Teenager depression prediction system

Abstract

The invention relates to the technical field of medical care informatics, in particular to a teenager depression prediction system. Comprising an acoustic characteristic characterization module, the acoustic characteristic characterization module comprises an ultrasonic transducer array and electroencephalogram electrode co-integration, an acoustic impedance sensor is embedded in an electrode substrate, and a temperature-sensitive acoustic coupling adhesive layer is arranged on the surface of the electrode substrate; the multi-layer medium acoustic modeling module is used for establishing an accurate skull tissue acoustic transmission model according to the characteristics of an individual anatomical structure, and is used for quantifying the frequency dispersion effect and the like generated when the electroencephalogram signals are transmitted in different tissue layers; the problem that depression-related network abnormity and development-related transmission change cannot be distinguished due to systematic influence of tissue-specific dispersion on a functional connection network due to teenager myelin sheath thickness change in the prior art is solved.

Description

Technical Field

[0001] This invention relates to the field of healthcare informatics technology, specifically to a system for predicting adolescent depression. Background Technology

[0002] Early prediction and intervention for adolescent depression are crucial to preventing the condition from worsening. Existing methods for predicting depression primarily rely on clinical assessments, questionnaires, or behavioral data analysis based on machine learning. However, these methods are highly subjective, susceptible to individual report bias, and lack objective physiological indicators. Research has found a correlation between early adolescent depression and electroencephalogram (EEG) signals. Using EEG signals as the basis for predicting adolescent depression offers greater objectivity. However, EEG signals experience dispersion effects and phase distortion as they penetrate different tissue layers. Particularly in the prefrontal cortex, due to uneven skull thickness and cerebrospinal fluid flow, frequency-related phase delays occur during EEG signal transmission. These delays can lead to spurious estimates of inter-brain synchrony in functional connectivity analysis. This dispersion effect exhibits individualized spatial distribution patterns and is closely related to brain development. Adolescence is a critical stage for myelination of the brain's white matter, and changes in myelin thickness further alter signal transmission characteristics. Existing technologies treat EEG signals as propagating in a homogeneous medium, completely ignoring the systemic impact of this tissue-specific dispersion on functional connectivity networks. This leads to an inability to distinguish between depression-related network abnormalities and development-related transmission changes, resulting in biased predictions. Summary of the Invention

[0003] The technical problem to be solved by this invention is: how to improve the adolescent depression prediction system and solve the problem in the prior art that the tissue-specific dispersion caused by changes in myelin sheath thickness in adolescents has a systematic impact on the functional connectivity network, which leads to the inability to distinguish between depression-related network abnormalities and development-related transmission changes, resulting in biased prediction results.

[0004] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A brain-computer interface-based adolescent depression prediction system includes: The acoustic characterization module includes an integrated ultrasonic transducer array and EEG electrodes, with an acoustic impedance sensor embedded in the electrode substrate and a temperature-sensitive acoustic coupling adhesive layer on the surface of the electrode substrate. Multilayer media acoustic modeling module: used to establish an accurate acoustic transmission model of cranial tissue based on individual anatomical structural characteristics, and to quantify the dispersion effect generated when EEG signals propagate in different tissue layers; The phase compensation filtering module is used to design and implement phase compensation based on acoustic modeling results to correct phase distortion of EEG signals during tissue transmission. The dispersion-invariant functional connection calculation module is used to calculate the functional connection index against dispersion interference based on the compensated signal output by the adaptive phase compensation algorithm. By introducing a dispersion stability index and a dynamic weighting mechanism, it ensures that the functional connection network analysis is not affected by residual phase distortion. The Bayesian network module is used to construct a dynamic Bayesian network model based on the results of dispersion invariant functional connection calculations, which is used for time-series prediction and uncertainty quantification of depression risk; by integrating dispersion stability indices and acoustic features, a more robust prediction box is established. Depression Intervention Decision Module: Based on the output of Bayesian network modeling, combined with dispersion stability index and acoustic features, a multi-level, adaptive decision system is constructed to accurately trigger and adjust depression intervention strategies.

[0005] Furthermore, in the aforementioned adolescent depression prediction system, the multi-layered acoustic modeling module specifically includes the following applications: obtaining individual skull thickness distribution based on CT image data; measuring cerebrospinal fluid pulsation velocity using Doppler ultrasound; and calculating the acoustic transmission matrix for each tissue layer. Ti = in, denoted as the complex wave number (rad / m) of the i-th layer of medium. ; For frequency; The phase velocity is frequency-dependent; The frequency-dependent attenuation coefficient (Np / m); The thickness of the i-th layer is obtained through CT image segmentation with an accuracy of ±0.1 mm. Let be the acoustic impedance of the i-th layer (Rayl); ; The density of the medium is (kg / m³). The imaginary unit; Derive the acoustic transfer function of the system from the total transfer matrix: in, It is a complex transfer function that includes amplitude and phase information; This is the reference impedance.

[0006] Furthermore, in the aforementioned adolescent depression prediction system, the phase compensation filtering module specifically includes the following uses: Acoustic transfer function obtained from multi-layer media acoustic modeling Calculate the relative phase difference between channels i and j: in, Let be the acoustic transfer function of channel i, derived from the acoustic modeling output, and be a complex function containing amplitude and phase information; The acoustic transfer function for channel j is also derived from the acoustic modeling output; Extract the phase angle of the complex number for the operator, in radians (rad), with an output range of [-π, π]. This represents the phase distortion of channel i relative to channel j, which varies with frequency. Construct a frequency domain compensation filter to correct phase distortion and dispersion effects: ) in, The imaginary unit; The dispersion compensation coefficient, in seconds (s), is used to adjust the strength of group delay compensation and is determined by maximum likelihood estimation. To correct static phase shift and ensure that the phase difference between channels is zero after compensation; To correct the dispersion effect caused by group delay variation and reduce frequency-dependent phase distortion.

[0007] Furthermore, in the aforementioned adolescent depression prediction system, the dispersive invariant functional connection calculation module specifically includes the following uses: Empirical mode decomposition is performed on the compensated signal; Extract the instantaneous phase of each IMF component; Calculate the dispersion stability index: in: Phase difference sequence Standard deviation (rad); This is the normalization factor, corresponding to the maximum possible range of phase difference variation; Calculate the dispersion-weighted phase lag exponent: in: The attenuation coefficient was determined through cross-validation. This is a dispersion stability index, with a value range of [0,1]. Construct a functional connectivity network that considers dispersion stability; Connection weights: in: Let be the directed connection weight from channel i to j; This represents the total number of channels.

[0008] Furthermore, in the aforementioned adolescent depression prediction system, the Bayesian network module is specifically used for the following purposes: State-space model definition: ; in: For dispersion adjustment coefficient, dimension ×1, estimated via variational Bayesian method; The average dispersion stability index is a scalar: This is the acoustic characteristic vector, which includes the acoustic impedance and transmission loss of each channel; This is the dispersion stability adjustment coefficient; This is the acoustic feature adjustment matrix; Q is the process noise covariance matrix, with dimensions DS×DS, assumed to be a diagonal matrix. ; Observation model: in: It is a nonlinear mapping function, implemented through a three-layer feedforward neural network; b1 is the hidden layer bias vector; b2 is the output layer bias vector; b3 is the output adjustment bias term; The observation noise adjustment vector; To observe noise; Initialize the model prior based on data from healthy populations: State Priors: in: The average functional connectivity vector of 100 healthy adolescents; For the functional connectivity covariance matrix of the healthy population; posterior distribution is calculated based on variational Bayesian inference; probability of depression risk is then calculated. in: , , These are the logistic regression coefficients, obtained through training with historical data.

[0009] Furthermore, in the aforementioned adolescent depression prediction system, the specific uses of the depression intervention decision-making module include: Constructing decision feature vectors: Feature vector composition: in: The probability of depression is derived from Bayesian network modeling (range 0-1). The average dispersion stability index (range 0-1); For acoustic feature vectors; : represents the rate of change of risk probability; A scalar for global efficiency metrics; The average clustering coefficient is a scalar. Modular scalar indicators; Calculate the dynamic weighted comprehensive intervention score: in, ; in: ; in, ; Let L2 be the norm of the acoustic vector; Intervention parameters are adaptively adjusted based on the comprehensive intervention score.

[0010] Furthermore, in the aforementioned adolescent depression prediction system, the use of the depression intervention decision-making module to adaptively adjust intervention parameters based on comprehensive intervention scores specifically includes: Adjusting the intensity of neurofeedback: in: ; ; Adjust the task difficulty level: in: Adjust the continuous feedback time: in: Seconds; based on duration; This is the sensitivity coefficient to change.

[0011] The beneficial effects of this invention are as follows: The synergy between the ultrasonic transducer and the acoustic impedance sensor enables real-time monitoring of tissue characteristics, reducing transmission path estimation errors; the synergy between acoustic modeling and phase compensation accurately corrects phase distortion caused by dispersion, improving the reliability of the connectivity network; and the synergy between the dispersion stability index and the Bayesian model enhances the model's ability to identify real neural connections. By improving the adolescent depression prediction system, this invention addresses the problem in existing technologies where tissue-specific dispersion due to changes in myelin sheath thickness in adolescents systematically affects functional connectivity networks, leading to the inability to distinguish between depression-related network abnormalities and developmental transmission changes. Attached Figure Description

[0012] Figure 1 This is a structural block diagram of a specific embodiment of the adolescent depression prediction system of the present invention. Detailed Implementation

[0013] To explain in detail the technical content, objectives, and effects of the present invention, the following description is provided in conjunction with the embodiments and accompanying drawings.

[0014] Please refer to Figure 1 Specific embodiments of the present invention relate to an adolescent depression prediction system, including: The acoustic characterization module includes an integrated ultrasonic transducer array and EEG electrodes, with an acoustic impedance sensor embedded in the electrode substrate and a temperature-sensitive acoustic coupling adhesive layer on the surface of the electrode substrate. Specifically, the ultrasonic transducer array uses PZT-5H piezoelectric material, with dimensions of 1mm×1mm×0.3mm, and is fabricated as an 8×8 array using photolithography, with a center frequency of 1MHz and a bandwidth of 0.5-1.5MHz. Specifically, the acoustic impedance sensor uses a MEMS surface acoustic wave sensor with a resonant frequency of 100MHz and a sensitivity of 0.001MRayl, and is connected to the signal processing circuit via gold wire bonding; the temperature-sensitive acoustic coupling adhesive layer is prepared by dissolving polyurethane-silicone copolymer (mass ratio 60:40) in xylene, with a solid content of 25%, and spin-coating to form a film with a thickness of 200μm, and an acoustic impedance of 1.55MRayl at 35℃.

[0015] Multilayer media acoustic modeling module: used to establish an accurate acoustic transmission model of cranial tissue based on individual anatomical structural characteristics, and to quantify the dispersion effect generated when EEG signals propagate in different tissue layers; Specifically, a 256-slice CT scanner was used to acquire axial images of the head. Scanning parameters were: voltage 120kV, current 250mA, slice thickness 0.5mm, and pixel size 0.4mm × 0.4mm. An adaptive threshold segmentation algorithm was used to extract the inner and outer boundaries of the skull and calculate the local thickness. - in, coordinates Thickness of the skull at that location; Let i be the coordinates of the i-th point on the outer surface; These are the coordinates of the corresponding point on the inner surface; This is the number of sampling points, typically 25 points / mm. 2 ; Transcranial Doppler ultrasound (probe frequency 2MHz) was used to measure the blood flow velocity in the middle cerebral artery, with a sampling frequency of 100Hz. The cerebrospinal fluid pulsation velocity was estimated using the following formula: in, (t) represents the cerebrospinal fluid flow velocity at time t; τ represents the blood flow velocity in the middle cerebral artery; k is the coupling coefficient, ranging from 0.3 to 0.5; τ is the phase delay, typically 0.1 s. The base flow velocity is 2-3 cm / s. The signal transmission path is discretized into a multi-layer structure, with each layer considered as a homogeneous medium, and its acoustic characteristics are calculated. For the i-th layer medium, its transmission matrix is ​​represented as: Ti = in, denoted as the complex wave number (rad / m) of the i-th layer of medium. ; For frequency; The phase velocity is frequency-dependent; The frequency-dependent attenuation coefficient (Np / m); The thickness of the i-th layer is obtained through CT image segmentation with an accuracy of ±0.1 mm. Let be the acoustic impedance of the i-th layer (Rayl); The density of the medium is (kg / m³). The imaginary unit; The parameter values ​​for each tissue layer are as follows: Skin layer: Thickness: 1.5-2.5mm (individual measurement); Density: 1100 kg / m³; Phase velocity: (m / s); Attenuation coefficient: (dB / cm); Skull layer: Thickness: 3-8mm (individual measurement); Density: 1900 kg / m³; Phase velocity: (m / s); Attenuation coefficient: (dB / cm); Cerebrospinal fluid layer: Thickness: 3-8mm (individual measurement); Density: 1005 kg / m³; Phase velocity: (m / s); Attenuation coefficient: (dB / cm); Brain tissue layer: Thickness: The remaining portion of the transmission path; Density: 1040 kg / m³; Phase velocity: (m / s); Attenuation coefficient: (dB / cm); For a transmission path containing N layers, the total transmission matrix is: Derive the acoustic transfer function of the system from the total transfer matrix: in, It is a complex transfer function that includes amplitude and phase information; As a reference impedance, the characteristic impedance of the water sample is 1.48 × 10⁻⁶. 6 Rayl); The phase compensation filtering module is used to design and implement phase compensation based on acoustic modeling results to correct phase distortion of EEG signals during tissue transmission. include: Acoustic transfer function obtained from multi-layer media acoustic modeling Calculate the relative phase difference between channels i and j: Channel i and channel j represent different electrode locations on the EEG acquisition device. These channels are placed on the scalp surface according to the international 10-20 system or a denser layout, with each channel recording the electrical activity of the brain region beneath it.

[0016] Extract the phase angle of the complex number for the operator, in radians (rad), with an output range of [-π, π]. This represents the phase distortion of channel i relative to channel j, which varies with frequency.

[0017] Since the phase angle may be wrapped in the interval [-π, π], it needs to be unwrapped to ensure continuity: For frequency sequences Calculate the phase difference increment: Frequency range: 0.5-45 Hz (covering the main frequency bands of EEG), frequency resolution Δf = 0.1 Hz, number of frequency points K = 445; like If ≥ π, then for Adding or subtracting integer multiples of 2π makes ≤π; repeat until all frequency points have been processed to obtain a continuous phase difference. ; Calculate the first derivative of phase distortion with respect to frequency, i.e., the group delay difference, to quantify the dispersion effect; in: =2 (Twice the frequency resolution).

[0018] For endpoints Use forward difference or backward difference, for example ; This represents the relative group delay between channels i and j, measured in seconds, reflecting the time delay differences between different frequency components of the signal.

[0019] Construct a frequency domain compensation filter to correct phase distortion and dispersion effects: in, The imaginary unit; The dispersion compensation coefficient, in seconds (s), is used to adjust the strength of group delay compensation and is determined by maximum likelihood estimation. To correct static phase shift and ensure that the phase difference between channels is zero after compensation; To correct the dispersion effect caused by group delay variation and reduce frequency-dependent phase distortion; In practical applications, to keep the signal amplitude constant, normalization can be used to make the filter amplitude close to 1, but the adjustment effect of β is retained. Beta values ​​were calibrated using EEG data from healthy adolescents to ensure optimal compensation. Collect EEG data from M healthy adolescents (without depressive symptoms), e.g., M = 50; simultaneously acquire their acoustic transfer functions. For each subject, calculate the compensated phase difference for all channels at (i, j): in, and The frequency representation of EEG signals, = It is a conjugate filter; Define ideal phase difference (Assuming there is no inherent phase difference between channels in a healthy state); Calculate the phase error: | | Assuming the error follows a Gaussian distribution, construct the likelihood function: in, The standard deviation of the error is initially set to 0.1 rad. Solve : Using grid search in Solve within the specified range, with a step size of 0.001s, the typical optimized value is approximately... ; The frequency-domain compensated filter is converted into a time-domain finite impulse response (FIR) filter for real-time processing: The time-domain impulse response is obtained by performing the inverse discrete Fourier transform (IDFT): Where: N=64: filter length, corresponding to a time length of 0.5 s (sampling rate 128 Hz); n=0,1,……,N-1: Time-domain index; To ensure For real sequences, Apply Hermitian symmetry constraints: Where * denotes complex conjugation; F s The sampling rate of the EEG signal (e.g., 128 Hz); Applying Hamming windows to reduce Gibbs phenomenon: Final filter coefficients: · , ; It is ultimately a real-coefficient FIR filter of length N that can be used to perform convolution operations with EEG signals; A compensation filter was applied in real-time computer signal processing. The EEG signal sampling rate was 128 Hz, the frame length was 4 seconds (512 points), and the overlap was 50%. Fourier transform was performed on each frame to obtain... and ; The compensated signal is: in To ensure consistency in compensation; The theoretical form is precisely Discrete-time Fourier transform; right Perform a Fourier transform to obtain the compensated result. ; This algorithm directly utilizes the transfer function output from acoustic modeling to calculate phase distortion, ensuring that compensation is based on individualized organizational characteristics. Group delay calculation relies on the phase dispersion relationship provided by acoustic modeling, while β estimation is calibrated using health data to adapt compensation to population variability. The entire process forms a closed loop, correcting signal transmission distortion in real time and improving the reliability of functional connectivity analysis.

[0020] The dispersion-invariant functional connection calculation module is used to calculate the functional connection index against dispersion interference based on the compensated signal output by the adaptive phase compensation algorithm. By introducing a dispersion stability index and a dynamic weighting mechanism, it ensures that the functional connection network analysis is not affected by residual phase distortion. The compensated signal output by the adaptive phase compensation algorithm is further processed: Each channel signal is decomposed into intrinsic mode functions using Empirical Mode Decomposition (EMD): in: The k-th eigenmode function satisfies the narrowband condition; This is the residual component; for quantity; For each The components are subjected to Hilbert transform to obtain the analytic signal: in: For Hilbert transform operators; The imaginary unit; Calculate the instantaneous phase: in: Let be the instantaneous phase (rad) of the k-th IMF in channel i; phase unwrapping processing ensures continuity; Calculate the compensated inter-channel phase difference: in: The weight of the k-th IMF is calculated based on its energy share. To determine the effective number of IMFs, higher-order components dominated by noise are excluded. in: Phase difference sequence Standard deviation (rad); This is the normalization factor, corresponding to the maximum possible range of phase difference variation; Calculate the anti-interference phase synchronization index based on the compensated phase difference: Traditional PLI calculation: in: This is a sign function that returns ±1. This represents the average value within the time window T. Time window T = 4 seconds, overlap 50%; Calculate weighted PLI: in: The attenuation coefficient, determined through cross-validation, has a typical value of λ = 1.5. This is a dispersion stability index, with a value range of [0,1]. Construct a functional connectivity network that considers dispersion stability; Connection weights: in: Let be the directed connection weight from channel i to j; Total number of channels (32); Statistically significant connections are preserved, and the threshold is based on a null hypothesis test. in: For all The mean; For all Standard deviation; Extracting key graph theory features from functional connectivity networks: Calculate the overall efficiency: in: Let i be the degree (number of connections) of node i. The network is divided into modules using the Louvain algorithm, and the modularity index Q is calculated: in: Total network connectivity; The module to which node i belongs; For the Kronecker delta function; The above modules directly utilize the compensated signal output by the adaptive phase compensation algorithm. The signal has passed through the filter. The phase distortion caused by tissue dispersion was corrected; the dispersion stability index DSI was calculated based on the variability of the compensated phase difference, reflecting the degree of residual distortion, and was incorporated into the PLI calculation as a weighting factor to form a dual protection mechanism. The Bayesian network module is used to construct a dynamic Bayesian network model based on the results of dispersion invariant functional connection calculations, which is used for time-series prediction and uncertainty quantification of depression risk; by integrating dispersion stability indices and acoustic features, a more robust prediction box is established. Construct a dynamic system that includes functional connectivity states and observed variables: Define state variables: in: Let be the dispersion-invariant functional connectivity matrix (N×N) at time t. For matrix vectorization operations, convert an N×N matrix into... ×1 vector; For the state vector, dimension (N=32 channels); State evolution equation: ; in: For dispersion adjustment coefficient, dimension ×1, estimated via variational Bayesian method; The average dispersion stability index is a scalar: This is the acoustic characteristic vector, which includes the acoustic impedance and transmission loss of each channel; This is the dispersion stability adjustment coefficient; This is the acoustic feature adjustment matrix; Q is the process noise covariance matrix, with dimensions DS×DS, assumed to be a diagonal matrix: Q=diag(q1, q2, ..., qDS); Let be the equivalent acoustic impedance (MRayl) of channel i; The transmission loss (dB) at the reference frequency of 10Hz; For acoustic dispersion stability; This is the normalization coefficient (typical value 0.01). Establish the relationship between state variables and multimodal observation data: Definition of observed variables: in: The phase lag exponential matrix after compensation; A scalar for global efficiency metrics; The average clustering coefficient is a scalar. Modular scalar indicators; The observation equation is: in: It is a nonlinear mapping function, implemented through a three-layer feedforward neural network; b1 is the hidden layer bias vector; b2 is the output layer bias vector; b3 is the output adjustment bias term; The observation noise adjustment vector; To observe noise; Initialize the model prior based on data from healthy populations: State Priors: in: The average functional connectivity vector of 100 healthy adolescents; The functional connectivity covariance matrix of the healthy population is regularized to a diagonal matrix. Parameter priors: in: ; The dimension of the observation vector; Using mean field variational inference to infer approximate posterior distribution: Variational distribution setting: The variational distributions are all Gaussian distributions: Lower bound of evidence: Coordinate Ascending Update Principle: Status Update: in, The Jacobian matrix of the neural network; Parameter update: Calculate the probability of depression risk: Calculate the real-time depression risk score based on the posterior distribution; Calculate the health reference distance: in The Mahalanobis distance; Calculate the probability of risk: in: , , These are the logistic regression coefficients, obtained through training with historical data; Typical value: ; ; ; Learning model hyperparameters using the EM algorithm: Step E: Calculate the posterior expectation using the variational inference described above; M-step: Maximize the likelihood update parameters based on complete data. Depression Intervention Decision Module: Based on the output of Bayesian network modeling, combined with dispersion stability index and acoustic features, a multi-level adaptive decision system is constructed to accurately trigger and adjust depression intervention strategies. Feature vector composition: in: The probability of depression is derived from Bayesian network modeling (range 0-1). The average dispersion stability index (range 0-1); For acoustic feature vectors; : represents the rate of change of risk probability; A scalar for global efficiency metrics; The average clustering coefficient is a scalar. Modular scalar indicators; Feature normalization: in: The mean of characteristics of the healthy group; Standard deviation of characteristics of the healthy population; Construct a dual-threshold decision system based on signal quality and depression risk; Includes: First layer: Signal quality assessment: like The signal quality is then considered excellent. like <0.85, A signal quality of ≥0.70 is considered good. like <0.70, If the value is ≥0.60, the signal quality is medium. like If the value is less than 0.60, the signal quality is poor. Second level: Depression risk classification: like If the value is less than 0.30, the risk level is low. like <0.50, If the value is ≥0.30, the risk level is considered low to medium. like <0.70, If the value is ≥0.50, the risk level is medium to high. like If the value is ≥0.70, the risk level is high; Third layer: Fusion decision matrix Refer to Table 1: Table 1 Calculate the comprehensive intervention score to quantify the intensity of the decision: Base score: in: ; Preferred, Network Feature Correction Score: in, Preferred, Overall intervention score: in, ; Preferred, Let L2 be the norm of the acoustic vector; Dynamically adjust intervention intensity parameters based on scores: Adjusting the intensity of neurofeedback: in: (Range Intensity); (Score range); Adjust the task difficulty level: in: Adjust the continuous feedback time: in: Seconds; based on duration; The coefficient of sensitivity to change; When signal quality is poor, prioritize acquisition optimization, including: Evaluation of optimization effect: Optimization success criteria: Emergency response activated for high-risk situations: Emergency condition detection: It is an exponential function; Emergency Intervention Protocol: Initiate high-intensity neurofeedback immediately ( Parallel alarms are sent to the monitoring system; the duration of a single intervention is extended to 300 seconds; continuous monitoring mode is activated (sampling interval is shortened to 1 second); To avoid frequent fluctuations in decision-making and maintain continuity of intervention, an index-weighted decision-making system is adopted: in: As the current decision weight; To smooth out the decision; Maintain a historical record of the most recent N=10 decisions, and only execute a major strategy change when M=5 consecutive decisions are consistent; As a case study, the following real-time data was recorded during the brain-computer interface-based depression prediction and intervention process for a 16-year-old female adolescent: Monitoring period: 20:00-20:30 on March 15, 2024; Features derived from Bayesian network modeling: (normalized) The value is 0.73; the signal quality is good. (32 channels) Details are as follows: Forehead area: Channel 1: FP1: 1.5523 (equivalent impedance 1.58 MRayl, transmission loss -11.2dB, DSI=0.88); Channel 2: FP2: 1.5481 (equivalent impedance 1.57 MRayl, transmission loss -11.1dB, DSI=0.89); Channel 3: Fpz: 1.5498 (equivalent impedance 1.575 MRayl, transmission loss -11.0dB, DSI=0.87); Channel 4: AF3: 1.5321 (equivalent impedance 1.56 MRayl, transmission loss -10.9dB, DSI=0.86); Channel 5: AF4: 1.5308 (equivalent impedance 1.55 MRayl, transmission loss -10.8dB, DSI=0.87); Channel 6: F7: 1.4983 (equivalent impedance 1.52 MRayl, transmission loss -10.5dB, DSI=0.85); Channel 7: F8: 1.5012 (equivalent impedance 1.53 MRayl, transmission loss -10.6dB, DSI=0.84); Channel 8: Fz: 1.5215 (equivalent impedance 1.545 MRayl, transmission loss -10.7dB, DSI=0.86); Forehead area: Channel 9: F3: 1.4789 (equivalent impedance 1.50 MRayl, transmission loss -10.4dB, DSI=0.83); Channel 10: F4: 1.4817 (equivalent impedance 1.51 MRayl, transmission loss -10.3dB, DSI=0.84); Channel 11: F1: 1.4923 (equivalent impedance 1.52 MRayl, transmission loss -10.5dB, DSI=0.82); Channel 12: F2: 1.4956 (equivalent impedance 1.525 MRayl, transmission loss -10.4dB, DSI=0.83); Channel 13: F5: 1.4682 (equivalent impedance 1.49 MRayl, transmission loss -10.2dB, DSI=0.81); Channel 14: F6: 1.4719 (equivalent impedance 1.495 MRayl, transmission loss -10.1dB, DSI=0.82); Channel 15: FC1: 1.4583 (equivalent impedance 1.48 MRayl, transmission loss -9.8dB, DSI=0.80); Channel 16: FC2: 1.4621 (equivalent impedance 1.485 MRayl, transmission loss -9.7dB, DSI=0.81); Central District: Channel 17: C3: 1.4238 (equivalent impedance 1.45 MRayl, transmission loss -9.5dB, DSI=0.78); Channel 18: C4: 1.4285 (equivalent impedance 1.455 MRayl, transmission loss -9.4dB, DSI=0.79); Channel 19: Cz: 1.4356 (equivalent impedance 1.46 MRayl, transmission loss -9.3dB, DSI=0.77); Channel 20: C1: 1.4189 (equivalent impedance 1.44 MRayl, transmission loss -9.6dB, DSI=0.76); Channel 21: C2: 1.4217 (equivalent impedance 1.445 MRayl, transmission loss -9.5dB, DSI=0.77); Channel 22: C5: 1.4083 (equivalent impedance 1.43 MRayl, transmission loss -9.7dB, DSI=0.75); Top leaf area: Channel 23: P3: 1.3956 (equivalent impedance 1.42 MRayl, transmission loss -9.8dB, DSI=0.74); Channel 24: P4: 1.3989 (equivalent impedance 1.425 MRayl, transmission loss -9.7dB, DSI=0.75); Channel 25: Pz: 1.4023 (equivalent impedance 1.43 MRayl, transmission loss -9.6dB, DSI=0.73); Channel 26: P1: 1.3887 (equivalent impedance 1.41 MRayl, transmission loss -9.9dB, DSI=0.72); Channel 27: P2: 1.3921 (equivalent impedance 1.415 MRayl, transmission loss -9.8dB, DSI=0.73); Channel 28: P7: 1.3785 (equivalent impedance 1.40 MRayl, transmission loss -10.1dB, DSI=0.71); Occipital lobe region: Channel 29: O1: 1.3658 (equivalent impedance 1.39 MRayl, transmission loss -10.2dB, DSI=0.70); Channel 30: O2: 1.3689 (equivalent impedance 1.395 MRayl, transmission loss -10.1dB, DSI=0.71); Channel 31: z: 1.3723 (equivalent impedance 1.40 MRayl, transmission loss -10.0dB, DSI=0.69); Channel 32: POz: 1.3856 (equivalent impedance 1.41 MRayl, transmission loss -9.9dB, DSI=0.72); ; The value is 0.68; the risk level is medium to high. It is 0.45; It is 0.38; It is 0.29; (Signal quality="Good", Risk level="Medium-high") → Decision="Standard neural feedback", Priority="High"; =0.6×0.68 + 0.25×(1-0.73) + 0.15×0.12 = 0.408 + 0.0675 +0.018 = 0.4935; =0.4×(1-0.45) + 0.3×0.38 + 0.3×0.29 = 0.22 + 0.114 + 0.087 =0.421; =0.5×0.4935 + 0.3×0.421 + 0.2×0.63 = 0.2468 + 0.1263 + 0.126 = 0.4991; =0.3 + (1.0-0.3) × (0.4991-0.2) / (0.8-0.2) = 0.3 + 0.7×0.499 =0.649; =1 / (1 + exp(-(-1.5 + 3.0×0.68 + 2.0×0.73)))=0.881; =180 × (1 + 2.0 × 0.12) = 180 × 1.24 = 223; Implement the intervention plan: Neurofeedback task: Prefrontal alpha wave enhancement training; Objective: Increase the prefrontal alpha wave power to 130% of the baseline value; Difficulty level: 0.881 (high difficulty); Feedback intensity: 0.649 (moderate to strong); Duration: 223 seconds; Intervention effectiveness evaluation: 60 seconds after intervention: Prefrontal alpha wave power: increased from baseline to 125%; Participation score: 0.72 (good); Signal quality maintained: DSI = 0.71; 5 minutes after the intervention ended: It shows a downward trend.

[0021] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent modifications made based on the content of the present invention specification and drawings, or direct or indirect applications in related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. Adolescent Depression Prediction System, characterized in that, include: The phase compensation filtering module is used to design and implement phase compensation based on acoustic modeling results to correct phase distortion of EEG signals during tissue transmission. The dispersion-invariant functional connection calculation module is used to calculate the functional connection index against dispersion interference based on the compensated signal output by the adaptive phase compensation algorithm. By introducing a dispersion stability index and a dynamic weighting mechanism, it ensures that the functional connection network analysis is not affected by residual phase distortion. The Bayesian network module is used to construct a dynamic Bayesian network model based on the results of dispersion-invariant functional connection calculations, for time-series prediction and uncertainty quantification of depression risk; a prediction box is established by integrating dispersion stability indices and acoustic features. Depression Intervention Decision Module: This module is used to construct a decision system based on the output of Bayesian network modeling, combined with dispersion stability indicators and acoustic features, to trigger and adjust depression intervention strategies.

2. The adolescent depression prediction system according to claim 1, characterized in that, Also includes: The acoustic characterization module includes an integrated ultrasonic transducer array and EEG electrodes, with an acoustic impedance sensor embedded in the electrode substrate and a temperature-sensitive acoustic coupling adhesive layer on the surface of the electrode substrate. Multilayer media acoustic modeling module: used to establish an accurate acoustic transmission model of cranial tissue based on individual anatomical structural characteristics, and to quantify the dispersion effect generated when EEG signals propagate in different tissue layers.

3. The adolescent depression prediction system according to claim 2, characterized in that, The applications of the multilayer media acoustic modeling module include: obtaining individual skull thickness distribution based on CT image data; measuring cerebrospinal fluid pulsation velocity via Doppler ultrasound; and calculating the acoustic transmission matrix for each tissue layer. If = in, denoted as the complex wave number (rad / m) of the i-th layer of medium. ; For frequency; The phase velocity is frequency-dependent; The attenuation coefficient is frequency-dependent. The thickness of the i-th layer is obtained through CT image segmentation. Let be the acoustic impedance of the i-th layer; The density of the medium; The imaginary unit; Derive the acoustic transfer function of the system from the total transfer matrix: in, It is a complex transfer function that includes amplitude and phase information; This is the reference impedance.

4. The adolescent depression prediction system according to claim 1, characterized in that, The phase compensation filtering module has the following specific applications: Acoustic transfer function obtained based on multi-layer media acoustic modeling Calculate the relative phase difference between channels i and j: in, Let be the acoustic transfer function of channel i, derived from the acoustic modeling output, and be a complex function containing amplitude and phase information; The acoustic transfer function for channel j is also derived from the acoustic modeling output; Extract the phase angle of the complex number for the operator, in radians, with an output range of [-π, π]. This represents the phase distortion of channel i relative to channel j, which varies with frequency. Construct a frequency domain compensation filter to correct phase distortion and dispersion effects: ) in, The imaginary unit; , is the dispersion compensation coefficient, in seconds, used to adjust the strength of group delay compensation, and is determined by maximum likelihood estimation; To correct static phase shift and ensure that the phase difference between channels is zero after compensation; To correct the dispersion effect caused by group delay variation and reduce frequency-dependent phase distortion.

5. The adolescent depression prediction system according to claim 1, characterized in that, The uses of the dispersion-invariant function connection calculation module specifically include: Empirical mode decomposition is performed on the compensated signal; Extract the instantaneous phase of each IMF component; Calculate the dispersion stability index: in: Phase difference sequence Standard deviation; This is the normalization factor, corresponding to the maximum possible range of phase difference variation; Calculate the dispersion-weighted phase lag exponent: in: The attenuation coefficient was determined through cross-validation. This is a dispersion stability index, with a value range of [0,1]. Construct a functional connectivity network that considers dispersion stability; Connection weights: in: Let be the directed connection weight from channel i to j; This represents the total number of channels.

6. The adolescent depression prediction system according to claim 1, characterized in that, The uses of Bayesian network modules include: State-space model definition: ； in: For dispersion adjustment coefficient, dimension ×1, estimated via variational Bayesian method; The average dispersion stability index is a scalar: This is the acoustic characteristic vector, which includes the acoustic impedance and transmission loss of each channel; This is the dispersion stability adjustment coefficient; This is the acoustic feature adjustment matrix; Q is the process noise covariance matrix; Observation model: in: It is a nonlinear mapping function, implemented through a three-layer feedforward neural network; b1 is the hidden layer bias vector; b2 is the output layer bias vector; b3 is the output adjustment bias term; The observation noise adjustment vector; To observe noise; Initialize the model prior based on data from healthy populations: State Priors: in: The average functional connectivity vector of 100 healthy adolescents; For the functional connectivity covariance matrix of the healthy population; posterior distribution is calculated based on variational Bayesian inference; probability of depression risk is then calculated. in: , , These are the logistic regression coefficients, obtained through training with historical data.

7. The adolescent depression prediction system according to claim 1, characterized in that, The uses of the depression intervention decision-making module specifically include: Constructing decision feature vectors: Feature vector composition: in: The probability of depression is derived from Bayesian network modeling; The average dispersion stability index; For acoustic feature vectors; : represents the rate of change of risk probability; A scalar for global efficiency metrics; The average clustering coefficient is a scalar. Modular scalar indicators; Calculate the dynamic weighted comprehensive intervention score: in, ; in: ; in, ; Let L2 be the norm of the acoustic vector; Intervention parameters are adaptively adjusted based on the comprehensive intervention score.

8. The adolescent depression prediction system according to claim 7, characterized in that, The uses of the depression intervention decision-making module, specifically the adaptive adjustment of intervention parameters based on comprehensive intervention scores, include: Adjusting the intensity of neurofeedback: in: ; ； Adjust the task difficulty level: in: Adjust the continuous feedback time: in: Seconds; based on duration; This is the sensitivity coefficient to change.

Images