A Method and System for Emotion Monitoring and Intervention Based on Multimodal Neurochemical Sensing

Abstract

This application belongs to the field of emotion monitoring technology, and provides a method and system for emotion monitoring and intervention based on multimodal neurochemical sensing. It acquires electroacoustic and optical fluorescence spectral signals from a user's target tissue region through a wearable device, and identifies neurotransmitter system activity information through signal fusion and decoding. This information is then input into an individualized dynamic emotion monitoring model to obtain the user's current emotional state type and intensity. Combining these two indicators, an emotion instability index is calculated in real time. When the index exceeds an individualized critical threshold, a personalized intervention plan matching the instability characteristics is generated and executed. The critical threshold can also be dynamically adjusted based on the intervention effect and real-time physiological state. This invention achieves non-invasive neurotransmitter monitoring and emotion monitoring, assessment, and intervention, improving monitoring accuracy and the timeliness and personalization of intervention, and is applicable to fields such as medicine and education.

Description

Technical Field

[0001] This invention relates to the fields of biomedical engineering and emotion computing technology, specifically, to a method and system for emotion monitoring and intervention based on multimodal neurochemical sensing. Background Technology

[0002] Accurate monitoring and timely intervention of emotional states are crucial for maintaining mental health and have significant application value in various fields such as medicine, education, and elderly care. Current emotion assessment and intervention technologies mainly fall into two categories: one is subjective evaluation methods based on scales and questionnaires, relying on user self-reporting to collect data. These methods are easily influenced by individual cognitive biases regarding emotions, resulting in highly subjective and inaccurate assessments, and they cannot achieve dynamic and continuous monitoring. The other category is objective assessment methods based on physiological or behavioral indicators such as electroencephalography (EEG), heart rate variability, and facial expressions. These methods infer emotional states through machine learning models, but they rely only on indirect manifestations of the autonomic nervous system and do not address the core driving factors of emotional changes, namely neurotransmitters. Therefore, they cannot directly and comprehensively reflect the dynamic process of emotional changes, leading to insufficient stability and applicability of the assessment results.

[0003] Meanwhile, existing emotion intervention techniques are mostly reactive, taking intervention measures only after negative emotions have clearly appeared. They lack the ability to predict and prevent the critical state before emotional instability, resulting in delayed intervention and difficulty in preventing the deterioration and generalization of emotions. In addition, existing intervention programs are mostly generalized, failing to fully consider individual user physiological characteristics, historical intervention effects, and real-time scenarios, resulting in low personalization and inconsistent intervention effects. Summary of the Invention

[0004] In response, this invention provides a method and system for emotion monitoring and intervention based on multimodal neurochemical sensing, to at least partially solve the above-mentioned technical problems.

[0005] This invention provides a method for emotion monitoring and intervention based on multimodal neurochemical sensing, comprising the following steps:

[0006] Acquire the acoustic-electrodynamic spectrum and optical-fluorescence spectrum of the user's target tissue region;

[0007] Based on the aforementioned acoustic-electrodynamic spectrum signal and optical fluorescence spectrum signal, the activity information of at least one neurotransmitter system related to emotional state is identified through signal fusion and decoding.

[0008] The activity information of the neurotransmitter system is input into the individualized dynamic emotion monitoring model to obtain and output the user's current emotion state type and emotion state intensity.

[0009] Based on the current emotional state type and intensity, and combined with the activity information of the neurotransmitter system, a quantitative emotional instability index is calculated in real time to characterize the trend of the emotional state towards a negative direction.

[0010] When the emotional instability index exceeds the individualized critical threshold set for the user, a personalized emotional intervention plan that matches the instability characteristics reflected by the emotional instability index is generated and executed.

[0011] Furthermore, the identification of neurotransmitter system activity information through signal fusion and decoding includes:

[0012] Temporal dynamic features and frequency domain relaxation features are extracted from the acoustic-electrodynamic spectrum signal, and optical features are extracted from the optical fluorescence spectrum signal;

[0013] The extracted acoustic and electronic features are combined with optical features to form a high-dimensional feature vector;

[0014] The high-dimensional feature vector is input into a neurotransmitter decoding model built based on a deep learning algorithm, and the model outputs the relative activity intensity of a specific neurotransmitter category.

[0015] Furthermore, during the decoding process, consistency verification between acoustic and optical signals is performed, and the weights of different signals in the final decoding result are dynamically adjusted or the re-acquisition of the trigger signal is initiated based on the verification results.

[0016] Furthermore, the consistency verification of the acoustic signal and the optical signal includes:

[0017] Calculate the statistical correlation between specific time-domain or frequency-domain characteristics of the acoustic-electrodynamic spectral signal and specific optical characteristics of the optical-fluorescence spectral signal;

[0018] When the statistical correlation is higher than the first validity threshold, the signal is determined to be valid and used for neurotransmitter decoding.

[0019] When the statistical correlation is lower than the first validity threshold but higher than the second validity threshold, the contribution weight of the optical fluorescence spectral signal in decoding is reduced, and the ambient light compensation mechanism of the optical module is triggered.

[0020] When the statistical correlation is lower than the second validity threshold, the data is determined to be invalid and the signal re-acquisition process is initiated.

[0021] Furthermore, the real-time calculation of the emotional instability index specifically includes:

[0022] Based on the activity information of the neurotransmitter system and the intensity of the current emotional state, a neurochemical state vector is constructed for the current moment.

[0023] By combining the historical neurochemical state vectors within a preset time window, a neurochemical state time sequence is constructed.

[0024] Complexity analysis and synchronicity analysis methods are used to extract multi-scale features characterizing the dynamic stability of the system from the time series.

[0025] The multi-scale features are input into a trained emotion instability prediction model, which outputs a value between zero and one as the emotion instability index.

[0026] Furthermore, the generation of a personalized emotion intervention program that matches the instability characteristics reflected by the emotion instability index includes:

[0027] Based on the activity information of the neurotransmitter system and the current emotional state type, determine the dominant type of emotional instability;

[0028] Based on the dominant type of emotional instability, the corresponding intervention content and frequency are matched from the pre-set intervention strategy library to form an initial intervention plan;

[0029] Based on the user's historical intervention effect data, the stimulus parameters in the initial intervention plan are individually modified to generate the final personalized emotion intervention plan.

[0030] Furthermore, the emotional instability index at the time of each intervention trigger and the emotional state recovery data after the intervention were recorded;

[0031] Based on the evaluation results of the intervention effect within a preset period, the individualized critical threshold is adaptively adjusted. When the intervention effect is consistently better than expected, the threshold is increased; when the intervention effect is consistently worse than expected, the threshold is decreased.

[0032] Simultaneously, by combining the user's real-time heart rate variability and exercise status data, the individualized critical threshold is adjusted instantaneously according to the specific scenario.

[0033] In another aspect, this application also provides an emotion monitoring and intervention system based on multimodal neurochemical sensing, comprising:

[0034] The signal acquisition module acquires the acoustic-electrodynamic spectrum and optical-fluorescence spectrum signals of the user's target tissue region through a wearable device;

[0035] The activity information recognition module is used to identify the activity information of at least one neurotransmitter system related to emotional state through signal fusion and decoding based on the acoustic-electrodynamic spectrum signal and the optical fluorescence spectrum signal.

[0036] The emotion information acquisition module is used to input the activity information of the neurotransmitter system into the individualized dynamic emotion monitoring model, and to acquire and output the user's current emotion state type and emotion state intensity.

[0037] The emotion quantification module is used to calculate a quantitative emotion instability index in real time, which is used to characterize the trend of the emotion state towards a negative development, based on the current emotion state type and intensity, and combined with the activity information of the neurotransmitter system.

[0038] The intervention plan generation module is used to generate and execute a personalized emotional intervention plan that matches the instability characteristics reflected by the emotional instability index when the emotional instability index exceeds the individualized critical threshold set for the user.

[0039] In another aspect, this application also provides an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the emotion monitoring and intervention method based on multimodal neurochemical sensing as described above.

[0040] In another aspect, this application provides a computer-readable storage medium having stored thereon computer program instructions that can be executed by a processor to implement the emotion monitoring and intervention method based on multimodal neurochemical sensing as described above.

[0041] Another aspect of this application provides a computer program product, including a computer program that, when executed by a processor, implements the emotion monitoring and intervention method based on multimodal neurochemical sensing as described above.

[0042] This invention overcomes the limitations of traditional indirect physiological indicators by employing focused ultrasound modulation combined with acoustic-electric and near-infrared optical synergistic sensing. It can non-invasively and specifically monitor dynamic changes in neurotransmitters, significantly improving the directness and accuracy of emotion monitoring. Simultaneously, it constructs a general model and a meta-learning-based individualized emotion monitoring model, combined with Monte Carlo Dropout confidence verification, balancing group universality and individual adaptability, addressing the low individualization level of existing models. Furthermore, through the quantification of the emotion instability index and the critical state prediction mechanism, it shifts from post-event intervention to pre-event preventative intervention. By dynamically adjusting the intervention plan and critical threshold based on the user's historical effects and real-time physiological state, it significantly improves the timeliness and accuracy of intervention. The entire process relies on wearable devices, balancing portability and real-time performance, and can be widely applied in medical, educational, and rehabilitation fields, providing an efficient and personalized solution for emotional health management. Attached Figure Description

[0043] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0044] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0045] Figure 1 This is a schematic diagram of an emotion monitoring and intervention method based on multimodal neurochemical sensing provided in an embodiment of the present invention.

[0046] Figure 2 This is a schematic diagram illustrating the implementation process of the personalized intervention scheme provided in an embodiment of the present invention.

[0047] Figure 3 This is a schematic diagram of the structure of an emotion monitoring and intervention system based on multimodal neurochemical sensing, provided in an embodiment of the present invention.

[0048] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0050] It should be noted that all user information (including but not limited to user device information, user personal information, object information corresponding to device usage data, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, device usage data, etc.) involved in all embodiments of this disclosure are information and data authorized by the user or fully authorized by all parties.

[0051] The technical solutions of this application will be described in detail below with reference to various embodiments.

[0052] like Figure 1 As shown in the diagram, this invention discloses a method for emotion monitoring and intervention based on multimodal neurochemical sensing, comprising the following steps:

[0053] S101, acquires the acoustic-electrodynamic spectrum signal and optical fluorescence spectrum signal of the user's target tissue area through a wearable device;

[0054] S102, based on the acoustic-electrodynamic spectrum signal and the optical fluorescence spectrum signal, the activity information of at least one neurotransmitter system related to the emotional state is identified through signal fusion and decoding;

[0055] S103, input the activity information of the neurotransmitter system into the individualized dynamic emotion monitoring model, and obtain and output the user's current emotion state type and emotion state intensity;

[0056] S104. Based on the current emotional state type and intensity, and combined with the activity information of the neurotransmitter system, calculate in real time a quantitative emotional instability index to characterize the trend of the emotional state toward a negative development.

[0057] S105, when the emotional instability index exceeds the individualized critical threshold set for the user, a personalized emotional intervention plan that matches the instability characteristics reflected by the emotional instability index is generated and executed.

[0058] In some embodiments, for step S101, for example, the wearable device on which this step depends, such as a smart headband or a wristband with sensing capabilities, integrates an integrated acoustic, electro-optical sensing module.

[0059] Specifically, the wearable device integrates multiple sensing units, such as a focused ultrasound modulation unit, a multi-band impedance analysis unit, an optical detection unit, and a coordinated sensing control unit.

[0060] For example, the focused ultrasound modulation unit may be a high-density miniature focused ultrasound transducer array manufactured using MEMS technology. The preferred number of array elements is 16×16, the operating frequency range is set to 500kHz–1MHz, and the output acoustic intensity of a single element is strictly controlled to <100mW / cm² to avoid excitation or inhibition of nerve tissue, serving only as a signal modulation carrier. The focusing depth of the transducer array can be dynamically adjusted from 0.5 to 3cm through element phase modulation to adapt to different target tissue regions. For example, the focusing depth is set to 2cm for monitoring the prefrontal cortex and 0.8cm for monitoring the median nerve at the wrist.

[0061] For example, a multi-band impedance analysis unit may include a miniature high-density high-frequency impedance analysis electrode array and a customized impedance analysis chip. The electrode array is made of biocompatible nickel-titanium alloy and has a four-electrode structure, with the two outer electrodes serving as current injection electrodes and the two inner electrodes as voltage measurement electrodes. The electrode spacing is set to 2 mm to eliminate interference from skin contact impedance on the measurement results. This unit has the ability to generate a swept-frequency alternating electric field from 1 kHz to 10 MHz, and the output current amplitude is controlled at 10 μA to avoid stimulating the human body. The impedance measurement accuracy reaches ±0.1% to capture subtle changes in the electrical properties of cell membranes caused by neurotransmitter activity.

[0062] For the optical detection unit, it may integrate, for example, a near-infrared II laser diode (NIR-II) in the 1000–1700 nm band, an ultra-narrowband optical filter with a bandwidth of <20 nm, and a high-sensitivity InGaAs photodetector array. The preferred center emission wavelength of the laser diode is 1310 nm, which allows for deep penetration and low scattering loss in biological tissues. The output power is controlled at 5 mW to avoid tissue damage. The ultra-narrowband optical filter is used to filter out interference from ambient light and tissue-scattered light, allowing only neurotransmitter fluorescence signals to pass through.

[0063] For example, the coordinated sensing control unit is responsible for synchronously scheduling the working sequence of the focused ultrasound modulation unit, the multi-band impedance analysis unit, and the optical detection unit to avoid mutual interference between different energy fields; it also has data preprocessing functions, which can perform preliminary filtering and format conversion on the acquired raw signals.

[0064] Specifically, in order to achieve the synergistic effect of the focused ultrasound energy field, the swept frequency alternating electric field and the near-infrared light signal, according to the embodiments of the present invention, the timing consistency of signal acquisition is ensured by coordinating the sensor control unit to adopt time-division multiplexing and phase-locked synchronization.

[0065] For example, the signal acquisition period is set to 1 second / time, and each period is divided into three stages, with a 10ms interval between each stage to avoid interference:

[0066] The ultrasound modulation phase lasts approximately 200ms: The focused ultrasound modulation unit emits a sequence of low-frequency, low-intensity focused ultrasound pulses with a pulse width of 10μs and a repetition frequency of 10Hz, forming an ultrasound energy field covering the target tissue area. This ultrasound energy field generates minute shear forces on the extracellular fluid and cell membrane through mechanical pressure waves, temporarily increasing the binding / dissociation rate of neurotransmitters and receptors, synchronizing numerous electrical changes in the postsynaptic membrane, and amplifying previously difficult-to-capture microscopic synaptic signals into measurable macroscopic impedance changes.

[0067] The electric field application and impedance detection phase lasts approximately 500ms: Simultaneously with the continuous application of the ultrasonic energy field, within the same time window of ultrasonic modulation, the multi-band impedance analysis unit applies a swept-frequency alternating electric field according to a preset frequency sequence, such as 1kHz to 10MHz, with a frequency step size of 10kHz, totaling 1000 sampling points. A known sinusoidal alternating current is injected into the two outer electrodes. The voltage drop generated by the synchronous measurement of the two inner electrodes Impedance spectra were recorded under the ultrasonic energy field on-state. Impedance spectrum in the off state The acoustic-electrodynamic spectrum signal is obtained by calculating the difference between the two:

[0068]

[0069] This difference signal can effectively filter out the static background impedance of tissues, retaining only the dynamic impedance component related to neurotransmitter activity.

[0070] The optical signal acquisition phase lasts approximately 300ms: In the same spatial region as ultrasonic modulation and impedance detection, the laser diode of the optical detection unit synchronously emits NIR-II laser pulses, exemplarily with a pulse width of 50ns and a repetition frequency of 20Hz. After the laser penetrates the target tissue region, neurotransmitters such as dopamine and serotonin produce weak endogenous fluorescence emission due to their molecular structure characteristics. An InGaAs photodetector array synchronously receives this fluorescence signal. After background interference is filtered out by an ultra-narrowband filter, the arrival time and intensity of the fluorescent photons are recorded by a time-gated single-photon counting module, ultimately forming an optical fluorescence spectrum signal with a wavelength resolution of 10nm and an intensity dynamic range of [missing information]. .

[0071] In some embodiments, for step S102, specifically, in order to eliminate the influence of noise and interference on the subsequent decoding results, the acoustic-electrodynamic spectrum signal and optical fluorescence spectrum signal acquired in step S101 are preprocessed.

[0072] For acoustic-electrodynamic spectrum signals, for example, a Butterworth bandpass filter, with a passband of 1kHz to 10MHz and an order of 8, can be used to remove high-frequency noise such as >10MHz and low-frequency drift such as <1kHz. This step is a conventional signal processing method and will not be described in detail here.

[0073] Independent component analysis (ICA) algorithm is used to separate motion artifacts from neurotransmitter-related signals. First, the electroacoustic spectrum signal is decomposed into multiple independent components. By calculating the correlation between each component and motion sensor data (exemplarily, data from a 3-axis accelerometer built into a wearable device), the independent components corresponding to motion artifacts are identified and removed, while the components related to neurotransmitter activity are retained.

[0074] The amplitude of the preprocessed acousto-electrodynamic spectrum signal is normalized to the interval [-1, 1], as shown in the following formula:

[0075]

[0076] in, The normalized acousto-electrodynamic spectrum signal. and These are the minimum and maximum values ​​of the original signal, respectively.

[0077] For optical fluorescence spectral signals, for example, the output signal of the photodetector when there is no laser irradiation, i.e., the dark current signal collected during the calibration stage, is subtracted to eliminate the interference of dark current on the fluorescence signal.

[0078] The fluorescence spectrum signal was fitted to the background using a fifth-order polynomial fitting method to obtain the background intensity of tissue scattered light. From the original fluorescence intensity Subtract the background intensity from the middle:

[0079]

[0080] Fluorescence intensity at the user's resting state Using this as a benchmark, the relative fluorescence intensity change rate is calculated to achieve signal standardization among different users:

[0081]

[0082] According to embodiments of the present invention, feature extraction extracts time-domain dynamic features and frequency-domain relaxation features from the acoustic-electrodynamic spectrum signal, and extracts optical features from the optical fluorescence spectrum signal.

[0083] Specifically, for the acoustic-electrodynamic spectral signal features, time-domain dynamic features are extracted, including:

[0084] Resonance frequency shift First, the normalized acousto-electrodynamic spectrum signal... A moving average filter is applied, with a window size of 5 frequency points, to smooth signal fluctuations. Then, a peak detection algorithm is used, with a threshold set to 80% of the maximum signal amplitude, to find the frequency corresponding to the peak value of the smoothed signal. Finally, the resonant frequency of this frequency and the user's resting state are calculated. The difference, i.e. Different neurotransmitters have different effects on the equivalent capacitance and resistance of the cell membrane, which can lead to specific shifts in the resonant frequency.

[0085] Phase angle trajectory Within a 200ms time window of ultrasonic modulation, the phase angle of the acoustic-electrodynamic spectrum signal is acquired every 10ms. To form a phase angle sequence The changing trend of this sequence is the phase angle trajectory.

[0086] The Cole-Cole model was used to fit the acoustic-electrodynamic spectrum signal. This model can accurately describe the impedance relaxation characteristics of biological tissues, and its expression is as follows:

[0087]

[0088] in, The low-frequency limiting impedance reflects the extracellular fluid ionic conductivity. This is the high-frequency limiting impedance, which reflects the capacitance characteristics of the cell membrane. The relaxation time constant reflects the rate of dynamic change in the electrical properties of the cell membrane. Let be the dispersion coefficient, where 0 < α < 1, which reflects the uniformity of impedance relaxation. The signal is fitted using the Levenberg-Marquardt nonlinear least squares method to obtain . and ,in As the key frequency domain feature, glutamic acid corresponds to Approximately 1ms, GABA corresponding Approximately 5ms, dopamine-related It takes about 3ms, which can be achieved through The numerical values ​​can be used to initially distinguish the types of neurotransmitters.

[0089] Specifically, the extraction of features from optical fluorescence spectral signals includes,

[0090] Fluorescence peak wavelength : For the standardized fluorescence spectral signal Perform peak detection to find the wavelength corresponding to the maximum signal value. The molecular structures of different neurotransmitters, such as the types of chromophores and the number of conjugated double bonds, differ, resulting in specific peak wavelengths in their fluorescence emission spectra.

[0091] fluorescence lifetime Fluorescence lifetime was calculated using time-resolved spectroscopy. A laser diode was controlled to emit laser pulses with a width of 50 ns, and the decay curve of the fluorescence signal over time was recorded. The curve was fitted using a double exponential decay model:

[0092]

[0093] in, and Let the amplitudes of the two attenuation components be respectively, and exemplarily satisfy the following conditions: , and These are the fluorescence lifetimes of the fast-decaying and slow-decaying components, respectively. The weighted average lifetime is taken as the final fluorescence lifetime characteristic.

[0094]

[0095] According to embodiments of the present invention, signal fusion and neurotransmitter decoding specifically involve fusing the extracted multimodal features, decoding neurotransmitter system activity information using a deep learning model, and simultaneously performing signal consistency verification. The specific implementation is as follows:

[0096] First, construct a high-dimensional feature vector: extract the acoustic and electrical features ( , Standard deviation ) and optical characteristics ( , The features are combined in a "dimensional concatenation" manner to form a high-dimensional feature vector. For example, if the acoustic-electric features contain 3 parameters and the optical features contain 2 parameters, then the high-dimensional feature vector is... The vector dimension is 5. To ensure a balanced contribution of features, each feature needs to be min-max normalized, i.e., mapped to the [0,1] interval, as shown in the following formula:

[0097]

[0098] in, These are the original eigenvalues. and These are the minimum and maximum values ​​of the feature in the training dataset, respectively.

[0099] For neurotransmitter decoding models, in some embodiments, the decoding model is built on graph neural networks (GNNs), an architecture that can effectively capture the correlations between multimodal features.

[0100] The model includes an input layer that receives high-dimensional feature vectors (5 dimensions), a graph convolutional layer that transforms the feature vectors into graph nodes and constructs an adjacency matrix using cosine similarity (64-dimensional output), a pooling layer (e.g., top-k pooling to retain 32 important node features), a fully connected layer (e.g., two hidden layers with output dimensions of 32 and 16 respectively), a ReLU activation function, and an output layer (optionally sigmoid activation function) that outputs the relative activity intensities of three emotion-related neurotransmitters: glutamate, GABA, and dopamine (3-dimensional output).

[0101] The training dataset contains a mapping between high-dimensional feature vectors and neurotransmitter concentrations. The model uses the Adam optimizer with a learning rate of 1e-4, a batch size of 32, and 100 training epochs. The loss function is mean squared error (MSE).

[0102]

[0103] in, For the sample size, For the first The first sample Normalized values ​​of the true concentrations of various neurotransmitters These are the model's predicted values. To adapt to the real-time requirements of wearable devices, the trained model is quantized, such as converting 32-bit floating-point numbers to 8-bit integers and pruning. For example, a pruning rate of 30% reduces the model size by 75% and increases inference speed by 3 times.

[0104] The high-dimensional feature vector is input into the optimized model, and the output is glutamic acid (…). ), GABA ( ), dopamine ( The relative activity intensity, ranging from [0,1], for example, output This indicates that glutamate activity is relatively strong, GABA activity is relatively weak, and dopamine activity is moderate.

[0105] According to an embodiment of the present invention, preferably, it includes acoustic-electrical-optical signal consistency verification; specifically, it selects acoustic-electrical features... In optical characteristics As a check feature pair (both are strongly correlated with neurotransmitter type), the Pearson correlation coefficient was calculated. :

[0106]

[0107] in, Optimization of sampling number for a single monitoring session , , These are the characteristic means.

[0108] Based on a large-scale calibration dataset, such as 1000 user data, a first validity threshold of 0.7 and a second validity threshold of 0.5 are determined.

[0109] when The signals are determined to be highly consistent and effective for decoding. The weights of both acoustic and optical features are set to 0.5.

[0110] when The signal consistency is determined to be moderate. The weight of optical features is reduced to 0.3, and the weight of acoustic-electric features is increased to 0.7. At the same time, ambient light compensation of the optical module is triggered, such as extending the detector integration time to 200μs.

[0111] when If the data is deemed invalid, the data collection process will be restarted. For example, a maximum of 3 retries will be made. If the process still fails, the user will be reminded via the app to "adjust the environment, such as moving away from strong light or adjusting the position of the device".

[0112] In some embodiments, for step S103, the training of the general emotion monitoring model, exemplarily, includes training a multi-task learning model using a large-scale population dataset, such as a sample size ≥1000 people. The inputs are neurotransmitter activity information and basic physiological data such as heart rate and HRV. The outputs are emotion state types such as happiness, sadness, anger, anxiety, depression, etc. (8 categories) and emotion state intensity, such as 1 to 5 levels. The model architecture includes a CNN-LSTM hybrid structure, which is used to process temporal features and a cross-individual activity LSTM multi-task autoencoder, which simultaneously optimizes type classification and intensity regression tasks. This part is a conventional deep learning model training process and will not be described in detail here.

[0113] For individual users, within a preset period (e.g., 7 days) of their first device use, emotional tags are obtained through Dynamic Ecosystem Immediate Assessment (EMA): the device randomly triggers EMAs 3 times per day (interval ≥ 4 hours), and users are prompted to self-assess their emotional type and intensity via an app pop-up window, such as "Current Emotion: Anxiety, Intensity: 3 / 5". Neurotransmitter activity information at that moment is simultaneously recorded to construct a longitudinal correlation dataset for the user. A Meta-Learning Algorithm (MAML) is used to optimize the parameters of the general model; the specific process is as follows:

[0114] Using the general model parameters as initial weights, fine-tuning is performed 10 times using 80% of the samples in the user longitudinal association dataset, such as a learning rate of 5e-5.

[0115] The remaining 20% ​​of the samples were used to validate the model performance. If the emotion type classification accuracy was ≥85% and the intensity prediction MAE was ≤0.5, it was determined to be an individualized dynamic emotion monitoring model.

[0116] If performance is not up to standard, increase the number of EMA triggers, such as 5 times per day, and readjust after supplementing the dataset until the performance requirements are met.

[0117] Specifically, the neurotransmitter activity information obtained from step S102 ( , , The process of inputting an individualized model and outputting the user's current emotional state type and intensity is as follows:

[0118] Input features include the intensity of neurotransmitter activity at the current time ( , , The activity intensity time series within the last 5 minutes, with one data point per minute and a total of 5 time steps, form a time series feature matrix with a dimension of 3×5.

[0119] For predicting the type of emotional state, the CNN layer of the CNN-LSTM hybrid structure extracts local patterns of temporal features, such as the rapid increase of glutamate within 5 minutes, while the LSTM layer captures long-term dependencies, such as the continuous decrease of dopamine. The output features are input into a softmax classifier to obtain the probability of each emotional type, such as anxiety probability 0.8, depression probability 0.1, and other probability 0.1. The type with the highest probability is selected as the current emotional state type.

[0120] For predicting the intensity of emotional state, specifically, the regression branch of the multi-task learning sub-model outputs continuous values ​​such as 1 to 5 based on temporal features. The larger the value, the higher the intensity of the emotion. For example, an output of 3.5 indicates moderate anxiety.

[0121] In some embodiments, for step S104, according to the present invention, this step constructs a time sequence based on neurotransmitter activity information and emotional state intensity.

[0122] Specifically, this includes constructing a neurochemical state vector, which incorporates neurotransmitter activity information from step S102 ( , , The intensity of emotional state in step S103 The combination of (normalized to the [0,1] interval) forms the neurochemical state vector at the current moment. For example, if , , , (The normalized value after intensity level 3.5), then .

[0123] In addition, a time series sequence of neurochemical states is generated. A preset time window of 30 minutes is set, which balances timeliness and stability. The time step is 1 minute. Historical neurochemical state vectors within this window are extracted to form a time series sequence. For example, there are 30 vectors in total, forming a time series matrix with a dimension of 30×4.

[0124] Multi-scale features representing emotion stability are extracted using complexity analysis and synchronicity analysis methods, as specifically implemented below:

[0125] Complexity Analysis: The multi-scale entropy (MSE) algorithm is used to calculate the entropy of the time series at different time scales, such as scales 1 to 10. A lower entropy value indicates a more stable neurochemical state, meaning emotions are less likely to become unstable; a higher entropy value indicates a more chaotic state, meaning the state is more prone to negative emotions. For example, before anxiety arises, the MSE value will significantly increase in the range of 5 to 8.

[0126] Synchronization analysis: Calculate the mutual information (MI) between the activity intensities of different neurotransmitters to quantify system synergy.

[0127] Calculate the mutual information between glutamate and GABA This reflects the balance between excitatory and inhibitory neurotransmitters;

[0128] Calculate the mutual information between dopamine and the intensity of emotional state. This reflects the degree of correlation between neuromodulation and emotional intensity;

[0129] Calculate the mutual information between glutamate and dopamine This reflects the synergistic relationship between excitatory neurotransmitters and neuromodulators.

[0130] Reduced mutual information indicates weakened synergy in the neurotransmitter system, making one more prone to emotional instability.

[0131] Multi-scale feature vector construction: Combine 10 MSE entropy values ​​with 3 mutual information values ​​to form a multi-scale feature vector with dimension 13. .

[0132] Specifically, the emotion instability prediction model is built on a temporal convolutional network (TCN), which can effectively capture the long-term dependencies of time series and achieve accurate quantification of emotion instability trends. The process is as follows:

[0133] For model training, the training dataset contains a correspondence between "multi-scale feature vectors and emotional instability labels". The label 1 indicates that the emotional intensity increases by ≥1 level within 30 minutes, and 0 indicates no increase. The model adopts the TCN architecture and can contain 3 convolutional blocks. Each block contains 2 convolutional layers, a batch normalization layer and a ReLU activation function. The output layer uses the sigmoid activation function and outputs the emotional instability probability, i.e., the emotional instability index.

[0134] During the model inference stage, multi-scale feature vectors are... Input a pre-trained TCN model and output an emotion instability index between 0 and 1. .For example, This indicates a very strong tendency for emotions to develop negatively. This indicates a stable emotional state.

[0135] In some embodiments, for step S105, according to the present invention, a personalized intervention plan is generated based on the characteristics of emotional instability, including instability type judgment and personalized correction of intervention parameters. Please refer to... Figure 2 , Figure 2 This is a schematic diagram illustrating the implementation process of the personalized intervention scheme provided in this embodiment of the invention. The specific implementation is as follows:

[0136] In S201, the dominant type of emotional instability is determined: combining the neurotransmitter activity information from step S102 with the emotional state type from step S103, the dominant type of instability is determined:

[0137] If the emotional state type is "anxiety" and the intensity of norepinephrine activity is elevated, such as by additional monitoring of norepinephrine characteristics, A decrease, i.e. an imbalance between excitability and inhibition, is classified as "anxiety-type instability";

[0138] If the mood state type is "depression" and the serotonin activity is reduced, If dopamine levels decrease, such as a weakened association between dopamine and mood intensity, it is diagnosed as "depressive instability";

[0139] If multiple neurotransmitter abnormalities are present simultaneously, such as elevated glutamate and decreased serotonin, it is classified as "mixed instability".

[0140] In S202, initial intervention program matching: The pre-set intervention strategy library contains various intervention methods, such as digital odor, microtouch, thermal feedback, and low-intensity ultrasound, and the initial program is matched according to the dominant instability type.

[0141] Anxiety-related instability: Match with "lavender scent for its calming effect and 5Hz low-frequency microtactile stimulation, once every 10 minutes, for a total of 3 times";

[0142] Depressive instability: Matched with "citrus scent for an uplifting effect and 38°C thermal feedback, once every 15 minutes, for a total of 2 times";

[0143] Mixed instability: Matched with "light lavender scent, 8Hz micro-tactile feedback and 36℃ thermal feedback, once every 12 minutes, for a total of 3 times".

[0144] In S203, personalized adjustment of intervention parameters: the initial plan is adjusted based on the user's historical intervention effect data.

[0145] Effectiveness evaluation indicators: defining the intervention effect , The larger the value, the better the intervention effect;

[0146] If historical data shows that users' preference for the "lavender scent" If the effect is poor, the scent type will be changed to "cedar scent";

[0147] If the user's subjective feedback score for the 5Hz micro-touch is less than 3 / 5 points, indicating low acceptance, then the vibration frequency will be adjusted to 8Hz.

[0148] If a user experiences discomfort when their HR rises by more than 15 bpm at 38℃, the temperature will be adjusted to 36℃.

[0149] In S204, the final intervention plan is determined. The revised plan includes intervention content such as "cedar scent and 8Hz microtouch", intervention frequency such as "once every 10 minutes" and stimulation parameters such as scent concentration of 60% and vibration intensity of 0.3g.

[0150] According to an embodiment of the present invention, preferably, specifically, when the emotional instability index When the threshold exceeds the user's individualized metastable critical threshold, this threshold can be adaptively adjusted based on the user's historical data and current situation. Within this metastable time window, preventative and timely multi-channel subconscious intervention is provided to the user. The critical threshold is dynamically adjusted based on the intervention effect and real-time physiological state to ensure the accuracy of the intervention timing, specifically including:

[0151] Periodic adaptive adjustment:

[0152] Record the trigger index for each intervention. Emotional instability index 30 minutes after intervention and changes in emotional intensity The average intervention effect is calculated every 7 days in a pre-set cycle. Number of interventions within 7 days;

[0153] Threshold adjustment:

[0154] like Results exceeded expectations: Individualized critical threshold Increasing it by 0.05, such as from 0.6 to 0.65, reduces unnecessary intervention;

[0155] like Results did not meet expectations: A reduction of 0.05, such as from 0.6 to 0.55, triggers intervention earlier;

[0156] like :maintain constant.

[0157] Contextualized instantaneous correction:

[0158] Combining real-time HRV and motion status data of users, Perform instantaneous correction:

[0159] HRV Correction: If the real-time HRV < 30ms, i.e., a high-pressure state, then That is, lower the threshold and intervene earlier; if HRV > 50ms, i.e., a low-stress state, then That is, raise the threshold and reduce disturbances;

[0160] Motion state correction: If the motion sensor detects that the user is in motion, such as acceleration > 0.5g, then... Because exercise can cause temporary fluctuations in neurotransmitters, raising the threshold can prevent false triggering; if in a resting state, it can maintain... constant.

[0161] Final instantaneous correction threshold This ensures that the threshold is adapted to the user's current scenario and physiological state.

[0162] Therefore, this invention achieves non-invasive monitoring of neurotransmitter activity through multimodal neurochemical sensing, and combines individualized emotion monitoring models with dynamic instability index calculation to form a complete process of emotion monitoring, assessment, intervention and optimization.

[0163] Figure 3 An emotion monitoring and intervention system 300 based on multimodal neurochemical sensing is illustrated. The system implementation is similar to... Figure 1 Corresponding to the method embodiments shown, this system can be specifically applied to various electronic devices, specifically including,

[0164] The signal acquisition module 301 acquires the acoustic-electrodynamic spectrum signal and optical-fluorescence spectrum signal of the user's target tissue area through a wearable device;

[0165] Activity information recognition module 302 is used to identify the activity information of at least one neurotransmitter system related to emotional state by signal fusion and decoding based on the acoustic-electrodynamic spectrum signal and optical fluorescence spectrum signal.

[0166] The emotion information acquisition module 303 is used to input the activity information of the neurotransmitter system into the individualized dynamic emotion monitoring model, and to acquire and output the user's current emotion state type and emotion state intensity.

[0167] The emotion quantification module 304 is used to calculate in real time a quantitative emotion instability index that characterizes the trend of the emotion state toward a negative development, based on the current emotion state type and emotion state intensity, and combined with the activity information of the neurotransmitter system.

[0168] The intervention plan generation module 305 is used to generate and execute a personalized emotional intervention plan that matches the instability characteristics reflected by the emotional instability index when the emotional instability index exceeds the individualized critical threshold set for the user.

[0169] Based on the same inventive concept, this application also provides an electronic device. The method corresponding to the electronic device can be the method in the foregoing embodiments, and its problem-solving principle is similar to that method. The electronic device provided in this application includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the methods and / or technical solutions of the foregoing embodiments of this application.

[0170] The electronic device can be a user device, or a device formed by integrating user devices and network devices through a network, or it can be an application running on the aforementioned devices. The user device includes, but is not limited to, various terminal devices such as computers, mobile phones, tablets, smartwatches, and smart bands. The network device includes, but is not limited to, network hosts, single network servers, multiple network server sets, or cloud computing-based computer sets, and can be used to implement some processing functions when setting an alarm clock. Here, the cloud consists of a large number of hosts or network servers based on cloud computing. Cloud computing is a type of distributed computing, consisting of a virtual computer composed of a group of loosely coupled computer sets.

[0171] Figure 4 The diagram illustrates the structure of an apparatus suitable for implementing the methods and / or technical solutions in the embodiments of this application. The apparatus 400 includes a central processing unit 401, which can perform various appropriate actions and processes based on a program stored in a read-only memory 402 or a program loaded from a storage portion 408 into a random access memory 403. Various programs and data required for system operation are also stored in the RAM 403. The CPU 401, ROM 402, and RAM 403 are interconnected via a bus 404. An input / output interface 405 is also connected to the bus 404.

[0172] The following components are connected to I / O interface 405: an input section 406 including a keyboard, mouse, touchscreen, microphone, infrared sensor, etc.; an output section 407 including cathode ray tubes, liquid crystal displays, LED displays, OLED displays, etc., and speakers, etc.; a storage section 408 including one or more computer-readable media such as hard disks, optical disks, magnetic disks, semiconductor memory, etc.; and a communication section 409 including network interface cards such as LAN cards, modems, etc. The communication section 409 performs communication processing via a network such as the Internet.

[0173] In particular, the methods and / or embodiments in this application can be implemented as computer software programs. For example, the embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowchart. When the computer program is executed by the central processing unit (CPU) 401, it performs the functions defined in the methods of this application.

[0174] Another embodiment of this application provides a computer-readable storage medium having computer program instructions stored thereon, which can be executed by a processor to implement the methods and / or technical solutions of any one or more embodiments of this application described above.

[0175] Specifically, this embodiment may employ any combination of one or more computer-readable media. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. In this document, a computer-readable storage medium may be any tangible medium that contains or stores a program that may be used by or in conjunction with an instruction execution system, apparatus, or device.

[0176] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0177] Program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0178] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0179] The flowcharts or block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of devices, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-specific system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0180] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0181] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or page components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between devices or units through some interfaces, and may be electrical, mechanical, or other forms.

[0182] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0183] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in a combination of hardware and software functional units.

[0184] The integrated units implemented as software functional units described above can be stored in a computer-readable storage medium. These software functional units, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks.

[0185] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

[0186] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices recited in a device claim may also be implemented by a single unit or device through software or hardware. The terms "first," "second," etc., are used to indicate names and do not indicate any specific order.

Claims

1. A method for emotion monitoring and intervention based on multimodal neurochemical sensing, characterized in that, The methods and steps include the following: Acquire the acoustic-electrodynamic spectrum and optical-fluorescence spectrum of the target tissue region of the user; wherein, the acoustic-electrodynamic spectrum is specifically the difference between the impedance spectrum in the ultrasonic energy field on state and the impedance spectrum in the ultrasonic energy field off state, the ultrasonic energy field is a sequence of focused ultrasonic pulses emitted by the focused ultrasonic modulation unit, forming an ultrasonic energy field covering the target tissue region, amplifying the synaptic signal into a measurable impedance change; the impedance spectrum in the ultrasonic energy field on state is obtained by a multi-band impedance analysis unit applying a swept-frequency alternating electric field according to a preset frequency sequence within the same time window of ultrasonic modulation while the ultrasonic energy field is continuously acting; Based on the aforementioned acoustic-electrodynamic spectrum signal and optical fluorescence spectrum signal, the activity information of at least one neurotransmitter system related to emotional state is identified through signal fusion and decoding. The activity information of the neurotransmitter system is input into the individualized dynamic emotion monitoring model to obtain and output the user's current emotion state type and emotion state intensity. Based on the current emotional state type and intensity, and combined with the activity information of the neurotransmitter system, a quantitative emotional instability index is calculated in real time to characterize the trend of the emotional state towards a negative direction. When the emotional instability index exceeds the individualized critical threshold set for the user, a personalized emotional intervention plan that matches the instability characteristics reflected by the emotional instability index is generated and executed.

2. The emotion monitoring and intervention method based on multimodal neurochemical sensing according to claim 1, characterized in that, The process of identifying neurotransmitter system activity information through signal fusion and decoding includes: Temporal dynamic features and frequency domain relaxation features are extracted from the acoustic-electrodynamic spectrum signal, and optical features are extracted from the optical fluorescence spectrum signal; The extracted acoustic and electronic features are combined with optical features to form a high-dimensional feature vector; The high-dimensional feature vector is input into a neurotransmitter decoding model built based on a deep learning algorithm, and the model outputs the relative activity intensity of a specific neurotransmitter category. Furthermore, during the decoding process, consistency verification between acoustic and optical signals is performed, and the weights of different signals in the final decoding result are dynamically adjusted or the re-acquisition of the trigger signal is initiated based on the verification results.

3. The emotion monitoring and intervention method based on multimodal neurochemical sensing according to claim 2, characterized in that, The consistency verification of the acoustic and optical signals includes: Calculate the statistical correlation between specific time-domain or frequency-domain characteristics of the acoustic-electrodynamic spectral signal and specific optical characteristics of the optical-fluorescence spectral signal; When the statistical correlation is higher than the first validity threshold, the signal is determined to be valid and used for neurotransmitter decoding. When the statistical correlation is lower than the first validity threshold but higher than the second validity threshold, the contribution weight of the optical fluorescence spectral signal in decoding is reduced, and the ambient light compensation mechanism of the optical module is triggered. When the statistical correlation is lower than the second validity threshold, the data is determined to be invalid and the signal re-acquisition process is initiated.

4. The emotion monitoring and intervention method based on multimodal neurochemical sensing according to claim 1, characterized in that, The real-time calculation of the emotional instability index specifically includes: Based on the activity information of the neurotransmitter system and the intensity of the current emotional state, a neurochemical state vector is constructed for the current moment. By combining the historical neurochemical state vectors within a preset time window, a neurochemical state time sequence is constructed. Complexity analysis and synchronicity analysis methods are used to extract multi-scale features characterizing the dynamic stability of the system from the time series. The multi-scale features are input into a trained emotion instability prediction model, which outputs a value between zero and one as the emotion instability index.

5. The emotion monitoring and intervention method based on multimodal neurochemical sensing according to claim 1, characterized in that, The generation of personalized emotion intervention programs that match the instability characteristics reflected by the emotion instability index includes: Based on the activity information of the neurotransmitter system and the current emotional state type, determine the dominant type of emotional instability; Based on the dominant type of emotional instability, the corresponding intervention content and frequency are matched from the pre-set intervention strategy library to form an initial intervention plan; Based on the user's historical intervention effect data, the stimulus parameters in the initial intervention plan are individually modified to generate the final personalized emotion intervention plan.

6. The emotion monitoring and intervention method based on multimodal neurochemical sensing according to claim 5, characterized in that, It also includes, Record the emotional instability index at the time of each intervention trigger and the emotional state recovery data after the intervention; Based on the evaluation results of the intervention effect within a preset period, the individualized critical threshold is adaptively adjusted. When the intervention effect is consistently better than expected, the threshold is increased; when the intervention effect is consistently worse than expected, the threshold is decreased. Simultaneously, real-time heart rate variability and exercise status data of users are acquired, and the individualized critical threshold is adjusted instantaneously based on the specific context.

7. An emotion monitoring and intervention system based on multimodal neurochemical sensing, characterized in that, include: The signal acquisition module acquires the acoustic-electrodynamic spectrum and optical fluorescence spectrum of the user's target tissue region via a wearable device. Specifically, the acoustic-electrodynamic spectrum is the difference between the impedance spectrum in the ultrasonic energy field's on state and its off state. The ultrasonic energy field is a sequence of focused ultrasonic pulses emitted by the focused ultrasonic modulation unit, forming an ultrasonic energy field covering the target tissue region and amplifying the synaptic signal into a measurable impedance change. The impedance spectrum in the ultrasonic energy field's on state is acquired by a multi-band impedance analysis unit applying a swept-frequency alternating electric field according to a preset frequency sequence within the same time window of ultrasonic modulation, while the ultrasonic energy field is continuously acting. The activity information recognition module is used to identify the activity information of at least one neurotransmitter system related to emotional state through signal fusion and decoding based on the acoustic-electrodynamic spectrum signal and the optical fluorescence spectrum signal. The emotion information acquisition module is used to input the activity information of the neurotransmitter system into the individualized dynamic emotion monitoring model, and to acquire and output the user's current emotion state type and emotion state intensity. The emotion quantification module is used to calculate a quantitative emotion instability index in real time, which is used to characterize the trend of the emotion state towards a negative development, based on the current emotion state type and intensity, and combined with the activity information of the neurotransmitter system. The intervention plan generation module is used to generate and execute a personalized emotional intervention plan that matches the instability characteristics reflected by the emotional instability index when the emotional instability index exceeds the individualized critical threshold set for the user.

8. An electronic device, the electronic device comprising: At least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-6.

9. A computer-readable medium having stored thereon computer program instructions that can be executed by a processor to implement the method as described in any one of claims 1-6.

10. A computer program product comprising a computer program that, when executed by a processor, implements the method as described in any one of claims 1-6.

Images