Search

CN-121997299-A - Multi-mode signal processing method and system based on functional ultrasound and electroencephalogram

CN121997299ACN 121997299 ACN121997299 ACN 121997299ACN-121997299-A

Abstract

The invention discloses a multimode signal processing method and system based on functional ultrasound and electroencephalogram, relates to the technical field of signal processing, solves the technical problems that the synchronous monitoring of ultrasound signals and electroencephalogram signals cannot be realized and the single-mode information of the existing brain function imaging technology is incomplete, the technical scheme is characterized in that the acquired ultrasonic signals and the electroencephalogram signals are synchronously processed and jointly analyzed, signal alignment is achieved through unified reference, hemodynamic characteristics and nerve electric activity characteristics are extracted, a coupling model is built, brain function task decoding and brain region activation assessment are achieved, the integrity and accuracy of brain function assessment are improved, and the brain blood flow imaging stability is improved.

Inventors

  • MA YINJIE
  • YIN FENG
  • YANG HANG

Assignees

  • 东南大学

Dates

Publication Date
20260508
Application Date
20260409

Claims (10)

  1. 1. A multimode signal processing method based on functional ultrasound and electroencephalogram is characterized by comprising the following steps: Acquiring an ultrasonic signal data stream and an electroencephalogram signal data stream; processing the ultrasonic signal data stream to obtain hemodynamic characteristics; processing the electroencephalogram signal data stream to obtain nerve electric activity characteristics; The method comprises the steps of carrying out time alignment and synchronization processing on nerve electric activity characteristics and blood flow dynamics characteristics based on a unified time reference, constructing a coupling model between the nerve electric activity characteristics and the blood flow dynamics characteristics after time alignment is completed, estimating the association strength and time-lapse relation between nerve electric activity changes and blood flow response changes through correlation analysis or regression analysis or model fitting mode by the coupling model to obtain nerve-blood flow coupling characteristics of a target brain region, and obtaining brain function joint analysis results reflecting the functional state, task related activation mode and nerve regulation characteristics of the target brain region based on the nerve-blood flow coupling characteristics.
  2. 2. The method for processing a multi-modal signal according to claim 1, wherein the processing the electroencephalogram data stream to obtain the neuroelectric activity characteristics includes: performing reference reconstruction preprocessing on the electroencephalogram signal data stream to obtain a first electroencephalogram signal, wherein the first electroencephalogram signal is expressed as: ; Wherein, the Representing the first brain electrical signal after the re-reference, Represent the first An electroencephalogram signal data stream of each electroencephalogram channel, Represent the first An electroencephalogram signal data stream of each electroencephalogram channel, Representing the number of effective electroencephalogram channels; Identifying and inhibiting non-brain-derived artifact components in the first electroencephalogram signals to obtain second electroencephalogram signals, wherein the second electroencephalogram signals are expressed as follows: ; Wherein, the Represents the second electroencephalogram signal after inhibition, Representing an artifact component estimated by an artifact modeling or separation method; performing time domain analysis or frequency domain analysis or time-frequency domain analysis on the second electroencephalogram signal to obtain nerve electric activity characteristics, wherein the power spectrum characteristics are obtained through frequency domain analysis and are expressed as follows: ; Wherein, the Representing the first electroencephalogram signal in the second electroencephalogram signal The signal of each brain electric channel is in a preset frequency band Internal power spectral features, i.e., neuroelectrical activity features; Is that Is a representation of the frequency domain of (c), Representing the first electroencephalogram signal in the second electroencephalogram signal And (5) an electroencephalogram channel signal.
  3. 3. The method of multi-modal signal processing as set forth in claim 2 wherein said processing the ultrasound signal data stream to obtain hemodynamic characteristics includes: Obtaining a corresponding channel-level slow time domain complex signal according to the ultrasonic signal data stream, and performing low-rank-sparse decomposition processing on the channel-level slow time domain complex signal to separate a low-rank component representing a tissue background from a sparse component representing a blood flow signal; When brain function monitoring is carried out, beam synthesis is carried out based on the low-rank component to obtain a tissue structure imaging result, and meanwhile, beam synthesis is carried out based on the sparse component to obtain a preliminary blood flow imaging result representing the blood flow distribution of the target brain microvasculature; and carrying out baseline normalization processing on the functional ultrasonic imaging result to obtain hemodynamic characteristics.
  4. 4. A multi-modal signal processing method as claimed in claim 3 wherein the channel-level slow time domain complex signal is represented as: ; Wherein, the , The representation of the hilbert transform is given, Representing the radio frequency signal corresponding to the ultrasound signal data stream, Indicating the number of the received array element, ; Representing the discrete-time sampling points of the sample, Indicating that the slow time frame index is to be used, ; The low rank-sparse decomposition process is expressed as: ; Wherein, the Representing a low-rank component of the signal, Representing sparse components; the low rank-sparse decomposition process is solved by an optimization problem expressed as: ; Wherein, the The number of kernels is represented by a kernel norm, Representation of The norm of the sample is calculated, Regularized weight coefficients; the tissue structure imaging results are expressed as: ; Wherein, the Representing low rank components Fourier transform from the space-time domain to the frequency-wavenumber domain, Representing the f-k migration operator, Representing an inverse fourier transform from the frequency-wavenumber domain to the spatial domain; The preliminary blood flow imaging results are expressed as: ; Wherein, the Representing the results of the preliminary blood flow imaging, Represents an angular coherence factor, an , A stability term to prevent denominator from being zero; Representing blood flow imaging results at different incidence angles; The compensated blood flow imaging results are expressed as: ; Wherein, the Representing the result of blood flow imaging after compensation, Representing two adjacent frames of structural images And Is a displacement field of (2); the functional ultrasound imaging results are expressed as: ; Wherein, the Representing the result of the functional ultrasound imaging, Representing the result of blood flow imaging after compensation.
  5. 5. The multi-modal signal processing method as claimed in claim 4 wherein the hemodynamic characteristics are expressed as: ; Wherein, the Indicating a baseline blood flow signal corresponding to the resting stage, Representing the mean blood flow signal in the brain region of interest, , Representing brain regions of interest The number of pixels in the pixel array is, Expressed in time index Any pixel position in target brain region Is a power doppler signal of (a); Representing hemodynamic characteristics.
  6. 6. The method of multi-modal signal processing as in claim 5 wherein said processing of the ultrasound signal data stream further comprises: And when the space alignment of the ultrasonic imaging area and the target brain area is carried out, carrying out beam synthesis based on the low-rank component so as to generate a tissue structure imaging result used for registering with a B-mode ultrasonic image or a magnetic resonance image, and carrying out space registration according to the tissue structure imaging result.
  7. 7. The multi-modal signal processing method as claimed in claim 6 wherein the coupling model is expressed as: ; Wherein, the Representing a time series of hemodynamic characteristics; representing a time series of characteristics of neural electrical activity; representing the neuro-blood flow response function, The residual term is represented as such, Representing a convolution operation; the correlation index of the correlation analysis or the regression analysis is expressed as: ; Wherein, the The covariance is represented by the sign of the covariance, Representing the standard deviation of the characteristics of the neural electrical activity, Represents the standard deviation of the hemodynamic characteristics of the brain.
  8. 8. A multi-modal signal processing system based on functional ultrasound and electroencephalogram for implementing the multi-modal signal processing method of any one of claims 1-7, characterized in that the multi-modal signal processing system comprises: The palm ultrasonic equipment is used for transmitting ultrasonic waves to a target brain region and receiving echo signals, sampling and digitizing the echo signals to generate ultrasonic signal data streams which are arranged in time sequence; the electroencephalogram equipment is used for collecting multichannel electroencephalogram signals of a target brain region, and performing time marking and data encapsulation on the collected electroencephalogram signals to generate an electroencephalogram signal data stream with an additional time stamp; the ultrasonic processing unit is used for processing the ultrasonic signal data stream to obtain hemodynamic characteristics; the electroencephalogram processing unit is used for processing the electroencephalogram signal data stream to obtain nerve electric activity characteristics; The system comprises a joint analysis module, a coupling model, a neural-blood flow coupling model and a brain function joint analysis result, wherein the joint analysis module is used for carrying out time alignment and synchronous processing on the neural electric activity characteristic and the blood flow dynamic characteristic based on a unified time reference, constructing the coupling model between the neural electric activity characteristic and the blood flow dynamic characteristic after the time alignment is completed, estimating the association strength and time-delay relation between the neural electric activity change and the blood flow response change through a correlation analysis or regression analysis or model fitting mode to obtain the neural-blood flow coupling characteristic of a target brain region, and obtaining the brain function joint analysis result reflecting the functional state, the task correlation activation mode and the nerve regulation characteristic of the target brain region based on the neural-blood flow coupling characteristic.
  9. 9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of the multi-modal signal processing method according to any one of claims 1 to 7 when the computer program is executed by the processor.
  10. 10. A computer storage medium storing a computer program, characterized in that the computer program when executed by a processor implements the steps of the multi-modal signal processing method according to any one of claims 1 to 7.

Description

Multi-mode signal processing method and system based on functional ultrasound and electroencephalogram Technical Field The application relates to the technical field of signal processing, in particular to a multi-mode signal processing method and system based on functional ultrasound and electroencephalogram. Background Brain function imaging is an important research tool for studying brain neural activity. Common brain functional imaging techniques include functional magnetic resonance imaging (fMRI), functional near infrared imaging (fNIRS), electroencephalogram (EEG), and the like. The fMRI has higher spatial resolution, but the equipment is huge and high in cost, bedside or mobile monitoring is difficult to realize, fNIRS equipment is portable, but the signal penetration depth is limited, only the shallow layer activity of the cortex can be detected, and EEG has high time resolution, but the space positioning capability is insufficient, and the brain function activity is difficult to comprehensively reflect. Functional ultrasound imaging (fUS, functional Ultrasound) is an emerging method of imaging neurological functions, indirectly reflecting neural activity by ultra-fast ultrasound detection of blood flow or blood volume changes in the brain. Compared with other brain function imaging modes, fUS has higher time resolution and spatial resolution, and simultaneously has the advantages of large imaging depth, high blood flow detection sensitivity and the like. However, because of the strong attenuation and distortion of the human skull to the ultrasonic signal, transcranial fUS signal quality is limited, and is currently used for animal experiments or research under craniotomy conditions. The palm ultrasonic equipment has the characteristics of small volume, low cost, portability and the like, and has been widely applied to bedside examination and monitoring in operation. If the palm ultrasonic equipment is combined with functional ultrasonic imaging, the blood flow dynamics detection of the local brain region can be realized in the bone window region of the patient after craniotomy, and a new approach is provided for clinical brain function assessment. How to use the palm ultrasonic equipment for functional ultrasonic imaging and combine with the electroencephalogram equipment to realize synchronous monitoring of electroencephalogram signals and ultrasonic signals is to be solved. Disclosure of Invention The application provides a multi-mode signal processing method and system based on functional ultrasound and electroencephalogram, which aims to reduce the volume of a brain functional imaging system by introducing palm ultrasound equipment, and perform cross-mode joint analysis on nerve electrical activity characteristics and blood flow dynamics characteristics so as to improve the integrity and accuracy of brain functional imaging and improve the stability of brain blood flow imaging. The technical aim of the application is realized by the following technical scheme: a multimode signal processing method based on functional ultrasound and electroencephalogram comprises the following steps: Acquiring an ultrasonic signal data stream and an electroencephalogram signal data stream; processing the ultrasonic signal data stream to obtain hemodynamic characteristics; processing the electroencephalogram signal data stream to obtain nerve electric activity characteristics; The method comprises the steps of carrying out time alignment and synchronization processing on nerve electric activity characteristics and blood flow dynamics characteristics based on a unified time reference, constructing a coupling model between the nerve electric activity characteristics and the blood flow dynamics characteristics after time alignment is completed, estimating the association strength and time-lapse relation between nerve electric activity changes and blood flow response changes through correlation analysis or regression analysis or model fitting mode by the coupling model to obtain nerve-blood flow coupling characteristics of a target brain region, and obtaining brain function joint analysis results reflecting the functional state, task related activation mode and nerve regulation characteristics of the target brain region based on the nerve-blood flow coupling characteristics. Further, the processing the electroencephalogram signal data stream to obtain the characteristics of the neuroelectric activity includes: performing reference reconstruction preprocessing on the electroencephalogram signal data stream to obtain a first electroencephalogram signal, wherein the first electroencephalogram signal is expressed as: ; Wherein, the Representing the first brain electrical signal after the re-reference,Represent the firstAn electroencephalogram signal data stream of each electroencephalogram channel,Represent the firstAn electroencephalogram signal data stream of each electroencephalogram channel,Representing the number of