CN110393525A - A kind of brain activity detection method based on deep-cycle self-encoding encoder - Google Patents
A kind of brain activity detection method based on deep-cycle self-encoding encoder Download PDFInfo
- Publication number
- CN110393525A CN110393525A CN201910526926.7A CN201910526926A CN110393525A CN 110393525 A CN110393525 A CN 110393525A CN 201910526926 A CN201910526926 A CN 201910526926A CN 110393525 A CN110393525 A CN 110393525A
- Authority
- CN
- China
- Prior art keywords
- brain
- deep
- signal
- cycle
- brain activity
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/05—Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves
- A61B5/055—Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves involving electronic [EMR] or nuclear [NMR] magnetic resonance, e.g. magnetic resonance imaging
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/242—Detecting biomagnetic fields, e.g. magnetic fields produced by bioelectric currents
- A61B5/245—Detecting biomagnetic fields, e.g. magnetic fields produced by bioelectric currents specially adapted for magnetoencephalographic [MEG] signals
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/369—Electroencephalography [EEG]
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/40—Detecting, measuring or recording for evaluating the nervous system
- A61B5/4058—Detecting, measuring or recording for evaluating the nervous system for evaluating the central nervous system
- A61B5/4064—Evaluating the brain
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus for radiation diagnosis, e.g. combined with radiation therapy equipment
- A61B6/02—Devices for diagnosis sequentially in different planes; Stereoscopic radiation diagnosis
- A61B6/03—Computerised tomographs
- A61B6/037—Emission tomography
Abstract
The invention discloses a kind of brain activity detection methods based on deep-cycle self-encoding encoder, comprising: receives brain signal, and pre-processes to brain signal；Using the brain signal at each moment as input matrix, dimension-reduction treatment is carried out using brain signal of the space characteristics extraction unit to input, extracts the space characteristics of brain signal；Memory, output timing behavioral characteristics are updated using space characteristics of the deep-cycle coding unit to each moment；Prediction is updated to timing behavioral characteristics using deep-cycle decoding unit, exports the time-sequential activity mode of prediction；Reconstructed mapped is carried out using time-sequential activity mode of the space characteristics reconstruction unit to prediction, restores brain signal.The brain activity detection method detects the temporal characteristics of brain activity and the spatial model of brain network by timing memory circular treatment, feature extraction and the automatic Reconstruction to brain signal, to provide effective ways without the brain activity detection on mark signal data.
Description
Technical field
The invention belongs to brain brain-computer interface fields, and in particular to a kind of brain activity based on deep-cycle self-encoding encoder
Detection method.
Background technique
The brain organ the most complicated as the mankind, is concerned the research and exploration of its structure and function.Always with
Come, the researcher of Neuscience is dedicated to the research to brain, using different Modeling Method brains to extraneous letter
The structure and movable mold for the brain network that reception, coding, response and the brain inside of breath or stimulus signal mutually cooperate
Formula.
Studies have shown that when brain receives environmental stimuli, the conduction for carrying out signal, generation movement consciousness, cerebral nerve
The electrical activity of metasystem can occur to change accordingly.Brain machine interface system is exactly the change by detecting the electroneurographic signal of brain
Change, the characteristic signal of analysis brain response, to detect the active state of brain.Further, it is stimulated generation in brain
When movable, brain magnetic field, blood oxygen concentration content etc. can all generate corresponding variation, then can by magneticencephalogram, magnetic resonance at
More various, the higher data acquisition means of resolution ratio such as picture, to detect more complicated brain activity.
More and more evidences show that cerebral function is by nervous activity that is varied while carrying out and brain net
Network is realized.Existing multiple hypotheses driving model and data-driven model model brain at present, including general linear
Model (GLM), principal component analysis (PCA), independent component analysis (ICA) and sparse expression method.Meanwhile in recent years, with mind
Continuous development through network and deep learning method, convolutional neural networks (CNN), depth confidence network (DBN), limited Bohr are hereby
The models such as graceful machine (RBM) also present advantage of the deep learning method in brain activity detection.However due to the activity of brain
Be essentially prolonged signal sequence, at present for most of the detection model of brain activity be it is simple according to current time or
The active state of collected brain signal prediction subsequent time in the very short time, to the long-term dependence in brain activity
There are no more effective research methods.On the other hand, under current medical condition, although a large amount of multiple types can be collected
The brain signal of type, but to the mark need of work of signal largely experienced image department doctor or researcher, so, it is right
Utilization without mark signal is particularly important in brain science research.
It is therefore proposed that it is a kind of based on without mark brain signal, and can efficiently use long-time brain signal sequence
Implicit information carry out brain activity detection method, to cerebral function and it is movable research will be very significant.
Summary of the invention
The present invention provides a kind of brain activity detection methods based on deep-cycle self-encoding encoder, by brain signal
Timing memory circular treatment, feature extraction and automatic Reconstruction, detect the temporal characteristics of brain activity and the spatial mode of brain network
Formula, to provide effective ways without the brain activity detection on mark signal data.
For achieving the above object, the present invention the following technical schemes are provided:
A kind of brain activity detection method based on deep-cycle self-encoding encoder, comprising the following steps:
Brain signal is received, and brain signal is pre-processed；
Using the brain signal at each moment as input matrix, using space characteristics extraction unit to the brain signal of input
Dimension-reduction treatment is carried out, the space characteristics of brain signal are extracted；
It is updated memory using space characteristics of the deep-cycle coding unit to each moment, output timing dynamic is special
Sign；
Prediction is updated to timing behavioral characteristics using deep-cycle decoding unit, exports the time-sequential activity mould of prediction
Formula；
Reconstructed mapped is carried out using time-sequential activity mode of the space characteristics reconstruction unit to prediction, restores brain signal.
Above-mentioned brain signal includes local field potentials (Local Field Potential, LFP), electroencephalogram
(Electroencephalogram, EEG), magneticencephalogram (Magnetoencephalography, MEG), positron emission fault
Be imaged (Positron Emission Tomography, PET), magnetic resonance imaging (Magnetic Resonance Imaging,
MRI) etc..
The brain signal of acquisition is influenced by heartbeat, breathing, head movement etc., can generate different degrees of noise and partially
It moves.Carrying out pretreatment to brain signal includes: skull removal, the dynamic correction of head, slice time adjustment, space smoothing, global drift
The processing such as elimination, filtering and noise reduction, normalization.
Space characteristics extraction unit mainly carries out dimension-reduction treatment to the brain signal at each moment of input, and it is special to extract space
Sign, reduces the data volume of subsequent processes.In the space characteristics extraction unit, all kinds of mathematical analysis sides can be used
Trained neural network can also be used such as principal component analysis, independent component analysis in method, such as full Connection Neural Network, convolution
Neural network, depth confidence network, limited Boltzmann machine carry out space characteristics extraction to input brain signal.Further,
In the space characteristics extraction unit, space characteristics extraction is carried out using the full Connection Neural Network through overfitting, to utilize
Lesser calculation amount obtains preferable feature extraction effect.
Deep-cycle coding unit mainly remembers not the time-sequential activity being hidden in each moment brain signal
It is disconnected to update, extract the time-series dynamics feature of brain activity.Preferably, in the deep-cycle coding unit, process is utilized
Trained Recognition with Recurrent Neural Network realizes that the memory to the time-series dynamics feature of brain activity is updated and extracted.
Deep-cycle decoding unit is true by simulating mainly using the time-series dynamics feature for having depth coding unit to extract
Brain activity and response process, the brain activity at each moment is predicted and is updated, the time-sequential activity mould of brain is obtained
Formula.Preferably, it in the deep-cycle decoding unit, is realized using trained Recognition with Recurrent Neural Network to brain activity
The update and prediction of time-sequential activity mode.
Space characteristics reconstruction unit is mainly used for per a period of time in the time-sequential activity mode for predicting deep-cycle decoding unit
The active state at quarter maps back practical brain signal space, obtains complete brain activity signal estimation in all time serieses.
Meanwhile each of space characteristics reconstruction unit element, the spatial distribution of a brain activity brain network is all represented, can be used
In visualization, the spatial character of brain activity brain network is intuitively shown.Preferably, in the space characteristics reconstruction unit
In, using trained full Connection Neural Network obtain brain activity brain network spatial distribution and complete brain activity signal
Prediction.
The device have the advantages that are as follows:
The present invention can use the brain signal data of no mark, and it is living sufficiently to excavate the feature implied in brain clock signal
Dynamic model formula, active state and brain network to brain are detected and are predicted.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to do simply to introduce, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art, can be with root under the premise of not making the creative labor
Other accompanying drawings are obtained according to these attached drawings.
Fig. 1 is that the system structure of the brain activity detection method provided by the invention based on deep-cycle self-encoding encoder is illustrated
Figure；
Fig. 2 is the working principle diagram of the brain activity detection method provided by the invention based on deep-cycle self-encoding encoder；
In fig. 1 and 2, (a) is brain signal, (b) is signal processing unit, (c) is space characteristics extraction unit,
(d) it is deep-cycle coding unit, is (e) the time-series dynamics feature extracted, (f) is deep-cycle decoding unit, (g) be space
Feature reconstruction unit, (h) brain signal to restore.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention more comprehensible, with reference to the accompanying drawings and embodiments to this
Invention is described in further detail.It should be appreciated that the specific embodiments described herein are only used to explain the present invention,
And the scope of protection of the present invention is not limited.
Fig. 1 is the realization system structure of the brain activity detection method provided by the invention based on deep-cycle self-encoding encoder
Schematic diagram, the system successively include signal processing unit, space characteristics extraction unit, deep-cycle coding unit, deep-cycle
Decoding unit and space characteristics reconstruction unit.Fig. 2 is the working principle of the invention figure, is expansion of the Fig. 1 in time dimension.
It can be observed from fig. 2 that deep-cycle coding unit and deep-cycle decoding unit are associated on time dimension, when current
Signal and the before behavioral characteristics co-determination at moment of the behavioral characteristics at quarter by current time.
Assuming that can collect signal from M brain locus, each position collects the signal of T time span altogether.It is right
Original brain signal carries out skull removal, head moves and corrects, be sliced time adjustment, space smoothing, global drift elimination, filter and drop
It makes an uproar, after normalized, signal matrix S can be obtained, the dimension of S is N × T, and wherein the signal vector of t moment is St, T moment
Signal vector be sequentially inputted to space characteristics extraction unit sequentially in time, extract the signal space feature at each moment
Vector At。
Later, spatial signature vectors A extraction obtainedtIt is input to deep-cycle coding unit sequentially in time, together
When, the behavioral characteristics vector F of previous moment deep-cycle coding unit outputt-1It is also input to deep-cycle coding unit.It answers
As explanation, the behavioral characteristics vector F of previous moment deep-cycle coding unit outputt-1It is by the letter at all moment before
Number accumulation codetermine.Each of the time-series dynamics eigenmatrix F being made of all moment time-series dynamics feature vectors
Dimension all represents a kind of time-series dynamics feature.
Similarly, the time-series dynamics feature of extraction is inputted into deep-cycle decoding unit and space characteristics reconstruction unit, it can
Restore brain signal.During restoring the signal of brain, space characteristics reconstruction unit, can be with by the update of constantly iteration
The optimal weights matrix W for restoring brain signal is found, while being exactly that brain is living represented by each of W weight vectors
The spatial distribution of a brain network during dynamic.
In the present embodiment, space characteristics extraction unit uses the full Connection Neural Network comprising 128 neurons, deep
It spends loop coding unit and deep-cycle decoding unit uses the Recognition with Recurrent Neural Network of 2 layers of structure, wherein every layer of circulation nerve net
Network includes 64 long-term short-term memory unit (LSTM) or gating cycle unit (GRU).To timing behavioral characteristics F in this example implementation
Dimension be set as 32, show in implementation process simultaneously to detect 32 behavioral characteristics.Space characteristics reconstruction unit
The full Connection Neural Network comprising 32~128 neurons is used, showing can be to 32~128 brain net in implementation process
Network is detected.
Technical solution of the present invention and beneficial effect is described in detail in the upper specific embodiment, it should be understood that
Be to be not intended to restrict the invention the foregoing is merely presently most preferred embodiment of the invention, it is all in spirit of the invention
Interior done any modification, supplementary, and equivalent replacement etc., should all be included in the protection scope of the present invention.
Claims (6)
1. a kind of brain activity detection method based on deep-cycle self-encoding encoder, comprising the following steps:
Brain signal is received, and brain signal is pre-processed；
Using the brain signal at each moment as input matrix, carried out using brain signal of the space characteristics extraction unit to input
The space characteristics of brain signal are extracted in dimension-reduction treatment；
Memory, output timing behavioral characteristics are updated using space characteristics of the deep-cycle coding unit to each moment；
Prediction is updated to timing behavioral characteristics using deep-cycle decoding unit, exports the time-sequential activity mode of prediction；
Reconstructed mapped is carried out using time-sequential activity mode of the space characteristics reconstruction unit to prediction, restores brain signal.
2. as described in claim 1 based on the brain activity detection method of deep-cycle self-encoding encoder, which is characterized in that big
It includes: that skull removes, head moves correction, slice time adjustment, space smoothing, global elimination of drifting about, filter that brain signal, which carries out pretreatment,
Wave and noise reduction, normalization.
3. as described in claim 1 based on the brain activity detection method of deep-cycle self-encoding encoder, which is characterized in that in institute
It states in space characteristics extraction unit, using principal component analysis, independent component analysis, trained full Connection Neural Network, volume
Product neural network, depth confidence network or limited Boltzmann machine carry out space characteristics extraction to input brain signal.
4. as described in claim 1 based on the brain activity detection method of deep-cycle self-encoding encoder, which is characterized in that in institute
Deep-cycle coding unit is stated, realizes the note to the time-series dynamics feature of brain activity using trained Recognition with Recurrent Neural Network
Recall update and extraction；
In deep-cycle decoding unit, the time-sequential activity mould to brain activity is realized using trained Recognition with Recurrent Neural Network
The update and prediction of formula.
5. as described in claim 1 based on the brain activity detection method of deep-cycle self-encoding encoder, which is characterized in that in institute
It states in space characteristics reconstruction unit, obtains the spatial distribution of brain activity brain network using trained full Connection Neural Network
With complete brain activity signal estimation.
6. the brain activity detection method as claimed in any one of claims 1 to 5 based on deep-cycle self-encoding encoder, feature
It is, the brain signal includes local field potentials, electroencephalogram, magneticencephalogram, positron emission tomography, magnetic resonance imaging.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910526926.7A CN110393525B (en) | 2019-06-18 | 2019-06-18 | Brain activity detection method based on deep cycle self-encoder |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910526926.7A CN110393525B (en) | 2019-06-18 | 2019-06-18 | Brain activity detection method based on deep cycle self-encoder |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110393525A true CN110393525A (en) | 2019-11-01 |
| CN110393525B CN110393525B (en) | 2020-12-15 |
Family
ID=68323260
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910526926.7A Active CN110393525B (en) | 2019-06-18 | 2019-06-18 | Brain activity detection method based on deep cycle self-encoder |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110393525B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110840411A (en) * | 2019-12-06 | 2020-02-28 | 深圳市德力凯医疗设备股份有限公司 | Anesthesia depth measuring method, storage medium and electronic device |
| CN112869754A (en) * | 2021-01-08 | 2021-06-01 | 浙江大学 | Brain-machine fusion neural signal lie detection method |
| CN113408697A (en) * | 2021-05-31 | 2021-09-17 | 杭州电子科技大学 | Electroencephalogram data reconstruction method based on reinforcement learning self-encoder |
| CN115359236A (en) * | 2022-10-19 | 2022-11-18 | 之江实验室 | Work memory task magnetoencephalogram classification system based on machine learning |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101101511A (en) * | 2007-07-13 | 2008-01-09 | 杭州电子科技大学 | Brain electrical signal recovery method in brain-machine interaction |
| CN102327116A (en) * | 2011-09-28 | 2012-01-25 | 杭州电子科技大学 | Network system random resonance restoration method of cortex electroencephalographic signal |
| US20120172743A1 (en) * | 2007-12-27 | 2012-07-05 | Teledyne Licensing, Llc | Coupling human neural response with computer pattern analysis for single-event detection of significant brain responses for task-relevant stimuli |
| CN102697493A (en) * | 2012-05-03 | 2012-10-03 | 北京工业大学 | Method for rapidly and automatically identifying and removing ocular artifacts in electroencephalogram signal |
| CN105816170A (en) * | 2016-05-10 | 2016-08-03 | 广东省医疗器械研究所 | Wearable NIRS-EEG based schizophrenia early detection evaluation system |
| CN107961007A (en) * | 2018-01-05 | 2018-04-27 | 重庆邮电大学 | A kind of electroencephalogramrecognition recognition method of combination convolutional neural networks and long memory network in short-term |
| CN109063639A (en) * | 2018-07-30 | 2018-12-21 | 浙江大学 | A kind of method of real-time prediction Brain behavior |
-
2019
- 2019-06-18 CN CN201910526926.7A patent/CN110393525B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101101511A (en) * | 2007-07-13 | 2008-01-09 | 杭州电子科技大学 | Brain electrical signal recovery method in brain-machine interaction |
| US20120172743A1 (en) * | 2007-12-27 | 2012-07-05 | Teledyne Licensing, Llc | Coupling human neural response with computer pattern analysis for single-event detection of significant brain responses for task-relevant stimuli |
| CN102327116A (en) * | 2011-09-28 | 2012-01-25 | 杭州电子科技大学 | Network system random resonance restoration method of cortex electroencephalographic signal |
| CN102697493A (en) * | 2012-05-03 | 2012-10-03 | 北京工业大学 | Method for rapidly and automatically identifying and removing ocular artifacts in electroencephalogram signal |
| CN105816170A (en) * | 2016-05-10 | 2016-08-03 | 广东省医疗器械研究所 | Wearable NIRS-EEG based schizophrenia early detection evaluation system |
| CN107961007A (en) * | 2018-01-05 | 2018-04-27 | 重庆邮电大学 | A kind of electroencephalogramrecognition recognition method of combination convolutional neural networks and long memory network in short-term |
| CN109063639A (en) * | 2018-07-30 | 2018-12-21 | 浙江大学 | A kind of method of real-time prediction Brain behavior |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110840411A (en) * | 2019-12-06 | 2020-02-28 | 深圳市德力凯医疗设备股份有限公司 | Anesthesia depth measuring method, storage medium and electronic device |
| CN110840411B (en) * | 2019-12-06 | 2022-03-11 | 深圳市德力凯医疗设备股份有限公司 | Measuring device, storage medium and electronic equipment of anesthesia degree of depth |
| CN112869754A (en) * | 2021-01-08 | 2021-06-01 | 浙江大学 | Brain-machine fusion neural signal lie detection method |
| CN113408697A (en) * | 2021-05-31 | 2021-09-17 | 杭州电子科技大学 | Electroencephalogram data reconstruction method based on reinforcement learning self-encoder |
| CN113408697B (en) * | 2021-05-31 | 2022-09-16 | 杭州电子科技大学 | Electroencephalogram data reconstruction method based on reinforcement learning self-encoder |
| CN115359236A (en) * | 2022-10-19 | 2022-11-18 | 之江实验室 | Work memory task magnetoencephalogram classification system based on machine learning |
| CN115359236B (en) * | 2022-10-19 | 2023-03-28 | 之江实验室 | Work memory task magnetoencephalogram classification system based on machine learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110393525B (en) | 2020-12-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110393525A (en) | A kind of brain activity detection method based on deep-cycle self-encoding encoder | |
| Wu et al. | Bayesian Machine Learning: EEG\/MEG signal processing measurements | |
| Bassett et al. | Emerging frontiers of neuroengineering: a network science of brain connectivity | |
| CN112861604B (en) | Myoelectric action recognition and control method irrelevant to user | |
| Cohen | A tutorial on generalized eigendecomposition for denoising, contrast enhancement, and dimension reduction in multichannel electrophysiology | |
| CN109614885A (en) | A kind of EEG signals Fast Classification recognition methods based on LSTM | |
| CN108135520B (en) | Generating natural language representations of psychological content from functional brain images | |
| CN112120694A (en) | Motor imagery electroencephalogram signal classification method based on neural network | |
| CN110090017A (en) | A kind of EEG signals source localization method based on LSTM | |
| CN109034263A (en) | The Alzheimer disease auxiliary diagnostic equipment and method of the brain network multi-frequency fusion kernel of graph | |
| Das et al. | Neuro-current response functions: A unified approach to MEG source analysis under the continuous stimuli paradigm | |
| Wei et al. | Edge sparse basis network: a deep learning framework for EEG source localization | |
| Zhao et al. | Two-stage spatial temporal deep learning framework for functional brain network modeling | |
| Liu et al. | Straps: A fully data-driven spatio-temporally regularized algorithm for M/EEG patch source imaging | |
| Sommer et al. | Exploratory analysis and data modeling in functional neuroimaging | |
| Cui et al. | Modeling brain diverse and complex hemodynamic response patterns via deep recurrent autoencoder | |
| Sahu et al. | Brain and Behavior Computing | |
| CN115359236B (en) | Work memory task magnetoencephalogram classification system based on machine learning | |
| CN116596046A (en) | Method for reconstructing image by utilizing electroencephalogram signals and visual features | |
| Bayrak et al. | From brain to body: learning low-frequency respiration and cardiac signals from fMRI dynamics | |
| Dong et al. | A novel fMRI representation learning framework with GAN | |
| Fordson et al. | A feature learning approach based on multimodal human body data for emotion recognition | |
| Morishige et al. | Estimation of hyper-parameters for a hierarchical model of combined cortical and extra-brain current sources in the MEG inverse problem | |
| Sweeney-Fanelli et al. | Automated Emotion Recognition Employing Wearable ECG Sensor and Deep-Learning | |
| CN113408697B (en) | Electroencephalogram data reconstruction method based on reinforcement learning self-encoder |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |