaining feature data based on the scan images.

[0059] That is to say, feature data is obtained by calculating the scan images.

[0060] Step S3, inputting the basic information and the feature data into a pre-trained ultra-early cognitive decline state recognition model to identify whether the subject will develop a cognitive decline state within a predetermined time period. Among them, the ultra-early cognitive decline state recognition model can be obtained by training samples.

[0061] The basic information and feature data of the subject are input into a pre-trained ultra-early cognitive state decline recognition model, so as to identify whether the subject will develop a cognitive decline state within a predetermined time period (for example, within 3 to 5 years).

[0062] The above method for identifying the ultra-early cognitive decline state involves collecting basic information of the subject and scanning the subject to obtain scanned images including Aβ images and Tau images. Feature data is obtained based on the scanned images. The basic information and feature data of the subject are input into a pre-trained ultra-early cognitive state decline recognition model to identify whether the subject will experience a cognitive decline state within a predetermined time period (e.g., within 3 to 5 years). This helps screen out subjects who benefit from ultra-early Aβ / Tau combined targeted drug intervention, improve the efficacy detection ability of such trials, and reduce trial costs. If the ultra-early Aβ / Tau combined targeted drug intervention is effective, it can also be used to screen the drug administration population. The ultra-early cognitive decline recognition model integrates basic information (demographics, genetic information) and feature data (imaging information), has high robustness, and good recognition effect. By combining the manifestations of images of two AD-related proteins (Aβ images and Tau images), as well as global and regional information, it captures possible ultra-early changes.

[0063] In some embodiments of the present invention, the feature data includes Centiloid value and CenTauR value.

[0064] Based on the scanned images, the obtained feature data includes: based on the Aβ image, obtaining the Centiloid value; based on the Tau image, obtaining the CenTauR value. The Centiloid value and CenTauR value can be obtained by existing technologies or as described below.

[0065] PET (Positron Emission Tomography) enables in-vivo visualization and quantitative analysis of Aβ / Tau, and can directly provide information on the total load and spatial distribution of Aβ / Tau pathology. There are various optional tracers for Aβ and Tau PET respectively. With the increasing use of different PET tracers in clinical and research, it is necessary to standardize the SUVr (standardised uptake value ratio) between tracers for unified comparison. For this reason, Centiloid (CL) for Aβ was first proposed in 2015. It is an unbounded value, with 0 anchored to the mean of healthy young control groups and 100 anchored to the mean of typical AD patient groups with mild to moderate symptoms. This standardization enables direct comparison between different devices, algorithms, and tracers, and Centiloid was immediately applied to many Aβ-related studies. Given the success of Centiloid, a similar concept for Tau, CenTauR (CTR), was also proposed in 2024.

[0066] Thus, the Centiloid value and CenTauR value facilitate unified comparison and can accurately anchor AD subjects.

[0067] Furthermore, based on the Aβ image, the Centiloid value is obtained according to the following formula (1):

[0068] (1)

[0069] Where , is the SUVr of the subject, and are the intercept and slope obtained by linear regression, respectively.

[0070] Public data on Centiloid (including data corresponding to five common Aβ tracers) can be downloaded, processed with a local process to obtain a standard space image, and the SUVr (Standardized Uptake Value Ratio) is calculated using the standard cortical regions and the whole cerebellar region recommended by the Centiloid project. This is linearly regressed with the reference SUVr (included in the public data, with the tracer fixed as PiB) to obtain the SUVr converted from the locally calculated SUVr to PiB-Clac for each tracer.

[0071]

[0072] Where is the SUVr obtained by processing public data with a local process, and are the intercept and slope obtained by linear regression, respectively. Then substitute the above formula into the formula for calculating Centiloid from the SUVr of PiB provided by the Centiloid project:

[0073] (1)

[0074] Thus, the formula for converting the locally calculated SUVr to Centiloid for each tracer is obtained. Note that formula (1) here can also be obtained by processing the PiB-Clac data in the public data with a local process, but the process is slightly more complex, so it is omitted.

[0075] Process the local Aβ image with the same local process, select the corresponding formula according to the tracer, and convert the SUVr to Centiloid.

[0076] Furthermore, based on the Tau image, the CenTauR value is obtained.

[0077] Using the JPM (joint propagation model), the local CenTauR calculation process can be established. Public data on CenTauR can be downloaded, including five sets of head-to-head data and five sets of anchor point data. Among them, the head-to-head data is the image data of two Tau tracers for each subject, containing the relevant information between the tracers; the anchor point data is the data of the non-cognitive decline population and the AD population, which are respectively anchored to the CenTauR values of 0 and 100, with a total of five sets, and each set corresponds to one Tau tracer. The standard space image is obtained by processing with the local process, and the SUVr is calculated using the Tau general ROI (region of interest) and the inferior cerebellar cortex.

[0078] JPM is a non-linear mixed effect model. The model itself is more complex than linear regression, but the advantage is that it does not require each tracer to have head-to-head data with another fixed tracer. The result will be a set of linear mapping equations for converting SUVr to CenTauR. JPM is based on a basic assumption that SUVr is a tracer-specific linear transformation with noise of the true CenTauR value of the measured object. Therefore, JPM poses the problem of calculating CenTauR as an inverse problem, that is, inferring the CenTauR value that is most likely to produce the observed SUVr. The model is:

[0079]

[0080]

[0081]

[0082]

[0083] Among them, the subscript t represents the tracer, and i represents the subject. indicates whether it belongs to the head-to-head data, 1 for yes and 0 for no. and represent whether it belongs to the non-cognitive decline population and the AD population. is the CenTauR value of the subject itself in the head-to-head data, and its SUVr has a linear relationship with CenTauR. is the error. is the CenTauR value of the non-cognitive decline population, following a normal distribution with a mean of 0 and a standard deviation of is the CenTauR value of the AD population, following a normal distribution with a mean of 100 and a standard deviation of Normal distribution. By specifying all relationships in a model, JPM can propagate information among all tracers through anchor point data and head-to-head data. According to the maximum likelihood estimation, a set of linear transformation formulas for calculating CenTauR from SUVr can be obtained. Applying this set of formulas to local SUVr data gives CenTauR.

[0084] In some embodiments of the present invention, the feature data further includes a first z-score of the Aβ image and a second z-score of the Tau image.

[0085] The identification method further includes: when the Centiloid value is lower than a predetermined Centiloid threshold, determining that the subject is an Aβ-negative individual, and for the subject being a negative individual, calculating a first z-score of the Aβ image and a second z-score of the Tau image.

[0086] In addition to the overall assessment of the Aβ / Tau status, the importance of its topological distribution is also a major focus of AD research. Studies have shown that this information can provide additional clinical information, especially in the early stages of Aβ / Tau accumulation and development, where certain regions (such as the posterior cingulate gyrus) may accumulate proteins earlier than other regions. And local information is exactly the information that can be provided by imaging methods such as PET, which cannot be achieved by conventional cerebrospinal fluid- and blood-based biochemical analyses. Similarly, there are also problems with differences in local SUVr among different devices, algorithms, and tracers, which can be solved by using z-scores (the first z-score and the second z-score).

[0087] Individuals who are clearly Aβ-negative in the local data are selected through visual assessment of the images by experts or by setting a Centiloid threshold (when the Centiloid value is lower than a predetermined Centiloid threshold, determining that the subject is an Aβ-negative individual) for use in subsequent calculations of regional z-scores (the first z-score and the second z-score).

[0088] Thus, the first z-score and the second z-score can solve the problem of differences in local SUVr among different devices, algorithms, and tracers.

[0089] Furthermore, the first z-score and the second z-score are obtained based on the following formula (2):

[0090] (2)

[0091] where is the of the subject, is the mean of the Aβ-negative population, is the standard deviation of the Aβ-negative population;

[0092] The first z-score includes the posterior cingulate gyrus z-score, anterior cingulate gyrus z-score, lateral prefrontal cortex z-score, orbital frontal cortex z-score, precuneus z-score, lateral parietal lobe z-score, lateral occipital lobe, and lateral temporal lobe z-scores, and the second z-score includes the medial temporal lobe z-score, lateral temporal lobe z-score, temporoparietal lobe z-score, and frontal lobe z-score.

[0093] Process the MR 3DT1w images with brain structure analysis software (such as FreeSurfer) to obtain the ROI masks for each subject. Set a threshold of 0.7 to reduce partial volume effects, and calculate the masks for a total of 8 ROIs including the posterior cingulate gyrus, anterior cingulate gyrus, lateral prefrontal cortex, orbital frontal cortex, precuneus, lateral parietal lobe, lateral occipital lobe, and lateral temporal lobe. Register the Aβ images in the local data to their respective 3DT1w images, and use the whole cerebellum as the reference region to calculate the SUVr of each ROI.

[0094] Calculate the mean and standard deviation of the SUVr values of each ROI in the Aβ-negative group, and calculate the respective first z-scores (Aβ ROI z-scores) for each data:

[0095] (2)

[0096] Note that each tracer has its own mean and standard deviation for each ROI here.

[0097] The Tau ROI masks can be obtained, which include 4 ROIs: the medial temporal lobe, lateral temporal lobe, temporoparietal lobe, and frontal lobe. Based on the standard space Tau images, calculate the mean and standard deviation of the SUVr values of each ROI in the Aβ-negative group. Similarly, calculate the respective second z-scores (Tau ROI z-scores) for each data using Equation (2).

[0098] In some embodiments of the present invention, the scanned images further include MR 3DT1w images, and the characteristic data further includes the hippocampal volume.

[0099] The calculation of the hippocampal volume includes:

[0100] Obtain the initial hippocampal volume and the total intracranial volume of the subject;

[0101] Correct the initial hippocampal volume based on the following formula (3):

[0102] (3)

[0103] Wherein, is the initial hippocampal volume of the subject, is the total intracranial volume of the subject, is the mean of the total intracranial volume, and b is the slope obtained from the linear regression of the hippocampal volume and the total intracranial volume.

[0104] Process the MR 3DT1w images using brain structure analysis software (such as FreeSurfer) to obtain the initial hippocampal volume and total intracranial volume of each subject, and then calculate the corrected hippocampal volume through Equation (3).

[0105] The hippocampal volume is a marker of neurodegenerative diseases. The hippocampus is one of the earliest brain regions affected in AD, and its volume atrophy is an important imaging index for diagnosing early AD (measured by MRI).

[0106] In some embodiments of the present invention, the training of the ultra-early cognitive decline state recognition model includes:

[0107] Construct a logistic regression model with basic information and feature data as independent variables and whether there is cognitive decline as the dependent variable.

[0108] The logistic regression model is shown in Equation (4) below:

[0109] (4)

[0110] Wherein, is the feature vector composed of basic information and feature data, is the parameter, is the bias, is the transpose, is whether there is a cognitive decline state;

[0111] Based on the sample data containing the label of whether there is a cognitive decline state, train the logistic regression model to obtain the ultra-early cognitive decline state recognition model.

[0112] That is to say, construct a logistic regression model with gender, age, APOE gene, Aβ Centiloid value, Tau CenTauR value, hippocampal volume, 8 Aβ ROI z (first z-score) scores, and 4 Tau ROI z-scores (second z-score) as independent variables and whether there is cognitive decline as the dependent variable, as shown in Equation (4). According to the data, the model parameters that minimize the loss function can be obtained, and finally a model with good prediction performance can be obtained.

[0113] After the model is determined, collect the gender, age, and APOE gene information of new subjects, and perform MR 3DT1w scans, PET Aβ scans, and Tau scans to obtain the corresponding images. Process the image data as described above, and then input the results into the model to obtain a prediction of whether the subject has a risk of cognitive decline in the next 3 to 5 years, so as to help screen subjects who are more likely to benefit from ultra-early Aβ / Tau combined targeted drug intervention, improve the efficacy detection ability of such trials, and reduce the trial cost. If the ultra-early Aβ / Tau combined targeted drug intervention is effective, this model can also be used to screen the drug administration population.

[0114] Furthermore, the sample data includes a first sample and a second sample. Among them, the first sample is a sample in which the score of the initial Mini-Mental State Examination reaches a first predetermined score, and after a predetermined time period, the score decrease of the preclinical AD cognitive composite scale compared to the initial is higher than a second predetermined score. The second sample is a sample in which the score of the initial Mini-Mental State Examination reaches a first predetermined score, and after a predetermined time period, the score decrease of the preclinical AD cognitive composite scale compared to the initial is lower than or equal to the second predetermined score. Both the first sample and the second sample are more than 200.

[0115] For example, collect data of at least 500 subjects, collect information on gender, age, and APOE gene (an important genetic risk factor for AD), require the MMSE (Mini-Mental State Scale) score to be 27 or above (i.e., normal cognitive function) at the first visit, conduct PACC (preclinical AD cognitive composite scale) scale assessment, obtain MR 3DT1w images, PET Aβ images, and Tau images (the tracers can be different), visit again after 3 - 5 years, and conduct PACC scale assessment. Subjects with a score decrease of more than 0.5 compared to the initial PACC score are determined to have cognitive decline, otherwise they have no cognitive decline. It is required that there are at least 200 cases in the cognitive decline group and at least 200 cases in the non-cognitive decline group.

[0116] Thus, the logistic regression model can be better trained, enabling the logistic regression model to accurately identify whether a subject will experience a cognitive decline state within a predetermined time period based on high accuracy.

[0117] Next, the identification device 500 for subjects in the ultra-early cognitive decline state according to the embodiments of the present invention will be described.

[0118] As Figure 2 shown, the device 500 may include an acquisition module 501, a feature data calculation module 502, and an identification module 503.

[0119] The acquisition module 501 is used to acquire the basic information and scan images of the subject. The basic information includes gender, age, and APOE gene information, and the scan images include Aβ images and Tau images.

[0120] The feature data calculation module 502 is used to obtain feature data based on the scan images.

[0121] The identification module 503 is used to input the basic information and the feature data into a pre-trained ultra-early cognitive decline state identification model to identify whether the subject will experience a cognitive decline state within a predetermined time period.

[0122] It should be noted that, when the device provided in the above embodiments realizes its functions, only the division of the above functional modules is used for illustration. In actual applications, the above functions can be allocated to different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device provided in the above embodiments and the corresponding method embodiments belong to the same concept. For the specific implementation process, please refer to the corresponding method embodiments and will not be elaborated here.

[0123] An embodiment of the present invention further provides an electronic device, which includes a processor and a memory. At least one instruction or at least one program segment is stored in the memory, and the at least one instruction or the at least one program segment is loaded and executed by the processor to implement the method for identifying the ultra-early cognitive decline state provided in the above method embodiments.

[0124] The memory can be used to store software programs and modules. The processor runs the software programs and modules stored in the memory to perform various functional applications and data processing. The memory mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for functions, etc.; the data storage area can store data created according to the use of the device. In addition, the memory can include high-speed random access memory, and can also include non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Correspondingly, the memory can also include a memory controller to provide the processor with access to the memory.

[0125] Combined with the accompanying Figure 3 figures, a block diagram of an electronic device 600 according to an embodiment of the present invention is shown. The electronic device 600 may include one or more processors 602, a system control logic 608 connected to at least one of the processors 602, a system memory 604 connected to the system control logic 608, a non-volatile memory (NVM) 606 connected to the system control logic 608, and a network interface 610 connected to the system control logic 608.

[0126] The processor 602 may include one or more single-core or multi-core processors. The processor 602 may include any combination of general-purpose processors and dedicated processors (for example, a graphics processor, an application processor, a baseband processor, etc.). In the embodiments herein, the processor 602 may be configured to execute according to various embodiments.

[0127] In some embodiments, the system control logic 608 may include any suitable interface controller to provide any suitable interface to at least one of the processors 602 and / or any suitable device or component communicating with the system control logic 608.

[0128] In some embodiments, the system control logic 608 may include one or more memory controllers to provide an interface to the system memory 604. The system memory 604 may be used to load and store data and / or instructions. In some embodiments, the memory 604 of the device 600 may include any suitable volatile memory, such as a suitable dynamic random access memory (DRAM).

[0129] The NVM / memory 606 may include one or more tangible, non-transitory computer-readable media for storing data and / or instructions. In some embodiments, the NVM / memory 606 may include any suitable non-volatile memory such as flash memory and / or any suitable non-volatile storage device, such as at least one of a HDD (Hard Disk Drive), a CD (Compact Disc) drive, and a DVD (Digital Versatile Disc) drive.

[0130] The NVM / memory 606 may include a portion of the storage resources installed on the device 600, or it may be accessible by the device but not necessarily part of the device. For example, the NVM / storage 606 may be accessed via the network interface 610 over a network.

[0131] Specifically, the system memory 604 and the NVM / memory 606 may respectively include: a temporary copy and a permanent copy of the instructions 620. The instructions 620 may include: instructions that, when executed by at least one of the processors 602, cause the device 600 to implement a method for identifying a state of ultra-early cognitive decline. In some embodiments, the instructions 620, hardware, firmware, and / or its software components may alternatively / additionally be located in the system control logic 608, the network interface 610, and / or the processor 602.

[0132] The network interface 610 may include a transceiver for providing a radio interface for the device 600 to communicate with any other suitable device (such as a front-end module, an antenna, etc.) over one or more networks. In some embodiments, the network interface 610 may be integrated with other components of the device 600. For example, the network interface 610 may be integrated with at least one of the communication module of the processor 602, the system memory 604, the NVM / memory 606, and a firmware device with instructions (not shown), and when at least one of the processors 602 executes the instructions, the device 600 implements various embodiments.

[0133] The network interface 610 may further include any suitable hardware and / or firmware to provide a multiple-input multiple-output radio interface. For example, the network interface 610 may be a network adapter, a wireless network adapter, a telephone modem, and / or a wireless modem.

[0134] In one embodiment, at least one of the processors 602 may be logically packaged with one or more controllers for the system control logic 608 to form a system-in-package (SiP). In one embodiment, at least one of the processors 602 may be integrated with the logic of one or more controllers for the system control logic 608 on the same die to form a system-on-chip (SoC).

[0135] The device 600 may further include: an input / output (I / O) device 612. The I / O device 612 may include a user interface that enables a user to interact with the device 600; the design of the peripheral component interface enables peripheral components to also interact with the device 600. In some embodiments, the device 600 further includes sensors for determining at least one of environmental conditions and location information related to the device 600.

[0136] In some embodiments, the user interface may include, but is not limited to, a display (e.g., a liquid crystal display, a touch screen display, etc.), a speaker, a microphone, one or more cameras (e.g., a still image camera and / or a video camera), a flashlight (e.g., a light-emitting diode flash), and a keyboard.

[0137] In some embodiments, the peripheral component interface may include, but is not limited to, a non-volatile memory port, an audio jack, and a power interface.

[0138] In some embodiments, the sensors may include, but are not limited to, a gyroscope sensor, an accelerometer, a proximity sensor, an ambient light sensor, and a positioning unit. The positioning unit may also be part of the network interface 610 or interact with the network interface 610 to communicate with components of a positioning network (e.g., Global Positioning System (GPS) satellites).

[0139] It can be understood that the structure illustrated in the embodiments of the present invention does not constitute a specific limitation on the electronic device 600. In other embodiments of the present invention, the electronic device 600 may include more or fewer components than shown in the figures, or combine certain components, or split certain components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0140] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for identifying a state of ultra-early cognitive decline, characterized in that, The recognition method includes: Obtaining the basic information and scan images of the subject, where the basic information includes gender, age, and APOE gene information, and the scan images include Aβ images, Tau images, and MR3DT1w images; Based on the scan images, obtaining feature data; Inputting the basic information and the feature data into a pre-trained ultra-early cognitive decline state recognition model to identify whether the subject will experience a cognitive decline state within a predetermined time period, where the feature data includes the Centiloid value based on the Aβ image, the CenTauR value based on the Tau image, the first z-score of the Aβ image, the second z-score of the Tau image, and the hippocampal volume of the MR3DT1w image, When the Centiloid value is lower than a predetermined Centiloid threshold, it is determined that the subject is an Aβ-negative individual, and for the subject being a negative individual, a first z-score of the Aβ image and a second z-score of the Tau map are calculated, the first z-score and the second z-score being based on the negative population's calculated value. the first z-score includes the posterior cingulate gyrus z-score, the anterior cingulate gyrus z-score, the lateral prefrontal cortex z-score, the orbital frontal cortex z-score, the precuneus z-score, the lateral parietal lobe z-score, the lateral occipital lobe, and the lateral temporal lobe z-score, and the second z-score includes the medial temporal lobe z-score, the lateral temporal lobe z-score, the temporoparietal lobe z-score, and the frontal lobe z-score; The calculation of the hippocampal volume includes: Obtaining the initial hippocampal volume and the total intracranial volume of the subject; Correcting the initial hippocampal volume based on the total intracranial volume of the subject and the mean total intracranial volume; The training of the ultra-early cognitive decline state recognition model includes: Constructing a logistic regression model with the basic information and the feature data as independent variables and whether there is cognitive decline as the dependent variable, Training the logistic regression model based on sample data with labels indicating whether there is a cognitive decline state to obtain the ultra-early cognitive decline state recognition model.

2. The recognition method of the ultra-early cognitive decline state according to claim 1, wherein Obtaining the Centiloid value based on the following formula (1): Among them, , and are the intercept and slope obtained by linear regression, respectively.

3. The method for identifying the ultra-early cognitive decline state according to claim 1, characterized in that, The first z-score and the second z-score are obtained based on the following formula (2): , where is that of the subject , is the mean of the Aβ-negative group and is the standard deviation of the Aβ-negative group 4. The method for identifying the ultra-early cognitive decline state according to claim 1, characterized in that The initial hippocampal volume is corrected based on the following formula (3): where is the initial hippocampal volume of the subject, is the total intracranial volume of the subject, is the mean of the total intracranial volumes, and b is the slope obtained from the linear regression of the hippocampal volume against the total intracranial volume.

5. The method for identifying the ultra-early cognitive decline state according to claim 1, wherein The logistic regression model is shown as the following formula (4): Among them, is a feature vector composed of basic information and feature data, is a parameter, is a bias, is a transpose, is whether the state of cognitive decline occurs.

6. The method for identifying the ultra-early cognitive decline state according to claim 1, characterized in that The sample data includes a first sample and a second sample; where the first sample is a sample in which the score of the initial Mini-Mental State Examination reaches a first predetermined score, and the score reduction of the preclinical AD cognitive composite scale is higher than a second predetermined score compared to the initial after a predetermined time period; the second sample is a sample in which the score of the initial Mini-Mental State Examination reaches a first predetermined score, and the score reduction of the preclinical AD cognitive composite scale is lower than or equal to the second predetermined score compared to the initial after a predetermined time period, and both the first sample and the second sample are more than 200.

7. An apparatus for identifying a subject in a state of ultra-early cognitive decline, characterized in that, Including: An acquisition module for obtaining the basic information and scan images of the subject, where the basic information includes gender, age, and APOE gene information, and the scan images include Aβ images, Tau images, and MR3DT1w images; A feature data calculation module for obtaining feature data based on the scan images; An identification module for inputting the basic information and the feature data into a pre-trained ultra-early cognitive decline state recognition model to identify whether the subject will experience a cognitive decline state within a predetermined time period, Among them, the characteristic data includes the Centiloid value based on the Aβ image, the CenTauR value based on the Tau image, the first z-score of the Aβ image, the second z-score of the Tau image, and the hippocampal volume of the MR3DT1w image. When the Centiloid value is below a predetermined Centiloid threshold, it is determined that the subject is an Aβ-negative individual, and for the subject being a negative individual, a first z-score of the Aβ image and a second z-score of the Tau map are calculated, where the first z-score and the second z-score are based on the negative population's calculated value. The first z-score includes the posterior cingulate gyrus z-score, the anterior cingulate gyrus z-score, the lateral prefrontal cortex z-score, the orbital frontal cortex z-score, the precuneus z-score, the lateral parietal lobe z-score, the lateral occipital lobe, and the lateral temporal lobe z-score. The second z-score includes the medial temporal lobe z-score, the lateral temporal lobe z-score, the temporoparietal lobe z-score, and the frontal lobe z-score. The calculation of the hippocampal volume includes: Obtaining the initial hippocampal volume and the total intracranial volume of the subject; Correcting the initial hippocampal volume based on the total intracranial volume of the subject and the mean value of the total intracranial volume; The training of the ultra-early cognitive decline state recognition model includes: Constructing a logistic regression model with the basic information and the characteristic data as independent variables and whether there is cognitive decline as the dependent variable. Based on the sample data containing the label of whether there is a cognitive decline state, training the logistic regression model to obtain the ultra-early cognitive decline state recognition model.

8. An electronic device, characterized in that, The electronic device includes a processor and a memory. At least one instruction or at least one program segment is stored in the memory. The at least one instruction or the at least one program segment is loaded and executed by the processor to implement the method for recognizing the ultra-early cognitive decline state according to any one of claims 1-6.

Images