Prefrontal lobe fNIRS entropy-based depression brain age prediction and auxiliary diagnosis method and system

Abstract

The invention relates to the technical field of intelligent medical diagnosis, and discloses a depression brain age prediction and auxiliary diagnosis method and system based on prefrontal lobe fNIRS entropy, and the method comprises the steps: collecting a prefrontal lobe static hemoglobin time sequence signal through a near infrared spectrum optical brain imaging device; according to the method, static brain entropy, a dynamic entropy mean value, a dynamic entropy standard deviation and a variable coefficient are extracted, a brain age prediction model is constructed in combination with a machine learning algorithm, and effective auxiliary diagnosis is carried out on depression based on a brain age deviation value obtained by the brain age prediction model. According to the depression brain age prediction and auxiliary diagnosis method based on the prefrontal lobe fNIRS entropy provided by the invention, the depression brain age prediction precision can be effectively improved, and the brain aging degree can be accurately quantified; the method provides basis and reference for establishing diagnostic biomarkers, and is suitable for early screening of mental diseases, evaluation of curative effects and monitoring of neurodegenerative diseases (such as Alzheimer's disease).

Description

Technical Field

[0001] This application relates to the field of intelligent medical diagnostic technology, and in particular to a method and system for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy. Background Technology

[0002] Current clinical diagnosis of depression (MDD) still relies on subjective assessment scales (such as the Hamilton Depression Rating Scale / HAMD-17), lacking objective biomarkers. Although studies based on structural magnetic resonance imaging (sMRI) have shown that patients with depression exhibit accelerated brain aging (predicted brain age is on average 4.43 years older than actual age), existing brain age models, due to their over-reliance on structural indicators (such as gray matter volume and cortical thickness) and neglect of dynamic functional parameters, cannot distinguish between physiological and pathological aging, thus limiting their clinical applicability.

[0003] Brain entropy can quantify the complexity and dynamism of brain function, and abnormal prefrontal brain entropy in patients with depression is significantly associated with cognitive decline. However, traditional signal analysis methods struggle to reliably extract brain entropy features from neuroimaging data, which greatly limits the application of brain entropy as a biomarker in the diagnosis of depression.

[0004] Functional near-infrared spectroscopy (fNIRS) can reflect neurovascular coupling activity by real-time detection of changes in the concentration of oxyhemoglobin and deoxyhemoglobin (HbO / HbR) in the cerebral cortex (with precision down to the micromolar level). This technology boasts high temporal resolution (capturing millisecond-level neural activity at a 10Hz sampling rate), strong resistance to motion artifacts (allowing for natural head movements within a 5mm range), silent and non-invasive detection (no radiation and no claustrophobia), portability and low cost (device size is only 1 / 20th that of MRI, reducing single-test cost by 90%), and multi-parameter simultaneous acquisition capability (simultaneously acquiring three-dimensional dynamic data of HbO, HbR, and total hemoglobin). However, existing fNIRS analysis methods only focus on linear time-domain indicators (such as HbO amplitude) and fail to capture nonlinear dynamic characteristics (such as phase synchronization and signal complexity), leading to insufficient identification of specific patterns in depression-related mental illness (MDD) and limiting the in-depth application of fNIRS technology in the diagnosis of mental disorders.

[0005] Therefore, there is an urgent need to develop a novel analytical method and system that integrates fNIRS multimodal data with brain entropy quantification models to overcome the following technical bottlenecks: integrating dynamic features of brain entropy with sMRI structural indicators to improve the discriminative power of brain age prediction for depression MDD; extracting depression-related parameters from fNIRS signals by improving entropy algorithms; and constructing a lightweight fNIRS brain age prediction system for depression to achieve rapid clinical detection and automated analysis. Summary of the Invention

[0006] This specification provides a method and system for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy, which addresses the following technical problems in the prior art: the lack of objective biomarkers, insufficient accuracy of brain age models, and insufficient depth of fNIRS analysis in existing depression diagnoses, providing a solution for the accurate prediction and diagnosis of mental illnesses.

[0007] The embodiments in this specification adopt the following technical solutions:

[0008] According to the first aspect, this application provides a method for predicting brain age and assisting in the diagnosis of depression based on prefrontal fNIRS entropy, which includes the following steps:

[0009] Step S1: Acquire the resting-state hemoglobin time series signal of the prefrontal cortex using a multi-channel near-infrared spectroscopy optical brain imaging device, and preprocess the resting-state hemoglobin time series signal of the prefrontal cortex;

[0010] Step S2: Extract the resting-state hemoglobin time series signal from the preprocessed prefrontal cortex, and extract at least the oxyhemoglobin time series signal, the deoxyhemoglobin time series signal, and the total hemoglobin time series signal.

[0011] Step S3: Calculate the static permutation entropy of the oxyhemoglobin time series signal, the static permutation entropy of the deoxyhemoglobin time series signal, and the static permutation entropy of the total hemoglobin time series signal according to the preset time series, wherein the preset time series includes all time series.

[0012] Step S4: Based on the preset window data, the prefrontal resting-state hemoglobin time series signal is dynamically segmented using the sliding window method to obtain prefrontal resting-state hemoglobin time series signals in N consecutive overlapping time segments. Based on the N prefrontal resting-state hemoglobin time series signals, the dynamic entropy characteristics of the oxyhemoglobin time series signal, the dynamic entropy characteristics of the deoxyhemoglobin time series signal, and the dynamic entropy characteristics of the total hemoglobin time series signal are calculated respectively.

[0013] Step S5: Obtain static brain entropy features, dynamic brain entropy features, and comprehensive brain entropy features;

[0014] Step S6: Construct a brain age prediction model;

[0015] Step S7: Input the static brain entropy features, dynamic brain entropy features, comprehensive brain entropy features and the individual's feature values ​​into the brain age prediction model to obtain the predicted brain age value of the individual. The individual's feature values ​​include at least the individual's actual age. Calculate the difference between the predicted brain age value and the actual age to obtain the brain age deviation value. Based on the brain age deviation value, perform auxiliary diagnosis of depression to obtain the depression prediction result.

[0016] Furthermore, step S1 involves preprocessing the prefrontal resting-state hemoglobin time-series signal, specifically including:

[0017] The preprocessing includes at least one or more of the following: head motion removal, filtering, and physiological noise cancellation. The filtering process uses bandpass filtering with a frequency of 0.01-0.5Hz.

[0018] Furthermore, the calculation of the static permutation entropy of the oxyhemoglobin time series signal in step S3 specifically includes:

[0019] Based on the preset embedding dimension and time delay, the time series is reconstructed into a phase space vector;

[0020] For each phase space vector, its elements are arranged in ascending order to obtain multiple arrangement patterns;

[0021] Calculate the probability of multiple permutation patterns occurring in a time series;

[0022] The static permutation entropy is calculated based on the probability using the information entropy formula.

[0023] Further, step S4, which calculates the dynamic entropy characteristics of the oxyhemoglobin time series signal, the deoxyhemoglobin time series signal, and the total hemoglobin time series signal based on N prefrontal resting-state hemoglobin time series signals, specifically includes:

[0024] Based on N prefrontal cortex resting-state hemoglobin time segment signals, oxyhemoglobin time segment signals in N consecutive overlapping time segments are obtained. The permutation entropy of the oxyhemoglobin time segment signal in each time segment is calculated to obtain the oxyhemoglobin permutation entropy sequence. The oxyhemoglobin permutation entropy sequence is statistically analyzed to extract the dynamic entropy characteristics of the oxyhemoglobin time sequence signal.

[0025] Based on N prefrontal cortex resting-state hemoglobin time segment signals, deoxyhemoglobin time segment signals in N consecutive overlapping time segments are obtained. The permutation entropy of the deoxyhemoglobin time segment signal in each time segment is calculated to obtain the deoxyhemoglobin permutation entropy sequence. Statistical analysis is performed on the deoxyhemoglobin permutation entropy sequence to extract the dynamic entropy characteristics of the deoxyhemoglobin time sequence signal.

[0026] Based on N prefrontal cortex resting-state hemoglobin time segment signals, total hemoglobin time segment signals in N consecutive overlapping time segments are obtained. The permutation entropy of the total hemoglobin time segment signal in each time segment is calculated to obtain the total hemoglobin permutation entropy sequence. Statistical analysis is performed on the total hemoglobin permutation entropy sequence to extract the dynamic entropy characteristics of the total hemoglobin time series signal.

[0027] Furthermore, dynamic entropy characteristics include at least the mean dynamic entropy, the standard deviation of dynamic entropy, and the coefficient of variation.

[0028] Furthermore, step S5, which involves obtaining static brain entropy features, dynamic brain entropy features, and comprehensive brain entropy features, specifically includes:

[0029] The static arrangement entropy of the oxyhemoglobin time series signal, the static arrangement entropy of the deoxyhemoglobin time series signal, and the static arrangement entropy of the total hemoglobin time series signal are spliced ​​together to obtain the static brain entropy features.

[0030] The dynamic entropy features of the oxyhemoglobin time series signal, the deoxyhemoglobin time series signal, and the total hemoglobin time series signal are spliced ​​together to obtain the dynamic brain entropy features.

[0031] The comprehensive brain entropy features are obtained by splicing together the static arrangement entropy of the oxyhemoglobin time series signal, the static arrangement entropy of the deoxyhemoglobin time series signal, the static arrangement entropy of the total hemoglobin time series signal, the dynamic entropy features of the oxyhemoglobin time series signal, the dynamic entropy features of the deoxyhemoglobin time series signal, and the dynamic entropy features of the total hemoglobin time series signal.

[0032] Furthermore, the construction of the brain age prediction model in step S6 specifically includes:

[0033] A brain age prediction model framework is constructed using support vector machines;

[0034] By employing a radial basis function kernel function and pre-setting parameters, and optimizing the brain age prediction model framework through a leave-one-out cross-validation strategy, a brain age prediction model is obtained.

[0035] Furthermore, the prediction of depression based on brain age bias in step S7 specifically includes:

[0036] The brain age prediction value is used as a single feature input to the support vector machine, and the classification threshold is determined based on the Youden index.

[0037] The brain age deviation value was integrated with the clinical scale to construct a composite feature, which was then classified using a support vector machine.

[0038] The depression classification result is output based on a pre-set decision threshold through weighted probability fusion.

[0039] According to the second aspect, this application provides a brain age prediction and auxiliary diagnosis system for depression based on prefrontal fNIRS entropy, wherein the brain age prediction and auxiliary diagnosis system for depression based on prefrontal fNIRS entropy is used to implement the brain age prediction and auxiliary diagnosis method for depression based on prefrontal fNIRS entropy as described in any of the above claims, and the system includes at least a multi-channel near-infrared spectroscopy optical brain imaging device.

[0040] Furthermore, the multi-channel near-infrared spectroscopy optical brain imaging device is a 26-channel near-infrared spectroscopy optical brain imaging device; and / or

[0041] The multi-channel near-infrared spectroscopy optical brain imaging device has a head shape curvature radius of 80-120 mm; and / or

[0042] The distance between the multi-channel near-infrared spectroscopy optical brain imaging device and the external light source is 25-35mm; and / or

[0043] The multi-channel near-infrared spectroscopy optical brain imaging equipment uses a frequency ≥5Hz.

[0044] Furthermore, the distance between the multi-channel near-infrared spectral optical brain imaging device and the external light source is 30mm.

[0045] Furthermore, the brain age prediction and auxiliary diagnostic system for depression based on prefrontal fNIRS entropy also includes:

[0046] The peripheral light source uses an 11-transmitter × 8-receiver optical array with a wavelength of 735nm-850nm, covering the prefrontal cortex region.

[0047] The real-time entropy analysis engine uses a field-programmable gate array to calculate one or more of the following: static entropy, dynamic entropy mean, dynamic entropy standard deviation, and coefficient of variation.

[0048] A visualization interface is used to dynamically display one or more of the following: brain age deviation value, area under the working characteristic curve of the individual under test, and prefrontal entropy change heatmap.

[0049] According to a third aspect, this application provides a device for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy, comprising: at least one processor; and,

[0050] A memory communicatively connected to the at least one processor; wherein,

[0051] The memory stores instructions that can be executed by the at least one processor, which enables the at least one processor to perform any of the foregoing methods for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy.

[0052] According to a third aspect, this application provides a non-volatile computer storage medium for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy, storing computer-executable instructions, which, when executed, can realize the method for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy as described above.

[0053] The above-described at least one technical solution adopted in the embodiments of this specification can achieve the following beneficial effects:

[0054] 1. The fNIRS entropy-based brain age-assisted diagnostic method for depression provided in the embodiments of this specification breaks through the limitation of traditional fNIRS which only analyzes linear changes in hemoglobin concentration. It innovatively integrates static permutation entropy (capturing signal complexity) and dynamic entropy features (CV value characterizing the stability of neural activity) to construct a highly discriminative biomarker.

[0055] 2. The brain age-assisted diagnostic method for depression based on fNIRS entropy provided in the embodiments of this specification constructs a four-level SVM classification system: at the micro level, it accurately captures local brain function abnormalities through single-channel dynamic entropy features (6 dimensions / channel); at the meso level, it adopts comprehensive brain entropy features (78-dimensional static entropy features, 78-dimensional dynamic entropy features, and 156-dimensional comprehensive brain entropy features); and at the macro level, it introduces brain age prediction values ​​and brain age deviation values ​​to quantify the degree of aging of the brain system, effectively solving the problem of insufficient integration of multi-scale features of brain function in existing technologies. Attached Figure Description

[0056] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0057] Figure 1 This is a flowchart illustrating a method for auxiliary diagnosis of depression based on fNIRS entropy using brain age in an embodiment of this specification.

[0058] Figure 2 This is a schematic diagram of another method for assisting in the diagnosis of depression based on fNIRS entropy, as provided in one embodiment of this specification.

[0059] Figure 3 This is a schematic diagram of the photoelectric arrangement of a brain age-assisted diagnostic system for depression based on fNIRS entropy provided in one embodiment of this specification.

[0060] Figure 4 The diagram shows the channel distribution characteristics of static arrangement entropy and dynamic entropy features obtained by using the brain age prediction and auxiliary diagnosis system for depression based on prefrontal fNIRS entropy provided in an embodiment of this specification, and by using the brain age prediction and auxiliary diagnosis method for depression based on prefrontal fNIRS entropy provided in an embodiment of this specification to predict and assist the brain age of multiple individuals to be tested.

[0061] Figure 5 This document describes the brain age prediction and auxiliary diagnosis system for depression based on prefrontal fNIRS entropy provided in an embodiment of this specification. It also describes the brain age prediction and auxiliary diagnosis method for depression based on prefrontal fNIRS entropy provided in an embodiment of this specification. The brain age prediction and auxiliary diagnosis of multiple individuals are performed, and the resulting brain entropy and brain age indicators are used to identify the efficacy of depression. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of this specification clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments in this specification without creative effort are within the scope of protection of this application.

[0063] The technical solutions provided by the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0064] Functional near-infrared spectroscopy (fNIRS) can reflect neurovascular coupling activity by real-time detection of changes in the concentration of oxyhemoglobin and deoxyhemoglobin (HbO / HbR) in the cerebral cortex (with precision down to the micromolar level). It boasts advantages such as high temporal resolution (capturing millisecond-level neural activity at a 10Hz sampling rate), strong resistance to motion artifacts (allowing for natural head movements within a 5mm range), silent and non-invasive detection (no radiation and no claustrophobia), portability and low cost (device size is only 1 / 20th that of MRI, reducing single-test cost by 90%), and multi-parameter simultaneous acquisition capability (simultaneously acquiring three-dimensional dynamic data of HbO, HbR, and total hemoglobin). However, existing fNIRS analysis methods only focus on linear time-domain indicators (such as HbO amplitude) and fail to capture nonlinear dynamic characteristics (such as phase synchronization and signal complexity), leading to insufficient identification of specific patterns in depression-related mental health disorders (MDD).

[0065] The technical solutions provided by the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0066] Reference Figure 1 and Figure 2 As shown, this embodiment provides a method for predicting brain age and assisting in the diagnosis of depression based on prefrontal fNIRS entropy, which includes the following steps:

[0067] Step S1: Acquire the resting-state hemoglobin time series signal of the prefrontal cortex using a multi-channel near-infrared spectroscopy optical brain imaging device, and preprocess the resting-state hemoglobin time series signal of the prefrontal cortex.

[0068] Preferably, a 26-channel near-infrared spectroscopy optical brain imaging device (fNIRS) is used to acquire the prefrontal cortex resting-state hemoglobin time-series signal of the individual under test at a sampling frequency of ≥5 Hz. The prefrontal cortex resting-state hemoglobin time-series signal is then preprocessed, including head motion removal, filtering, and physiological noise reduction. Preferably, bandpass filtering is used with a frequency of 0.01-0.5 Hz.

[0069] Step S2: Extract the preprocessed resting-state hemoglobin time series signal from the prefrontal cortex, and extract at least the oxyhemoglobin time series signal, the deoxyhemoglobin time series signal, and the total hemoglobin time series signal.

[0070] After preprocessing, the time series signals of oxyhemoglobin (HbO), deoxyhemoglobin (HbR), and total hemoglobin (HbT) from 26 channels were extracted.

[0071] Step S3: Calculate the static permutation entropy of the oxyhemoglobin time series signal, the static permutation entropy of the deoxyhemoglobin time series signal, and the static permutation entropy of the total hemoglobin time series signal according to the preset time series, wherein the preset time series includes all time series.

[0072] Step S3 specifically includes: reconstructing the time series into a phase space vector according to the preset embedding dimension and time delay;

[0073] For each phase space vector, its elements are arranged in ascending order to obtain multiple arrangement patterns;

[0074] Calculate the probability of multiple permutation patterns occurring in a time series;

[0075] The static permutation entropy is calculated based on the probability using the information entropy formula.

[0076] The static permutation entropy was calculated for each of the 26 extracted time-series signals of the three types of hemoglobin (a total of 78 signals). The calculation steps for the static permutation entropy are further explained in detail below:

[0077] Step S3.1 Phase Space Reconstruction

[0078] A time series X = {x1, x2, ..., x3} is pre-defined. N}, selecting the embedding dimension m (usually 3-7) and time delay τ (usually 1), the pre-set time series is reconstructed into a phase space vector: Xi = (x i x i+τ , ..., x i+(m-1)τ ), where: i=1, 2,..., N-(m-1)τ.

[0079] Step S3.2 Calculate the permutation pattern

[0080] For each phase space vector X i Arrange its elements in ascending order to obtain the permutation pattern π. For example, if X i = (x3, x1, x2), then the permutation pattern is π = (3, 2, 1).

[0081] Step S3.3 Statistical distribution of permutation patterns

[0082] Calculate all possible permutation patterns π k The probability P(π) of (m!) kinds appearing in the time series k ).

[0083] Step S3.4 Calculate the static permutation entropy

[0084] The static permutation entropy is calculated using the information entropy formula, where the formula is:

[0085]

[0086] For normalization, it is usually divided by ln(m!), as shown in the formula:

[0087] PE norm =PE(m) / ln(m!)

[0088] Among them, PE norm The range of values ​​for is: 0 ≤ PE norm ≤1, PE norm The closer the value is to 1, the more random the signal; the closer it is to 0, the more regular the signal.

[0089] Step S3.5 Calculate the static permutation entropy of all time series.

[0090] Select the preset time series as the entire time series, and calculate the static permutation entropy (labeled as HbO_Static) of the oxyhemoglobin time series signal, the static permutation entropy (labeled as HbR_Static) of the deoxyhemoglobin time series signal, and the static permutation entropy (labeled as HbT_Static) of the total hemoglobin time series signal, respectively, according to steps S3.1 to S3.4 above.

[0091] Step S4: Based on the preset window data, the prefrontal resting-state hemoglobin time series signal is dynamically segmented using the sliding window method to obtain prefrontal resting-state hemoglobin time series signals in N consecutive overlapping time segments. Based on the N prefrontal resting-state hemoglobin time series signals, the dynamic entropy characteristics of the oxyhemoglobin time series signal, the dynamic entropy characteristics of the deoxyhemoglobin time series signal, and the dynamic entropy characteristics of the total hemoglobin time series signal are calculated respectively.

[0092] Step S4 specifically includes:

[0093] Based on N prefrontal cortex resting-state hemoglobin time segment signals, oxyhemoglobin time segment signals in N consecutive overlapping time segments are obtained. The permutation entropy of the oxyhemoglobin time segment signal in each time segment is calculated to obtain the oxyhemoglobin permutation entropy sequence. The oxyhemoglobin permutation entropy sequence is statistically analyzed to extract the dynamic entropy characteristics of the oxyhemoglobin time sequence signal.

[0094] Based on N prefrontal cortex resting-state hemoglobin time segment signals, deoxyhemoglobin time segment signals in N consecutive overlapping time segments are obtained. The permutation entropy of the deoxyhemoglobin time segment signal in each time segment is calculated to obtain the deoxyhemoglobin permutation entropy sequence. Statistical analysis is performed on the deoxyhemoglobin permutation entropy sequence to extract the dynamic entropy characteristics of the deoxyhemoglobin time sequence signal.

[0095] Based on N prefrontal cortex resting-state hemoglobin time segment signals, total hemoglobin time segment signals in N consecutive overlapping time segments are obtained. The permutation entropy of the total hemoglobin time segment signal in each time segment is calculated to obtain the total hemoglobin permutation entropy sequence. Statistical analysis is performed on the total hemoglobin permutation entropy sequence to extract the dynamic entropy characteristics of the total hemoglobin time series signal.

[0096] Further details on the calculation process of the dynamic entropy characteristics of oxyhemoglobin time-series signals:

[0097] First, the oxyhemoglobin (HbO) time series is dynamically segmented based on the sliding window method. For example, the window length is set to 100ms and the step size is 50ms, and the complete time series signal is divided into N consecutive overlapping time segments.

[0098] Then, the static permutation entropy is calculated independently for each time segment to obtain the entropy value sequence (PE1, PE2, ..., PE). N ).

[0099] Finally, three dynamic characteristic indicators were extracted through statistical analysis:

[0100] 1) The mean dynamic entropy (μPE) is used to reflect the overall signal complexity. The formula for calculating the mean dynamic entropy is:

[0101] μPE=ΣPE i / N;

[0102] 2) The standard deviation of dynamic entropy (σPE) is used to characterize the intensity of dynamic fluctuations in brain function;

[0103] 3) Coefficient of variation, the formula for calculating the coefficient of variation is:

[0104] CV = σPE / μPE

[0105] The dynamic entropy mean, dynamic entropy standard deviation and coefficient of variation are combined as dynamic entropy features to obtain the dynamic entropy features of the oxyhemoglobin time series signal (labeled as HbO_dynamic).

[0106] The above calculation process is repeated synchronously to calculate the dynamic entropy characteristics (labeled as HbR_dynamic) of the deoxyhemoglobin time series signal and the dynamic entropy characteristics (labeled as HbT_dynamic) of the total hemoglobin time series signal.

[0107] Step S5: Obtain static brain entropy features, dynamic brain entropy features, and comprehensive brain entropy features.

[0108] Step S5 specifically includes:

[0109] The static arrangement entropy of the oxyhemoglobin time series signal, the static arrangement entropy of the deoxyhemoglobin time series signal, and the static arrangement entropy of the total hemoglobin time series signal are spliced ​​together to obtain the static brain entropy features.

[0110] The dynamic entropy features of the oxyhemoglobin time series signal, the deoxyhemoglobin time series signal, and the total hemoglobin time series signal are spliced ​​together to obtain the dynamic brain entropy features.

[0111] The comprehensive brain entropy features are obtained by splicing together the static arrangement entropy of the oxyhemoglobin time series signal, the static arrangement entropy of the deoxyhemoglobin time series signal, the static arrangement entropy of the total hemoglobin time series signal, the dynamic entropy features of the oxyhemoglobin time series signal, the dynamic entropy features of the deoxyhemoglobin time series signal, and the dynamic entropy features of the total hemoglobin time series signal.

[0112] Further details are as follows:

[0113] By concatenating HbO_Static, HbR_Static, and HbT_Static, we obtain the static brain entropy feature (labeled as Static_al l).

[0114] By concatenating HbO_Dynamic, HbR_Dynamic, and HbT_Dynamic, we obtain the dynamic brain entropy feature (labeled as Dynamic_al l).

[0115] By concatenating HbO_Static, HbR_Static, HbT_Static, HbO_Dynamic, HbR_Dynamic, and HbT_Dynamic, we obtain the comprehensive brain entropy feature (labeled as ALL).

[0116] Step S6 involves constructing a brain age prediction model. Step S6 specifically includes:

[0117] A brain age prediction model framework is constructed using support vector machines;

[0118] By employing a radial basis function kernel function and pre-setting parameters, and optimizing the brain age prediction model framework through a leave-one-out cross-validation strategy, a brain age prediction model is obtained.

[0119] Further details on the construction method of the brain age prediction model:

[0120] The Support Vector Regression (SVR) algorithm is used as the core prediction framework. The radial basis function (RBF) kernel function is used and the optimal hyperparameters are set (penalty coefficient C = 1.0, kernel function parameter γ = 0.01, insensitive band ε = 0.1). The robustness and generalization ability of the model are ensured by leave-one-out cross-validation (LOOCV) strategy.

[0121] Step S7: Input the static brain entropy features, dynamic brain entropy features, comprehensive brain entropy features and the individual's feature values ​​into the brain age prediction model to obtain the predicted brain age value of the individual. The individual's feature values ​​include at least the individual's actual age. Calculate the difference between the predicted brain age value and the actual age to obtain the brain age deviation value. Based on the brain age deviation value, perform auxiliary diagnosis of depression to obtain the depression prediction result.

[0122] Step S7, which involves predicting depression based on brain age bias, specifically includes:

[0123] The brain age prediction value is used as a single feature input to the support vector machine, and the classification threshold is determined based on the Youden index.

[0124] The brain age deviation value was integrated with the clinical scale to construct a composite feature, which was then classified using a support vector machine.

[0125] The depression classification result is output based on a pre-set decision threshold through weighted probability fusion.

[0126] Multidimensional entropy features extracted from the fNIRS signal (HbO_Static, HbR_Static, HbT_Static, HbO_Dynamic, HbR_Dynamic, and HbT_Dynamic for 26 channels, and combined features Static_all, Dynamic_all, and ALL) are input into the brain age prediction model, along with the individual's characteristic values ​​(such as age and gender). All features are standardized and preprocessed before input. The brain age prediction model outputs the predicted brain age for the individual, and the difference between the predicted brain age and the actual age is calculated to obtain the brain age deviation value (labeled as BrainAGE).

[0127] A four-class support vector machine (SVM) classification model was constructed, and the specific implementation process is as follows: First, based on single-channel brain entropy features (including six-dimensional features: HbO_Static, HbR_Static, HbT_Static, HbO_Dynamic, HbR_Dynamic, and HbT_Dynamic), channels with significant differences were screened through a two-sample t-test (p<0.05, FDR correction). An RBF kernel SVM (γ=0.1, C=1.0) was independently trained for each channel, and a hard voting mechanism was used to fuse the channel-level classification results. Second, for combined brain entropy features (Static_al l is a 78-dimensional statically arranged entropy direction, Dynamic_al l...),... The 78-dimensional dynamic entropy feature (l) and the 156-dimensional comprehensive brain entropy feature (ALL) were dimensionality-reduced using PCA (preserving 95% variance) and then input into a linear kernel SVM (C=0.5) for modeling. Next, the predicted brain age value was used as a single feature input into the linear SVM (C=1.0), and the classification threshold was determined based on the Youden index. Finally, the brain age deviation value (BrainAGE) and clinical scales were fused to construct a composite feature (Z=0.6×BrainAGE+0.4×HAMD_score), which was then classified using an RBF kernel SVM (γ=0.05, C=0.8). All model features were preprocessed using standardization (x-μ) / σ, and the brain age feature values ​​underwent Box-Cox transformation (λ=0.3). Finally, a weighted probability fusion (weight coefficients 0.3 / 0.4 / 0.1 / 0.2) was used to output a comprehensive classification result with a decision threshold of 0.65, providing auxiliary diagnosis for depression and obtaining depression prediction results.

[0128] The method for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy provided in this embodiment breaks through the limitation of traditional fNIRS that only analyzes linear changes in hemoglobin concentration. It innovatively integrates static permutation entropy (capturing signal complexity) and dynamic entropy features (CV value characterizing the stability of neural activity) to construct a highly discriminative biomarker.

[0129] The brain age prediction and auxiliary diagnosis method for depression based on prefrontal fNIRS entropy provided in this embodiment constructs a four-level SVM classification system: at the micro level, it accurately captures local brain functional abnormalities through single-channel dynamic entropy features (6 dimensions / channel); at the meso level, it adopts comprehensive brain entropy features (78-dimensional static entropy features, 78-dimensional dynamic entropy features, and 156-dimensional comprehensive brain entropy features); and at the macro level, it introduces brain age prediction values ​​and brain age deviation values ​​to quantify the degree of aging of the brain system, effectively solving the problem of insufficient integration of multi-scale features of brain function in existing technologies.

[0130] This embodiment also provides a brain age prediction and auxiliary diagnosis system for depression based on prefrontal fNIRS entropy, which is used to realize the above-mentioned brain age auxiliary diagnosis method for depression based on fNIRS entropy. The system includes at least a multi-channel near-infrared spectroscopy optical brain imaging device, an external light source and a real-time entropy analysis engine.

[0131] A schematic diagram of the light pole arrangement in a brain age prediction and auxiliary diagnostic system for depression based on prefrontal fNIRS entropy, as shown below. Figure 3 As shown, the external light source uses an array of 11 emitters and 8 receivers with wavelengths of 735nm-850nm, covering the prefrontal cortex region and forming 26 measurement channels (S represents the light source pole, and D represents the detector pole); the spatial distribution of the 26 prefrontal cortex channels (purple indicates the light source pole, and blue indicates the detector pole).

[0132] The near-infrared spectroscopy optical brain imaging device has a head shape curvature radius of 80-120mm, a distance of 25-35mm between the device and the external light source, and an operating frequency of ≥5Hz. The optimal distance between the device and the external light source is 30mm.

[0133] The real-time entropy analysis engine uses a field-programmable gate array (FPGA) to calculate static entropy, dynamic entropy mean, dynamic entropy standard deviation, and coefficient of variation.

[0134] The brain age prediction and auxiliary diagnosis system for depression based on prefrontal fN I RS entropy may also include a visualization interface, which dynamically displays one or more of the brain age deviation value, the area under the working characteristic curve of the individual under test, and the heat map of prefrontal entropy change.

[0135] The brain age prediction and auxiliary diagnostic system for depression based on prefrontal cortex fNIR entropy provided in this embodiment represents a significant upgrade in clinical applicability: First, the detection process is significantly optimized, supporting rapid screening in 5 minutes (traditional MRI requires 45 minutes), greatly improving accessibility to primary healthcare; second, it innovatively employs dynamic entropy indicators to achieve precise efficacy monitoring. These innovations make this technology suitable for the entire clinical cycle from early screening and efficacy assessment to long-term prognostic monitoring.

[0136] The brain age prediction and auxiliary diagnosis system for depression based on prefrontal fNIRS entropy provided in this embodiment has several advantages. In terms of hardware compatibility, the system is compatible with various fNIRS devices with 26-256 channels (the minimum configuration requirement is a sampling rate ≥10Hz), ensuring seamless integration with devices from different medical institutions. At the algorithm level, it innovatively provides a PyTorch-based federated learning framework, enabling secure sharing of multi-center data through distributed machine learning. For clinical integration, the system can automatically generate structured diagnostic reports containing entropy-varying brain function maps, accurate brain age deviation values, and depression risk probability scores (0-100%). It supports seamless integration with hospital information systems via the DICOM standard interface, achieving intelligent processing from data acquisition to clinical decision-making. These technological breakthroughs give this invention broad clinical application adaptability and scalability.

[0137] The brain age prediction and auxiliary diagnosis system for depression based on prefrontal fNIR entropy provided in this embodiment was used, and the brain age prediction and auxiliary diagnosis method for depression based on prefrontal fNIR entropy provided in this embodiment were used to predict and diagnose the brain age of multiple individuals. Among them, 45 individuals were in the healthy control group and 39 individuals were in the depression group. Static permutation entropy and dynamic entropy characteristics were obtained, such as... Figure 4 As shown, this represents the channel distribution characteristics of static arrangement entropy and dynamic entropy. Figure 4 In the table, HC represents the healthy control group; MDD represents the depression group; AC represents the static permutation entropy of HbO / HbR / HbT in the healthy control group; DF represents the mean static permutation entropy of the healthy control group; GI represents the static permutation entropy of the depression group; KM represents the mean static permutation entropy of the depression group; NP represents the dynamic entropy feature of the healthy control group; QS represents the mean dynamic entropy feature of the healthy control group; TV represents the dynamic entropy feature of the depression group; and WY represents the mean dynamic entropy feature of the depression group.

[0138] This embodiment employs the brain age prediction and auxiliary diagnosis system for depression based on prefrontal fNIRS entropy, and uses the brain age prediction and auxiliary diagnosis method for depression based on prefrontal fNIRS entropy provided in this embodiment to predict and assist in the diagnosis of brain age in multiple individuals. The resulting brain entropy and brain age indicators demonstrate their diagnostic efficacy for depression, such as... Figure 5 As shown in the figure, Figure A shows the six entropy features with the highest discrimination (AUC range: 0.70-0.73) in single-channel analysis, with CH2_HbO_static exhibiting the best classification accuracy; Figure B shows the improvement in diagnostic efficacy of combined hemodynamic parameters: a progressive improvement from single features (AUC = 0.82-0.95) to comprehensive combinations (AUC = 0.99-1.00); Figure C shows the classification performance of the brain age prediction model: complete discrimination is achieved with multiple feature combinations (AUC = 1.00); Figure D shows the diagnostic value of brain age difference (BAG) analysis: complete separation between groups is achieved when all entropy features are integrated (AUC = 1.00). In the figures, AUC represents the area under the curve; MDD represents major depressive disorder; HbO represents oxyhemoglobin; HbR represents deoxyhemoglobin; HbT represents total hemoglobin; and BAG represents the brain age deviation value.

[0139] This embodiment also provides a device for predicting and assisting in the diagnosis of depression based on prefrontal cortex fNIR entropy, including: at least one processor; and,

[0140] A memory communicatively connected to the at least one processor; wherein,

[0141] The memory stores instructions that can be executed by the at least one processor, which enables the at least one processor to perform the above-described method for predicting and assisting in the diagnosis of depression based on prefrontal fN I RS entropy.

[0142] This embodiment also provides a non-volatile computer storage medium for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy, which stores computer-executable instructions. When the computer-executable instructions are executed, the above-mentioned method for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy can be implemented.

[0143] The various embodiments in this application are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the device and medium embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the description of the method embodiments.

[0144] The devices and media provided in this application are one-to-one with the methods. Therefore, the devices and media also have similar beneficial technical effects as their corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the devices and media will not be repeated here.

[0145] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0146] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0147] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0148] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0149] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0150] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0151] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0152] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0153] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for predicting brain age and assisting in the diagnosis of depression based on prefrontal fNIRS entropy, characterized in that, Includes the following steps: Step S1: Acquire the resting-state hemoglobin time series signal of the prefrontal cortex using a multi-channel near-infrared spectroscopy optical brain imaging device, and preprocess the resting-state hemoglobin time series signal of the prefrontal cortex. Step S2: Extract the preprocessed resting-state hemoglobin time series signal from the prefrontal cortex, and extract at least the oxyhemoglobin time series signal, the deoxyhemoglobin time series signal, and the total hemoglobin time series signal. Step S3: Calculate the static permutation entropy of the oxyhemoglobin time series signal, the static permutation entropy of the deoxyhemoglobin time series signal, and the static permutation entropy of the total hemoglobin time series signal according to the preset time series, wherein the preset time series includes all time series. Step S4: Based on the preset window data, the prefrontal resting-state hemoglobin time series signal is dynamically segmented using the sliding window method to obtain prefrontal resting-state hemoglobin time series signals in N consecutive overlapping time segments. Based on the N prefrontal resting-state hemoglobin time series signals, the dynamic entropy characteristics of the oxyhemoglobin time series signal, the dynamic entropy characteristics of the deoxyhemoglobin time series signal, and the dynamic entropy characteristics of the total hemoglobin time series signal are calculated respectively. Step S5: Obtain static brain entropy features, dynamic brain entropy features, and comprehensive brain entropy features; Step S6: Construct a brain age prediction model; Step S7: Input the static brain entropy features, the dynamic brain entropy features, the comprehensive brain entropy features, and the individual's feature values ​​into the brain age prediction model to obtain the predicted brain age value of the individual. The individual's feature values ​​include at least the individual's actual age. Calculate the difference between the predicted brain age value and the actual age to obtain the brain age deviation value, and use the brain age deviation value to assist in the diagnosis of depression.

2. The method for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy according to claim 1, characterized in that, The preprocessing of the prefrontal resting-state hemoglobin time-series signal in step S1 specifically includes: The preprocessing includes at least one or more of the following: head motion removal processing, filtering processing, and physiological noise cancellation processing. The filtering processing uses bandpass filtering with a frequency of 0.01-0.5Hz.

3. The method for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy according to claim 1, characterized in that, The calculation of the static permutation entropy of the oxyhemoglobin time series signal in step S3 specifically includes: Based on the preset embedding dimension and time delay, the time series is reconstructed into a phase space vector; For each phase space vector, its elements are arranged in ascending order to obtain multiple arrangement patterns; Calculate the probability of the occurrence of the multiple permutation patterns in the time series; The static permutation entropy is calculated based on the probability using the information entropy formula.

4. The method for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy according to claim 1, characterized in that, Step S4, which involves calculating the dynamic entropy characteristics of the oxyhemoglobin time series signal, the deoxyhemoglobin time series signal, and the total hemoglobin time series signal based on the N prefrontal resting-state hemoglobin time series signals, specifically includes: Based on N prefrontal resting-state hemoglobin time segment signals, oxyhemoglobin time segment signals in N consecutive overlapping time segments are obtained. The permutation entropy of the oxyhemoglobin time segment signal in each time segment is calculated to obtain the oxyhemoglobin permutation entropy sequence. The oxyhemoglobin permutation entropy sequence is statistically analyzed to extract the dynamic entropy characteristics of the oxyhemoglobin time sequence signal. Based on N prefrontal resting-state hemoglobin time segment signals, deoxyhemoglobin time segment signals in N consecutive overlapping time segments are obtained. The permutation entropy of the deoxyhemoglobin time segment signal in each time segment is calculated to obtain the deoxyhemoglobin permutation entropy sequence. The deoxyhemoglobin permutation entropy sequence is statistically analyzed to extract the dynamic entropy characteristics of the deoxyhemoglobin time sequence signal. Based on N prefrontal resting-state hemoglobin time segment signals, total hemoglobin time segment signals in N consecutive overlapping time segments are obtained. The permutation entropy of the total hemoglobin time segment signal in each time segment is calculated to obtain the total hemoglobin permutation entropy sequence. The total hemoglobin permutation entropy sequence is statistically analyzed to extract the dynamic entropy features of the total hemoglobin time series signal.

5. The method for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy according to claim 1, characterized in that, The dynamic entropy features include at least the dynamic entropy mean, dynamic entropy standard deviation, and coefficient of variation.

6. The method for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy according to claim 1, characterized in that, The acquisition of static brain entropy features, dynamic brain entropy features, and comprehensive brain entropy features in step S5 specifically includes: The static arrangement entropy of the oxyhemoglobin time series signal, the static arrangement entropy of the deoxyhemoglobin time series signal, and the static arrangement entropy of the total hemoglobin time series signal are spliced ​​together to obtain the static brain entropy feature. The dynamic entropy features of the oxyhemoglobin time series signal, the dynamic entropy features of the deoxyhemoglobin time series signal, and the dynamic entropy features of the total hemoglobin time series signal are concatenated to obtain dynamic brain entropy features. The static arrangement entropy of the oxyhemoglobin time series signal, the static arrangement entropy of the deoxyhemoglobin time series signal, the static arrangement entropy of the total hemoglobin time series signal, the dynamic entropy feature of the oxyhemoglobin time series signal, the dynamic entropy feature of the deoxyhemoglobin time series signal, and the dynamic entropy feature of the total hemoglobin time series signal are concatenated to obtain the comprehensive brain entropy feature.

7. The method for predicting and assisting the diagnosis of depression based on prefrontal fNIRS entropy according to claim 1, characterized in that, The construction of the brain age prediction model in step S6 specifically includes: A brain age prediction model framework is constructed using support vector machines; By employing a radial basis function kernel function and pre-setting parameters, the brain age prediction model framework is optimized using a leave-one-out cross-validation strategy to obtain the brain age prediction model.

8. The method for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy according to claim 1, characterized in that, The step S7, which involves predicting depression based on the brain age deviation value, specifically includes: The brain age prediction value is used as a single feature input to the support vector machine, and the classification threshold is determined based on the Youden index. The brain age deviation value was fused with a clinical scale to construct a composite feature, which was then classified using a support vector machine. The depression classification result is output based on a pre-set decision threshold through weighted probability fusion.

9. A system for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy, characterized in that: The system for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy is used to implement the method for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy as described in any one of claims 1-8, wherein the system includes at least a multi-channel near-infrared spectroscopy optical brain imaging device.

10. The system for predicting and assisting in the diagnosis of depression based on prefrontal fNIRS entropy according to claim 9, characterized in that: The multi-channel near-infrared spectroscopy optical brain imaging device is a 26-channel near-infrared spectroscopy optical brain imaging device; and / or The multi-channel near-infrared spectroscopy optical brain imaging device has a head shape curvature radius of 80-120 mm; and / or The distance between the multi-channel near-infrared spectroscopy optical brain imaging device and the external light source is 25-35mm; and / or The multi-channel near-infrared spectral optical brain imaging device uses a frequency ≥5Hz.

Images