Social function prediction system for depression patients

Abstract

The invention belongs to the technical field of intelligent diagnosis of mental diseases, and particularly relates to a social function prediction system for depression patients. The system comprises an input module used for inputting indexes of a patient; the feature extraction module is used for extracting an attention-related grey matter feature ICSAGMV of the brain network from the input indexes; the prediction module is integrated with a depression patient social function prediction model, and the indexes are calculated to obtain a patient social function prediction result; and the output module is used for outputting the prediction result. The system provided by the invention can accurately predict the social function of the depression patient, and has a good application prospect.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent diagnosis of mental illness, and in particular relates to a social function prediction system for patients with depression. Background Art

[0002] Major Depressive Disorder (MDD), also known as clinical depression, major depressive disorder, and unipolar depression, is a mental illness that is characterized by a depressed emotional state, low self-esteem, and loss of interest in activities that were previously enjoyed.

[0003] Social function refers to an individual's ability in occupation, life, marriage, social interaction, etc. MDD often presents with cognitive impairment. Changes in cognitive function, such as attention impairment, may seriously affect the patient's clinical remission and long-term social function. However, the neural mechanism of cognitive impairment in MDD remains unclear.

[0004] In the prior art, the diagnosis of social function of MDD patients is mainly based on the social function status of individuals assessed by the Social Function Deficit Screening Scale (SDSS). The SDSS contains 10 items and uses a 0-2 approximately 3-level scoring method: (0) no abnormality, or only very slight defects that do not cause complaints / problems; (1) there are indeed functional defects; (2) severe functional defects. Specifically, it includes 10 items: occupation and work, marital function, parental function, social withdrawal, social activities outside the family, too few activities within the family, family function, personal life, interest and concern for the outside world, responsibility and planning. However, there is currently a lack of objective biological predictive indicators for evaluating individual social function.

[0005] Artificial intelligence has been widely used in the medical field and has been applied to the diagnosis of various diseases. For example, "CN202010170779.7 A Depression Intelligent Diagnosis Device and System Based on Knowledge Graph" has attempted to diagnose depression through machine learning methods. However, the neural mechanism of cognitive impairment in MDD is still unclear. Therefore, it is not clear what kind of indicators can be used as input features of the machine learning model to achieve accurate social function diagnosis. Summary of the invention

[0006] In view of the problems of the prior art, the purpose of the present invention is to provide a social function prediction system for patients with depression.

[0007] A social function prediction system for patients with depression, comprising:

[0008] An input module, used to input the patient's indicators;

[0009] The feature extraction module is used to extract the gray matter feature IC_SA_GMV of the attention-related brain network from the input indicators;

[0010] The prediction module integrates a social function prediction model for depression patients, and obtains the social function prediction results of the patients by calculating the indicators;

[0011] An output module is used to output the prediction result.

[0012] Preferably, the gray matter feature IC_SA_GMV of the attention-related brain network is obtained according to the following method:

[0013] Step 1, input the patient's attention test to obtain the cognitive test score of attention;

[0014] Step 2, input the patient's structural MRI and resting-state functional MRI data, and extract ALFF and GMV images;

[0015] Step 3, using the supervised multivariate canonical correlation analysis method MCCAR and the joint independent component analysis jICA technology, the attention cognitive test scores obtained in step 1 are used as reference information to guide the data fusion of ALFF and GMV to obtain the gray matter characteristics IC_SA_GMV of the brain network related to attention.

[0016] Preferably, the attention test (RVP test, Rapid Visual Processing) is an RVP test from the Cambridge Automated Cognitive Test Battery (CANTAB).

[0017] Preferably, the social function prediction model for patients with depression is a linear regression model.

[0018] Preferably, the linear regression model is expressed as:

[0019] y=-14.391X+2.213

[0020] Among them, X is the gray matter characteristic IC_SA_GMV of the brain network related to attention, and y is the social function score.

[0021] Preferably, a higher social function score indicates a worse social function of the patient, and a lower social function score indicates a better social function of the patient.

[0022] The present invention also provides a computer-readable storage medium, on which is stored: a computer program for implementing the above-mentioned social function prediction system for depression patients.

[0023] The present invention has discovered multimodal brain imaging markers related to attention by analyzing a large number of MDD patient sample data. Using these multimodal brain imaging markers, the present invention has for the first time constructed a system that uses a model to predict the social function of MDD patients, which can accurately predict the social function of patients and has good application prospects.

[0024] Obviously, according to the above contents of the present invention, in accordance with common technical knowledge and customary means in the art, without departing from the above basic technical ideas of the present invention, other various forms of modification, replacement or change may be made.

[0025] The following is a further detailed description of the above contents of the present invention through specific implementation methods in the form of embodiments. However, this should not be understood as the scope of the above subject matter of the present invention being limited to the following examples. All technologies implemented based on the above contents of the present invention belong to the scope of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 Figure 1 is a flow chart of Experimental Example 1, in which (a) in the discovery dataset (16-60 years old), the RVP test cognitive score from CANTAB was used as a reference to guide the fusion of two MRI features. (b) The analysis of cognitive-guided fusion was repeated in the age-restricted cohort (18-45 years old). (c) Group comparisons of the identified components between MDD and HC were performed, and linear regression models were used to identify predictive markers of social functioning. DETAILED DESCRIPTION

[0027] It should be noted that the algorithms of data collection, transmission, storage and processing steps not specifically described in the embodiments, as well as the hardware structure, circuit connection, etc. not specifically described can all be implemented through the contents disclosed in the prior art.

[0028] Example 1: Social function prediction system for patients with depression

[0029] The system of this embodiment includes:

[0030] An input module, used to input the patient's indicators;

[0031] The feature extraction module is used to extract the gray matter feature IC_SA_GMV of the attention-related brain network from the input indicators;

[0032] The prediction module integrates a social function prediction model for depression patients, and obtains the social function prediction results of the patients by calculating the indicators;

[0033] An output module is used to output the prediction result.

[0034] The gray matter feature IC_SA_GMV of the attention-related brain network is obtained according to the following method:

[0035] Step 1: Perform the RVP test of the Cambridge Automated Cognitive Function Test Battery on the patient to obtain the attention-related cognitive test score;

[0036] Step 2, collecting the patient's structural MRI and resting-state functional MRI data, and extracting brain imaging images, including ALFF and GMV images;

[0037] Step 3, using the supervised multivariate canonical correlation analysis method MCCAR and the joint independent component analysis jICA technology, the cognitive test scores of the five dimensions obtained in step 1 are used as reference information to guide the data fusion of ALFF and GMV.

[0038] The social function prediction model for patients with depression is a linear regression model. The expression of the linear regression model is:

[0039] y=-14.391X+2.213

[0040] Among them, X is the gray matter feature IC_SA_GMV of the brain network related to attention, and y is the social function score. The meaning of the social function score is similar to the meaning of the SDSS scale assessment result. The higher the score, the worse the patient's social function, and vice versa, the lower the score, the better the patient's social function.

[0041] The technical solution of the present invention is further illustrated by experiments below.

[0042] Experimental Example 1: Screening Study of Predictive Indicators of Social Function

[0043] 1. Research Objects and Methods

[0044] The process of this experiment is as follows Figure 1As shown in the figure, a total of 131 MDD patients who met the diagnostic criteria of DMS-IV depression and 145 matched normal controls were included in this study. The general information of all subjects was collected, and the attention function (RVP) of all subjects was assessed using the Cambridge Automated Neurocognitive Test Battery (CANTAB). At the same time, the structural MRI and resting-state functional MRI data of all subjects were collected. The Hamilton Depression Rating Scale (HAMD-17) was used to assess the severity of individual illness. After 2 years, the patients were followed up, and the social functioning status of the individuals was assessed using the Social Functioning Deficit Screening Scale (SDSS). The brain imaging data were processed and analyzed using FSL and DPARSF software, and the low-frequency oscillation (ALFF) and gray matter volume (GMV) images of each subject were extracted as representative features of brain imaging data fusion analysis, among which ALFF can reflect the intrinsic functional activity of the individual brain, and GMV reflects the changes in the gray matter volume of the individual brain.

[0045] Furthermore, supervised multivariate canonical correlation analysis (MCCAR) and joint independent component analysis (jICA) techniques were used to guide the fusion of individual brain low-frequency amplitude (ALFF) and gray matter volume (GMV) data using cognitive test scores of sustained attention RVP as reference information, with the aim of:

[0046] (1) To identify multimodal brain imaging biomarkers associated with cognitive function, especially sustained attention impairment, in MDD.

[0047] (2) Based on the identified differential multimodal brain imaging components between the MDD patient group and the control group, potential predictive markers of social function were identified using linear regression models.

[0048] 2. Experimental Results

[0049] In this experimental case, a t-test was performed to explore the inter-group differences in the two imaging feature loads, with significance set at p < 0.05. The extracted components were then compared in two ways: First, the inter-group differences in attention-related brain imaging features were compared ( Figure 1 c). Secondly, Pearson correlation analysis was used to explore the correlation between cognitive-related imaging features (ALFF_ICref for rsfMRI and GMV_ICref for sMRI) and the subjects' attention scores (significance was p<0.05). Linear regression analysis was then performed to explore brain image features that could predict social function in MDD patients two years later ( Figure 1 d). In addition, the following variables were included in the linear regression analysis to construct a stepwise regression model: age, gender, years of education, age at onset, number of relapses, total course of illness, presence of suicidal ideation, presence of suicidal behavior, and clinical symptoms. Statistical significance was set at p value < 0.05.

[0050] Finally, only the gray matter feature IC_SA_GMV of the brain network was significantly correlated with the patient's social function, and the expression of the established linear regression model was:

[0051] y=-14.391X+2.213

[0052] Among them, X is the gray matter characteristic IC_SA_GMV of the brain network related to attention, and y is the social function score.

[0053] It can be seen from the above embodiments and experimental examples that the present invention has discovered attention-related multimodal brain imaging markers that are significantly correlated with patients' social functions. The use of this multimodal brain imaging marker can accurately predict the social functions of MDD patients, and has good application prospects.

Claims

1. A social function prediction system for patients with depression, characterized in that: include: An input module, used to input the patient's indicators; The feature extraction module is used to extract the gray matter feature IC_SA_GMV of the attention-related brain network from the input indicators; The prediction module integrates a social function prediction model for depression patients, and obtains the social function prediction results of the patients by calculating the indicators; An output module is used to output the prediction result.

2. The social function prediction system for depression patients according to claim 1, characterized in that: The gray matter feature IC_SA_GMV of the attention-related brain network is obtained as follows: Step 1, input the patient's attention test to obtain the cognitive test score of attention; Step 2, input the patient's structural MRI and resting-state functional MRI data, and extract ALFF and GMV images; Step 3, using the supervised multivariate canonical correlation analysis method MCCAR and the joint independent component analysis jICA technology, the attention cognitive test scores obtained in step 1 are used as reference information to guide the data fusion of ALFF and GMV to obtain the gray matter characteristics IC_SA_GMV of the brain network related to attention.

3. The social function prediction system for depression patients according to claim 2, characterized in that: The attention test was the RVP test from the Cambridge Automated Cognitive Battery.

4. The social function prediction system for depression patients according to claim 1, characterized in that: The social function prediction model for patients with depression is a linear regression model.

5. The system for predicting social function of patients with depression according to claim 4, characterized in that: The linear regression model is expressed as: y=-14.391X+2.213 Among them, X is the gray matter characteristic IC_SA_GMV of the brain network related to attention, and y is the social function score.

6. The system for predicting social function of patients with depression according to claim 5, characterized in that: A higher social function score indicates that the patient's social function is worse, and a lower social function score indicates that the patient's social function is better.

7. A computer-readable storage medium, characterized in that: Stored thereon is: a computer program for implementing the social function prediction system for depression patients as described in any one of claims 1-6.

Images