Localization of Physiological Signals

Abstract

Methods and apparatus are provided for acquiring and analyzing data, such as electrophysiological signal data from sensors, such as data suitable for EEG, MEG, ECG, etc., where the data exhibits a linear relationship or may be converted to a linearized relationship. The method, which may be implemented in computer software, includes calculating a cortical current vector field or distribution of values ​​indicative of activity at cortical locations according to an existing selected method, determining which currents are not flowing inward, defining a diagonal weighting matrix, where entries of the diagonal weighting matrix representing locations where current is not flowing inward are small compared to other entries of the diagonal weighting matrix, and recalculating the cortical current vector according to the selected method but incorporating the diagonal weighting matrix, or modifying the distribution of values ​​indicative of activity, where values ​​representing locations where current is not flowing inward exhibit less activity than before the modification. Output of the method may be stored in a computer file for display on a suitable monitor.

Description

Technical Field

[0001] The present invention relates to a physiological method, in particular, an apparatus for acquiring electrophysiological signals related to electroencephalogram (EEG) and magnetoencephalogram (MEG) measurements, and a method for analyzing electrical signals generated during the measurements by the apparatus.

Background Art

[0002] Brain activity can be represented by data from EEG and MEG, and EEG and MEG are composed of measuring electrical signals from electrode sensors (EEG) placed adjacent to the head or coils (MEG) placed on the head surface. In the analysis of EEG and MEG data obtained from sensor outputs, brain activity can be represented as a discrete three-dimensional vector field, and each vector represents a bipolar current source hereinafter referred to as an "electric current source". This result provides an expression of the potential synaptic activity of neurons over time at a certain point in the active brain.

[0003] EEG and MEG recordings of interictal epileptic brain activity often contain waveform patterns known as spikes. By using source localization techniques at the onset or peak of such spikes, the locations of the brain involved in the epileptic network can be revealed. The waveform pattern at the onset or peak of a spike is at least 10 cm 2 (EEG) or 6 cm 2Generated by the simultaneous activity of a type of neuron called pyramidal cells from an extended patch of cortical gray matter having the size of (MEG). Since the sources of EEG and MEG signals are the activities of pyramidal cell neurons, and due to the dominant orientation of this particular cell type, it is known that the direction of the cerebral current is orthogonal to the local cortical gray matter surface. Since the onset and peak of spikes mostly represent activities originating from the depolarized part of the neuron activity cycle, it is further known that the direction of the cerebral current is inward and towards the gray matter - white matter boundary. In the current state of the art, as constraints, the fact that brain activity occurs only from the cortical gray matter, the direction of the cortical current is orthogonal to the cortical gray matter surface, and adjacent cortical locations have similar activities are incorporated into the source location determination algorithm. What is needed is to improve the use of the range of characteristics of measurable physiological signals to represent physiological functions.

[0004] It is known in the art that the relationship between the above - mentioned vector field and the measurement signal is linear. This relationship is uniquely determined by the layout of the (EEG) sensors adjacent to the head or (MEG) sensors on the head, the choice of reference (ground), measurement noise, and the conductivity of the head, and is known as the "forward model". Between any time points, this linear relationship can be written as Ax + n = b, where A (the lead - field matrix) represents the forward model and the choice of reference, n represents the measurement noise, b represents the measurement data, x represents the vector of the intensities of the current sources "currents", and the vector contains 1 to 3 inputs for each discrete point. The above - mentioned vector field is composed of unit vectors used in the calculation of A, and each unit vector is multiplied by the corresponding scalar input x. For convenience, in this document, the symbols representing matrices are written in bold capital letters, the symbols representing vectors are written in bold lowercase letters, the N - th input x of the vector is identified by x N and the (M,N) - th input A of the matrix is identified by A M,N .

[0005] In the case of a specific distribution of current sources, sensor layout, reference, and forward model, assuming no electrical noise, the measured physiological electrical signal data can be uniquely predicted. This is known as the "forward problem."

[0006] However, when the measured physiological electrical signal data, sensor layout, reference, and forward model are an arbitrary set of choices, the distribution of current sources has the following reasons: · The number of sensors is limited, or · The noise is unknown, or · Typically, there are more unknown values (currents) than known values (sensors), or · There are current configurations (silent sources) that do not generate measurable signals For any of these reasons, there is a problem that it cannot be uniquely calculated.

[0007] Such problematic situations, which are common in electrophysiological measurements, are known in the art as ill-posed, adverse-condition inverse problems. However, such current estimation is an important goal in EEG and MEG analysis, for example, to make sense of the measured output.

[0008] Methods known in the art for calculating currents utilize a data model that includes the noise characteristics of the data and a source model that includes the assumed characteristics of the currents, assuming measurements of physiological signal data. The noise characteristics of the data (data model) are typically represented using a noise covariance matrix C n and the noise covariance matrix C n can be estimated from the measured signal data using the assumptions. The data and lead field matrix can also be "whitened beforehand," resulting in a noise covariance matrix of 1.

[0009] Assumptions widely made in the art regarding the characteristics of the current (source model) are that most of the current is small or zero. This assumption is obtained, for example, from the nature of the observed brain state assuming that one local type of activity is dominant, or from the nature of the experiment averaging a large number of cases of data sharing common features of interest, and as a result, it is assumed that, except for the observed features, it is suppressed by the averaging method. The corresponding source model is the minimum norm least squares method model, with the L2-norm x T C S -1 assumed to be minimum, where C S is the source covariance of x. If information about the source covariance is not available, C S = 1. Regularization is used to balance the effects of the data model and the source model. After this inference strategy, the equation, x opt = arg min[(Ax - b) T C n -1 (Ax - b)+λx T C S -1 x] is solved by minimizing the linear inverse problem to obtain a unique x opt (where x opt is the optimal vector defined above), where λ is the regularization parameter. It is well known in the art that an analytical solution for x opt can be obtained. Furthermore, it is well known in the art that the optimal value of λ can be obtained without further information.

[0010] It is well known in the art that the representation of the middle layer of an individual cortical gray - white matter sheet (the "gray - white matter surface"), where cone cell neurons are present, can be obtained from magnetic resonance imaging (MRI) data. Since the orientation of the developing neurons is locally orthogonal to the gray - white matter surface, when estimating cortical currents, and when the discrete points already mentioned are dense enough to sample the cortical gray - white matter surface taking into account the variability of the orientation within the gray matter, it is also well known in the art that the vector x of the current source can contain only one input for each discrete point. When the lead - field matrix A is generated based on a unit current that is consistently either inward or outward, the sign of x N can serve as an indicator of whether the current at location N is flowing inward (depolarization) or outward (repolarization). In this context, "inward" means "towards the white matter", while "outward" means "towards the pial surface".

[0011] Rather than calculating the vector x that represents the current opt there are methods known in the art for calculating a vector s of measurement criteria that indicate locations in the cortex that may be involved in the generation of the event of interest. opt One example of these is the sLORETA method.

[0012] In this document, the terms "comprising" or "including" are used interchangeably with the same meaning and are not limited to any list of features or features described.

Prior Art Documents

Non - Patent Documents

[0013]

Non - Patent Document 1

Non-Patent Document 7

Non-Patent Document 8

Non-Patent Document 9

Non-Patent Document 10

Non-Patent Document 11

Non-Patent Document 12

Non-Patent Document 13

Non-Patent Document 14

Non-Patent Document 15

Summary of the Invention

Problems to be Solved by the Invention

[0014] The cortical current in the inward direction has not yet been used as a constraint condition in the analysis of MEG and EEG measurements. Such a constraint condition is the subject of the present invention. Using the cortical current in the inward direction as a constraint condition for the source location identification algorithm is not an obvious extension of the current state of the art. This is because most of the many types of brain activities typically subject to source location identification do not predominantly originate from depolarizing neuron activities and, therefore, cannot be characterized by the cortical current in the inward direction. Epileptic spikes are an example of a notable, clinically relevant special case.

[0015] It is an object of the present invention to provide a method for converting data including EEG and / or MEG signal measurements in order to represent brain activity solely from depolarizing neurons (inward current). It is a further object to provide an algorithm that can be implemented by computer software for analyzing electrophysiological signal data and providing results including a representation of physiological activity for interpretation and analysis. It is a further object to provide an apparatus for obtaining physiological signal measurements involving the linear relationships described herein and converting the signal measurements into a representation of physiological activity.

Means for Solving the Problems

[0016] The present invention provides a method for analyzing electrophysiological signal data so as to enable a physiological interpretation of the measurement signals. According to the present invention, physiological signal data b is measured and a lead field matrix A is calculated. Further, a discrete cortical source current vector x opt , or a discrete measurement criterion s indicating the trend of the cortical current opt is calculated.

[0017] In one aspect, the present invention provides a method for converting electrical signal data from sensors using a microprocessor, the method comprising collecting the electrical signal data and storing it in a computer file, preprocessing the data, marking one or more time points of interest, applying an averaging step, calculating or obtaining the location of the cortex and the orientation of the corresponding neurons, calculating the location weights and / or the cortical current, determining which current is not flowing inwards, modifying the weights according to the determination, calculating the current according to the weights, calculating the distribution of values indicating activity at the location of the cortex, calculating, extracting or estimating the cortical current direction, determining which current is not flowing inwards, modifying the value distribution according to the determination, and storing the obtained data in at least one computer file. Preferably, the method includes applying a data imaging technique to the stored and obtained data to convert the data into a form suitable for visual representation of the data, and displaying the converted data for visual inspection.

[0018] In another aspect, the present invention provides an apparatus for collecting, converting and displaying electrical signal data, the apparatus comprising a sensor for collecting electrical signals, means for storing the electrical signal data, and at least one microprocessor having a computer program, the computer program performing preprocessing the data, marking one or more time points of interest or causing the user to mark one or more time points of interest, applying an averaging step, calculating or obtaining the location of the cortex and the orientation of the corresponding neurons, calculating the location weights and / or the cortical current, determining which current is not flowing inwards, modifying the weights according to the determination, calculating the current according to the weights, calculating the distribution of values indicating activity at the location of the cortex, calculating, extracting or estimating the cortical current direction, determining which current is not flowing inwards, and modifying the value distribution according to the determination. Preferably, the apparatus includes means for storing the converted data. BRIEF DESCRIPTION OF THE DRAWINGS

[0019]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Embodiments for Carrying Out the Invention

[0020] The method is most advantageously applicable to EEG and MEG measurement signals in order to provide results indicative of the representation of brain activity. The present invention is most advantageously applicable to the collection and analysis of EEG and MEG data, but the method is not limited to the analysis of EEG and MEG data, and the present invention is more generally applicable, for example, in the application to electrocorticogram (ECoG) measurement of brain activity, intracranial (iEEG) measurement of brain activity, electrocardiogram (ECG) measurement and magnetocardiogram (MCG) measurement of cardiac activity, etc. The present invention provides a method for analyzing data including electrophysiological data, and the data either exhibits the linear relationships described herein or can be linearized (e.g., using Newton's method) to exhibit linear relationships. The present invention is useful in all cases where it is known that the sign of the value of x or s is only zero or positive, or only zero or negative.

[0021] According to the present invention, the method can be used to enhance either an existing method for calculating cortical currents or an existing method for calculating the distribution of values providing a measure indicative of the location of the cortex that may be involved in the generation of an event of interest, and further enables the calculation of the current direction for each cortical source or the extraction or estimation of the current direction for each cortical source. Both options are described below.

[0022] When used to enhance existing methods for calculating cortical currents, if the method is capable of incorporating a weight matrix or other mechanism that indirectly adjusts the strength of the calculated cortical currents, for the purposes of the present invention, this mechanism is used to assign weights to cortical sources according to the already calculated current directions. This assignment is made with the desired intention that cortical sources without an associated calculated inward direction become less active. If the present invention is implemented such that these weights are determined iteratively based on several repetitions of a weighted inverse calculation for each specific algorithm definition, the further weighting implemented for the purposes of the present invention can be incorporated into the existing algorithm, for example, after each repetition, or in the final step after the last repetition of the existing method. If the method is not implemented as an iterative weighting scheme after executing the existing method, the same or a similar method is repeated, but this time incorporating the weighting implemented for the purposes of the present invention based on the cortical currents obtained in the first execution.

[0023] A transformation technique known as "Source Weighting" uses the equation: C s =W -2 C p where C p is the source covariance matrix of x. C p encodes external prior knowledge about the source distribution. If such information is not available, C p = 1. The diagonal weight matrix W is determined by the source weighting method itself. Different values of x are obtained depending on W when A, b, and C p are given. To determine W, the value of x calculated by the existing method is used such that W N = f(x N ), where the weighting function f is designed such that the function value never becomes negative, but the value x N is smaller when the case where the current is not flowing inwards is indicated compared to the case where the value x N indicates that the current is flowing inwards.

[0024] For example, if x < 0, then f(x) = 1, otherwise, f(x) = 0, assuming that the inward current at location N is specified by a negative value of x N In this case, the cortical source current vector x opt is recalculated using the weight matrix W. In actual calculations, the cortical source current vector x opt typically does not actually require finding the inverse of W, and from this, it is established that a negative value of W N is not a problem. Even if it is necessary to explicitly find the inverse of W according to the implementation of existing selection methods, 1 / 0 becomes a large, positive number.

[0025] The method of the present invention advantageously implements the techniques previously described in computer software and converts electrical signal data into a representation by a method that was not previously known to be useful.

[0026] The use of weight matrices is known in the art. However, weight matrices are used in the art to achieve a desired amount of locality in the source distribution or to effectively minimize a norm other than the L 2 -norm of x. According to the present invention, the weight matrix is used to suppress currents that are not flowing inward, resulting in the discovery of surprisingly useful results. The method of the present invention when used in electrophysiological signal measurements, for example, EEG or MEG measurements, or other suitable measurements, has not been shown heretofore.

[0027] The present invention includes a device having electrodes for acquiring electrophysiological signal data, means for storing the data, means for converting the data, a microprocessor for performing calculations during conversion, computer software for implementing the algorithm of the method, means for storing the converted data, and means for displaying the converted data. In one embodiment, the present invention comprises an EEG device, electrodes for measuring EEG, means for electronically storing the EEG data, means for storing computer software and executing the computer software for implementing the present invention, means for electronically storing the converted data, and a screen for displaying the converted data. The screen can be any suitable screen capable of displaying images. This screen may include a screen on an analog monitor or a digital monitor. It will be understood that the scope of the present invention includes many embodiments for achieving the object.

[0028] Method embodiments include combinations of data collection and conversion steps shown within boxes in the flowchart shown in FIG. 1. First, sensor electrodes are placed adjacent to the subject's head, for example, in the case of EEG and MEG 1, and a computer is set to collect the output and convert it into a computer data file 2. It will be understood that the scope of the present invention includes any and all types of physiological signals suitable for use in the methods described herein. The conversion data representing the electrophysiological signals are collected and / or stored for further processing 3. When processing the data, a determination is made as to whether the data should be preprocessed 4. The data can be preprocessed 5, or alternatively, time points or points of interest can be marked without preprocessing 6. When further processing the data, a determination is made as to whether one or more points of interest are marked 7. The data can be averaged 8, or alternatively, without averaging, the location of the cortex and the orientation of the corresponding neurons can be calculated or obtained, and the noise covariance, lead field, and prior source covariance can be calculated 9. Existing selection methods are methods that calculate cortical currents and enable location weighting 10. Next, the location weights and / or cortical currents are calculated according to existing methods 12. It is determined which currents are not flowing inwards 13. The weight W is defined or modified in response to this determination 14. The cortical current is calculated taking into account the weight W 15. A determination is made as to whether further iterations are required 16. The resulting data is stored in random access memory (RAM) and further converted by appropriate data imaging techniques that represent the data for visual display, or output to a computer file for later use 21.

[0029] More specifically, as an existing selection method, a method using the Minimum Norm Least Squares (MNLS) or Focal Underdetermined System Solution (FOCUSS) or sLORETA-Weighted Accurate Minimum-Norm (SWARM) together with an iterative or any other weighted linear inverse solver determines the cortical source current vector x opt in the following steps: a) Step of collecting electrical signal data into a computer file. Optionally, step of applying preprocessing such as filtering. b) Step of marking the time point of interest. Optionally, averaging step. c) Location of the cortex corresponding to the orientation of the neuron, noise covariance C n , lead field A and prior source covariance C p Determining step. d) Measurement data b, noise covariance C n , lead field A and prior source covariance C p Based on, by executing an existing selection method, until successfully repeated or until step e) is continued, current density vector x opt and final weight matrix W final Calculating step. In the case of MNLS, the number of iterations is 1 and W final = 1. e) The input of the diagonal weight matrix W (one for each location) is determined by a function of the corresponding value of x opt , and calculating the diagonal weight matrix W such that the locations of non-inward currents obtain smaller weights than the locations of inward currents, for example, W N, N = sgn(x opt,N ) × 0.5 + 1. f) By solving the relevant weighted linear inverse problem, measurement data b, noise covariance C n , lead field A and diagonal weight matrix W and W final and weighted source covariance Cs = W -2 W final -2 Based on Cp, recalculating the updated current density vector x opt . g) If the selection method is an iterative method and a non-iterative selection is made in step d), continue step d) until successfully repeated.

[0030] As an alternative to the use of a weight matrix W in which some W N,N are set to zero, the method often excludes the locations of the corresponding sources, and thus, x and xopt By reducing the dimensionality and by recalculating the lead field A and the presource covariance C p it can also be implemented by simply deleting the corresponding rows and columns.

[0031] When used to enhance an existing method for calculating the distribution of values providing a measure s indicating locations in the cortex that may be involved in the generation of an event of interest, and further when calculating the current direction for each cortical source or enabling the extraction or estimation of the current direction for each cortical source, or when supplementing by a method that calculates the current direction for each cortical source or enables the extraction or estimation of the current direction for each cortical source, this mechanism is used to modify the distribution of values such that locations without an inward current direction are shown to be less likely to be involved in the generation of the event of interest.

[0032] According to the invention, the obtained measure s opt is calculated based on the result s of an existing method and information about whether the current direction at a given location N is inward, and in s opt values are obtained such that locations where the current is not flowing inward are shown to be less likely to be involved in the generation of the event of interest compared to s. For example, if the current at location N is flowing inward, s opt,N = s N and otherwise s opt,N = 0.

[0033] The method of the invention advantageously implements the previously described techniques in computer software and converts electrical signal data into a representation in a way that was not previously considered useful.

[0034] According to the invention, information about the cortical current direction, together with the modification of the obtained measure s, leads to the discovery of surprisingly useful results. The method of the invention when used in electrophysiological signal measurements, for example EEG or MEG measurements, or other suitable measurements, has not been shown heretofore.

[0035] The present invention includes a device having electrodes for acquiring electrophysiological signal data, means for storing the data, means for converting the data, a microprocessor that performs calculations during conversion, computer software for implementing the algorithm of the method, means for storing the converted data, and means for displaying the converted data. In one embodiment, the present invention includes an EEG device, electrodes for measuring EEG, means for electronically storing the EEG data, means for storing computer software and executing the computer software for implementing the present invention, means for electronically storing the converted data, and a screen for displaying the converted data. The screen can be any suitable screen capable of displaying an image. This screen can include a screen on an analog monitor or a digital monitor. It will be understood that the scope of the present invention includes many embodiments for achieving the objective.

[0036] Method embodiments include combinations of data collection and conversion steps shown within boxes in the flowchart shown in FIG. 1. First, sensor electrodes are placed adjacent to the subject's head, for example, in the case of EEG and MEG 1, and a computer is set to collect the output and convert it into a computer data file 2. It will be understood that the scope of the present invention includes any and all types of physiological signals suitable for use in the methods described herein. The conversion data representing the electrophysiological signals is collected and / or stored for further processing 3. When processing the data, a determination is made as to whether the data should be preprocessed 4. The data can be preprocessed 5, or without preprocessing, a time point or points of interest can be marked 6. When further processing the data, a determination is made as to whether one or more points of interest have been marked 7. The data can be averaged 8, or without averaging, the location of the cortex and the orientation of the corresponding neurons can be calculated or obtained, and the noise covariance, lead field, and prior source covariance can be calculated 9. Existing selection methods are methods that calculate the location of likely cortical currents and enable the calculation, extraction, or estimation of the cortical current direction 10. Next, the distribution of values indicating the activity at the location of the cortex is calculated according to existing methods 17. The cortical current direction is calculated 18. It is determined which currents are flowing inwards 19. The distribution of values indicating the activity is corrected based on the current direction 20. The resulting data is stored in random access memory (RAM) and further converted by an appropriate data imaging technique that represents the data for visual display, or output to a computer file for later use 21.

[0037] More specifically, the method of using sLORETA as an existing selection method is a measure s that indicates the location of the cortex that may be involved in the generation of the event of interest opt is determined in the following steps: a) The step of collecting the electrical signal data into a computer file. Optionally, the step of applying preprocessing such as filtering. b) The step of marking the point or points of interest. Optionally, an averaging step. c) The location of the cortex corresponding to the orientation of the neuron, the noise covariance C n , the lead field A, and the prior source covariance C p Determining step. d) By solving the relevant unweighted linear inverse problem, the measurement data b, the noise covariance C n , the lead field A, and the prior source covariance C p Based on, calculating the current density vector x opt Calculating step. e) Based on the current density vector x opt Calculating the sLORETA result s. f) For the location of the cortex, determining that the current direction stored in x opt is not inwards. g) By assigning a value indicating a lower likelihood of being involved in the generation of the event of interest to the location where the current direction is not inwards, the measurement criterion s opt Calculating step based on the sLORETA result s.

[0038] As an existing selection method, the method of using SWARM without iteration uses the measurement criterion s opt before calculating the cortical current, contrary to the measurement criterion s. Alternatively, by assigning a value indicating that the likelihood of being involved in the generation of the event of interest is zero, the measurement criterion s opt Based on the sLORETA result s, the method of using SWARM without iteration can also be implemented by reducing the dimensionality of the corresponding source location, thus s opt Can also be implemented by reducing the number of dimensions.

[0039] The method of the present invention is most conveniently executed by implementing the method within a computer algorithm. In particular, there is a large amount of signal data that must be converted by the method of the present invention, obtained from EEG or MEG measurements, in order to provide meaningful results.

[0040] Example Simulated EEG data containing a point source is shown in Figure 2, and the time evolution of the source intensity models the depolarization phase followed by the repolarization phase. In the left Figure 2a, the output 2 of 25 sensors located on the head in the EEG, scale 4, and the magnitude 5 of each channel in μV at the time point indicated by the vertical time cursor 3 are shown. The time point indicated by the vertical time cursor 3 indicates the time point used for the analysis and is the peak of the depolarization phase. Further, each sensor (channel) is labeled according to the sequence 1 on the left hand side. In the right Figure 2b, the rendering 2 of the sensors generated by the computer (identified by the sensor labeling), and the equipotential lines 3 of the voltage for the selected time point are shown along with the use of scale 1. The noise covariance matrix C n is, in this example, a diagonal weight matrix, and all non-zero inputs of the diagonal weight matrix are (0.5 μV) 2 which corresponds to a signal-to-noise ratio of 10. The source prior covariance matrix C p is 1.

[0041] Figures 3 to 6 show the analysis results applied to the EEG signal data. In all of these drawings, in the parts from a to c, three orthogonal cuts through the 3-D solution space show the analysis result 2. The analysis result is shown as an arrow indicating the location, orientation, and intensity of the analysis result. The location indicated by each arrow is the center of the arrow, i.e., the middle between the rear and the tip. The intensity represented by each arrow is also indicated by the color and size of the arrow. The tip of each arrow indicates the cortical current direction. Labels indicating right (“R”) 1 and left (“L”), the anatomical background 5, and the surface 4 representing the middle layer of the cortical sheet where the source locations are distributed are also shown. In the orthogonal cuts, the black crosshairs 3 indicate the location of the simulated point source. In part d, an enlarged view of the area around the crosshairs from part c is shown. In part e, a scale showing the colors used for the display of the analysis results is shown.

[0042] Figure 3 shows the results of the existing method SWARM with iteration. The unit that appears on the scale is μAmm, which is the current dipole moment.

[0043] Figure 4 shows the results of the proposed method where the existing method is the SWARM method with iterations. The unit visible on the scale is μAmm, the current dipole moment.

[0044] Figure 5 shows the results of the existing method sLORETA. The unit visible on the scale shows the unitless F-distribution statistical score.

[0045] Figure 6 shows the results of the proposed method where the existing method is the sLORETA method. The unit visible on the scale shows the unitless F-distribution statistical score.

[0046] References (Non-Patent Documents 1 - 15): Dale A.M., Sereno M.l. Improved Localization of Cortical Activity by Combining EEG and MEG with MRI Cortical Surface Reconstruction: A Linear Approach. Journal of Cognitive Neuroscience 5, pp. 162 - 176, 1993 Fuchs M., Wagner M., Kohler T., Wischmann H.A. Linear and Nonlinear Current Density Reconstructions. J Clin Neurophysiol 16, pp. 267 - 295, 1999 Gorodnitsky I.F., George J.S., Rao B.D. Neuromagnetic source imaging with FOCUSS: a recursive weighted minimum norm algorithm. Electroencephalogry and Clinical Neurophysiology 4, pp. 231 - 51, 1995 Hamalainen M., Ilmoniemi R. Interpreting magnetic fields of the brain: minimum norm estimates. Report TKK-F-A559, Helsinki University of Technology, Espoo, 1984 Kohler T. Losungen des bioelektromagnetischen inversen Problems. PhD-Thesis, University of Hamburg, Hamburg, 1998 Pascual-Marqui R.D. Review of methods for solving the EEG inverse problem, International Journal of Bioelectromagnetism 1, pp. 75 - 86, 1999 Pascual-Marqui R.D. Standardized low resolution brain electromagnetic tomography (sLORETA): technical details. Methods & Findings in Experimental & Clinical Pharmacology 24D, pp. 5 - 12, 2002 Pascual-Marqui R.D., Michel C.M., Lehmann D., Low resolution electromagnetic tomography: a new method for localizing electrical activity in the brain, International Journal of Psychophysiology 18, pp. 49 - 65, 1994 Sekihara K., Sahani M., Nagarajan S.S., Localization bias and spatial resolution of adaptive and non-adaptive spatial filters for MEG source reconstruction. NeuroImage 25, pp. 1056 - 1067, 2005 Tao JX, Ray A, Hawes-Ebersole S, Ebersole JS., Intracranial EEG substrates of scalp EEG interical spikes, Epilepsia 2005;46: pp. 669 - 676 Tarantola A., Inverse Problem Theory (2nd Edition), Elsevier, Amsterdam, 1994 Wagner M., Rekonstruktion Neuronaler Strome, Shaker Verlag, Aachen, 1998 Wagner M, Fuchs M, Kastner J., Current Density Reconstructions and Deviation Scans Using Extended Sources, Biomag 2002, Editors: H. Nowak, J. Haueisen, F. Giessler, R. Huonker, VDE Verlag, Berlin, Offenbach 2002, pp. 804 - 806 Wagner M., Fuchs M., Kastner J., Evaluation of sLORETA in the presence of noise and multiple sources, Brain Topogr. 16, pp. 277 - 80, 2004 Wagner M, Fuchs M, Kastner J. (2007) SWARM: sLORETA-weighted accurate minimum-norm inverse solutions. In (Eds.) Cheyne D, Ross B, Stroink G, Weinberg H. New Frontiers in Biomagnetism. Amsterdam, Elsevier, pp. 185 - 188

Claims

1. A method for converting electrical signal data from a sensor using a microprocessor, comprising: a) collecting the electrical signal data and storing it in a computer file; b) calculating a cortical current vector; c) determining which current is not flowing inwards; d) calculating a diagonal weight matrix, wherein the input of the diagonal weight matrix representing the location where the current is not flowing inwards is smaller compared to the other inputs of the diagonal weight matrix; e) calculating the current vector, incorporating the diagonal weight matrix determined within the calculation step; f) storing the obtained data in at least one computer file, wherein to obtain the cortical source current vector xopt of the cortical current vector, determining the location of the cortex corresponding to the orientation of the neuron, the noise covariance Cn, the lead field A, and the prior source covariance Cp; calculating the current density vector xopt and the final weight matrix Wfinal based on the measurement data b, the noise covariance Cn, the lead field A, and the prior source covariance Cp; calculating the diagonal weight matrix W, wherein the input of the diagonal weight matrix W is determined by a function of the corresponding value of xopt, and the location of the non-inward current obtains a smaller weight compared to the location of the inward current. A method characterized by the above.

2. A method for converting electrical signal data from a sensor using a microprocessor, comprising: a) collecting the electrical signal data and storing it in a computer file; b) calculating a distribution of values indicating the activity of the location of the cortex; c) calculating, extracting, or estimating the cortical current direction; d) determining which current is not flowing inwards; e) modifying the distribution of values indicating the activity, such that the value representing the location where the current is not flowing inwards indicates less activity than before the modification; f) storing the obtained data in at least one computer file, wherein to obtain the measurement criterion sopt of the distribution of values indicating the activity of the location of the cortex, determining the location of the cortex corresponding to the orientation of the neuron, the noise covariance Cn, the lead field A, and the prior source covariance Cp; Calculating a current density vector x_opt based on measurement data b, noise covariance C_n, lead field A, and prior source covariance C_p by solving an associated unweighted linear inverse problem A method characterized by the above. **Claim 3** Further comprising applying a data imaging technique to the stored data obtained to convert the data into a form suitable for a visual representation of the data The method according to claim 1 or 2. **Claim 4** Further comprising displaying the converted data for visual inspection The method according to claim 3. **Claim 5** An apparatus for collecting, converting, and displaying the electrical signal data to implement the method according to claim 1 or 2, Comprising a sensor for collecting the electrical signal, means for storing the electrical signal data, and at least one microprocessor for converting the electrical signal data An apparatus characterized by the above. **Claim 6** Further comprising means for storing the converted data The apparatus according to claim 5. **Claim 7** Further comprising means for displaying the converted data The apparatus according to claim 5 or 6.

Images