Personalized electrical stimulation signal generation system and otDCS intervention method thereof

Abstract

The invention discloses a personalized electrical stimulation signal generation system and a space-frequency domain adaptive otDCS intervention method for modulating gamma oscillation, and the method comprises the steps: collecting EEG signal data of an individual, carrying out the re-reference, filtering, artifact removal and data segment interception of the collected gamma-ASSR data, extracting the resonant frequency of the gamma oscillation of the individual as the stimulation frequency of otDCS through spectrum analysis, and carrying out the spatial frequency domain adaptive otDCS intervention of the individual. The instantaneous phase of individual gamma oscillation is extracted through phase estimation to serve as the stimulation phase of the otDCS, and key nodes in an individual gamma oscillation brain network are extracted through a network analysis method to serve as stimulation sites of the otDCS; and according to the resonant frequency, the instantaneous phase and the key node of the gamma oscillation of the individual obtained by the personalized stimulation parameter extraction module, locking the stimulation frequency, the stimulation phase and the stimulation site, and generating an otDCS waveform for subsequent electrical stimulation output of the individual.

Description

Technical Field

[0001] The present invention relates to the field of neural regulation, and in particular to a personalized electrical stimulation signal generation system and a spatial-frequency domain adaptive otDCS (oscillatory transcranial direct current stimulation) intervention method for modulating gamma oscillations. Background Art

[0002] Major depressive disorder (MDD) is a common clinical mental disorder with high morbidity, high relapse rate, high disability rate and low cure rate. It has become a major public health problem threatening human mental health. Clinically, the first-line treatment of MDD is still mainly drug therapy, but its effectiveness is less than 50%, the effect is slow and there are certain side effects.

[0003] In recent years, scientists have confirmed that brain functional activities are closely related to the dynamic interaction of local and large-scale networks, which is reflected in neural oscillation activities in different frequency bands, and has promoted intervention strategies targeting EEG oscillation activities in order to accurately improve brain function. Among them, EEG gamma oscillations, as a research focus, are leading a new wave of brain function exploration. Previous studies have revealed that the coherence of gamma oscillations in the right hemisphere of the brain of MDD patients is significantly reduced when performing emotional face tasks. The Liu Shuang team of Tianjin University found that there is spontaneous gamma activity inhibition in local brain areas of MDD patients (Xiaoya Liu, Shuang Liu, Meijuan Li, et al. Altered gamma oscillations and beta–gamma coupling in drug-naive first-episode major depressive disorder: association with sleep and cognitive disturbance [J]. Journal of Affective Disorders, 2022, 316: 99-108). Subsequently, the team designed a 40Hz auditory steady-state response (ASSR) induced paradigm and found for the first time that the consistency of 40Hz-ASSR trials in the right frontal and temporal lobe of MDD patients was weakened, suggesting that MDD patients have abnormal gamma oscillation information encoding, and gamma oscillation is expected to become a potential biomarker for identifying MDD (Shuang Liu, Xiaoya Liu, Sitong Chen, et al. Neurophysiological markers of depression detection and severity prediction in first-episode major depressive disorder [J]. Journal of Affective Disorders, 2023, 331: 8-16). In addition, prospective studies in the field of MDD treatment have focused on the gamma oscillation activity of patients and found that the EEG gamma oscillations of MDD patients also change during the symptom relief process.For example, Noda et al. used transcranial magnetic stimulation to treat MDD and found that during the recovery process, the patient's prefrontal gamma activity was enhanced, suggesting that gamma oscillations may become a new target for the treatment of MDD (Noda Y, Zomorrodi R, Saeki T, et al. Resting-state EEG gamma power and theta–gamma coupling enhancement following high-frequency left dorsolateral prefrontal rTMS in patients with depression [J]. Clinical Neurophysiology, 2017, 128 (3): 424-432).

[0004] Transcranial electrical stimulation techniques include transcranial direct current stimulation (tDCS), transcranial alternating current stimulation (tACS), etc. Given their safety, portability, ease of operation and low cost, they have broad application prospects in the treatment of MDD. Among them, tDCS regulates neuronal excitability through cathode or anode stimulation, while tACS uses alternating current of a specific frequency to regulate neural oscillation activity. In recent years, oscillatory tDCS (otDCS) has been developed as a special tDCS technology that simulates or enhances endogenous neural oscillations in the brain by applying a current intensity (frequency modulated current) that changes periodically over time to the scalp. This regulatory technology can not only cause changes in neuronal excitability, but also induce an entrainment effect to modulate brain neural oscillation activity, providing a new way to effectively regulate EEG gamma oscillations in MDD patients.

[0005] The regulatory effect of transcranial electrical stimulation mainly depends on the configuration mode of different stimulation parameters such as stimulation frequency and stimulation site, and there is significant heterogeneity of effects between individuals, that is, there are significant differences in the responses of different individuals to stimulation. Therefore, it is particularly important to establish a personalized stimulation parameter configuration mode and combine closed-loop control to construct an adaptive intervention method to improve the intervention effect. At present, research in related fields has made certain progress. For example, in terms of stimulation frequency, Kudo D et al. revealed that when the stimulation frequency is close to the resonant frequency of individual neural oscillations, its cortical oscillations show a more obvious regulatory effect (Kudo D, Koseki T, Katagiri N, et al. Individualized beta-band oscillatory transcranial direct current stimulation over the primary motor cortex enhances corticomuscular coherence and corticospinal excitability in healthy individuals [J]. Brain Stimulation. 2022, 15 (1): 46-52). In terms of stimulation phase, research by Chen Ni et al. shows that when stimulation occurs at different phases of the EEG signal, it can produce excitatory or inhibitory neuromodulatory effects. However, exogenous stimulation needs to maintain a stable phase lock with the EEG signal to synchronize the excitatory state of oscillatory activity in order to effectively regulate neural oscillations (Chen Ni, Qin Yurong, Xiong Yanting, et al. EEG phase-locked stimulation method based on variational mode decomposition [J]. Chinese Journal of Scientific Instrument, 2020, 41(5): 205-213). In terms of stimulation sites, given the spatial specificity of brain regions in MDD patients, the positioning of the optimal stimulation target plays a core role in the selection of stimulation sites. Yang Jun et al. proposed that brain network functional connectivity can predict the efficacy of magnetic stimulation therapy (Yang Jun, Yang Chunxia, ​​Liu Penghong, et al. Research progress on brain network mechanisms of repetitive transcranial magnetic stimulation for antidepressant treatment based on magnetic resonance imaging [J]. Chinese Journal of Psychiatry, 2023, 56(2): 144-149), suggesting that key nodes in the network are expected to become lesion targets for MDD patients to improve network connectivity and thus alleviate depressive symptoms, providing a new idea for the positioning of MDD lesion targets.

[0006] In summary, given the urgent need for new technologies in the treatment of MDD, a method is needed that can effectively modulate gamma oscillations and provide precise matching and adaptive intervention around the three key dimensions of frequency, phase, and location. Summary of the invention

[0007] In view of the deficiencies of the prior art, the present invention provides a personalized electrical stimulation signal generation system and a spatial-frequency domain adaptive otDCS intervention method for modulating gamma oscillations. The generation system and intervention method are intended to effectively modulate gamma oscillations, by designing a frequency-stepped gamma band ASSR induced paradigm, relying on electroencephalogram (EEG) technology, using spectrum analysis to extract the resonant frequency of individual gamma oscillations, then using an autoregressive-based phase estimation algorithm to extract the instantaneous phase of individual gamma oscillations, then constructing a brain function network model of individual gamma oscillations based on Granger causality analysis, and locating the key nodes of the network by calculating indicators such as betweenness centrality; combining EEG and otDCS technology to achieve closed-loop adaptive control, and regulating the gamma oscillation activity of MDD patients' brain waves in a personalized and precise control method, improving their high-frequency auditory perception function, and alleviating their clinical symptoms, in order to provide a new framework and technical support for targeted and precise treatment of MDD.

[0008] The objective of the present invention is achieved through the following technical solutions:

[0009] A personalized electrical stimulation signal generation system, comprising:

[0010] A signal acquisition module for collecting EEG signal data of an individual; in the signal acquisition module, the EEG signal data is subjected to a preset frequency-stepping gamma-ASSR paradigm as an auditory stimulation paradigm, thereby collecting gamma-ASSR data; wherein the gamma frequency band range of the gamma-ASSR paradigm is 30 Hz to 50 Hz;

[0011] The signal preprocessing module re-references, filters, and removes artifacts from the gamma-ASSR data collected by the signal acquisition module, extracts ASSR signals of different gamma frequencies from the data without artifact interference, and intercepts multiple data segments;

[0012] A personalized stimulation parameter extraction module extracts personalized stimulation parameters based on the data segments obtained by the signal preprocessing module, wherein the resonance frequency of the individual gamma oscillation is extracted by spectrum analysis as the stimulation frequency of otDCS, the instantaneous phase of the individual gamma oscillation is extracted by phase estimation as the stimulation phase of otDCS, and the key nodes in the individual gamma oscillation brain network are extracted by network analysis method as the stimulation sites of otDCS; and an otDCS stimulation module locks the stimulation frequency, stimulation phase and stimulation site based on the resonance frequency, instantaneous phase and key nodes of the individual gamma oscillation obtained by the personalized stimulation parameter extraction module, and generates an otDCS waveform for subsequent individual electrical stimulation output.

[0013] Specifically, the signal preprocessing module further includes:

[0014] The obtained EEG data were re-referenced, and the average value of the two leads of the bilateral mastoid M1 and M2 was selected as the reference;

[0015] The EEG data after re-reference was filtered, and the band-pass filter was selected as 30-50 Hz;

[0016] Independent component analysis was used to detect artifacts in the data to remove artifacts such as electrooculography and electromyography.

[0017] The ASSR signals of different gamma frequencies are extracted from the data after artifact interference is removed and subdivided into epoch data segments of preset time periods.

[0018] Furthermore, extracting the resonant frequency of individual gamma oscillations includes:

[0019] Fast Fourier transform is used to calculate the power spectrum curve of the gamma frequency band of each epoch data segment in the gamma-ASSR signal under each lead, and the average ASSR energy value of the gamma frequency of the individual is obtained by superposition and averaging at the lead level and the epoch data segment level;

[0020] Use the built-in peak finding function findpeaks() and maximum finding function max() in MATLAB to obtain the resonant frequency F of the individual gamma oscillation IGR .

[0021] Furthermore, extracting the instantaneous phase of individual gamma oscillations includes:

[0022] Construct a phase estimation model based on autoregression and select F IGR -ASSR data are modeled, and the least squares method is used to estimate the AR model coefficients; a voting decision method is used to obtain the instantaneous phase of the individual gamma oscillation in the phase estimated by the data; wherein the F IGR -ASSR data refers to the extracted individual gamma resonance frequency F IGR is the center frequency, with F IGR ±2 Hz is the bandwidth, and the data is obtained after the ASSR data is band-pass filtered using a Butterworth filter.

[0023] Furthermore, the key nodes of the brain network for extracting individual gamma oscillations include:

[0024] Constructing F of MDD patients based on Granger causality analysis IGR-ASSR brain functional network model, extract the betweenness centrality of the individual brain functional network, and select the two electrodes with the highest betweenness centrality as the key nodes of the brain network of the individual gamma oscillation.

[0025] Preferably, the individual is an MDD patient, and the obtained individualized gamma resonance frequency is determined as the individual's therapeutic stimulation frequency; the obtained individualized gamma oscillation instantaneous phase is determined as the individual's therapeutic stimulation phase, and the obtained individualized gamma oscillation brain network key nodes are determined as the individual's therapeutic targets.

[0026] Preferably, the frequency-stepped gamma-ASSR paradigm uses 1 Hz as a step size, sets multiple gamma frequencies within the frequency band, and selects a chirp signal as the sound stimulus.

[0027] The spatial-frequency domain adaptive otDCS intervention method of modulating gamma oscillation using the personalized electrical stimulation signal generation system comprises:

[0028] Step 1: Collect individual EEG signal data according to the preset frequency-stepping gamma-ASSR paradigm to obtain gamma-ASSR data;

[0029] Step 2: After re-referencing, filtering, and removing artifacts from the collected gamma-ASSR data, extract ASSR signals of different gamma frequencies from the artifact-free data and intercept multiple data segments;

[0030] Step 3: extract the resonance frequency of individual gamma oscillations as the stimulation frequency of otDCS using spectrum analysis, extract the instantaneous phase of individual gamma oscillations as the stimulation phase of otDCS using phase estimation, and extract the key nodes in the individual gamma oscillation brain network as the stimulation sites of otDCS using network analysis methods; and

[0031] Step 4: Based on the resonant frequency, instantaneous phase and key nodes of the individual gamma oscillation obtained in step 3, lock the stimulation frequency, stimulation phase and stimulation site, and generate a personalized stimulation parameter otDCS waveform for subsequent individual electrical stimulation output.

[0032] Furthermore, step four specifically includes:

[0033] S401: According to the individualized gamma resonance frequency F obtained in step 3 IGR , which is determined as the stimulation frequency of the stimulation current. Based on the relationship between frequency and angular velocity, ω=2πf, it is converted into a sinusoidal waveform expression, and we can get:

[0034] s(t)=A×sin(2πF IGR·t)+DC

[0035] Wherein, DC is the current intensity of the preset DC component of the stimulation current, and A is the current intensity of the preset AC component superimposed on the DC component.

[0036] S402: Phase value estimated according to step 3 The initial output phase of the stimulation current is selected according to the value selection rules of the up and down curves. The waveform formula is as follows:

[0037]

[0038] in, Represents the initial phase of the AC component, based on the phase value estimated in step 3 Sure.

[0039] S403: Select the betweenness centrality N from the betweenness centralities obtained in step 3 BC (k) The top two key nodes were used as stimulation sites for electrical stimulation.

[0040] S404: According to the stimulation frequency, stimulation phase and stimulation site obtained in steps S401-403, an otDCS waveform with personalized stimulation parameter configuration is obtained for subsequent individual electrical stimulation output.

[0041] A device includes a processor and a memory, wherein the memory stores program instructions, and when the instructions are executed by the processor, the instructions cause the processor to execute the steps of the space-frequency domain adaptive otDCS intervention method for modulating gamma oscillations.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] 1. The present invention designs a frequency-stepped gamma-band auditory steady-state response paradigm and records the EEG signals under the individual gamma-ASSR induced paradigm, thereby providing a key paradigm and data support for obtaining electrical stimulation parameters such as personalized stimulation frequency, stimulation phase, and stimulation site for modulating gamma oscillations;

[0044] 2. Utilize spectrum analysis, phase estimation and network analysis methods to extract the resonance frequency of individual gamma oscillations as the stimulation frequency of otDCS, extract the instantaneous phase of individual gamma oscillations as the stimulation phase of otDCS, and extract the key nodes in the individual gamma oscillation brain network as the stimulation sites of otDCS. Combine the rapid feedback and closed-loop control of EEG-otDCS technology to achieve adaptive parameter configuration, thereby achieving personalized and precise regulation of individual EEG gamma oscillations, effectively improving the clinical symptoms of MDD patients, and assisting in precise clinical policy implementation. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 is a flow chart of the space-frequency domain adaptive otDCS intervention method for modulating gamma oscillations according to the present invention;

[0046] Figure 2 It is a frequency-stepping gamma-ASSR paradigm flow chart designed in step 1 of the adaptive otDCS intervention method of the present invention;

[0047] Figure 3 The F in step 3 S301 of the adaptive otDCS intervention method of the present invention IGR Extract sample graph;

[0048] Figure 4 It is an example diagram of instantaneous phase extraction in step 3 S302 of the adaptive otDCS intervention method of the present invention;

[0049] Figure 5 ; This is an example diagram of key node extraction in step 3 S303 of the adaptive otDCS intervention method of the present invention;

[0050] Figure 6 It is an otDCS waveform diagram of a personalized stimulation parameter configuration mode obtained according to the adaptive otDCS intervention method of the present invention. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution, beneficial effects and significant progress of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention is clearly and completely described below in conjunction with the drawings provided in the examples of the present invention. Obviously, all the described embodiments are only partial embodiments of the present invention, rather than all embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0052] The personalized electrical stimulation signal generation system and the spatial-frequency domain adaptive otDCS intervention method using the generation device to modulate gamma oscillations described in the present invention can achieve personalized and precise regulation of EEG gamma oscillations, thereby alleviating the clinical symptoms of patients with depression, and can assist in clinical precision policy implementation. It is personalized, targeted, and intelligent, and can obtain considerable social and economic benefits. The best implementation plan is to adopt patent transfer, technical cooperation or product development. The working process of this method is further described below in conjunction with the accompanying drawings.

[0053] A personalized electrical stimulation signal generation system, comprising:

[0054] A signal acquisition module for collecting EEG signal data of an individual; in the signal acquisition module, the EEG signal data is subjected to a preset frequency-stepping gamma-ASSR paradigm as an auditory stimulation paradigm, thereby collecting gamma-ASSR data; wherein the gamma frequency band range of the gamma-ASSR paradigm is 30 Hz to 50 Hz;

[0055] The signal preprocessing module re-references, filters, and removes artifacts from the gamma-ASSR data collected by the signal acquisition module, extracts ASSR signals of different gamma frequencies from the data without artifact interference, and intercepts multiple data segments;

[0056] A personalized stimulation parameter extraction module extracts personalized stimulation parameters based on the data segments obtained by the signal preprocessing module, wherein the resonance frequency of the individual gamma oscillation is extracted by spectrum analysis as the stimulation frequency of otDCS, the instantaneous phase of the individual gamma oscillation is extracted by phase estimation as the stimulation phase of otDCS, and the key nodes in the individual gamma oscillation brain network are extracted by network analysis method as the stimulation sites of otDCS; and an OtDCS stimulation module locks the stimulation frequency, stimulation phase and stimulation site based on the resonance frequency, instantaneous phase and key nodes of the individual gamma oscillation obtained by the personalized stimulation parameter extraction module, and generates an otDCS waveform for subsequent individual electrical stimulation output.

[0057] like Figure 1 As shown, a spatial-frequency domain adaptive otDCS intervention method for modulating gamma oscillation using the personalized electrical stimulation signal generation system includes:

[0058] Step 1: Collect EEG signals using the gamma-ASSR evoked paradigm

[0059] S101: Design an auditory stimulation paradigm. The present invention selects a frequency-stepping gamma-ASSR paradigm, wherein the gamma frequency band ranges from 30Hz to 50Hz, with a step length of 1Hz, and multiple gamma frequencies are set, including: 31Hz, 32Hz, ..., 49Hz. At the same time, a chirp signal is selected as a sound stimulus. The chirp signal is designed based on the cochlear traveling wave theorem, which can enhance the synchronous discharge of neurons in the cochlea, improve the ASSR response intensity, and induce a more stable and effective ASSR.

[0060] like Figure 2As shown in the figure, the gamma-ASSR paradigm process is as follows: this part defaults to 3 groups of chirp sound stimulation tasks (the number of groups can be increased according to clinical needs). In each group of stimulation tasks, the frequency of chirp sound presentation is different gamma frequencies (31Hz, 32Hz, ..., 49Hz), and the presentation order of chirp sound stimulation with different gamma frequencies is random. Each sound stimulation is presented 5 times, and each time is a trial. In a single trial, a 2s chirp stimulus and a 1s stimulus interval are presented. The 1s stimulus interval is set to ensure that the information between each trial does not have a cross-influence. Among them, the generation and control of the induced paradigm are implemented by writing codes in the Matlab special toolbox (Psychtoolbox), and synchronous event information is sent to the EEG amplifier to ensure data synchronization.

[0061] S102: Collect EEG signals. MDD patients were asked to wear wired noise-cancelling headphones (model: Edifier W820NB) to provide sound to them, with dual-channel playback and a sound pressure level of 45dB. Relying on the 10-20 international standard lead system, a 62-channel EEG acquisition system and its dedicated acquisition software were used to continuously record 62-lead EEG data under the gamma-ASSR paradigm, with a sampling rate of 1000Hz and the reference electrode being the left mastoid M1. During the stimulation, the user only needed to stay as still as possible and no other operations were required.

[0062] Step 2: Preprocess the collected gamma-ASSR data. Specifically include:

[0063] S201: bilateral mastoid average re-reference (left and right mastoid M1, M2), obtain 60-lead EEG data (excluding the remaining leads of M1 and M2), and bandpass filter to 30-50Hz;

[0064] S202: Use independent components analysis (ICA) to perform artifact detection on the data to remove artifact interference such as electrooculography and electromyography;

[0065] S203: Extract ASSR signals of different gamma frequencies from the data after artifact interference is removed and subdivide them into 2s epoch data segments.

[0066] Step 3: Calculate individual stimulation parameters based on preprocessed data

[0067] S301: Calculate the individualized gamma resonance frequency (F IGR )

[0068] The specific steps are as follows: use fast Fourier transform to calculate the power spectrum curve of the gamma (30-50Hz) frequency band of each epoch data segment in the gamma-ASSR signal under each lead, and perform superposition and averaging at the lead level and epoch level to obtain the individual average gamma frequency ASSR energy value;

[0069] Use the built-in peak finding function findpeaks() and maximum finding function max() in MATLAB to obtain the resonant frequency of the individual gamma oscillation, that is, the gamma frequency corresponding to the maximum response. The schematic diagram of the individualized gamma resonant frequency is as follows: Figure 3 As shown, the individualized gamma resonance frequency F obtained in this example IGR It is 37Hz.

[0070] S302: Construct an autoregressive (AR)-based phase estimation model to calculate the instantaneous phase of individual gamma oscillations

[0071] The specific steps are as follows: based on step S301, extract F IGR - epoch data segment of ASSR, i.e., the individualized gamma resonance frequency F obtained by S301 IGR As the center frequency of narrowband filtering, F IGR The bandwidth was ±2 Hz, and the ASSR data was processed by narrow-band filtering using a Butterworth filter to extract the epoch data segment under the 2-s sound stimulation. A total of at least 5 trials*3 groups*60 leads of F IGR -ASSR data. According to the principle of autoregressive modeling, the EEG signal time series {x n The corresponding relationship between the nth x value and the first m x values ​​in} can be expressed as:

[0072] x n =a1x n-1 +a2x n-2 +…+a m x n-m +ε n (1)

[0073] Among them, a1, a2, …, a m is the autoregressive model coefficient, m is the order of the AR model, the default model order is 10 (determined based on experimental tests and can be adjusted accordingly according to needs), ε n is the residual of the AR model and is a white noise sequence.

[0074] Construct an AR model. Since the least squares method has the advantages of high accuracy, simplicity and unbiased estimation when estimating parameters, the AR model uses the least squares method to estimate the AR model coefficients; extract the F of 2s IGR -ASSR data segment, a total of 5 trials * 3 groups * 60 leads = 900 data segments, and then extract the first 2 / 3 of the data of each 2s data segment for AR modeling to obtain the least squares estimate, that is, the model coefficient. Then predict the remaining 1 / 3 of the data.

[0075] Next, the predicted data is smoothed and four phase feature points are identified, namely, the upper zero crossing point, the peak, the lower zero crossing point, and the trough (such as Figure 4 The phase values ​​are 0, π / 2, π, and 3π / 2 (–π / 2).

[0076] Then, the time difference and phase difference between adjacent phase feature points are used to calculate T i Phase of time The specific calculation formula is:

[0077]

[0078] in, T n+1 The phase of the moment, T n Phase.

[0079] like Figure 4 As shown, assuming T n is the zero crossing point, T n+1 is the peak, then is π / 2.

[0080] The remaining 1 / 3 of the data is used to verify the reliability of the model and optimize the model coefficients. Furthermore, the phase of the next moment of the 2s data is predicted, that is, the phase of each data segment at the next moment is estimated in 900 data segments. Then, the value is taken according to the rules of the upward and downward curves, that is, when the predicted phase value is in the upward curve of "trough (–π / 2)→peak (π / 2)" (around the zero point), the output phase is 0°; when the predicted phase value is in the downward curve of "peak (π / 2)→trough (3π / 2)" (around the zero point), the output phase is 180°.

[0081] Finally, the voting decision method was used to extract the instantaneous phase of the individual gamma oscillation at the next moment after the sound stimulation ended as the estimated phase value according to the majority vote rule.

[0082] S303: Calculate the key node information of the brain functional network of individual gamma oscillations

[0083] The specific steps are as follows: take the 60-lead electrodes as network nodes, and construct the F-network of MDD patients based on Granger causality analysis. IGR -ASSR brain function network model, extracting the network's betweenness centrality. The betweenness centrality is a key indicator for measuring the importance of nodes in a network. Studies have shown that nodes with high betweenness centrality mean that they play an important "bridge" role in the network and are key locations for the flow and integration of information, resources, etc. The calculation formula for betweenness centrality is:

[0084]

[0085] Among them, σ ij is the number of all shortest paths from node i to node j, and σ ij (k) is the number of paths passing through node k among these shortest paths. The two electrodes with the highest betweenness centrality are extracted as the key nodes of the brain network of individual gamma oscillations.

[0086] Step 4: Output otDCS waveform according to the personalized stimulation parameters obtained in step 3

[0087] S401: Stimulation frequency: According to the individualized gamma resonance frequency F of different individuals obtained in S301 IGR , determine the stimulation frequency of the stimulation current. Based on the relationship between frequency and angular velocity, ω=2πf, taking a sinusoidal waveform as an example, we can get:

[0088] s(t)=A×sin(2πF IGR ·t)+DC (4)

[0089] Where A is the current intensity of the AC component of the stimulation current, and DC is the current intensity of the DC component of the stimulation current. According to the preset framework commonly used in existing studies ( K, J, D,et al.Theta-modulated oscil-latory transcranial direct current stimulation over posteriorparietal cortex improves associative memory[J].Scientific Reports,2021,11(1):3013-3020.), set A to a constant value of 1mA (peak-to-peak value of 2mA), and DC to a constant value of 1.5mA, which can also be adjusted accordingly according to needs.

[0090] S402: Stimulation phase: The phase value of each data segment at the next moment estimated in S302 The initial phase of the stimulation current is selected, mainly including two types of stimulation phases of 0° and 180°. The waveform formula is as follows:

[0091]

[0092] in, Indicates the initial phase of the AC component, which is set as follows:

[0093]

[0094] Finally, the voting decision method is used to obtain the initial phase value of the stimulation current according to the majority rule.

[0095] S403: Stimulation site: Select the betweenness centrality N from the betweenness centrality obtained in S303 BC (k) The top two key nodes serve as specific targets for modulating individual gamma oscillations ( Figure 5 The two darker dots in the middle are the stimulation sites for electrical stimulation.

[0096] S404: According to the individualized gamma resonance frequency, instantaneous phase and key nodes obtained in steps S401-403, the stimulation frequency is locked to F IGR =37Hz, stimulation phase is The stimulation sites were F4-P4, so as to obtain Figure 6 The otDCS waveform of the personalized stimulation parameter configuration mode shown is used to obtain the otDCS waveform corresponding to the MDD treatment object. Moreover, the otDCS waveform can enable physicians or caregivers to visualize its stimulation frequency, phase, and site information through computer devices such as displays, and incorporate it into clinical decision-making, which is beneficial for the treatment object to obtain its personalized stimulation treatment.

[0097] The intervention method described in the present invention proposes a new otDCS adaptive intervention method with precise configuration of personalized parameters around the three key dimensions of frequency, phase and site. In view of the large individual differences in brain gamma oscillations in MDD patients and the lack of targeting of otDCS regulation technology, an innovative targeted closed-loop otDCS adaptive intervention method with precise matching of personalized stimulation parameters is proposed to solve the existing problems of large heterogeneity of MDD intervention responses and insufficient regulation targeting, which helps to achieve efficient and precise intervention of EEG gamma oscillations and provide new methods and new approaches for improving cognitive function and alleviating depressive symptoms.

Claims

1. A personalized electrical stimulation signal generation system, comprising: Signal acquisition module, collecting individual EEG signal data; In the signal acquisition module, the EEG signal data is collected by using a preset frequency-stepping gamma-ASSR paradigm as an auditory stimulation paradigm, thereby acquiring gamma-ASSR data; wherein the gamma frequency band range of the gamma-ASSR paradigm is 30 Hz to 50 Hz; The signal preprocessing module re-references, filters, and removes artifacts from the gamma-ASSR data collected by the signal acquisition module, extracts ASSR signals of different gamma frequencies from the data without artifact interference, and intercepts multiple data segments; A personalized stimulation parameter extraction module extracts personalized stimulation parameters based on the data segments obtained by the signal preprocessing module, wherein the resonance frequency of the individual gamma oscillation is extracted by spectrum analysis as the stimulation frequency of otDCS, the instantaneous phase of the individual gamma oscillation is extracted by phase estimation as the stimulation phase of otDCS, and the key nodes in the individual gamma oscillation brain network are extracted by network analysis method as the stimulation sites of otDCS; and an otDCS stimulation module locks the stimulation frequency, stimulation phase and stimulation site based on the resonance frequency, instantaneous phase and key nodes of the individual gamma oscillation obtained by the personalized stimulation parameter extraction module, and generates an otDCS waveform for subsequent individual electrical stimulation output.

2. The personalized electrical stimulation signal generation system according to claim 1, characterized in that: The signal preprocessing module further comprises: The obtained EEG data were re-referenced, and the average value of the two leads of the bilateral mastoid M1 and M2 was selected as the reference; The EEG data after re-reference was filtered, and the band-pass filter was selected as 30-50 Hz; Independent component analysis was used to detect artifacts in the data to remove artifacts such as electrooculography and electromyography. The ASSR signals of different gamma frequencies are extracted from the data after artifact interference is removed and subdivided into epoch data segments of preset time periods.

3. The personalized electrical stimulation signal generation system according to claim 1, characterized in that: Extracting the resonant frequency of an individual gamma oscillation involves: Fast Fourier transform is used to calculate the power spectrum curve of the gamma frequency band of each epoch data segment in the gamma-ASSR signal under each lead, and the average ASSR energy value of the gamma frequency of the individual is obtained by superposition and averaging at the lead level and the epoch data segment level; Use the built-in peak finding function findpeaks() and maximum finding function max() in MATLAB to obtain the resonant frequency F of the individual gamma oscillation IGR .

4. The personalized electrical stimulation signal generation system according to claim 3, characterized in that: Extracting the instantaneous phase of an individual gamma oscillation involves: Construct a phase estimation model based on autoregression and select F IGR -ASSR data were modeled, and the least squares method was used to estimate the AR model coefficients; the voting decision method was used to extract the instantaneous phase of the individual gamma oscillation in the phase estimated by the data; wherein the F IGR -ASSR data refers to the extracted individual gamma resonance frequency F IGR is the center frequency, with F IGR ±2 Hz is the bandwidth, and the data is obtained after the ASSR data is band-pass filtered using a Butterworth filter.

5. The personalized electrical stimulation signal generation system according to claim 4, characterized in that: The key nodes of the brain network for extracting individual gamma oscillations include: Constructing F of MDD patients based on Granger causality analysis IGR -ASSR brain functional network model, extract the betweenness centrality of individual brain functional network, and select the two electrodes with the highest betweenness centrality as the key nodes of the brain network.

6. The personalized electrical stimulation signal generation system according to claim 1, characterized in that: The individual is an MDD patient, and the obtained individualized gamma resonance frequency is determined as the individual's treatment stimulation frequency; the obtained individualized gamma oscillation instantaneous phase is determined as the individual's MDD treatment stimulation phase, and the obtained key nodes in the individualized gamma oscillation brain network are determined as the individual's treatment targets.

7. The personalized electrical stimulation signal generation system according to claim 1, characterized in that: The frequency-stepped gamma-ASSR paradigm uses 1 Hz as a step length, sets multiple gamma frequencies within the frequency band, and selects a chirp signal as a sound stimulus.

8. A spatial-frequency domain adaptive otDCS intervention method for modulating gamma oscillations using the personalized electrical stimulation signal generation system as claimed in claim 1, comprising: Step 1: Collect individual EEG signal data according to the preset frequency-stepping gamma-ASSR paradigm to obtain gamma-ASSR data; Step 2: After re-referencing, filtering, and removing artifacts from the collected gamma-ASSR data, extract ASSR signals of different gamma frequencies from the artifact-free data and intercept multiple data segments; Step 3: Use spectrum analysis to extract the resonance frequency of individual gamma oscillations as the stimulation frequency of otDCS, use phase estimation to extract the instantaneous phase of individual gamma oscillations as the stimulation phase of otDCS, and use network analysis methods to extract key nodes in the individual gamma oscillation brain network as the stimulation sites of otDCS; as well as Step 4: Based on the resonant frequency, instantaneous phase and key nodes of the individual gamma oscillation obtained in step 3, lock the stimulation frequency, stimulation phase and stimulation site, and generate a personalized stimulation parameter otDCS waveform for subsequent individual electrical stimulation output.

9. The space-frequency domain adaptive otDCS intervention method for modulated gamma oscillation according to claim 8, characterized in that: Step 4 specifically includes: S401: According to the individualized gamma resonance frequency F obtained in step 3 IGR , determine the stimulation frequency of the stimulation current, based on the relationship between frequency and angular velocity ω=2πf, convert it into a sinusoidal waveform expression, and we can get: s(t)=A×sin(2πF IGR ·t)+DC Wherein, DC is the current intensity of the DC component of the preset stimulation current, and A is the current intensity of the AC component superimposed on the DC component; S402: Phase value estimated according to step 3 The initial output phase of the stimulation current is selected according to the value selection rules of the up and down curves. The waveform formula is as follows: in, Represents the initial phase of the AC component, based on the phase value estimated in step 3 Sure; S403: Select the betweenness centrality N from the betweenness centralities obtained in step 3 BC (k) The top two key nodes were used as stimulation sites for electrical stimulation; S404: According to the resonance frequency, instantaneous phase and key nodes of the individual gamma oscillation obtained in steps S401-403, the stimulation frequency, stimulation phase and stimulation site are locked to obtain the otDCS waveform with personalized stimulation parameter configuration for subsequent individual electrical stimulation output.

10. A device comprising a processor and a memory, wherein the memory stores program instructions, and when the instructions are executed by the processor, the instructions cause the processor to perform the steps of the space-frequency domain adaptive otDCS intervention method for modulating gamma oscillations as described in any one of claims 8 to 9.

Images