CN-122006123-A - Transcranial magnetic stimulation system and method controlled by electroencephalogram signals

CN122006123ACN 122006123 ACN122006123 ACN 122006123ACN-122006123-A

Abstract

The invention relates to the technical field of transcranial magnetic stimulation, in particular to a transcranial magnetic stimulation system and method controlled by brain electrical signals. The system comprises an electroencephalogram acquisition module, a self-adaptive signal processing module, a dynamic stimulation control module and a reconfigurable transcranial magnetic stimulation and closed-loop feedback adjustment module. The method comprises the steps of collecting and preprocessing brain electrical signals of a subject, reducing noise by means of an improved U-Net model fused with physiological priori, extracting multidimensional brain electrical characteristics by means of wavelet packet decomposition and Hilbert-Huang transform, dynamically optimizing stimulation parameters based on a particle swarm optimization algorithm, achieving magnetic field focusing by means of a flexible coil array, outputting synchronous stimulation pulses by means of a phase locking mechanism, and enabling closed loop feedback adjustment to ensure that the stimulation effect reaches the standard. The invention solves the problems of low brain electrolysis precision, poor parameter adaptability, insufficient focusing and low synchronization precision of the traditional system, improves the nerve regulation effect, reduces the damage risk of non-target brain areas, and is suitable for personalized treatment of nerve diseases such as depression, cerebral apoplexy and the like.

Inventors

ZHANG YAO
LIANG YU
WANG WEI
GUO JIANYING
LI BAOJUAN
LIU JIAN
HAO WANTING
HUANG PENG
CAI MIN
WU JUAN
MU NAN

Assignees

中国人民解放军空军军医大学

Dates

Publication Date: 20260512
Application Date: 20260129

Claims (10)

1. A transcranial magnetic stimulation system controlled by brain electrical signals, comprising: the brain electricity acquisition module is used for acquiring scalp brain electricity signals of a subject and converting the scalp brain electricity signals into digital signals; The self-adaptive signal processing module is in communication connection with the electroencephalogram acquisition module, performs self-adaptive noise reduction processing on an electroencephalogram digital signal by adopting a deep learning model fused with an electroencephalogram physiological priori, and extracts multidimensional electroencephalogram characteristic parameters by combining wavelet packet decomposition with Hilbert-Huang transform, wherein the electroencephalogram characteristic parameters comprise rhythmic frequency, power spectral density, instantaneous phase and brain region channel coherence; The dynamic stimulation control module is in communication connection with the self-adaptive signal processing module, a brain region-parameter mapping database is built in, a stimulation parameter optimization objective function is constructed based on the electroencephalogram characteristic parameters, and optimal stimulation parameters are obtained through solving a particle swarm optimization algorithm, wherein the optimal stimulation parameters comprise stimulation frequency, intensity, pulse width and trigger phase; The reconfigurable transcranial magnetic stimulation module is electrically connected with the dynamic stimulation control module and comprises a flexible coil array and an array driving circuit, and the dynamic stimulation control module realizes magnetic field focusing by adjusting the current magnitude and direction of each unit in the coil array; The closed-loop feedback regulation module is respectively in communication connection with the electroencephalogram acquisition module and the dynamic stimulation control module, acquires stimulated electroencephalogram signals and transmits the stimulated electroencephalogram signals to the dynamic stimulation control module to form a closed-loop regulation loop; The dynamic stimulation control module adopts a phase locking mechanism, extracts the instantaneous phase of the brain characteristic wave through Hilbert transformation, and triggers the stimulation pulse output when the instantaneous phase is less than 5 degrees from a preset trigger phase difference.
2. The transcranial magnetic stimulation system controlled by electroencephalogram signals according to claim 1, wherein the deep learning model fused with the electroencephalogram physiological priors is an improved U-Net architecture, an electroencephalogram rhythm frequency band constraint module is embedded in an encoder end, an attention mechanism layer is arranged on a decoder end, and the attention mechanism layer focuses on a theta wave of 4-8Hz, an alpha wave of 8-13Hz and a beta wave characteristic frequency band of 13-30 Hz.
3. The transcranial magnetic stimulation system controlled by brain electrical signals according to claim 1, wherein the brain electrical acquisition module adopts 64-channel silver/silver chloride electrodes, the sampling rate is 250-1000Hz, and a reference electrode are arranged in the brain electrical acquisition module, so that the brain electrical signal control transcranial magnetic stimulation system has a power frequency interference suppression function.
4. The transcranial magnetic stimulation system controlled by brain electrical signals according to claim 1, wherein the stimulation parameter optimization objective function targets "target brain region characteristic wave adjustment amplitude maximization" under the constraint that the stimulation intensity is not more than 2.5T and the stimulation frequency is not more than 50Hz.
5. The transcranial magnetic stimulation system controlled by electroencephalogram signals according to claim 1, wherein the flexible coil array is of an 8X 8 array structure, each coil unit is 5mm in diameter, litz wire winding is adopted, and an array driving circuit is a DRV8833 type current driving chip and supports 0-5A current regulation.
6. A transcranial magnetic stimulation method controlled by an electroencephalogram, the method being implemented using the transcranial magnetic stimulation system controlled by an electroencephalogram according to any one of claims 1 to 5, comprising the steps of: s1, acquiring an electroencephalogram signal, namely acquiring an original electroencephalogram signal of a target brain region of a subject through an electroencephalogram acquisition module, and preprocessing to obtain a digital electroencephalogram signal; S2, self-adaptive noise reduction and feature extraction, namely inputting the digital brain electrical signals into a self-adaptive signal processing module, carrying out noise reduction by improving a U-Net model, adopting wavelet packet decomposition to obtain 3-layer decomposition coefficients, and extracting brain electrical feature parameters by combining Hilbert-Huang transform; S3, dynamically optimizing the stimulation parameters, calling a brain region-parameter mapping database by a dynamic stimulation control module, constructing an optimization objective function based on the brain electrical characteristic parameters, and obtaining optimal stimulation parameters through a particle swarm optimization algorithm; S4, synchronous focusing stimulation, wherein a dynamic stimulation control module generates a control signal according to the optimal stimulation parameter, adjusts current distribution of the reconfigurable coil array, and outputs synchronous stimulation pulses based on a phase locking mechanism; And S5, closed-loop feedback adjustment, wherein the closed-loop feedback adjustment module is used for collecting the stimulated electroencephalogram signals, comparing the stimulated electroencephalogram signals with the pre-stimulated electroencephalogram characteristic parameters, and if the characteristic wave adjustment amplitude does not reach a preset threshold value, returning to the step S3 to re-optimize the stimulation parameters until the threshold value requirement is met.
7. The transcranial magnetic stimulation method according to claim 6, wherein the preprocessing in step S1 comprises baseline correction, electro-oculography artifact removal and 50Hz power frequency notch processing.
8. The method of brain magnetic stimulation controlled by brain electrical signals according to claim 6, characterized in that the particle swarm optimization algorithm in step S3 sets the population size to 30, the iteration number to 50, and the learning factor c1=1.5, c2=1.7.
9. The transcranial magnetic stimulation method according to claim 6, wherein the preset trigger phase in step S4 is a theta wave rising edge with a phase of 0 ° or an alpha wave peak with a phase of 90 °.
10. The transcranial magnetic stimulation method controlled by electroencephalogram signals according to claim 6, wherein the preset threshold in the step S5 is characterized in that the variation of the power spectral density of the characteristic wave is more than or equal to 20%.

Description

Transcranial magnetic stimulation system and method controlled by electroencephalogram signals Technical Field The invention relates to the technical field of transcranial magnetic stimulation, in particular to a transcranial magnetic stimulation system and method controlled by brain electrical signals. Background Transcranial magnetic stimulation is a noninvasive nerve regulation and control technology based on an electromagnetic induction principle, and a time-varying magnetic field generated by a stimulation coil penetrates through the skull to induce induced current in brain tissues so as to regulate the excitability of neurons and realize regulation and control on nerve functions. The technology has the advantages of noninvasive, safe, repeatable and the like, and is widely applied to diagnosis and treatment of various neurological diseases such as depression, schizophrenia, dyskinesia after cerebral apoplexy, mild cognitive impairment and the like, and becomes an important means for research hot spot and clinical treatment in the field of neurosis. Along with the continuous improvement of clinical demands, the limitations of the traditional transcranial magnetic stimulation system are gradually highlighted, and the traditional transcranial magnetic stimulation system mainly comprises the following aspects of insufficient electroencephalogram signal analysis precision and poor interference suppression effect. The traditional filtering method (such as FIR filtering and IIR filtering) is adopted for noise reduction treatment in the electroencephalogram signal acquisition of the traditional TMS system, and the method has limited inhibiting effect on non-stationary noise (such as myoelectric artifacts, ocular electric artifacts and power frequency interference superimposed noise) and is easy to lose weak characteristic components in the electroencephalogram signal. Meanwhile, the feature extraction mostly adopts a single dimension index (for example, only extracts the power value of a certain frequency band), the dynamic change rule of the brain electrical activity cannot be comprehensively reflected, the suitability of the brain electrical signal and the stimulation parameter is poor, and the accurate regulation and control are difficult to realize. For example, in the treatment of depression, the abnormality of the theta wave (4-8 Hz) of a patient is closely related to the emotion regulating function, but the traditional method is difficult to accurately extract the instantaneous phase and power spectrum change of the theta wave, so that the stimulation parameters cannot be pertinently matched with the abnormal characteristics of the theta wave, and the treatment effect is limited. Secondly, stimulus parameter regulation lacks dynamics and personalization. The existing closed-loop TMS system mostly adopts a preset parameter template, can only carry out sectional parameter adjustment according to a fixed electroencephalogram index (such as the amplitude of a single frequency band), and cannot dynamically adapt to individual differences (such as the sulcus structure, the skull thickness and the nerve excitability) of different subjects and the dynamic change of electroencephalogram signals in the treatment process in real time. In addition, existing systems lack parameter mapping mechanisms for different brain region functions, and the same stimulation parameters are applied to different brain regions (such as the dorsum of the forehead, the motor cortex), which easily causes poor stimulation effects or causes untoward reactions (such as headache and muscle twitches) of non-target brain regions. For example, for the motor cortex stimulation of a patient with post-stroke motor dysfunction, the characteristics of a motor related mu rhythm (8-13 Hz) need to be matched, but the existing system cannot optimize the stimulation frequency and intensity in real time according to the dynamic change of the mu rhythm, so that the rehabilitation effect is affected. Thirdly, the stimulation focusing performance is poor, and the risk of damage to non-target brain areas is high. Traditional TMS system adopts single coil or fixed array coil, and magnetic field distribution scope is great (focus area diameter is generally not less than 10 mm), is difficult to accurately focus on target brain district (like forehead She Yaou of small volume), and the healthy brain district in every side is easily stimulated, causes unnecessary neural response. Meanwhile, the rigid structure of the coil cannot be adapted to the skull contours of different subjects, so that gaps exist between the coil and the scalp, and the magnetic field penetration efficiency and the focusing precision are further reduced. For example, in the treatment of cognitive disorders, the target brain region is the dorsum-lateral (smaller volume) of the forehead, and the magnetic field of the conventional coil is easily diffused to the orbital cortex, resulting in si