WO2005060830A1 - Procede d'analyse des fluctuations des signaux d'un eeg - Google Patents

Procede d'analyse des fluctuations des signaux d'un eeg Download PDF

Info

Publication number
WO2005060830A1
WO2005060830A1 PCT/CN2004/001493 CN2004001493W WO2005060830A1 WO 2005060830 A1 WO2005060830 A1 WO 2005060830A1 CN 2004001493 W CN2004001493 W CN 2004001493W WO 2005060830 A1 WO2005060830 A1 WO 2005060830A1
Authority
WO
WIPO (PCT)
Prior art keywords
analysis
frequency
power
spectrum
curve
Prior art date
Application number
PCT/CN2004/001493
Other languages
English (en)
French (fr)
Inventor
Jianlan Xu
Enhong Liu
Original Assignee
Guangzhou Kefu Medical Technology Co., Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Guangzhou Kefu Medical Technology Co., Ltd filed Critical Guangzhou Kefu Medical Technology Co., Ltd
Priority to US10/596,537 priority Critical patent/US7801597B2/en
Publication of WO2005060830A1 publication Critical patent/WO2005060830A1/zh

Links

Classifications

    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/72Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235Details of waveform analysis
    • A61B5/7253Details of waveform analysis characterised by using transforms
    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/24Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316Modalities, i.e. specific diagnostic methods
    • A61B5/369Electroencephalography [EEG]
    • A61B5/372Analysis of electroencephalograms
    • A61B5/374Detecting the frequency distribution of signals, e.g. detecting delta, theta, alpha, beta or gamma waves
    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/72Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235Details of waveform analysis
    • A61B5/7253Details of waveform analysis characterised by using transforms
    • A61B5/7257Details of waveform analysis characterised by using transforms using Fourier transforms

Definitions

  • the present invention relates to the field of medical technology equipment for diagnosis, and in particular, to a method and equipment for analyzing fluctuations in EEG signals. Background technique
  • EEG signals like the ECG.
  • electroencephalographs in medical use: analog signals and digital signals.
  • analog signals digital signals.
  • digital EEG signals are very weak and complicated, even the digital EEG with strong anti-interference is far from meeting the needs of clinical medical treatment, which makes the diagnostic significance of EEG far behind that of ECG.
  • the purpose of the present invention is to provide a method and related equipment for analyzing EEG signals by using contemporary computer technology to obtain a series of data parameters and display them in a variety of ways to provide a basis for brain function testing and disease diagnosis.
  • S spectrum The frequency range of the EEG detected by the present invention is in the ultra-slow wave range in mHz. This spectrum The line is called the super slow spectrum (Supra-slow pedigree :), which is abbreviated as the S spectrum.
  • the spectral lines that make up the S spectrum are called Sl, S2, S3 according to their frequency, respectively. For example, the spectral line corresponding to the lmHz frequency is Sl, and the frequency is 2mHz.
  • the X-inch spectrum should be S2, and so on.
  • Fundamental frequency Several commonly used EEG frequencies that are closely related to neurotransmitters in the brain, such as 1, 2, 3, 4, 5, 6, 7, 11, and 13mHz are called fundamental frequencies.
  • Basic pedigree The pedigree corresponding to the fundamental frequency is called the basic pedigree.
  • Dominant frequency, dominant spectral line, optimal value The power value of each frequency under each lead is ordered from large to small.
  • the frequency of the first ⁇ value, that is, Dl-Dn is called the dominant frequency;
  • the spectral line corresponding to the dominant frequency is Dominant spectral line;
  • the power value of the dominant frequency is the optimal value.
  • Optimal frequency and optimal spectral line The maximum value D1 of the dominant frequency per lead is called the optimal frequency, and the spectral line corresponding to the optimal frequency is called the optimal spectral line.
  • Continuous frequency The spectral lines entering the predominant frequency zone sometimes have continuous values, such as 2, 3, 4 ⁇ mHz, which is called continuous frequency, that is, continuous frequency.
  • Out-of-Frequency The spectrum formed by the non-resonant frequencies (frequency values greater than 13) and their multiples in the S-spectrum is called out-of-frequency.
  • the frequencies in the S spectrum are 23mHz, 27mHz, 28mHz, 29 mHz and their harmonic frequencies (such as 46mHz,
  • a series of characteristic frequencies of 54mHz,...) is called special frequency.
  • Characteristic spectral lines The spectral lines corresponding to continuous frequency, inter-frequency, special frequency, and optimal frequency are collectively referred to as characteristic spectral lines.
  • a / P, L / R According to the spatial distribution of the leads, calculate the ratio of the power values of the two leads before and after each frequency (such as lead F3 / C3), which is called the front-to-back ratio, and write A / P. Leads in the same position on the left and right brain
  • the invention applies computer technology to segment the EEG signal data according to a certain length of time, perform power spectrum analysis on each piece of data, select the maximum power amplitude in the range of 0.5-50HZ, and perform multiple power spectrum analysis and Spectrum analysis to obtain the power spectrum fluctuation chart in the ultra-slow wave range, and then perform a series of analysis on the fluctuation chart to obtain a series of data parameters. It is displayed in the form of numerical values, graphs and curves.
  • the method for analyzing the electroencephalogram fluctuation signal of the present invention includes at least a conventional power spectrum analysis. Secondly, it can also include the analysis of the EEG power fluctuation signal, the EEG fluctuation map analysis, and the S spectrum analysis in this order. Can also increase Set the error processing operation to correct the spectral lines where errors occur.
  • the method for analyzing the electroencephalogram fluctuation signal of the present invention further includes a collection of electroencephalogram signals.
  • the acquisition method and electrode placement can be applied to any lead system or any combination of leads.
  • the 12 lead of the international standard lead system is preferred.
  • the preferred positions for electrode placement are F3, F4, C3, C4, P3, P4, 01, 02, F7, F8, T5, T6.
  • the conventional power spectrum analysis includes the following steps:
  • the analysis of the EEG power fluctuation signal includes the following steps ⁇
  • the time-domain EEG signal with a total time length of N seconds is segmented by T seconds in chronological order, and the above-mentioned conventional power spectrum analysis and EEG power fluctuation signal analysis are sequentially performed on the N / T segment data to obtain the maximum power.
  • the analysis of the EEG fluctuation map includes the following steps: CI) Analyze the maximum power amplitude fluctuation signal P (n) with a length of n points; C 2) Multiply the Harming window with a length of n and then perform power spectrum analysis.
  • the unit of data time length is sand, so the frequency domain
  • the resolution is l / N Hz, and the spectral lines in a certain frequency range in the power spectrum analysis result are used to form a brain fluctuation map;
  • Household analysis S pedigree analysis includes the following steps:
  • the above analysis method is referred to as the first-level analysis method in the series of methods for analyzing EEG fluctuation signals of the present invention. Based on the results (data) obtained from the first-level analysis method, a number of further analyses can be performed, which is called the second-level analysis method, and includes the following 24 items.
  • each item of the second-level analysis method is numbered separately-based on the analysis of EEG power fluctuation signals, (1) entropy calculation and (2) single-frequency competition analysis; analysis in the S spectrum (3) S-spectrum analysis; (4) basic pedigree analysis; (5) optimal value analysis; (6) A / P reversal; L / R imbalance analysis; (7) special frequency analysis (8) Inter-frequency analysis; (9) Continuous frequency analysis; 0 ⁇ optimal frequency analysis; CD S spectrum power spatial distribution analysis; (1 Single-frequency power and relative value (L / R) distribution analysis; (13) Average power Distribution analysis; (14) Relative power value A / P, L / R analysis; (15) Long-term S spectrum curve analysis; (16) Long-term dominant spectrum curve analysis; (17) Long-time basic spectrum curve Analysis; (18) Analysis of long-term power spatial distribution curve; (19) Analysis of long-term entropy curve; (2 ⁇ Long-time special frequency curve analysis: (21) Long-time continuous frequency curve analysis; (22) Long Analysis of the power spatial distribution curve of the time-
  • the three items (1 single-frequency power distribution analysis, (13) average power distribution analysis, and (14) relative power value A / P, L / R analysis) can be collectively referred to as power distribution analysis.
  • long-term S spectrum curve analysis (16) long-term advantage spectrum curve analysis, (1 long-time basic spectrum curve analysis, (18) long-time power spatial distribution curve analysis, (19) long Time history entropy curve analysis, (20) Long-term special frequency curve analysis, (21) Long-time continuous frequency curve analysis, (22) Long-time basic spectrum power space distribution curve analysis, (23) Long-time conventional power spectrum curve analysis, (24 The long-term event marker i has identified ten items that can be called a long-term dynamic curve analysis.
  • the (1) entropy calculation is performed on the basis of analysis of power fluctuation signals of the EEG, and the method is:
  • the total entropy is calculated by integrating the total probability distributions of all N leads (total n * N :).
  • the (2) single-frequency competition analysis is also performed on the basis of the analysis of the power fluctuation signal of the EEG.
  • the method is that the same optimal frequency number in the frequency fluctuation graph f (n) of the EEG fluctuation signal is The time change process is accumulated to obtain the probability curve of the optimal frequency.
  • the (3) S-spectrum total spectrum analysis method is to represent the S-spectrum total spectrum data obtained by the S-spectrum analysis with a graph.
  • the (4) basic pedigree analysis method based on the frequency spectrum S corresponds to S lineage statistical analysis, greater than the frequency of 3 mHz times the period which begins accumulating the value of the frequency (e.g., when the statistics should be accumulated 6mHz 3 ⁇ ⁇ , 9mHz ,...); At the same time, all leads are divided into front and back, left and right sections for statistics according to their placement on the head.
  • the (5) optimal value analysis method is to display the power value of the dominant frequency and the corresponding frequency according to the spatial position distribution of the lead.
  • a / P reversal and L / R imbalance is: calculating the front-to-back ratio A / P of the power value of each frequency according to the spatial distribution position of the lead, and the frequency where the A / P value is greater than a certain limit Display; Calculate the left / right ratio L / R at the same time, and display the frequency with L / R value greater than a certain limit.
  • the four items of (7) special frequency analysis, (8) inter-frequency analysis, (9) continuous frequency analysis, and (1 ⁇ optimal frequency analysis) are performed by separately dividing the special frequency, inter-frequency, and continuous frequency of each lead. Frequency and optimal frequency are displayed according to the spatial distribution position of the lead.
  • the method for analyzing the spatial distribution of power in the IDS spectrum is to arrange the power value of each spectral line in the EEG fluctuation map according to the spatial lead position distribution, and open a window in a "picture-in-picture" manner on the display interface. With the selection of spectral lines, the selected spectral lines are displayed with the power value of each lead according to the spatial position distribution of the leads.
  • the (12) single-frequency power and relative value (L / R) distribution analysis method is: adding the corresponding power values of the dominant spectral lines D1-Dn of each lead to obtain the total power value of each lead; The values of the power and left / right ratio (L / R) greater than the limit value or less than 1 / limit value are displayed according to the spatial distribution of the leads.
  • the (13) average power distribution analysis method is to display the average power of each lead according to the spatial distribution position of the lead.
  • the (M) power relative value A / P, L / R analysis method is to calculate the front-to-back ratio and left-to-right ratio of the power value according to the spatial distribution position of the lead.
  • the (15) long-term S spectrum curve analysis method is based on the vertical axis of the fluctuation value of each pedigree or each spectrum line of each lead or all leads as the vertical axis and time as the horizontal axis. curve. Open a window in picture-in-picture mode on the display interface to select line or lineage.
  • the (16) long-term dominant spectral line curve analysis method is to make a curve with the frequency of the spectral line entering the dominant spectral line area as the vertical axis and the horizontal axis as the horizontal axis. Open a window on the display interface in "picture-in-picture” mode, and select the order of dominant spectral line ranking (Dl-Dn). .
  • the (17) long-term basic pedigree curve analysis method is to use the fluctuation values of all leads or basic pedigrees under each lead for a time period as the vertical axis, and time as the horizontal axis to make the basic pedigree. curve. Open a window in the picture-in-picture mode on the display interface and select the pedigree.
  • the (18) long-term power spatial distribution curve analysis method is: for each lead, for each spectral line, a curve is drawn with the power value on the vertical axis and time as the horizontal axis. Open a window in the picture-in-picture mode on the display interface and select the spectrum.
  • the (19) long-term entropy change curve analysis method is to use the vertical axis as the entropy value for all leads or each lead, and use the time as the horizontal axis to make curves separately to reflect the change of entropy value with time. .
  • the (20) long-term special frequency curve analysis method is to use the number of occurrences of the special frequency in all leads or each lead as the vertical axis, and use time as the horizontal axis to make a curve for observing the special frequency. Situations that change dynamically over time.
  • the (21) long-term continuous-frequency curve analysis method is to use the number of consecutive frequencies in all leads or each lead as the vertical axis, and use time as the horizontal axis to make a curve, respectively. Dynamic changes in time.
  • the method for analyzing the (22) long-term basic spectrum power spatial distribution curve is: reading the power value of each lead of the basic spectrum from the single-frequency power and relative value distribution (L / R) analysis, and using the power value
  • the vertical axis uses time as the horizontal axis to make a curve, and each curve is displayed according to the spatial position distribution of the lead.
  • the (23) long-term conventional power spectrum curve analysis method is to find the n frequencies Dl-Dn with the largest amplitude from the conventional power spectrum, sort them from large to small, and use the power values of these frequencies respectively. For the vertical axis and time for the horizontal axis, make n dynamic curves.
  • the (24) event mark recognition method is to identify the event mark signals recorded by the EEG recording box, and display the mark signals on the EEG signal playback and the corresponding positions on the time axis of various dynamic curves.
  • Event marker recognition is combined with any one or several of the other nine items of long-term dynamic curve analysis.
  • the above-mentioned analysis method can be applied to the data processing of the EEG signals collected by any one or several lead combinations, and the operation results of one or several lead combinations in the analysis result can be optionally output to the terminal processor for processing. Display, print or store.
  • the invention also discloses a device for analyzing the electroencephalogram fluctuation signal. It includes electrodes, digital EEG signal amplifier or EEG recording box, personal computer, data processor, terminal processor. They are connected in turn.
  • the electrodes collect EEG signals, and the signals are transmitted to a digital EEG signal amplifier and / or an EEG signal recording box, which receives, amplifies, digital-to-analog converts, digital filters, or / and data stores the signals, and digitizes the EEG signal amplifier.
  • / or the data in the EEG signal recording box is uploaded to the personal computer, the data processor connected to the computer completes the data processing and fluctuation analysis, and the analysis result is transmitted to the terminal processor for storage, display or printing.
  • the method of placing the electrodes may be selected from one or a combination of several leads.
  • the data processor includes a conventional power spectrum analysis module, an EEG power fluctuation signal analysis module, an EEG fluctuation map analysis module, and an S-spectrum analysis module for performing a first-level analysis. They are connected in sequence, and the data generated by the previous module is transferred to the next module for operation analysis.
  • the data processor may further include any one of the following 24 modules for performing the second-level analysis.
  • the second level analysis module accepts the first
  • the analysis results (data) of the AHP module are further analyzed.
  • the second-level analysis module is numbered respectively-(1) entropy calculation module and (2) single-frequency competition analysis module connected to the EEG power fluctuation signal analysis module and receiving its data;
  • S-spectrum analysis module connected to the S-spectrum analysis module and receiving its data; (4) basic pedigree analysis module; (5) optimal value analysis module; (6) A / P reversal, L / R imbalance Situation analysis module; (7) special frequency analysis module; (8) inter-frequency analysis module; (9) continuous frequency analysis module; (1 ⁇ optimal frequency analysis module; (11) S-spectrum power spatial distribution analysis module; (12) Single-frequency power distribution analysis module; (13) Average power distribution analysis module; (14) Relative power value A / P, L / R analysis module; (15) Long-term history S spectrum curve analysis module; (16) Long-term history Dominant spectral curve analysis module; (17) Long-term basic spectrum curve analysis module; (18) Long-time power spatial distribution curve analysis module; (1 ⁇ long-time entropy curve analysis module; (2 ⁇ long-time special frequency) Curve analysis module; (21) Long-term continuous-frequency curve analysis module; (22) Long-time basic spectrum system power spatial distribution curve analysis module; (23) Long-time conventional power spectrum curve analysis module; (24) Long
  • the conventional power spectrum analysis module is configured to perform power spectrum analysis on a time-domain EEG signal of a given time length to obtain a power spectrum, which can be expressed as ;
  • the sampling time is T seconds, and calculate the frequency domain resolution of the spectrum obtained by the FFT to be 1 / T.
  • the power amplitude ⁇ and the corresponding frequency f with the greatest energy are selected;
  • the time-domain EEG signal with a total time length of N seconds is segmented by T seconds, and the above-mentioned analysis of the conventional power spectrum analysis module and the analysis of the EEG power fluctuation signal analysis module are performed on each piece of data in order to obtain the maximum power amplitude.
  • the EEG fluctuation map analysis module is configured to analyze a maximum power amplitude fluctuation signal P (n) having a length of n points. Multiply the Harming window of length n, then do the power spectrum analysis, the data time length The unit is N seconds, so the resolution in the frequency domain is 1 / N Hz. Taking a certain range of spectral lines in the power spectrum analysis result to form an EEG fluctuation map. When the total sampling time is greater than N seconds, divide the paragraph in N seconds, and perform the above-mentioned conventional power spectrum analysis module, EEG power fluctuation signal analysis module, and EEG fluctuation graph analysis module ⁇ ⁇ for each section. .
  • the S-pedigree analysis module is used to find several dominant spectral lines Dl-Dn with the largest amplitude from each lead EEG fluctuation map, and sort the amplitudes from large to small to obtain a single lead data.
  • the second-level analysis modules are described below.
  • the data analyzed by the EEG power fluctuation signal analysis module in the first layer is transmitted to the ⁇ entropy calculation module and (2) the single-frequency competition analysis module for the second-level analysis. among them,
  • the entropy value calculation module is configured to calculate entropy according to ⁇ , where ⁇ is a probability that each frequency in the brain wave is dominant; synthesize the total probability distribution of all N leads to calculate the total entropy (total n * N) .
  • the single-frequency competition analysis module is used to accumulate the same optimal frequency number in the EEG fluctuation signal f (n) with time (1-n data segment) to obtain the optimal frequency. Probability curve.
  • the data obtained from the analysis of the S lineage analysis module in the first layer is transmitted to the following second layer analysis module:
  • the (3) S-spectrum analysis module displays the S-spectrum data obtained from the S-spectrum analysis module graphically. '
  • the (4) basic pedigree analysis module is configured to perform statistical analysis on the S pedigree corresponding to the fundamental frequency in the S spectrum, and accumulate values of multiples thereof starting from a frequency greater than 3 mHz; at the same time, all leads are distributed according to the distribution in the brain The position is divided into several parts before, after, and left, and statistics are made.
  • the (5) optimal value analysis module is configured to display the power value and the corresponding frequency of the Dl-Dn dominant frequency in each lead according to the spatial position distribution of the lead.
  • the (6) A / P reversal and L / R imbalance analysis module are used to calculate the front-to-back ratio A / P of each frequency power according to the spatial distribution position of the lead, and the A / P value is greater than a certain limit.
  • the frequency is displayed; the left / right ratio L / R is calculated at the same time, and the frequency whose L / R value is greater than a certain limit is displayed.
  • the (7) special frequency analysis module (8) different frequency analysis module, (9) continuous frequency analysis module, (1 ⁇ optimal frequency analysis
  • the modules are respectively used to display the special frequency, inter-frequency, continuous frequency, and each optimal frequency of each lead according to the spatial distribution position of the lead.
  • the (IDS spectrum power spatial distribution analysis module is used to arrange the power value of each spectral line in the EEG fluctuation map according to the spatial lead position distribution, and open a window with "picture in picture” on the display interface Select the spectral line, and display the power value under each lead according to the spatial position distribution of the selected frequency spectral line.
  • the (12) single-frequency power and relative value distribution L / R analysis module adds the corresponding power values of the dominant spectral lines D1-Dn of each lead to obtain the total power value of each lead; the power of the fundamental frequency and Values of the left-right ratio (L / R) that are greater than the limit value or less than the V limit value are displayed according to the spatial distribution of the leads.
  • the (13) average power distribution analysis module displays the average power of each lead according to the spatial distribution position of the lead.
  • the (14) power relative value A / P, L / R analysis module calculates the front-to-back ratio and left-to-right ratio of the power value according to the spatial distribution position of the lead. ⁇
  • the (15) long-term S-line curve analysis module uses the fluctuation values of each lineage or each line under each lead or all leads as the vertical axis and time as the horizontal axis, respectively. Curve; Open a window in picture-in-picture mode on the display interface to select the line or lineage.
  • the (16) long-term superiority spectral line curve analysis module is used to make a curve with the frequency of the spectral lines entering the dominant spectral line region as the vertical axis and time as the horizontal axis. Open a window on the display interface in "picture-in-picture” mode, and select the ranking order of superior spectral lines (Dl-Dn).
  • the (17) long-term basic pedigree curve analysis module is configured to use the fluctuation values of all leads or each time period of the basic pedigree as the vertical axis, and time as the horizontal axis to make the basic pedigree. Curve; Open a window in picture-in-picture mode on the display interface to select the pedigree.
  • the (18) long-term power spatial distribution curve analysis module is used to draw a curve on the vertical axis and time axis as the horizontal axis of each spectral line under each lead. Open a window in the picture-in-picture mode on the display interface to select the spectral line.
  • the (19) long-term entropy value change curve analysis module is configured to make the curve of the entropy value of the whole brain or each lead with the vertical axis and time as the horizontal axis.
  • the (2 ⁇ long-time special frequency curve analysis module is used to make a curve based on the number of occurrences of the special frequencies of all leads or each lead, and time as the horizontal axis.
  • the (21) long-term continuous-frequency curve analysis module outputs the signals in all leads or under each lead.
  • the current number is the vertical axis
  • the horizontal axis is the horizontal axis.
  • the (22) long-term basic spectrum power spatial distribution curve analysis module is configured to read the power value of each lead of the basic spectrum from the single-frequency power and relative value distribution (L / R) analysis, and use the power
  • the value is the vertical axis, and the curve is plotted with time as the horizontal axis.
  • Each curve is displayed according to the spatial position distribution of the lead. Open a window in picture-in-picture mode on the display interface to select the pedigree.
  • the (23) long-time-range conventional power spectrum curve analysis module is configured to find n frequencies D1-Dn with the largest amplitude from the conventional power spectrum, and sort them according to the amplitude from large to small, and respectively use the power of these frequencies
  • the value is the vertical axis
  • time is the horizontal axis.
  • N dynamic curves are made.
  • the (24) long-term event labeling module is configured to identify event label signals recorded in the EEG recording box, and display the label signals on the EEG signal playback and corresponding positions on the time axis of various dynamic curves. It is connected to each long-term analysis module.
  • the data processor can process the EEG signals collected by the electrodes by using any one or several lead combinations.
  • modules responsible for the second-level analysis described above can be assembled and used simultaneously, or (24) the event mark recognition module can be excluded, and all other modules (1) to (23) can be assembled and used in any combination.
  • the above-mentioned analysis method can be applied to data processing of the EEG signals collected by any one or several lead combinations, and can also choose to output the calculation results of one or more lead combinations in the analysis result to the terminal processor.
  • the terminal processor is composed of a display, a printer, and a storage device (such as a hard disk, a floppy disk, an optical disk, etc.), and is used for receiving signals from the data processor and selecting one or several lead combinations in the operation result of the data processor. As a result, it is stored, displayed, or printed.
  • a storage device such as a hard disk, a floppy disk, an optical disk, etc.
  • the functional status of the brain can be analyzed, the status of neurotransmitters in the brain of diseased patients can be analyzed, and the functional changes of the brain that are powerless by CT and MRI, Provide direct objective indicators for the diagnosis of functional encephalopathy in the medical community, and make up for the lack of objective detection indicators for the diagnosis of functional encephalopathy (such as mental illness) in the medical community.
  • Fig. 1 is a block diagram of a first embodiment of the present invention.
  • FIG. 2 is a block diagram of a second embodiment of the present invention.
  • FIG. 3 is a block diagram of a third embodiment of the present invention.
  • FIG. 4 is a block diagram showing the structure of a data processor of the present invention.
  • the logical connection is indicated by the arrow connector between the boxes: the data obtained by the previous module operation is the basis of the operation of the latter module, and the operation of the latter module depends on the data obtained by the operation of the previous module.
  • the connection relationship between the boxes is indicated by a straight line connection: the module shown in the previous box is composed of the following multiple modules. If the previous box is just the general name of the module category to which multiple modules belong, the box is shown with a dashed line. detailed description
  • the implementation process of analyzing the electroencephalogram fluctuations of the present invention is to extract the electroencephalogram fluctuation signals from the electroencephalogram using computer fluctuation scanning technology, and then perform spectrum analysis on the fluctuation signals to obtain a power spectrum in the range of l_255mH Z .
  • the specific analysis process can be summarized as follows: The 1024-second EEG data is segmented by 2 seconds per segment and divided into 512 segments; the power spectrum analysis is performed on each segment of data to select the maximum power amplitude in the range of 0.5-50Hz The power spectrum analysis is performed on the time fluctuation process of the maximum power amplitude (that is, the maximum value of the corresponding power spectrum of 512 data segments) to obtain a power spectrum fluctuation chart in the range of 1 to 255 mHz.
  • composition of the device of the present invention can be of the following three types:
  • FIG. 1 it includes an electrode A, a digital EEG signal amplifier B, a USB interface J, a personal computer C, a data processor D, and a terminal processor E connected in this order.
  • FIG. 2 it includes an electrode A, an EEG recording box F, a USB interface J, a personal computer C, a data processor D, and a terminal processor E connected in this order.
  • FIG. 3 it includes an electrode eight, a digital EEG signal amplifier B and an EEG recording box F, a USB interface J, a personal computer C, a data processor D, and a terminal processor E.
  • the electrode A is connected to the digital EEG signal amplifier B and the EEG recording box F at the same time, the digital EEG signal amplifier B is directly connected to the personal computer C, and the EEG recording box F is connected to the personal computer C through the USB interface J.
  • the personal computer C is in turn connected to the data processor D and the terminal processor E.
  • Electrode A Used to collect EEG signals.
  • the electrode placement uses 12 leads of the international standard lead system, and the positions are F3, F4, C3, C4, P3, P4, 01, 02, F7, F8, T5, and T6.
  • the EEG sampling rate is 128Hz.
  • Digital EEG signal amplifier B signal reception, signal amplification, digital-to-analog conversion, digital filtering And other functions.
  • EEG recording box F It is used to collect and analyze long-term (more than 18 minutes) EEG data. It has functions of signal acquisition, signal amplification, digital-to-analog conversion, digital filtering, data storage, and data playback. The data in the EEG recording box was uploaded to a personal computer for fluctuation analysis.
  • Personal computer C The host processor adopts PIV type, and the memory is 256M.
  • Terminal processor E It consists of a display, a printer, and a storage device (such as a hard disk, a floppy disk, or an optical disk). It accepts the calculation results of the data processor D and stores, displays, or prints.
  • the data processor includes a conventional power spectrum analysis module D1, an electroencephalogram fluctuation signal analysis module D2, an electroencephalogram fluctuation map analysis module D3, and an S-spectrum analysis module D4 for performing a first-level analysis. They are connected in sequence, and the data generated by the previous module is transferred to the next module.
  • the data processor may also include any of the following 24 modules for second-level analysis. For the convenience of description, the second-level analysis modules are numbered separately:
  • S spectral spectrum analysis module D4a connected to the D4 module and receiving its data; (4) basic pedigree analysis module D4b; (5) optimal value analysis module D4c; (6) A / P reversal, L / R Imbalance analysis module D4d; (7) Special frequency analysis module D4ea; (8) Inter-frequency analysis module D4eb ; (9) Connected frequency analysis module D4ec ; (10) Optimal frequency analysis module D4ed; (11) S-spectrum power space Distribution analysis module D4f; (12) Single-frequency power distribution analysis module D4ga; (13) Average power distribution analysis module D4gb ; (14) Relative power value A / P, L / R analysis module D4gc; (15) Long-term history S Pedigree curve analysis module D4ha; (16) Long-term advantage spectral line curve analysis module D4hb; (17) Long-time basic spectrum curve analysis module D4hc; (18) Long-time power spatial distribution curve analysis module Mhd; (19) Long-
  • the above four modules (7) special frequency analysis module D4ea, (8) inter-frequency analysis module D4eb, (9) continuous frequency analysis module D4ec, (1Q) optimal frequency analysis module D4ed are called a feature spectrum Line analysis module D4e.
  • long-term S spectrum curve analysis module D4ha (16) long-term dominant spectrum curve analysis module D4hb, (17) long-term basic spectrum curve analysis module D4hc, (18) long-term power spatial distribution Curve analysis module D4hd, 09) Long-time history entropy change curve analysis module D4he, (20) Long-time history special frequency curve analysis module D4hf, (21) Long-time history continuous frequency curve analysis module D4hg, (22) Long-time history Basic pedigree power spatial distribution curve analysis module D4hh, (23) Long time history conventional power spectrum curve analysis module D4hi, (24) Long time history event marker recognition module D4hj Ten modules are integrated into one, which is called long time history dynamic curve analysis Module D4h.
  • the conventional power spectrum analysis module D1 directly performs the Fourier transform on the N-point observation data (") of the random signal to obtain ⁇ ( ⁇ ⁇ ), and then takes the square of its amplitude and divides it by N as the pair.
  • the true power spectrum) is estimated.
  • the calculation of ⁇ is calculated by the fast Fourier transform (FFT):
  • a power spectrum analysis of a time-domain EEG signal with a given time length of 8 seconds is performed to obtain its energy distribution in the frequency domain, that is, a conventional power spectrum, and the power spectrum is transmitted to the terminal processor E in a graphical manner.
  • Power spectrum analysis is performed on the time-domain EEG signal with a given time length of 2 seconds after multiplying it by the Harming window.
  • the time-domain EEG signal with a total length of 1024 seconds is segmented in chronological order by 2 seconds per segment, and is divided into 512 segments.
  • the time fluctuation signal is transmitted to the EEG fluctuation graph analysis module D3, the entropy calculation analysis module D5, and the single-frequency competition analysis module D6, and the signal is transmitted to the terminal processor at the same time.
  • the maximum power amplitude fluctuation signal p (n) with a length of 512 points (1024 seconds) is analyzed. Multiply the Harming window with a length of 512 and do power spectrum analysis.
  • the unit of data time length is 1024 seconds, so the frequency domain resolution is 1/1024 Hz.
  • the EEG fluctuation map analysis module D3 transmits the EEG fluctuation map signal to the S-lineage analysis module D4, and also transmits it to the terminal processor.
  • the S-spectrum analysis module D4 transmits the single-lead S-spectrum and S-spectrum signals to the S-spectrum analysis module D4a, the basic pedigree analysis module D4b, the optimal value D (1-8) analysis module D4c, and A / P reversal , L / R imbalance analysis module D4d, characteristic spectral line analysis module D4e, S-spectrum power spatial distribution analysis module D4f, power distribution analysis module D4g, and long-term dynamic curve analysis module D4h.
  • the S-spectrum signal is transmitted to the terminal processor E.
  • Entropy calculation module D5 Calculate the entropy according to ⁇ , which is the probability that each frequency prevails in the brainwave (the total is 512, and the probability is the number of dominant single frequencies divided by 512). The total entropy is calculated by integrating the total probability distribution of all 12 leads to obtain the total entropy (the total is 512 * 12). The data is transmitted to the terminal processor E.
  • the S-spectrum signal generated in the S-spectrum analysis module D4 is made into a graph, and the graph and data are transmitted to the terminal processor E.
  • the power values and corresponding frequencies of the Dl-D8 optimal frequencies under each lead are displayed, and the results are transmitted to the terminal processor E.
  • Characteristic spectral line analysis module Me includes special frequency analysis module D4ea, inter-frequency analysis module Meb, connected frequency analysis module D4ec, and optimal frequency analysis module D4ed. among them,
  • Special frequency analysis module D4ea Display the special frequency of each lead according to the spatial distribution position of the lead, and transmit the result to the terminal processor.
  • Inter-frequency analysis module D4eb The trans-frequency of each lead is displayed according to the spatial distribution position of the lead, and the result is transmitted to the terminal processor E.
  • Link analysis module D4ec Display the link frequency of each lead according to the spatial distribution position of the lead, and transmit the result to the terminal processor E.
  • Optimal frequency analysis module D4ed Display the optimal frequency of each optimal frequency according to the lead spatial distribution position, and transmit the result to the terminal processor E.
  • Power distribution analysis module D4g including single-frequency power and relative value distribution (L / R) analysis module D4ga, average power distribution analysis module D4 g b, A / P, L / R analysis module D4gc.
  • Average power distribution analysis module D4gb The average power of each lead is displayed according to the spatial distribution position of the lead, and the result is transmitted to the terminal processor E.
  • a / P, L / R analysis module D4gc calculate the front-to-back ratio (such as F3 / C3, C3 / P3) and the left-to-right ratio (such as F3 / F4, C3 / C4) of the power value according to the spatial distribution position of the lead; The result is transmitted to the terminal processor E.
  • S lineage curve analysis module D4ha includes S lineage curve analysis module D4ha, dominant line curve analysis module D4hb, basic lineage curve analysis module D4hc, power spatial distribution curve analysis module D4hd, entropy curve analysis module D4he, special frequency curve analysis module D4hf, connected frequency curve analysis module D4h g , basic spectrum power distribution curve analysis module D4hh, conventional power spectrum curve analysis module D4hi, and event mark recognition module D4hj.
  • the EEG data of long duration (collection time greater than 18 minutes) is segmented according to the length of 18 minutes, and monthly If fluctuation graph analysis is performed on each piece of data (repeat steps D1-D4), and the analysis results are transmitted to the following Analysis module.
  • Dominant spectral curve analysis module D4ha It is used to take the frequency of the spectral lines into the dominant spectral region For the vertical axis and time as the horizontal axis, make a curve; open a window on the display interface in "picture-in-picture” mode, and select the order of dominant spectral line ranking (Dl-Dn);
  • S lineage curve analysis module D4hb Use the fluctuation value of each lineage or each line under each lead or all leads as the vertical axis and time as the horizontal axis to make a curve; On the interface, a window is opened in "picture-in-picture” mode to select line or lineage. Use this to observe how each frequency line or lineage changes over time throughout the brain.
  • Basic pedigree curve analysis module D4hc Make the basic pedigree curve with the fluctuation values of all leads or each time period of the basic pedigree under each lead as the vertical axis, and time as the horizontal axis to make the basic pedigree curve, on the display interface Open a window in "picture-in-picture” mode and select the pedigree. It is used to grasp the dynamics of several basic pedigrees that are most closely related to brain function.
  • each spectrum line uses its power value as the vertical axis and time as the horizontal axis to draw the curve. Open a window in the picture-in-picture mode on the display interface to select the pedigree. Based on this, observe the change of the power value of each spectral line with time under each lead.
  • Entropy curve analysis module D4he the entropy value of all leads or each lead is plotted on the vertical axis, and the time is plotted on the horizontal axis to reflect the change of entropy with time, and observe different times accordingly Brain energy distribution.
  • Special frequency curve analysis module D4hf Use the number of all lead special frequencies as the vertical axis and time as the horizontal axis to make a curve, and observe the dynamic changes of the special frequency over time.
  • Continuous frequency curve analysis module D4hg Use the number of occurrences of all leads or the frequency of each lead as the vertical axis, and use the inch as the horizontal axis to make a curve to observe the decline in brain function of patients at different times happening.
  • Spatial distribution curve of basic spectrum power D4hh Read the power value of each lead of the basic spectrum from the single-frequency power and relative value distribution (L / R) analysis module D4ga, and use the power value as the ordinate and Time is plotted for the abscissa, and each curve is displayed according to the spatial position distribution of the leads. Open a window in picture-in-picture mode on the display interface to select the pedigree. Based on this, observe the changes in time and spatial distribution of each basic pedigree.
  • Conventional power spectrum curve analysis module D4hi find the eight largest frequency values D1-D8 from the conventional power spectrum, and sort them from large to small, and use the power values of these frequencies as the vertical axis, respectively. With time as the horizontal axis, 8 dynamic curves were made, and the power-dominated frequency in EEG was observed to change with time.
  • Event marker recognition module D4hj It is connected to each long-term analysis module to identify the event marker signals recorded by the EEG recording box F, and these marker signals are played back on the EEG signals and various dynamic curves. A marker symbol is displayed at the corresponding position on the timeline.
  • the functional status of the brain can be analyzed, the status of neurotransmitters in the brain of patients with brain diseases, and the brain can be analyzed.
  • the changes provide a direct objective indicator for the diagnosis of functional encephalopathy in the medical community, and make up for the lack of objective detection indicators in the diagnosis of functional ⁇ encephalopathy (such as mental illness) in the medical community.

Description

一种脑电涨落信号分析方法及其设备 技术领域
本发明涉及诊断用医疗技术设备领域,尤其涉及一种对脑电涨落信号进行 分析的方法及其设备。 背景技术
理论上说,人和动物脑的生理和病理状态、脑的功能活动是可以如同心电 图那样通过对脑电信号的检测获知的。目前医疗上在用的脑电图仪有模拟信号 和数字信号两种。然而由于脑电信号非常微弱而复杂, 即便是抗干扰性较强的 数字化脑电图仪也远不能满足临床医疗的需要,致使脑电图的诊断意义较之心 电图落后很多。
对脑电涨落信号进行检测分析以判断脑功能状况和所患疾病,是中国航天 医学的研究成果, 其临床意义远高于普通脑电图。 因此, 能够对脑电涨落信号 进行全面、深入、准确分析的方法和相应的仪器设备便成为医学界的需要。 中 国专利 ZL96244175. 9 "脑电超慢涨落分析仪"的功能是对脑电信号进行采集、 放大、定时釆样、数模转换及数据预处理,将处理过的信号传输给个人计算机。 该专利技术将脑电信号的采集技术提高到一个新的水平,但是该设备并没有将 所采集的脑电信号进行分析的功能,所能呈现给医生的还只是脑电的波型即脑 电图。至于医生能从脑电图中获取多少对诊断有用的信息, 就凭医生个人对脑 电波型的辨识能力, 依然没有满足为临床医疗提供依据的需要。 发明内容
本发明的目的在于提出一种应用当代计算机技术对脑电信号进行分析的 方法和相关设备, 用以获 #一系列数据参数, 并以多种方式显示, 为脑功能检 测和疾病诊断提供依据。
以下先对实现本发明目的的技术方案所涉及、使用的若干技术名词、术语,. 参考梅磊著《ET-脑功能研究新技术》 (国防工业出版社, 1994)作如下定义和 解释。
S谱: 本发明所检测脑电的频率范围在以 mHz计的超慢波范围内。 这个谱 系被称为超慢谱 (Supra-slow pedigree:), 简称 S谱。 组成 S谱的谱线按其频 率高低, 分别称为 Sl、 S2、 S3 ······。 如 lmHz频率对应的谱线为 Sl, 2mHz频率
X寸应的谱线为 S2,以此类推。
基频: 比较常用的,与脑内神经递质关系比较密切的几个脑电频率,如 1、 2、 3、 4、 5、 6、 7、 11、 13mHz等频率称为基频。
基本谱系: 基频所对应的谱系称为基本谱系。
优势频率、优势谱线、最优值: 将每个导联下各个频率的功率值从大到小 ί 序, 前 η个数值的频率即 Dl- Dn称优势频率;优势频率对应的谱线为优势谱 线; 优势频率的功率值为最优值。
最优频、 最优谱线: 每导联优势频率中的最大值 D1称为最优频, 最优频 所对应的谱线称为最优谱线。
连频: 进入优势频率区的谱线有时出现其数值相连续现象, 如 2、 3、 4〜 mHz, 称之为连续频率, 即连频。
异频: S谱中非谐振频率 (频率值为大于 13的质数)及其倍频所形成的谱 线称为异频。
特频: S谱中频率为 23mHz、 27mHz、 28mHz、 29 mHz及其谐频 (如 46mHz、
54mHz、 …) 的一系列特征频率称为特频。
特征谱线: 连频、 异频、 特频、 最优频所对应的谱线统称为特征谱线。 A/P、 L/R: 按导联空间分布位置, 计算每一频率前后两个导联功率值比值 (如 F3导联 /C3导联), 称为前后比值, 写作 A/P。 将左右脑相同位置的导联
Θ勺功率值进行比较, 计算其比值 (如 F3导联 /F4导联), 称为左右比值, 写作 涨落数值: 泛指 S谱系分析中获得的各个谱系的数值。
本发明的目的是按以下技术方案实现的。
本发明应用计算机技术先对脑电信号数据按一定时间长度进行分段,对每 段数据进行功率谱分析, 选择出 0. 5-50HZ范围内的最大功率幅值, 进行多次 功率谱分析及频谱分析,获得超慢波范围内的功率谱涨落图, 再对涨落图进行 一系列分析,获得一系列数据参数. 并以数值、 图形和曲线的方¾显示出来。
本发明脑电涨落信号分析方法最少包括常规功率谱分析。其次还可以包括 依次进行的脑电功率涨落信号分析、脑电涨落图分析和 S谱系分析。还可增加 设置误差处理运算, 用于修正发生误差的谱线。
本发明脑电涨落信号分析方法还包括对脑电信号的釆集。采集方法及电极 置放可应用任意导联系统或任意导联的组合, 优选国际标准导联系统的 12导 联, 电极置放优选位置为 F3、 F4、 C3、 C4、 P3、 P4、 01、 02、 F7、 F8、 T5、 T6。
以下对每一分析方法的具体步骤和算法分别叙述。
所述常规功率谱分析包括如下步骤:
( 1 )对给定时间长度的时域脑电信号做功率谱分析,即对脑电信号 的
N点观察数据 直接做傅里叶变换, 得到^
(2) 取其幅值的平方, 并除以 Ν, 作为对 的真实的功率 i普 的估 计, 用周期图法估计出的功率谱可以表示为, (e; ) = ^Ιχ^ )Ι ;
( 3) 的计算由快速傅里叶变换计算得到:
Figure imgf000005_0001
所述脑电功率涨落信号分析包括如下步骤-
( 1 ) 选择边瓣幅值小且衰减快的窗函数, 它表 ω{ή) = 0.5 - 0.5 cos(— ), « = 0,1,...N - 1.
对脑电信号 数据进行截短, 得到
(2)对上述信号 = M")做功率谱分析,采样时间为 T秒,计算 FFT 得到频谱的频域分辨率为 1/T,选择出能量最大的功率幅值 ρ及相应的频率 f;
(3) 按时间顺序对总时间长度为 N秒的时域脑电信号按 T秒进行分段, 依次对 N/T段数据进行上述常规功率谱分析和脑电功率涨落信号分析,得到最 大功率幅值的涨落信号 P (n)和相应频率涨落信号 f (n),n=l…… n, n=N/T。 它 表示了总时间内功率及相应最大幅值的涨落过程。
所述脑电涨落图分析包括如下步骤: C I ) 对长度为 n点的最大功率幅值涨落信号 P (n)进行分析; C 2 ) 乘上长度为 n的 Harming窗, 再做功率谱分析, 数据时间长度单位 为 沙, 因而频域分辨率为 l/N Hz, 取功率谱分析结果中一定频率范围的谱 线组成脑涨落图; ·
C 3 ) 当总采样时间大于 N秒时, 以 N秒为单位进行段落划分, 对每一段 分别进行上述常规功率谱分析、 脑电功率涨落信号分析和脑电涨落图分析。
戶斤述 S谱系分析包括如下步骤:
( 1 ) 从每一个导联脑电涨落图中找出幅值最大的几条优势谱线 Dl-Dn, 按幅值从大到小排序, 得到单一导联数据的 S谱, 共 n个数值;
( 2 ) 综合所有 N导联共得到 N*n条优势谱线, 把频率相同的优势谱线的 数目累加, 得到 S谱总谱。
以上分析方法称为本发明脑电涨落信号分析系列方法中第一层次的分析 方法。 在第一层次分析方法所得结果(数据)的基础上还可以做多项进一步分 析, 称为第二层次分析方法, 包括以下 24个项目。 为叙述方便, 给第二层次 分析方法的每个项目分别冠以序号- 在脑电功率涨落信号分析基础上进行的, (1)熵值运算和 (2)单频竞争分析; 在 S谱系分析基础上进行的, (3)S谱总谱分析; (4)基本谱系分析; (5)最优 值分析; (6) A/P逆转; L/R失衡情况分析; (7)特频分析; (8)异频分析; (9)连频 分析; 0Φ最优频分析; CD S谱系功率空间分布分析; (1单频功率及相对值(L/R) 分布分析; (13)平均功率分布分析; (14)功率相对值 A/P、 L/R分析; (15)长时程 S 谱系曲线分析; (16)长时程优势谱线曲线分析; (17)长时程基本谱系曲线分析; (18) 长时程功率空间分布曲线分析;(19)长时程熵值曲线分析;(2Φ长时程特频曲线分 析: (21 )长时程连频曲线分析; (22)长时程基本谱系功率空间分布曲线分析; ( 23 ) 长时程常规功率谱曲线分析; (24) 长时程事件标记识别。
上述 (7)特频分析、(8)异频分析、(9)连频分析、(1Φ最优频分析四项目可集称 为特征谱线分析。
上述 (1 单频功率分布分析、 (13)平均功率分布分析、 (14)功率相对值 A/P、 L/R 分析三项目可集称为功率分布分析。
上述 (15)长时程 S谱系曲线分析、(16)长时程优势谱线曲线分析、(1长时程基 本谱系曲线分析、 (18)长时程功率空间分布曲线分析、 (19)长时程熵值曲线分析、 (20)长时程特频曲线分析、 (21 )长时程连频曲线分析、 (22)长时程基本谱系功 率空间分布曲线分析、 (23)长时程常规功率谱曲线分析、 (24)长时程事件标 i己识别十个项目可集称为长时程动态曲线分析。
以下对第二层次各分析方法作叙述。
所述 (1)熵值运算是在脑电功率涨落信号分析基础上进行的, 其方法为,
13
( 1 ) 根据 计算熵, 为脑电波中各个频率占优的概率;
(2)综合所有 N个导联总概率分布进行熵值计算得到总熵(总数为 n*N:)。 所述 (2)单频竞争分析也是在脑电功率涨落信号分析基础上进行的,其方法 为, 把脑电涨落信号中的频率涨落图 f (n)中相同的最优频率数目随时间变化 过程累加, 得到最优频的概率曲线。
以下所述序号为 (3)〜(24)第二层次分析方法都是在 S谱系分析基础上进 行的。
所述 (3)S谱总谱分析方法为, 将 S谱系分析中得到的 S谱总谱数据用图形 表示出来。
所述 (4)基本谱系分析方法为, 对 S谱中基频对应的 S谱系进行统计分析, 从频率大于 3 mHz 开始累加其倍周期频率的数值(如统计 3πιΗΖ 时还应累加 6mHz、 9mHz, ……的数值); 同时把所有导联按照在头部的放置位置分成分成 前后、 左右若干部分分别进行统计。
所述 (5)最优值分析方法为,将优势频率的功率数值和相应频率按导联的空 间位置分布显示。
所述 (6)A/P逆转、 L/R失衡情况分析方法为, 按导联空间分布位置计算每 一频率功率值的前后比值 A/P, 并将 A/P值大于一定限值的频率显示出来; 同 吋计算左右比值 L/R, 将 L/R值大于一定限值的频率显示出来。
所述 (7)特频分析、(8)异频分析、(9)连频分析、(1Φ最优频分析四个项目, 其 方法为,分别将每导联的特频、异频、连频、最优频按导联空间分布位置显示。
所述 (IDS谱系功率空间分布分析方法为, 将脑电涨落图中的每条谱线的功 率值按空间导联位置分布排列, 在显示界面上以 "画中画 "方式开一窗口, 进 ί亍谱线的选择,将被选定的谱线,按导联的空间位置分布显示其在每个导联下 的功率数值。 所述 (12)单频功率及相对值 (L/R) 分布分析方法为, 将每导联优势谱线 Dl-Dn相应的功率值相加, 得到每导联的总功率值; 将基频的功率及左右比值 (L/R) 中大于限定值或小于 1/限定值的数值按导联的空间分布显示出来。
所述 (13)平均功率分布分析方法为,将每个导联平均功率按导联空间分布位 置显示出来。
所述 (M)功率相对值 A/P、 L/R分析方法为, 按导联空间分布位置计算功率 值的前后比值和左右比值。
所述 (15)长时程 S谱系曲线分析方法为,分别以每个导联下或全部导联的每 一个谱系或每一条谱线的涨落数值为纵轴, 以时间为横轴做出曲线。在显示界 面上以 "画中画"方式开一窗口, 进行谱线或谱系选择。
所述 (16)长时程优势谱线曲线分析方法为,以进入优势谱线区的谱线的频率 为纵轴,以日寸间为横轴做出曲线。 在显示界面上以 "画中画"方式开一窗口, 进行优势谱线排名序次 (Dl- Dn)选择。 .
所述 (17)长时程基本谱系曲线分析方法为,分别以所有导联或每个导联下的 基本谱系备个时间段的涨落数值为纵轴, 以时间为横轴做出基本谱系曲线。在 显示界面上以 "画中画"方式开一窗口, 进行谱系选择。
所述 (18)长时程功率空间分布曲线分析方法为,每一个导联下,每一条谱线, 以其功率值为纵轴, 以时间为横轴做出曲线。在显示界面上以 "画中画"方式 开一窗口, 进行谱线选择。
所述 (19)长时程熵值变化曲线分析方法为,以全部导联或每个导联下的熵值 为纵轴, 以时间为横轴分别做出曲线, 反映熵值随时间的变化。
所述 (20)长时程特频曲线分析方法为,分别以所有导联或每个导联下特频出 现的个数为纵轴, 以时间为横轴做出曲线,用于观察特频随时间动态变化的情 况。
所述(21 )长时程连频曲线分析方法为, 分别以所有导联或每个导联下连 频出现的个数为纵轴, 以时间为横轴分别做出曲线,显示连频随时间动态变化 的情况。
所述(22)长时程基本谱系功率空间分布曲线分析方法为, 从单频功率及 相对值分布(L/R)分析中读取基本谱系每个导联下的功率值,以功率值为纵轴, 以时间为横轴做出曲线,按导联的空间位置分布显示出每条曲线。在显示界面 上以 "画中画"方式开一窗口, 进行谱系选择。
所述(23 )长时程常规功率谱曲线分析方法为, 从常规功率谱中找出幅值 最大的 n个频率 Dl-Dn, 按幅值从大到小排序, 分别以这些频率的功率值为纵 轴, 以时间为横轴, 作出 n条动态曲线。
所述(24 )事件标记识别方法为,辨认脑电记录盒所记录的事件标记信号, 将这些标记信号在脑电信号回放及各种动态曲线时间轴的相应位置显示标记 符号。
以上所述第二层次分析方法中, 共列举 24种分析方法。 各方法的选择和 组合应用可有如下方案:
( 1 )所有方法同时应用;
(2) 除 "事件标记识别"夕卜, 选择任何一个项目单独应用;
( 3) 除 "事件标记识别"夕卜, 其它项目作任意组合应用;
(4)事件标记识别与长时程动态曲线分析的其它九个项目中的任何一个 或几个项目结合进行。
以上所述分析方法可应用于对任意一个或几个导联组合所采集的脑 电信号进行数据处理,并且可以选择将分析结果中一个或几个导联组合的运算 结果输出到终端处理器进行显示、 打印或存储。
本发明还公开了一种对脑电涨落信号进行分析的设备。它包括电极、数字 化脑电信号放大器或脑电记录盒、个人计算机、 数据处理器、终端处理器。它 们依次连接。 其中, 电极采集脑电信号, 信号传送到数字化脑电信号放大器和 /或脑电信号 i己录盒,对信号接受、放大、数模转换、数字滤波或 /和数据存储, 数字化脑电信号放大器和 /或脑电信号记录盒中的数据上传至个人计算机, 由 与计算机相连的数据处理器完成数据处理和涨落分析,分析结果传输到终端处 理器进行存储、 显示或打印。
所述电极的放置方法可选取任意一种导联连接方法中的一个或几个导联 的组合。
所述数据处理器包括进行第一层次分析的常规功率谱分析模块、脑电功率 涨落信号分析模块、 脑电涨落图分析模块和 S谱系分析模块。 它们依次相连, 上一模块产生的数据传送到下一模块进行运算分析。数据处理器还可包括进行 第二层次分析的以下 24个模块中的任意个模块。 第二层次分析模块接受第一 层次分析模块的分析结果(数据)作进一步分析。 为叙述方便, 给第二层次分 析模块分别冠以序号- 与脑电功率涨落信号分析模块相连并接受其数据的 (1)熵值运算模块和 (2) 单频竞争分析模块;
与 S谱系分析模块相连并接受其数据的 (3)S谱总谱分析模块; (4)基本谱系 分析模块; (5)最优值分析模块; (6) A/P逆转、 L/R失衡情况分析模块; (7)特频 分析模块; (8)异频分析模块; (9)连频分析模块; (1Φ最优频分析模块; (11) S谱 系功率空间分布分析模块; (12)单频功率分布分析模块;(13)平均功率分布分析模 块; (14)功率相对值 A/P、 L/R分析模块; (15)长时程 S谱系曲线分析模块; (16)长 时程优势谱线曲线分析模块;(17)长时程基本谱系曲线分析模块;(18)长时程功率 空间分布曲线分析模块;(1Φ长时程熵值曲线分析模块;(2Φ长时程特频曲线分析 模块; (21 )长时程连频曲线分析模块; (22 )长时程基本谱系功率空间分布曲 线分析模块; (23 )长时程常规功率谱曲线分析模块; (24)长时程事件标记识 别模块。
以下对两层次各分析模块分别叙述。 在第一层次分析模块中,
所述常规功率谱分析模块,用于对给定时间长度的时域脑电信号做功率谱 分析, 得到功率谱, 可表示为
Figure imgf000010_0001
所述脑电功率涨落信号分析模块, 用于选择边瓣幅值小且衰减快的窗函 ω(77) = 0.5— 0.5 cos( )," = 0,l ..N— l. t、 数,它表示为 N ,对脑电信号 数据进行截短, 得到 ΧΝ (") = χ(η)ω{ );
对上述信号 (")做功率谱分析, 采样时间为 Τ秒, 计算 FFT得 到频谱的频域分辨率为 1/T, 选择出能量最大的功率幅值 ρ及相应的频率 f ; 按时间顺序对总的时间长度为 N秒的时域脑电信号按 T秒进行分段,依次对每 段数据进行上述常规功率谱分析模块的分析、脑电功率涨落信号分析模块的分 析, 得到最大功率幅值的时间涨落信号 p (n)和相应频率涨落信号 f (n), n=l n, n=N/T。
所述脑电涨落图分析模块, 用于对长度为 n 点的最大功率幅值涨落信号 P (n)进行分析。 乘上长度为 n的 Harming窗,再做功率谱分析, 数据时间长度 单位为 N秒, 因而频域分辨率为 1/N Hz, 取功率谱分析结果中频带一定范围 的谱线组成脑电涨落图。当总采样时间大于 N秒时, 以 N秒为单位进行段落划 分,对每一段分别进行上述常规功率谱分析模块、脑电功率涨落信号分析模块、 脑电涨落图分析模 ±夬的运算过程。
所述 S谱系分析模块,用于从每一个导联脑电涨落图中找出幅值最大的几 条优势谱线 Dl-Dn, 按幅值从大到小排序, 得到单一导联数据的 S谱, 共 n个 数值;综合所有 N导联共得到 N*n条优势谱线,把频率相同的优势谱线的数目 累加, 得到 S谱总谱。
以下对第二层次分析模块分别叙述。
第一层次中的脑电功率涨落信号分析模块分析所得数据传输给 ω熵值运 算模块和 (2)单频竞争分析模块, 做第二层次分析。 其中,
13
H = - V pk lg2 pk
所述熵值运算模块, 用于根据 ^ 来计算熵, ^为脑电波中各 个频率占优的概率; 综合所有 N个导联总概率分布进行熵值计算得到总熵(总 数为 n*N)。
所述单频竞争分析模块用于把脑电涨落信号中的频率涨落图 f (n)中相同 的最优频率数目随时间变化过程(1-n数据段)累加,得到最优频的概率曲线。
第一层次中的 S 谱系分析模块分析所得数据传输给以下第二层次分析模 块:
所述 (3)S谱总谱分析模块, 将从 S谱系分析模块中得到的 S谱总谱数据用 图形表示出来。 '
所述 (4)基本谱系分析模块,用于对 S谱中基频对应的 S谱系进行统计分析, 从频率大于 3 mHz开始累加其倍数频率的数值; 同时把所有导联按照在脑部的 分布位置分成前后左右若干部分分别进行统计。
所述 (5)最优值分析模块, 用于按导联的空间位置分布显示每个导联下的 Dl-Dn优势频率的功率数值和相应频率。
所述 (6)A/P逆转、 L/R失衡情况分析模块, 用于按导联空间分布位置计算 每一个频率功率俊的前后比值 A/P,并将 A/P值大于一定限值的频率显示出来; 同时计算左右比值 L/R, 将 L/R值大于一定限值的频率显示出来。
所述 (7)特频分析模块、(8)异频分析模块、(9)连频分析模块、(1Φ最优频分析 模块, 分别用于将每导联的特频、异频、连频、每个最优频按导联空间分布位 置显示出来。
所述 (IDS谱系功率空间分布分析模块, 用于将脑电涨落图中的每条谱线的 功率值按空间导联位置分布排列, 在显示界面上以 "画中画"方 开一窗口, 进行谱线的选择, 将被选定频率的谱线,按导联的空间位置分布显示出每个导 联下的功率数值。
所述 (12)单频功率及相对值分布 L/R分析模块, 将每导联优势谱线 Dl-Dn 相应的功率值相加, 得到每导联的总功率值; 将基频的功率及左右比值(L/R) 中大于限定值或小于 V限定值的数值按导联的空间分布显示出来。
所述 (13)平均功率分布分析模块,将每个导联平均功率按导联空间分布位置 显示出来。
所述 (14)功率相对值 A/P、 L/R分析模块, 按导联空间分布位置计算功率值 的前后比值和左右比值。 ·
所述 (15)长时程 S谱系曲线分析模块,分别以每个导联下或全部导联的每一 个谱系或每一条谱线的涨落数值为纵轴, 以时间为横轴, 做成曲线; 在显示界 面上以 "画中画"方式开一窗口, 进行谱线或谱系选择。
所述 (16)长时程优势谱线曲线分析模块,用于以进入优势谱线区的谱线的频 率为纵轴,以时间为横轴做成曲线。在显示界面上以 "画中画"方式开一窗口, 进行优势谱线排名序次 (Dl- Dn)的选择。
所述 (17)长时程基本谱系曲线分析模块,用于分别以所有导联或每个导联下 的基本谱系各个时间段的涨落数值为纵轴, 以时间为横轴做成基本谱系曲线; 在显示界面上以 "画中画"方式开一窗口, 进行谱系的选择。
所述 (18)长时程功率空间分布曲线分析模块,用于将每一个导联下的每一条 谱线,以其功率值为纵轴,以时间为横轴做出曲线。在显示界面上以 "画中画" 方式开一窗口, 进行谱线的选择。
所述 (19)长时程熵值变化曲线分析模块,用于将全脑或每个导联下的熵值为 纵轴, 以时间为横轴分别做出曲线。
所述 (2Φ长时程特频曲线分析模块,用于以所有导联或每个导联的特频出现 的个数为纵车由, 以时间为横轴做出曲线。
所述(21 )长时程连频曲线分析模块, 以所有导联或每个导联下的连频出 现的个数为纵轴, 以日寸间为横轴分别做出曲线。
所述(22)长时程基本谱系功率空间分布曲线分析模块, 用于从单频功率 及相对值分布(L/R)分析中读取基本谱系的每个导联下的功率值,以功率值为 纵轴, 以时间为横轴作出曲线, 按导联的空间位置分布显示出每条曲线。在显 示界面上以 "画中画" 方式开一窗口, 进行谱系的选择。
所述(23 )长时程常规功率谱曲线分析模块, 用于从常规功率谱中找出幅 值最大的 n个频率 D1 - Dn, 按幅值从大到小排序, 分别以这些频率的功率值为 纵轴, 以时间为横轴, 作出 n条动态曲线。
所述(24)长时程事件标记模块, 用于辨认脑电记录盒所记录的事件标记 信号,将这些标记信号在脑电信号回放及各种动态曲线时间轴的相应位置显示 标记符号。 它分别与各长时程分析模块连接。
所述数据处理器对电极所采集的脑电信号,可选择其中任意一个或几个导 联组合的数据进行处理。
以上所述担负第二层次分析的模块, 可以同时装配同时使用, 也可以将 ( 24)事件标记识别模块除外,其它 (1)〜(23 )所有模块作任意组合装配使用。
以上所述分析方法可应用于对任意一个或几个导联组合所采集的脑 电信号进行数据处理,也可以选择将分析结果中一个或几个导联组合的运算结 果输出到终端处理器。
所述终端处理器由显示器、 打印机和存储设备 (如硬盘、 软盘、 光盘等) 组成,用于接受数据处理器的信号, 并选择数据处理器的运算结果中一个或几 个导联组合的运算结果, 进行存储、 显示或打印。
根据本发明的方法和装置经过检测和运算显示出来的数据和曲线,可分析 脑的功能状态, 分析 ¾病患者脑内神经递质的状况, 能够分析 CT和核磁共振 无能为力的大脑功能性变化,为医学界对功能性脑病的诊断提供直接的客观指 标,弥补医学界对于功能性脑病(如精神病)的诊断缺少客观检测指标的空白。 附图说明
图 1为本发明的第一个实施例的方框图。
图 2为本发明的第二个实施例的方框图;
图 3为本发明的第三个实施例的方框图; 图 4为本发明的数据处理器的结构方框图。
其中方框之间以箭头连接者表示逻辑关系:在前模块运算所得数据是在后 模块运算的基础,在后模块的运算依赖于在前模块运算所得的数据。方框之间 以直线连接者表示包含关系:在前方框所示模块由在后多个模块构成。如在前 方框仅仅是在后方框所示多个模块所属模块类别的概全名称, 则以虚线框之。 具体实施方式
以下结合附图对本发明作进一步说明。
本发明对脑电涨落进行分析的实施过程为,用计算机涨落扫描技术从脑电 波中提取出脑电涨落信号, 再对涨落信号进行频谱分析, 从而获得 l_255mHZ 范围内的功率谱。 具体的分析过程可以概括为: 对 1024秒脑电数据按每段 2 秒进行分段, 共分成 512段; 对每段数据进行功率谱分析, 选择出 0. 5-50Hz 范围内的最大功率幅值; 对最大功率幅值的时间涨落过程(即 512个数据段相 应的功率谱的最大值) 进行功率谱分析, 获得 1- 255mHz范围内的功率谱涨落 图。
本发明设备的构成可以有以下三种类型:
如附图 1所示, 包括依次连接的电极 A、 数字化脑电信号放大器 B、 USB 接口 J、 个人计算机 C、 数据处理器 D、 终端处理器 E。
如附图 2所示, 包括依次连接的电极 A、 脑电记录盒 F、 USB接口 J、 个人 计算机 C、 数据处理器 D、 终端处理器 E。
如附图 3所示, 包括电极八、 数字化脑电信号放大器 B和脑电记录盒 F、 USB接口 J、 个人计算机 C、 数据处理器 D、 终端处理器 E。 其中电极 A同时连 接数字化脑电信号放大器 B和脑电记录盒 F, 数字化脑电信号放大器 B直接与 个人计算机 C连接, 而脑电记录盒 F则通过 USB接口 J与个人计算机 C连接。 个人计算机 C再顺序连接数据处理器 D和终端处理器 E。
对本发明设备各部件的功能作用分别叙述如下。
电极 A: 用于采集脑电信号。 电极放置采用国际标准导联系统的 12导联, 位置分别为 F3、 F4、 C3、 C4、 P3、 P4、 01、 02、 F7、 F8、 T5、 T6。 脑电釆样 率为 128Hz。
数字化脑电信号放大器 B: 有信号接受、 信号放大、 数模转换、 数字滤波 等功能。
脑电记录盒 F: 用于采集、 分析长时程 (超过 18分钟) 的脑电数据, 有 信号采集、信号放大、 数模转换、 数字滤波、 数据存储、 数据回放的功能。 脑 电记录盒中的数据上传至个人计算机进行涨落分析。
个人计算机 C: 主机处理器采用 PIV型, 内存 256M。
终端处理器 E: 由显示器、 打印机和存储设备 (如硬盘、 软盘、 光盘等) 组成, 接受数据处理器 D的运算结果并进行存储、 显示或打印。
数据处理器 D: 所述数据处理器包括进行第一层次分析的常规功率谱分析 模块 Dl、脑电涨落信号分析模块 D2、脑电涨落图分析模块 D3和 S谱系分析模 块 D4。 它们依次相连, 上一模块产生的数据传送到下一模块。 数据处理器还 可包括进行第二层次分析的以下 24个模块中的任意个模块。 为叙述方便, 给 第二层次分析模块分别冠以序号:
与 D2模块相连并接受其数据的 (1)熵值运算模块 D5和 (2)单频竞争分析模块
D6;
与 D4模块相连并接受其数据的 (3)S谱总谱分析模块 D4a; (4)基本谱系分析 模块 D4b; (5)最优值分析模块 D4c; (6) A/P逆转、 L/R失衡情况分析模块 D4d; (7)特频分析模块 D4ea; (8)异频分析模块 D4eb ; (9)连频分析模块 D4ec; (10)最优 频分析模块 D4ed; (11) S谱系功率空间分布分析模块 D4f; (12)单频功率分布分 析模块 D4ga; (13)平均功率分布分析模块 D4gb; (14)功率相对值 A/P、 L/R分析模 块 D4gc; (15)长时程 S谱系曲线分析模块 D4ha; (16)长时程优势谱线曲线分析模 块 D4hb; (17)长时程基本谱系曲线分析模块 D4hc; (18)长时程功率空间分布曲线 分析模块 Mhd; (19)长时程熵值变化曲线分析模块 D4he; (20)长时程特频曲线分 析模块 D4hf; ( 21 ) 长时程连频曲线分析模块 D4hg; (22) 长时程基本谱系功 率空间分布曲线分析模块 D4hh; (23 ) 长时程常规功率谱曲线分析模块 D4hi ; ( 24) 长时程事件标记识别模块 D4hj。
其中, 上述 (7)特频分析模块 D4ea、 (8)异频分析模块 D4eb、 (9)连频分析模 块 D4ec、 (1Q)最优频分析模块 D4ed四个模块集为一体, 称为特征谱线分析模块 D4e。
上述 (12)单/频功率分布分析模块 D4ga、 (13)平均功率分布分析模块 D4gb、 (14) 功率相对值 A/P、 L/R分析模块 D4gc三个模块集为一体,称为功率分布分析模 块 D4g0
上述 (15)长时程 S谱系曲线分析模块 D4ha、 (16)长时程优势谱线曲线分析模 块 D4hb、 (17)长时程基本谱系曲线分析模块 D4hc、 (18)长时程功率空间分布曲线 分析模块 D4hd、 09)长时程熵值变化曲线分析模块 D4he、 (20)长时程特频曲线分 析模块 D4hf、 (21 ) 长时程连频曲线分析模块 D4hg、 (22) 长时程基本谱系功 率空间分布曲线分析模块 D4hh、 (23)长时程常规功率谱曲线分析模块 D4hi、 (24) 长时程事件标记识别模块 D4hj十个模块集为一体, 称为长时程动态曲 线分析模块 D4h。
对所述各模块的功能作用分别叙述如下。
1. 常规功率谱分析模块 D1 把随机信号 的 N点观察数据 (")直接做傅里叶变换,得到 ^^(^ω),然 后再取其幅值的平方, 并除以 Ν, 作为对 的真实的功率谱 )的估计。 用周期图法估计出的功率谱可以表示为: ^^) = ^^^)1 。 ^ 的计算由快速傅里叶变换 (FFT) 计算得到:
XN (k) = V (n)WN , k = 0,1, ..., N - \, WN = e N
"=。 , 功 率 谱 可 表 示 为 :
Figure imgf000016_0001
对给定时间长度为 8秒的时域脑电信号做功率谱分析,得到其在频域的能 量分布情况即常规功率谱, 并将功率谱以图形方式传输给终端处理器 E。
2. 脑电功率涨落信号分析模块 D2
在实际估计功率谱过程中, 在选择窗函数时, 选取主瓣窗, 边瓣幅值小且 衰减快的窗函数。 Hanning (汉宁窗)主瓣稍宽, 但有较小的边瓣和较大衰减 ω(η) = 0.5 - O.5 cos(— ), « = 0,1,...N - 1. _
速度, 它表示为: N , 用 Harming窗对数据进 行截短, 得到 ΧΜ(") = λ»(")。
对给定时间长度为 2秒的时域脑电信号乘上 Harming窗后做功率谱分析, 采样时间为 2秒, 因而计算 FFT得到频谱的频域分辨率为 l/T=l/2=0. 5Hz (T 为采样时间),在 8-13Hz的频带范围内(8Hz, 8. 5Hz, 9Hz, 9. 5Hz, 10Hz, 10. 5Hz, 11Hz, 11. 5Hz, 12Hz, 12. 5Hz, 13Hz, 共 11个数值) 选择出能量最大的功率 幅值 P及相应的频率 f。
按时间顺序对总长度为 1024秒的时域脑电信号按每段 2秒进行分段, 共 分成 512段。 依次对 512段数据进行常规功率谱分析模块 Dl、 脑电涨落信号 分析模块 D2的运算, 得到最大功率幅值的时间涨落信号 p (n)和相应频率涨落 信号 f (n),n=l…… 512。 它表示了 1024秒时间内功率及相应最大幅值的涨落 过程。
将时间涨落信号传输给脑电涨落图分析模块 D3、 熵值运算分析模块 D5、 单频竞争分析模块 D6, 同时将信号传输给终端处理器£。
3. 脑电涨落图分析模块 D3
对长度为 512点(1024秒)的最大功率幅值涨落信号 p (n)进行分析。 乘上 长度为 512的 Harming窗, 再做功率谱分析。 数据时间长度单位为 1024秒, 因而频域分辨率为 1/1024 Hz , 取功率谱分析结果中频带范围 1/1024* (1-255) Hz的谱线组成脑电涨落图。 如果总采样时间大于 1024秒, 则 以 1024秒为单位进行段落划分, 对每一段重复进行上述常规功率谱分析模块 Dl、 脑电功率涨落信号分析模块 D2、 脑电涨落图分析模块 D3运算过程。
脑电涨落图分析模块 D3将脑电涨落图信号传输给 S谱系分析模块 D4, 同 时也传输给终端处理器£。
4. S谱系分析模块 D4
从每一个导联脑电涨落图中找出幅值最大的 8条优势谱线 Dl- D8, 按幅值 从大到小排序, 得到单一导联数据的 S谱, 共 8个数值。 综合 12导联共得到 12 X 8=96条优势谱线, 把频率相同的优势谱线的数目累加, 得到 S谱总谱。
S谱系分析模块 D4将单一导联 S谱及 S谱总谱信号传输给 S谱总谱分析 模块 D4a、 基本谱系分析模块 D4b、 最优值 D ( 1-8) 分析模块 D4c、 A/P逆转、 L/R失衡情况分析模块 D4d、特征谱线分析模块 D4e、 S谱系功率空间分布分析 模块 D4f、功率分布分析模块 D4g、 长时程动态曲线分析模块 D4h。 同时也将 S 谱总谱信号传输给终端处理器 E。
5. 熵值运算模块 D5 根据 ^ 来计算熵, 为脑电波中各个频率占优的概率 (总数 为 512, 概率为单个频率占优数目除以 512)。 综合所有 12个导联总概率分布 进行熵值计算得到总熵 (总数为 512*12)。 将数据传输给终端处理器 E。
6. 单频竞争分析模块 D6
把脑电涨落信号中的频率涨落图 f (n)中相同的最优频率数目随时间变化 过程(1-512数据段) 累加, 得到最优频的概率曲线。
7. S谱总谱分析模块 D4a
将 S谱系分析模块 D4中所产生的 S谱总谱信号做成图形, 并将图形及数 据传输给终端处理器 E。
8. 基本谱系分析模块 D4b
对 S谱中 lmHz、 2mHz、 3mHz、 4mHz、 5mHz、 6mHz、 7mHz、 llmHz、 13mHz 共 9个基频对应的谱系 (Sl、 S2、 S3、 S4、 S5、 S6、 S7、 Sll、 S13)进行统计 分析, 从频率大于 3 mHz幵始还应累加其倍数频率的数值 (如统计 3mHz时还应 累加 6mHz、 9mHz, ……的数值)。 同时把 12导联按照大脑位置分成左前 (F3, F7, C3)、 左后 (T5, Ρ3, 01)、 右前 (F4, C4, F8)、 右后 (P4, T6, 02) 四 个部分分别进行统计。 将统 if结果传输给终端处理器 E。
9. 最优值 (D1-8) 分析模块 D4c
按导联的空间位置分布显示出每个导联下的 Dl- D8 最优频率的功率数值 和相应频率, 并将结果传输给终端处理器 E。
10. A/P逆转、 L/R失衡情况分析模块 D4d
按导联空间分布位置计算每一个频率的功率值的前后比值 (如 F3/C3, C3/P3 ) A/P,将 A/P值大于 10的频率显示出来; 同时计算左右比值 (如 F3/F4, C3/C4) L/R, 将 L/R值大于 10的频率显示出来。
11.特征谱线分析模块 Me:包括特频分析模块 D4ea、异频分析模块 Meb、 连频分析模块 D4ec、 最优频分析模块 D4ed。 其中,
( 1 )特频分析模块 D4ea: 将每导联的特频按导联空间分布位置显示, 将 结果传输给终端处理器£。
(2)异频分析模块 D4eb : 将每导联的异频按导联空间分布位置显示, 将 结果传输给终端处理器 E。 ( 3)连频分析模块 D4ec: 将每导联的连频按导联空间分布位置显示, 将 结果传输给终端处理器E。
(4)最优频分析模块 D4ed: 将每个最优频按导联空间分布位置显示其相 应的频率, 将结果传输给终端处理器 E。
12. S谱系功率空间分布分析模块 D4f
从脑电涨落图中选择出指定频率的某一谱线,按导联的空间位置分布显示 出每个导联下的功率数值。 将功率空间分布情况传输给终端处理器 E。
13.功率分布分析模块 D4g: 包括单频功率及相对值分布(L/R)分析模块 D4ga、 平均功率分布分析模块 D4gb、 A/P、 L/R分析模块 D4gc。
将每导联最优谱线 Dl- D8相应的功率值相加,得到每导联的总功率值,将 结果分别传输给单步页功率及相对值分布 (L/R)分析模块 D4ga、 平均功率分布 分析模块 D4gb、 A/P、 L/R分析模块 D4gc。
( 1 )单频功率及相对值分布(L/R)分析模块 D4ga: 将基频的功率及 L/R 比值中大于 10或小于 0. 1的数值按导联的空间分布显示出来, 将结果传输给 终端处理器 E。
( 2)平均功率分布分析模块 D4gb: 将每个导联平均功率按导联空间分布 位置显示出来, 将结果传输给终端处理器 E。
( 3 ) A/P、 L/R分析模块 D4gc: 按导联空间分布位置计算功率值的前后比 值 (如 F3/C3, C3/P3 ) 和左右比值 (如 F3/F4, C3/C4) , 将结果传输给终端处 理器 E。
14.长时程动态曲线分析模块 D4h
包括 S谱系曲线分析模块 D4ha、 优势谱线曲线分析模块 D4hb、 基本谱系 曲线分析模块 D4hc、 功率空间分布曲线分析模块 D4hd、 熵值曲线分析模块 D4he、 特频曲线分析模块 D4hf、 连频曲线分析模块 D4hg、 基本谱系功率空间 分布曲线分析模块 D4hh、 常规功率谱曲线分析模块 D4hi、 事件标记识别模块 D4hj。
把长时程(采集时间大于 18分钟)的脑电数据按照 18分钟的长度进行分 段, 对每段数据进行月 If涨落图分析 (重复步骤 D1-D4) , 将分析结果传输给以 下的分析模块。
( 1 )优势谱线曲线分析模块 D4ha: 用于以进入优势谱线区的谱线的频率 为纵轴,以时间为横轴做成曲线;在显示界面上以 "画中画"方式开一窗口,进 行优势谱线排名序次 (Dl- Dn)的选择;
(2) S谱系曲线分析模块 D4hb: 分别以每个导联下或全部导联的每一个 谱系或每一条谱线的涨落数值为纵轴, 以时间为横轴, 做成曲线; 在显示界面 上以 "画中画" 方式开一窗口,进行谱线或谱系选择。据此观察每一频率谱线 或谱系在整个大脑中随时间变化的情况。
(3)基本谱系曲线分析模块 D4hc: 分别以所有导联或每个导联下的基本 谱系各个时间段的涨落数值为纵轴, 以时间为横轴做成基本谱系曲线,在显示 界面上以 "画中画"方式开一窗口,进行谱系的选择。用以掌握与大脑功能关 系最密切的几个基本谱系的动态情况。
(4) 功率空间分布曲线分析模块 D4hd: 每一个导联下, 每一条谱线, 以 其功率值为纵 $由, 以时间为横轴描绘曲线。在显示界面上以 "画中画"方式开 一窗口, 进行谱系的选择。据此观察每一个导联下, 每一条谱线功率值随时间 变化的情况。
( 5) 熵值曲线分析模块 D4he: 将所有导联或每个导联下的熵值为纵轴, 以时间为横轴分别做出曲线, 反映熵值随时间的变化, 据此观察不同时间大脑 能量分布状况。
(6)特频曲线分析模块 D4hf: 以所有导联特频出现的个数为纵轴, 以时 间为横轴做出曲线, 据此观察特频随时间动态变化的情况。
(7)连频曲线分析模块 D4hg: 以所有导联或每个导联下的联频出现的个 数为纵轴,以吋间为横轴分别做出曲线,观察病人不同时间脑功能下降的情况。
(8)基本谱系功率空间分布曲线 D4hh: 从单频功率及相对值分布 (L/R) 分析模块 D4ga中读取基本谱系的每个导联下的功率值,以功率值为纵坐标,以 时间为横坐标作出曲线, 按导联的空间位置分布显示出每条曲线。在显示界面 上以 "画中画" 方式开一窗口,进行谱系的选择。据此观察每个基本谱系随时 间变化及空间分布的情况。
(9) 常规功率谱曲线分析模块 D4hi : 从常规功率谱中找出幅值最大的 8 个频率值 D1-D8, 按幅值从大到小排序, 分别以这些频率的功率值为纵轴, 以 时间为横轴,作出 8条动态曲线,据此观察脑电中功率占优的频率随时间变化 的情况。 ( 10) 事件标记识别模块 D4hj : 它分别与各长时程分析模块连接, 用以 辨认脑电记录盒 F所记录的事件标记信号,将这些标记信号在脑电信号回放及 各种动态曲线上时间轴的相应位置显示标记符号。 工业应用性
根据本发明的方法和装置经过检测和运算显示出来的数据和曲线,可分析 脑的功能状态, 分析脑病患、者脑内神经递质的状况, 能够分析 CT和核磁共振 无能为力的大脑功能性变化,为医学界对功能性脑病的诊断提供直接的客观指 标,弥补医学界对于功能†生脑病(如精神病)的诊断缺少客观检测指标的空白。

Claims

权 利 要 求 书
1. 一种应用计算机技术对脑电涨落信号进行分析的方法, 其特征在于, 先对脑电信号数据按一定时间长度进行分段,对每段数据进行功率谱分析,选 择出 0. 5- 50Hz范围内的最大功率幅值,进行功率谱分析及频谱分析,获得超慢 波范围内的功率谱涨落图,再对涨落图进行一系列分析,获得一系列数据参数, 以数值、图形和曲线的方式显示出来,其具体分析方法最少包括常规功率谱分 析的如下步骤:
( 1 )对给定吋间长度的时域脑电信号做功率谱分析,即对脑电信号 的 N点观察数据 直接做傅里叶变换, 得到 W ) ;
(2)取其幅值的平方, 并除以 N, 作为对 的真实的功率谱 )的估 计, 用周期图法估计出的功率谱可以表示为, ) = Ν ^ Αω)^;
(3) 的计算由快速傅里叶变换计算得到:
X k = x n "k k = 0 1 . N - e
Figure imgf000022_0001
2. 如权利要求 1所述的脑电涨落信号分析方法, 其特征在于, 还包括脑 电功率涨落信号分析, 其具体方法包括如下步骤:
( 1 ) 选择边瓣幅值小且衰减快的窗函数, 它表示为 ,
Figure imgf000022_0002
对脑电信号 XW数据进行截短, 得到
(2)对信号 做功率谱分析, 采样时间为丁秒, 计算 FFT得到 频谱的频域分辨率为 1/T, 选择出能量最大的功率幅值 ρ及相应的频率 f;
( 3) 按时间顺序对总时间长度为 N秒的时域脑电信号按 T秒进行分段, 依 次对 N/T段数据进行常规功率谱分析和脑电涨落信号分析,得到最大功率幅值 涨落信号 P (n)和木目应频率涨落信号 f (n), n=l…… n, n=N/T。
3. 如权利要求 2所述的脑电涨落信号分析方法, 其特征在于, 还包括脑 电涨落图分析, 其具体方法包括如下步骤:
( 1 ) 对长度为 n点的最大功率幅值涨落信号 p (n)进行分析;
(2)乘上长度为 n的 Hanning窗, 再做功率谱分析, 数据时间长度单位为 N秒, 因而频域分辨率为 1/N HZ, 取功率谱分析结果中一定频率范围的谱线组 成脑涨落图;
(3) 当总采样时间大于 N秒时, 以 N秒为单位进行段落划分, 对每一段分 别进行上述常规功率谱分析、 脑电涨落信号分析和脑电涨落图分析。
4. 如权利要求 3所述的脑电涨落信号分析方法, 其特征在于, 还包括 S 谱系分析, 其具体方法包括如下步骤:
( 1 ) 从每一个导联脑电张落图中找出幅值最大的几条优势谱线 Dl- Dn, 按 幅值从大到小排序, 得到单一导联数据的 S谱, 共 n个数值;
(2) 综合所有 N导联共得到 N*n条优势谱线, 把频率相同的优势谱线的数 目累加, 得 S谱总谱;
(3) 当总采样时间大于 N秒时, 以 N秒为单位进行段落划分, 对每一段分 别进行常规功率谱分析、脑电功率涨落信号分析、脑电涨落图分析和 S谱系分 析, 分别以各项分析结果的数值为纵轴, 以时间为横轴做出曲线。
5. 如权利要求 1所述的脑电涨落信号分析方法, 其特征在于, 脑电信号 釆集途径包括应用任何导联或导联组合。
6. 如权利要求 5所述的脑电涨落信号分析方法, 其特征在于, 脑电信号 的采集采用国际标准导联系统的 12导联, 电极置放位置分别为 F3、 F4、 C3、 C4、 P3、 P4、 01、 02、 F7、 F8、 T5、 T6。
7. 如权利要求 2所述的脑电涨落信号分析方法, 其特征在于, 还包括在 脑电功率涨落信号分析基 ί出上所进行的第二层次分析项目一一熵值运算和单 频竞争分析方法中的一种或两种, 其中, 所述熵值运算包括如下步骤:
13
( 1 )根据 ^ 计算熵, A为脑电波中各个频率占优的概率;
(2) 综合所有 N个导联 、概率分布进行熵值计算得到总熵, 总数为 n*N; 所述单频竞争分析包括把脑电涨落信号中的频率涨落图 f (n)中相同的最 优频率数目随时间变化过程累加, 得到最优频的概率曲线。
8.如权利要求 4所述的脑电涨落信号分析方法, 其特征在于, 还包括在 S 谱系分析基础上进行的第二层次 21个分析项目中的任何一个项 或任意组合 的项目, 所述 21个项目为, (3)S谱总谱分析、 基本谱系分析、 (5)最优值分 析、(6) A/P逆转、 L/R失衡情况分析、(7)特频分析、(8)异频分析、(9)连频分析、 (1Q)最优频分析、 (11) S谱系功率空间分布分析、 (12)单频功率及相对值(L/R)分 布分析、 (13)平均功率分布分析、 (14)功率相对值 A/P、 L/R分析、 (15)长时程 S谱 系曲线分析、(16)长时程优势谱线曲线分析、(1 长时程基本谱系曲线分析、(18)长 时程功率空间分布曲线分析、(19)长时程熵值曲线分析、0))长时程特频曲线分析、 ( 21 ) 长时程连频曲线分析、 (22 ) 长时程基本谱系功率空间分布曲线分析、 ( 23 ) 长时程常规功率谱曲线分析, 其中,
所述 s谱总谱分析方法为,将 S谱系分析中得到的 S谱总谱数据用图形表 示出来;
所述基本谱系分析方法为,对 S谱中基频对应的 S谱系进行统计分析,从 频率大于 3 mHz开始累加其倍周期频率的数值, 同时把所有导联按照在头部的 放置位置分成左前、 左后、 右前、 右后四个部分分别进行统计;
所述最优值分析方法为,将优势频率的功率数值和相应频率按导联的空间 位置分布显示;
所述 A/P逆转、 L/R失衡情况分析方法为, 按导联空间分布位置计算每一 频率功率值的前后比值 A/P, 将 A/P值大于一定限值的频率显示; 同时计算左 右比值 L/R, 将 L/R值大于一定限值的频率显示;
所述特频分析、 异频分析、 连频分析、 最优频分析四个项目, 其方法为, 分别将每导联的特频、 异步员、 连频、 最优频按导联空间分布位置显示;
所述 S谱系功率空间分布分析方法为,将脑电涨落图中的每条谱线的功率 值按空间导联位置分布排^ J, 在显示界面上以 "画中画"方式开一窗口, 进行 谱线的选择,将被选定的谱线,按导联的空间位置分布显示其在每个导联下的 功率数值;
所述单频功率及相对值 ( L/R)分布分析方法为,将每导联优势谱线 D1 - Dn 相应的功率值相加, 得到每导联的总功率值; 将基频的功率及左右比值(L/R) 中大于限定值或小于 1/限定值的数值按导联的空间分布显示;
所述平均功率分布分析方法为,将每个导联平均功率按导联空间分布位置 显不;
所述功率相对值 A/P、 L/R分析方法为, 按导联空间分布位置计算功率值 的前后比值和左右比值;
所述长时程 S谱系曲线分析方法为,分别以每个导联下或全部导联的每一 个谱系或每一条谱线的长落数值为纵轴, 以时间为横轴做出曲线,在显示界面 上以 "画中画"方式开一窗口, 进行谱线或谱系的选择;
所述长时程优势谱线曲线分析方法为,以进入优势谱线区的谱线的频率为 纵轴,以时间为横轴做出曲线, 在显示界面上以 "画中画"方式开一窗口, 进 行优势谱线排名序次 (Dl-Dn)的选择;
所述长时程基本谱系曲线分析方法为,分别以所有导联或每个导联下的基 本谱系各个时间段的涨落数值为纵轴, 以时间为横轴做出基本谱系曲线,在显 示界面上以 "画中画"方式开一窗口, 进行谱系选择;
所述长时程功率空间分布曲线分析方法为, 每一个导联下, 每一条谱线, 以其功率值为纵轴, 以时间为横轴做出曲线,在显示界面上以 "画中画"方式 开一窗口, 进行谱线选择;
所述长时程熵值变化曲线分析方法为,将全部导联或每个导联下的熵值为 纵轴, 以时间为横轴分别做出曲线;
所述长时程特频曲线分析方法为,分别以所有导联或每个导联下特频出现 的个数为纵轴、 时间为横轴做出曲线;
所述长时程连频曲线分析方法为,分别以所有导联或每个导联下的连频出 现的个数为纵轴、 时间为横轴分别做出曲线;
所述长时程基本谱系功率空间分布曲线分析方法为,从单频功率及相对值 分布(L/R)分析中读取基本谱系的每个导联下的功率值,以功率值为纵轴, 以 时间为横轴作出曲线,按导联的空间位置分布显示出每条曲线。在显示界面上 以 "画中画"方式开一窗口, 进行谱系选择;
所述长时程熵值变化曲线分析方法为,以全部导联或每个导联下的熵值为 纵轴, 以时间为横轴分别做出曲线;
所述长时程特频曲线分析方法为,分别以所有导联或每个导联下特频出现 的个数为纵轴, 以时间为横轴做出曲线;
所述长时程连频曲线分析方法为,分别以所有导联或每个导联下连频出现 的个数为纵轴, 以时间为横轴分另 U做出曲线;
. 所述长时程基本谱系功率空 ί 分布曲线分析方法为,从单频功率及相对值 分布(L/R)分析中读取基本谱系每个导联下的功率值,以功率值为纵轴, 以时 间为横轴做出曲线,按导联的空间位置分布显示出每条曲线,在显示界面上以 "画中画"方式开一窗口, 进行谱系选择;
所述长时程常规功率谱曲线分析方法为,从常规功率谱中找出幅值最大的 η个频率 Dl- Dn, 按幅值从大到小 ίΙ序, 分别以这些频率的功率值为纵轴, 以 时间为横轴, 作出 η条动态曲线;
9.如权利要求 8所述的脑电张落信号分析方法,其特征在于,还包括(24) 长时程事件标记识别分析,其方法为,辨认脑电记录盒所记录的事件标记信号, 将这些标记信号在脑电信号回放及各种动态曲线时间轴的相应位置显示标记 符号, 它与 (15)长时程 S谱系曲线分析、(1©长时程优势谱线曲线分析、(17)长时程 基本谱系曲线分析、 (18)长时程功率空间分布曲线分析、 (19)长时程熵值曲线分析、 (20)长时程特频曲线分析、 (21 )长 ΕΙ寸程连频曲线分析、 (22)长时程基本谱系功 率空间分布曲线分析、 (23) 长时程常规功率谱曲线分析九个项目中的任何一 个或任意几个项目结合应用。
10. 如权利要求 1、 2、 3、 4、 7、 8、 9任一权利要求所述的脑电涨落信号 分析方法,其特征在于,所述各项分析方法可应用于对任意一个或几个导联组 合所采集的脑电信号进行数据处理,并且可以选择将分析结果中一个或几个导 联组合的运算结果输出到终端处理器进行显示、 打印或存储。
11. 一种脑电涨落信号的分析设备, 其特征在于, 它包括电极、 数字化脑 电信号放大器或脑电记录盒、 个人计算机、 数据处理器、 终端处理器, 并依次 连接, 其中, 电极采集脑电信号, 传送到数字化脑电信号放大器和 /或脑电信 号记录盒, 对信号接受、 放大、 数模转换、 数字滤波或 /和数据存储, 信号再 上传至个人计算机, 由与计算机目连的数据处理器完成数据处理和涨落分析, 分析结果传输到终端处理器进行存储、 显示或打印。
12. 如权利要求 11所述的脑电涨落信号分析设备, 其特征在于, 所述数 据处理器包括常规功率谱分析模: t夬,用于对给定时间长度的时域脑电信号做功
Figure imgf000026_0001
13. 如权利要求 12所述的脑电涨落信号分析设备, 其特征在于, 所述数 据处理器还包括脑电功率涨落信号分析模块,用于选择边瓣幅值小且衰减快的
、 _ = 0.5 - 0.5 cos(—— ), « = 0,1,...N - 1. ,、 窗函数, 它表示为, W ) , 对脑电信号 数据 进行截短, 得到 = 对上述信号 W = " ")做功率谱分析, 采 样时间为 T秒, 计算 FFT得到频谱的频域分辨率为 1/T, 选择出能量最大的功 率幅值 Ρ及相应的频率 f ; 按时间顺序对总的时域脑电信号进行分段, 依次对 每段数据进行常规功率谱分析和脑电涨落信号分析,得到最大功率幅值涨落信 号 p (n)和相应频率涨落信号 f (n), n=l…… n。
14. 如权利要求 13所述的脑电涨落信号分析设备, 其特征在于, 所述数 据处理器还包括脑电涨落图分析模块,用于对长度为 n点的最大功率幅值涨落 信号 p (n)进行分析; 乘上长度为 n的 Harming窗, 再做功率谱分析, 数据时 间长度单位为 N秒, 因而频域分辨率为 l/N Hz, 取功率谱分析结果中频带一 定范围的谱线组成脑电涨落图; 当总采样时间大于 N秒时, 以 N秒为单位进行 段落划分,对每一段重复进行常规功率谱分析、脑电功率涨落信号分析和脑电 涨落图分析。
15. 如权利要求 11所述的脑电涨落信号分析设备, 其特征在于, 所述数 据处理器还包括 S谱系分析模块,用于从每一个导联脑电涨落图中找出幅值最 大的几条最优谱线 Dl- Dn, 按幅值从大到小排序, 得到单一导联数据的 S谱, 共 n个数值; 综合所有 N导联共得到 N*n条优势谱线, 把频率相同的优势谱线 的数目累加, 得到 S谱总谱。
16. 如权利要求 13所述的脑电涨落信号分析设备, 其特征在于, 所述脑 电功率涨落信号分析模块数据传输给熵值运算模块和 /或单频竞争分析模块, 其中,
13
所述熵值运算模块用于根据 = 1§2 Ρ"来计算熵, 综合所有 N个导联 总概率分布进行熵值计算得到总熵;
所述单频竞争分析模块用于把脑电涨落信号中的频率涨落图 f (n)中相同 的最优频率数目随时间变化过程累加, 得到最优频的概率曲线。
17. 如权利要求 15 所述的脑电涨落信号分析设备, 其特征在于, 所述 S 谱系分析模块数据传输给以下第二层欢 21个分析模块中的任何一个模块或任 意个模块组合, 所述 21个模块为, (3)S谱总谱分析模块、 (4)基本谱系分析模 块、(5)最优值分析模块、(6) A/P逆转、 L/R失衡情况分析模块、(7)特频分析模 块、 (8)异频分析模块、 (9)连频分析模块、 (10)最优频分析模块、 (11) S谱系功率 空间分布分析模块、(12)单频功率分布分析模块、 (13)平均功率分布分析模块、 (14) 功率相对值 A/P、 L/R分析模块、 (15)长时程 S谱系曲线分析模块、 (16)长时程优 势谱线曲线分析模块、(17)长时程基本谱系曲线分析模块、(18)长时程功率空间分 布曲线分析模块、 (19)长时程熵值曲线分析模块、 (20)长时程特频曲线分析模块、 ( 21 ) 长时程连频曲线分析模块、 (22 ) 长时程基本谱系功率空间分布曲线分 析模块、 (23 ) 长时程常规功率谱曲线分析模块, 其中,
所述 (3)S谱总谱分析模块, 将 S谱总谱数据用图形表示出来;
所述 (4)基本谱系分析模块,用于对 S谱中基频对应的 S谱系进行统计分析, 从频率大于 3 mHz开始累加其倍数频率的数值, 同时把所有导联按照在脑部的 分布位置分成前后左右若干部分分别进行统计;
所述 (5)最优值分析模块, 用于按导联的空间位置分布显示每个导联下的
Dl-Dn优势频率的功率数值和相应频率;
所述 (6)A/P逆转、 L/R失衡情况分析模块, 用于按导联空间分布位置计算 每一个频率功率值的前后比值 A/P和左右比值 L/R, 并将 A/P值或 L/R值大于 一定限值的频率显示出来;
所述 (7)特频分析模块、(8)异频分析模块、(9)连频分析模块、(10)最优频分析 模块, 分别用于将每导联的特频、异歩 、连频、每个最优频按导联空间分布位 置显示出来; ·
所述 (IDS谱系功率空间分布分析模块, 用于将脑电涨落图中每条谱线的功 率值按空间导联位置分布排列, 在显示界面上以 "画中画 "方式开一窗口, 进 行谱线的选择, 将被选定频率的谱线,按导联的空间位置分布显示出每个导联 下的功率数值;
所述 (12)单频功率及相对值分布 L/R 分析模块, 用于将每导联优势谱线 Dl-Dn相应的功率值相加, 得到每导联的总功率值; 将基频的功率及左右比值 (L/R) 中大于限定值或小于 1/限定值的数值按导联的空间分布显示出来; 所述 (13)平均功率分布分析模块, |各每个导联平均功率按导联空间分布位置 显示出来;
所述 (14)功率相对值 A/P、 L/R分析模块, 按导联空间分布位置计算功率值 的前后比值和左右比值;
所述 (15)长时程 S谱系曲线分析模块,分别以每个导联下或全部导联的每一 个谱系或每一条谱线的涨落数值为纵轴, 以时间为横轴, 做成曲线, 在显示界 面上以 "画中画"方式开一窗口, 进行谱线或谱系选择;
所述 (16)长时程优势谱线曲线分析模块,用于以进入优势谱线区的谱线的频 率为纵轴,以时间为横轴做成曲线。在显示界面上以 "画中画"方式开一窗口, 进行优势谱线排名序次 (Dl- Dn)的选择;
所述 (17)长时程基本谱系曲线分析模块,用于分别以所有导联或每个导联下 的基本谱系各个时间段的涨落数值为纵轴, 以时间为横轴做成基本谱系曲线; 在显示界面上以 "画中画"方式开一窗口, 进行谱系的选择;
所述 (18)长时程功率空间分布曲线分析模块,用于将每一个导联下的每一条 谱线,以其功率值为纵轴,以时间为横轴做出曲线。在显示界面上以 "画中画" 方式开一窗口, 进行谱线的选择;
所述 (19)长时程熵值变化曲线分析模块,用于将全脑或每个导联下的熵值为 纵轴, 以时间为横轴分别做出曲线;
所述 (20)长时程特频曲线分析模块,用于 (^所有导联或每个导联的特频出现 的个数为纵轴以时间为横轴做出曲线; ;
所述(21 )长时程连频曲线分析模块, 以所有导联或每个导联下的连频出 现的个数为纵轴, 以时间为横轴分别做出曲线;
所述(22)长时程基本谱系功率空间分布曲线分析模块, 用于从单频功率 及相对值分布(L/R)分析中读取基本谱系的每个导联下的功率值,以功率值为 纵轴, 以时间为横轴作出曲线, 按导联的空间位置分布显示出每条曲线。在显 示界面上以 "画中画"方式开一窗口, 进行谱系的选择;
所述(23)长时程常规功率谱曲线分析模块, 用于从常规功率谱中找出幅 值最大的 n个频率 Dl-Dn, 按幅值从大到小排序, 分别以这些频率的功率值为 纵轴, 以时间为横轴, 作出 n条动态曲线;
18. 如权利要求 15所述的 SSi电涨落信号分析设备, 其特征在于, 所述 S 谱系分析模块数据传输给(24)长时程事件标记识别模块, 用于辨认脑电记录 盒所记录的事件标记信号,将这些标记信号在脑电信号回放及各种动态曲线时 间轴的相应位置显示标记符号, 它与 (15)长时程 S谱系曲线分析模块、(16)长时程 优势谱线曲线分析模块、(17)长时程基本谱系曲线分析模块、(18)长时程功率空间 分布曲线分析模块、 (19)长时程熵值曲线分析模块、 (¾))长时程特频曲线分析模块、 ( 21 ) 长时程连频曲线分析模块、 ( 22)长时程基本谱系功率空间分布曲线分 析模块、 (23) 长时程常规功率谱曲线分析模块九个模块中的任何一个或任意 几个结合应用。
PCT/CN2004/001493 2003-12-23 2004-12-22 Procede d'analyse des fluctuations des signaux d'un eeg WO2005060830A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US10/596,537 US7801597B2 (en) 2003-12-23 2004-12-22 Method and apparatus for brain wave fluctuations analysis

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN200310122416.2 2003-12-23
CNB2003101224162A CN100538713C (zh) 2003-12-23 2003-12-23 一种脑电涨落信号分析设备

Publications (1)

Publication Number Publication Date
WO2005060830A1 true WO2005060830A1 (fr) 2005-07-07

Family

ID=34706062

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2004/001493 WO2005060830A1 (fr) 2003-12-23 2004-12-22 Procede d'analyse des fluctuations des signaux d'un eeg

Country Status (3)

Country Link
US (1) US7801597B2 (zh)
CN (1) CN100538713C (zh)
WO (1) WO2005060830A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103876736A (zh) * 2014-04-11 2014-06-25 北京工业大学 一种基于功率谱划分的复杂度谱脑电预测和诊断方法
CN104739378A (zh) * 2015-04-02 2015-07-01 吕少萍 Ra脑神经分离检测装置

Families Citing this family (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101254098B (zh) * 2007-10-19 2010-12-15 北京舒普生工贸有限公司 一种脑电超慢波分析处理装置
CN102397068A (zh) * 2010-09-15 2012-04-04 梅磊 搜索人脑神经递质信息和人脑复杂网络信息的测量装置
US9775545B2 (en) 2010-09-28 2017-10-03 Masimo Corporation Magnetic electrical connector for patient monitors
WO2012050847A2 (en) 2010-09-28 2012-04-19 Masimo Corporation Depth of consciousness monitor including oximeter
CN104281264A (zh) * 2014-09-26 2015-01-14 上海闻泰电子科技有限公司 一种智能数据传输系统及方法
US10154815B2 (en) 2014-10-07 2018-12-18 Masimo Corporation Modular physiological sensors
CN106388811B (zh) * 2016-09-18 2020-09-22 惠州Tcl移动通信有限公司 一种提醒vr用户头晕的方法、系统及装置
CN107169462B (zh) * 2017-05-19 2020-12-22 山东建筑大学 一种基于层次分析的脑电信号特征分类的二分类方法
CN111670002A (zh) * 2017-12-05 2020-09-15 神经索引公司 麻醉管理的系统及方法
CN108538310B (zh) * 2018-03-28 2021-06-25 天津大学 一种基于长时信号功率谱变化的语音端点检测方法
CN108836325A (zh) * 2018-04-02 2018-11-20 东北电力大学 一种基于嗅觉脑电波和随机森林进行感官物质分类的方法
CN109805945B (zh) * 2019-01-30 2021-12-14 北京津发科技股份有限公司 记录回放装置及方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH06261873A (ja) * 1993-03-15 1994-09-20 Kanebo Ltd 脳波を用いた生理状態評価方法及び装置
US5458117A (en) * 1991-10-25 1995-10-17 Aspect Medical Systems, Inc. Cerebral biopotential analysis system and method
CN2256702Y (zh) * 1996-11-21 1997-06-25 中国华阳国际技术公司 脑电超慢涨落分析仪
US5678560A (en) * 1992-08-19 1997-10-21 D F C Co., Ltd. Brain-wave analysis method and apparatus
CN1289915A (zh) * 2000-11-09 2001-04-04 上海市气功研究所 练功程度评估方法及其装置

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4579125A (en) * 1984-01-23 1986-04-01 Cns, Inc. Real-time EEG spectral analyzer
US5357976A (en) * 1989-10-30 1994-10-25 Genquan Feng Method of and arrangement for diagnosing brain disease
DE19538925C2 (de) * 1995-10-19 2000-07-27 Wieland Friedmund Vorrichtung zur Auswertung eines Narkose- oder Intensiv-EEG
US6731975B1 (en) * 2000-10-16 2004-05-04 Instrumentarium Corp. Method and apparatus for determining the cerebral state of a patient with fast response
US7367949B2 (en) * 2003-07-07 2008-05-06 Instrumentarium Corp. Method and apparatus based on combination of physiological parameters for assessment of analgesia during anesthesia or sedation

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5458117A (en) * 1991-10-25 1995-10-17 Aspect Medical Systems, Inc. Cerebral biopotential analysis system and method
US5678560A (en) * 1992-08-19 1997-10-21 D F C Co., Ltd. Brain-wave analysis method and apparatus
JPH06261873A (ja) * 1993-03-15 1994-09-20 Kanebo Ltd 脳波を用いた生理状態評価方法及び装置
CN2256702Y (zh) * 1996-11-21 1997-06-25 中国华阳国际技术公司 脑电超慢涨落分析仪
CN1289915A (zh) * 2000-11-09 2001-04-04 上海市气功研究所 练功程度评估方法及其装置

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103876736A (zh) * 2014-04-11 2014-06-25 北京工业大学 一种基于功率谱划分的复杂度谱脑电预测和诊断方法
CN104739378A (zh) * 2015-04-02 2015-07-01 吕少萍 Ra脑神经分离检测装置

Also Published As

Publication number Publication date
US20070185407A1 (en) 2007-08-09
CN100538713C (zh) 2009-09-09
US7801597B2 (en) 2010-09-21
CN1632816A (zh) 2005-06-29

Similar Documents

Publication Publication Date Title
Tripathy et al. A novel approach for detection of myocardial infarction from ECG signals of multiple electrodes
Banerjee et al. Application of cross wavelet transform for ECG pattern analysis and classification
Matonia et al. Fetal electrocardiograms, direct and abdominal with reference heartbeat annotations
US7239988B2 (en) Apparatus and method for efficient representation of periodic and nearly periodic signals for analysis
JP5604451B2 (ja) 臨床的に意味のある生命徴候を自動的に捕獲及び記録保存するシステム及び方法
EP3698709A1 (en) Electrocardiogram information processing method and electrocardiogram workstation system
WO2005060830A1 (fr) Procede d'analyse des fluctuations des signaux d'un eeg
EP3698708A1 (en) Ecg information processing method and ecg workstation
JP2009541863A (ja) 痴呆および痴呆型障害を分析および評価するためのシステムおよび方法
CACERES Electrocardiographic analysis by a computer system
US3841309A (en) Method of analyzing cerebral electrical activity
CN106073760A (zh) 基于威尔逊导联的18导联动态心电图分析方法及系统
CN106901727A (zh) 一种基于脑电信号的抑郁症风险筛查装置
Kazemnejad et al. An open-access simultaneous electrocardiogram and phonocardiogram database
Mollakazemi et al. Fetal QRS detection in noninvasive abdominal electrocardiograms using principal component analysis and discrete wavelet transforms with signal quality estimation
Wołk et al. Early and remote detection of possible heartbeat problems with convolutional neural networks and multipart interactive training
Yu et al. Automatic detection of atrial fibrillation from ballistocardiogram (BCG) using wavelet features and machine learning
Escalona-Vargas et al. A comparison of five algorithms for fetal magnetocardiography signal extraction
CN112617833A (zh) 一种基于静息态脑电波检测抑郁症的装置
CN115630290B (zh) 基于同步挤压变换的心肺耦合特征提取方法与系统
Liu et al. Automatic Arrhythmia Detection Based on Convolutional Neural Networks.
Sanamdikar et al. Extraction of different features of ECG signal for detection of cardiac arrhythmias by using wavelet transformation Db 6
Gupta et al. Pre-processing based ECG signal analysis using emerging tools
CN110811673A (zh) 基于概率神经网络模型的心音分析系统
Hayek et al. Wavelet processing of systolic murmurs to assist with clinical diagnosis of heart disease

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BW BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE EG ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NA NI NO NZ OM PG PH PL PT RO RU SC SD SE SG SK SL SY TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): BW GH GM KE LS MW MZ NA SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LT LU MC NL PL PT RO SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG

121 Ep: the epo has been informed by wipo that ep was designated in this application
NENP Non-entry into the national phase

Ref country code: DE

WWW Wipo information: withdrawn in national office

Ref document number: DE

WWE Wipo information: entry into national phase

Ref document number: 2007185407

Country of ref document: US

Ref document number: 10596537

Country of ref document: US

122 Ep: pct application non-entry in european phase
WWP Wipo information: published in national office

Ref document number: 10596537

Country of ref document: US