CA2319227C - System and method for measuring, estimating and displaying rms current density maps - Google Patents
System and method for measuring, estimating and displaying rms current density maps Download PDFInfo
- Publication number
- CA2319227C CA2319227C CA002319227A CA2319227A CA2319227C CA 2319227 C CA2319227 C CA 2319227C CA 002319227 A CA002319227 A CA 002319227A CA 2319227 A CA2319227 A CA 2319227A CA 2319227 C CA2319227 C CA 2319227C
- Authority
- CA
- Canada
- Prior art keywords
- signals
- sensors
- array
- current density
- organ
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/242—Detecting biomagnetic fields, e.g. magnetic fields produced by bioelectric currents
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/242—Detecting biomagnetic fields, e.g. magnetic fields produced by bioelectric currents
- A61B5/245—Detecting biomagnetic fields, e.g. magnetic fields produced by bioelectric currents specially adapted for magnetoencephalographic [MEG] signals
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/22—Source localisation; Inverse modelling
Abstract
A method of performing synthetic aperture magnetometery on the signals from a target organ using an array of biomagnetic sensors positioned in a predetermined manner around the target organ, each sensor in the array having a position vector and an orientation vector relative to a common coordinate system encompassing the target organ. The method comprises the steps of: (i) simultaneously measuring EM signals from each sensor positioned in the array for a selected time interval; (ii) computing a covariance matrix of the measured EM signals over a user-selected time sub-interval within the selected time interval; (iii) selecting a set of coordinates for a region of interest to be imaged and a distance between voxels to form a grid of voxels; (iv) computing a forward solution for a current element at each of the voxels for each of the sensors in the array sensors; (v) computing an RMS current density estimate for each voxel from the covariance matrix and the forward solution for that voxel; and (vi) displaying the voxels estimating RMS current density as a first image. The method is particularly well suited for evaluating brain activity in a human subject.
Description
SYSTEM AND METHOD FOR MEASURING. ESTIMATING
AND DISPLAYING RMS CURRENT DENSITY MAPS
Technical Field ' The present invention relates to the measurement and analysis of bioelectro-magnetic activity in electrically active organs. More spe-cifically, the present invention relates to a method of transforming the measurements into a corresponding current distribution map estimating the original sources and combining current distribution maps of two or more states of organ activity into difference maps that reveal regions of the organ in which activity differs for the different states.
Background The ion currents of electrically active organs such as the brain and heart can generate magnetic fields that can be measured outside the surface of the body. Further, the corresponding electrical poten-tials in the organs themselves, when conducted through the body, can be measured on the skin, using surface electrodes or in the interior of the body by means of invasive depth electrodes. The process of computing the biological source current or currents giving rise to the observed magnetic and electrical measurements is generally referred to as "the bioelectromagnetic inverse problem" . The importance of having a biomagnetic or bioelectric inverse solution is that it can be used to correlate electrophysiological function with a particular coordinate within the body. This, in turn, can be used to associate normal function and dysfunction with sic cific anatomic structures_ It can be shown that, in three dimensions, there can be no unique bio-electromagnetic inverse solution without applying constraints to the solution, such as assuming the number and configuration of possible sources. Notwithstanding this, it is possible to calculate useful esti-mates or approximations of the distribution and intensity of source activity from electrophysiological measurements.
In the present art, Magnetoencephalographic (MEG) and Electroencephalographic (EEG) signals may be examined for waveform morphology in independent channels, characterized, for example, by their frequency and amplitude. In addition, MEG and/or EEG measurements, recorded from a plurality of sites, are often represented as topographic distributions of either spontaneous or evoked signals in the form of signal intensity maps about the head.
Such topographic maps are also commonly presented for MEG/EEG in distinct frequency bands.
It is also known to average MEG and EEG signals synchronously with a stimulus presented to the subject or to a volun-tary motor movement from the subject. Signal averaging can improve the signal-to-noise ratio (SNR) of the brain activity underlying a particular sensory or motor event. The resulting averaged signal is conventionally known as the event-related potential (ERP) or the event-related field (ERF). The averaged evoked response is most useful for improving the SNR or activity in the primary cerebral cortex, in which the time delay between stimulus and response has low variability. However, the evoked response relating to higher cognitive functions, which involve associative cerebral cortex, can be more variable in time delay and duration relative to the driving stimulus.
Thus, signal averaging is less useful for evaluating higher cognitive functions. The application of signal averaging to EEG and MEG brain signals is predicated upon the notion that the underlying neural events are identical with each and every stimulus event. Common sense and personal experience dictate that this is not necessarily the case for higher levels of brain functioning. The delay between external events and related thought processes is known to vary greatly. Brain activity associated with critical higher mental processes, such as the production and understanding of language, are therefore not adequately repre-sented by the averaged evoked response.
Another known representation of the MEG and EEG signals is the Equivalent Current Dipole (ECD). The ECD can be computed by fitting a simplified model of a current dipole (or multiple dipoles), each characterized by a location, current vector, and magnitude, to the MEG and/or EEG measurements at some selected instant of time, usually in the least squares sense. In the Minimum Norm Current Distribution method, a more complex, often under-determined model is fitted to the measurements at some instant of time by a least squares method. Both the ECD and Minimum Norm methods can yield erroneous results (e.g., inaccurate localization and magnitude of cortical generators) when noise is present in the EEG or MEG signal.
When the Minimum Norm solution is underdetermined (and it almost always is), the non-uniqueness of the inverse problem implies that the result is only one of many possible source configurations that can explain the measurements. Thus, only spontaneous EEG or MEG
signals having high signal-to-noise ratio and a source characterized by few parameters, such as epileptic spikes and abnormal high-amplitude "slow waves" (a sign of closed-head injury) can be localized accurately by these two methods. Normal (non-pathological) events within the Ontvanpen: 22/12/00 22:61 ; c....., ...... ~.....
13/22/99 WED 13:48 FAZ B04 B81 4081 OPEN WIGGS GREEN MUTALA ~ 011 brain are of much lower amplitude; when possible, signal averaging is conventionally used to improve the signal-to-noise ratio of such eveats.
Much of the above-mentioned prior art is described in SQUID-Based Measuring Techniques by Manfried Hoke in THE ART OF
MEASUREMENT ME?IeOLOGY IN FUNDAMENTAL AND APPLIED
PHYSICS, editod by B. Kramer (1988).
The activity of electrically active organs, such as the brain, may also be monitored and imaged using Positron Emission Tomography (PET) and functional magnetic resonance imaging (fMR1). Neither of these imaging modalities are direct measures of the electrochemical events that comprise neural activity. Instead, they detect local changes in metabolism, metabolic products or blood flow within the brain.
These changes are consequent to the energy requirements of the electrochemical events. Although electrochemical events can occur in less than one millisecond, corresponding local changes in metabolism and blood flow are much slower, having time constants of several seconds. Hence, PET and fMRI lack the time resolution of EEG and MEG, as. they are indirect measures of brain activity.
Lead Field Synthesis (LFS) departs from previous methods for analyzing bioelectromagnetic measurements. LFS is disclosed by S.E.
Robinson and W.C. Black in United States patents 4,977,896 (Robinson et al.) and United States (continuation-in-part) patent 5,269,325 (Robinson et aL) entitled "Analysis of Biological Signals using Data from Arrays of Sensorsp . Instead of localizing brain activity, LFS increases the spatial selectivity of an array of MEG
sensors by summing the weighted observations. The weights are selected to impart higher spatial selectivity to a specified coordinate in CA 02319227 2000-07-21 AM~NJ~D SHEET
iPCAlct~
Ontvanflon: 22/12/99 22:5: .....
1'2/22/99 WED 13:49 FAX 604 881 4081 OYEN WIGGS GREEN MUTALA ~ ut;
~ 5 the head. The sum of products of the measured signal and these weights results in a "virtual sensor" that estimates electrical activity as a function of time at the selected location.
It is also known that the bioelectromagnetic inverse solution can be improved by constraining the location of source currents to the cortex of the brain, since it is the electrical currents flowing between the dendrites and cell bodies of the neurons that are the primary contributors to the measured magnetic fields and electrical potentials.
Furthermore, the source current is known to flow in a direction approximately normal to each point on the cortical surface which provides an additional constraint for the inverse solution. The .
coordinates and vectors describing the cortical surface can be extracted from anatomical images of the brain. These images can be obtained, for example, using magnetic resonance imaging (MRn or computed tomography (CT) scanning of the head.
While certain advances have been made in this art, there is still .
much room for improvement. For example, heretofore, the prior art approach has been unable to localize brain activity in a manner which can adequately represent spontaneous (unaveraged) activity (e.g., brain activity), particularly that of normal higher cognitive functions.
Specifically, certain prior art approaches (e.g.; the ECD and Minimum Norm methods discussed above) rely on "model-fitting"
techniques generally involving the following steps:
(i) the signals are initially observed at some instant or time sample;
(ii) a parametric model is used to predict a forward solution for the measurements; and CA 02319227 2000-07-21 APII=iy~i''= S~~~ ~
Ire.'' Ont~tanpana 22/»/ery ec:ae, 12/22/99 ' WED 13: 49 FAZ B04 B81 4081 OPEN WIGGS GREEN MUTALA tm O1;
AND DISPLAYING RMS CURRENT DENSITY MAPS
Technical Field ' The present invention relates to the measurement and analysis of bioelectro-magnetic activity in electrically active organs. More spe-cifically, the present invention relates to a method of transforming the measurements into a corresponding current distribution map estimating the original sources and combining current distribution maps of two or more states of organ activity into difference maps that reveal regions of the organ in which activity differs for the different states.
Background The ion currents of electrically active organs such as the brain and heart can generate magnetic fields that can be measured outside the surface of the body. Further, the corresponding electrical poten-tials in the organs themselves, when conducted through the body, can be measured on the skin, using surface electrodes or in the interior of the body by means of invasive depth electrodes. The process of computing the biological source current or currents giving rise to the observed magnetic and electrical measurements is generally referred to as "the bioelectromagnetic inverse problem" . The importance of having a biomagnetic or bioelectric inverse solution is that it can be used to correlate electrophysiological function with a particular coordinate within the body. This, in turn, can be used to associate normal function and dysfunction with sic cific anatomic structures_ It can be shown that, in three dimensions, there can be no unique bio-electromagnetic inverse solution without applying constraints to the solution, such as assuming the number and configuration of possible sources. Notwithstanding this, it is possible to calculate useful esti-mates or approximations of the distribution and intensity of source activity from electrophysiological measurements.
In the present art, Magnetoencephalographic (MEG) and Electroencephalographic (EEG) signals may be examined for waveform morphology in independent channels, characterized, for example, by their frequency and amplitude. In addition, MEG and/or EEG measurements, recorded from a plurality of sites, are often represented as topographic distributions of either spontaneous or evoked signals in the form of signal intensity maps about the head.
Such topographic maps are also commonly presented for MEG/EEG in distinct frequency bands.
It is also known to average MEG and EEG signals synchronously with a stimulus presented to the subject or to a volun-tary motor movement from the subject. Signal averaging can improve the signal-to-noise ratio (SNR) of the brain activity underlying a particular sensory or motor event. The resulting averaged signal is conventionally known as the event-related potential (ERP) or the event-related field (ERF). The averaged evoked response is most useful for improving the SNR or activity in the primary cerebral cortex, in which the time delay between stimulus and response has low variability. However, the evoked response relating to higher cognitive functions, which involve associative cerebral cortex, can be more variable in time delay and duration relative to the driving stimulus.
Thus, signal averaging is less useful for evaluating higher cognitive functions. The application of signal averaging to EEG and MEG brain signals is predicated upon the notion that the underlying neural events are identical with each and every stimulus event. Common sense and personal experience dictate that this is not necessarily the case for higher levels of brain functioning. The delay between external events and related thought processes is known to vary greatly. Brain activity associated with critical higher mental processes, such as the production and understanding of language, are therefore not adequately repre-sented by the averaged evoked response.
Another known representation of the MEG and EEG signals is the Equivalent Current Dipole (ECD). The ECD can be computed by fitting a simplified model of a current dipole (or multiple dipoles), each characterized by a location, current vector, and magnitude, to the MEG and/or EEG measurements at some selected instant of time, usually in the least squares sense. In the Minimum Norm Current Distribution method, a more complex, often under-determined model is fitted to the measurements at some instant of time by a least squares method. Both the ECD and Minimum Norm methods can yield erroneous results (e.g., inaccurate localization and magnitude of cortical generators) when noise is present in the EEG or MEG signal.
When the Minimum Norm solution is underdetermined (and it almost always is), the non-uniqueness of the inverse problem implies that the result is only one of many possible source configurations that can explain the measurements. Thus, only spontaneous EEG or MEG
signals having high signal-to-noise ratio and a source characterized by few parameters, such as epileptic spikes and abnormal high-amplitude "slow waves" (a sign of closed-head injury) can be localized accurately by these two methods. Normal (non-pathological) events within the Ontvanpen: 22/12/00 22:61 ; c....., ...... ~.....
13/22/99 WED 13:48 FAZ B04 B81 4081 OPEN WIGGS GREEN MUTALA ~ 011 brain are of much lower amplitude; when possible, signal averaging is conventionally used to improve the signal-to-noise ratio of such eveats.
Much of the above-mentioned prior art is described in SQUID-Based Measuring Techniques by Manfried Hoke in THE ART OF
MEASUREMENT ME?IeOLOGY IN FUNDAMENTAL AND APPLIED
PHYSICS, editod by B. Kramer (1988).
The activity of electrically active organs, such as the brain, may also be monitored and imaged using Positron Emission Tomography (PET) and functional magnetic resonance imaging (fMR1). Neither of these imaging modalities are direct measures of the electrochemical events that comprise neural activity. Instead, they detect local changes in metabolism, metabolic products or blood flow within the brain.
These changes are consequent to the energy requirements of the electrochemical events. Although electrochemical events can occur in less than one millisecond, corresponding local changes in metabolism and blood flow are much slower, having time constants of several seconds. Hence, PET and fMRI lack the time resolution of EEG and MEG, as. they are indirect measures of brain activity.
Lead Field Synthesis (LFS) departs from previous methods for analyzing bioelectromagnetic measurements. LFS is disclosed by S.E.
Robinson and W.C. Black in United States patents 4,977,896 (Robinson et al.) and United States (continuation-in-part) patent 5,269,325 (Robinson et aL) entitled "Analysis of Biological Signals using Data from Arrays of Sensorsp . Instead of localizing brain activity, LFS increases the spatial selectivity of an array of MEG
sensors by summing the weighted observations. The weights are selected to impart higher spatial selectivity to a specified coordinate in CA 02319227 2000-07-21 AM~NJ~D SHEET
iPCAlct~
Ontvanflon: 22/12/99 22:5: .....
1'2/22/99 WED 13:49 FAX 604 881 4081 OYEN WIGGS GREEN MUTALA ~ ut;
~ 5 the head. The sum of products of the measured signal and these weights results in a "virtual sensor" that estimates electrical activity as a function of time at the selected location.
It is also known that the bioelectromagnetic inverse solution can be improved by constraining the location of source currents to the cortex of the brain, since it is the electrical currents flowing between the dendrites and cell bodies of the neurons that are the primary contributors to the measured magnetic fields and electrical potentials.
Furthermore, the source current is known to flow in a direction approximately normal to each point on the cortical surface which provides an additional constraint for the inverse solution. The .
coordinates and vectors describing the cortical surface can be extracted from anatomical images of the brain. These images can be obtained, for example, using magnetic resonance imaging (MRn or computed tomography (CT) scanning of the head.
While certain advances have been made in this art, there is still .
much room for improvement. For example, heretofore, the prior art approach has been unable to localize brain activity in a manner which can adequately represent spontaneous (unaveraged) activity (e.g., brain activity), particularly that of normal higher cognitive functions.
Specifically, certain prior art approaches (e.g.; the ECD and Minimum Norm methods discussed above) rely on "model-fitting"
techniques generally involving the following steps:
(i) the signals are initially observed at some instant or time sample;
(ii) a parametric model is used to predict a forward solution for the measurements; and CA 02319227 2000-07-21 APII=iy~i''= S~~~ ~
Ire.'' Ont~tanpana 22/»/ery ec:ae, 12/22/99 ' WED 13: 49 FAZ B04 B81 4081 OPEN WIGGS GREEN MUTALA tm O1;
(iii) the parameters of the model are adjusted so as to simulta-neously minimize the differences between measured and predicted signals at each of the sensors - usually in the least-squares sense.
As an example, a single ECD may be described using five free para-meters for magnetic measurements - three for position, one describing the tangential dipole-moment vector (radial currents are "silent" in magnetic measurement) and one describing the dipole-moment magni-tude. Further, for ECD representation to be operable, a relatively high signal-to-noise (SNR) ratio is needed. However, for certain organs such as the brain, the spontaneous signals generated by small functional regions do not provide an adequate SNR ratio (in the case of the brain, this is due to the fact that brain functions are also being carried out in areas which are not of interest). Accordingly, it is necessary to use signal averaging techniques to improve the SN'R ratio.
For signal averaging techniques to be useful, it is necessary that the regions of interest in, for example, the brain are in precise synchrony with external events resulting in poor vision of associative areas.
Accordingly, practical imaging of high brain function from MEG and EEG signals has heretofore been at least difficult (if not possible) to achieve. In the Minimum Norm solutions, there are many times more free parameters. When there are more parameters than sensors (measurements) the problem becomes underdetermined.
Certain other prior art approaches (e.g., the LFS method dis-cussed above) rely generally involve the following steps (i) the signals are initially observed;
l't~CAI~P
Ontvanpen: 22/~2/aa cc :~.., lE/22/99 WED 13:50 FAZ B04 B81 4081 OPEN wIGGS GREEN MUTALA ~ u14 (ii) observed signals are weighted by some coefficient; and (iii) derivation of an additional signal which is an estimate of activity (e.g., brain activity).
This approach has limited value in evaluating higher cognitive fune-lions due to the rapid fluctuations (liability) of the activity of certain organs (e.g., the brain).
It would be desirable to have a system and method for measur-ing, estimating and displaying root mean square (RMS) current density maps which obviates or mitigates the above-mentioned limitations of the prior art.
It is an object of the present invention to provide a novel appar-atus and method for measuring, estimating, and displaying RMS
current density maps of brain activity which obviates or mitigates at least one of the disadvantages of the prior art methods for calculating estimates of the distribution and intensity of source activity from electrophysiological measurements.
The system and method of the present invention, referred to herein as Synthetic Aperture Magnetometery (SAM) methodology, permits tomographic imaging of brain activity and represents a radical departure from the previously described prior art methods for analysing MEG and/or EEG data. In contrast to the conventional methods (e.g., ECD, minimum norm and LFS taught in Robinson et al. patents), SAM converts the measured data from, for example when brain activity is being evaluated, an MEG and/or EEG sensor array over a segment of time (rather than at a single instant), into an esti-Af~m~a,~~~~ . : ~~~T
CA 02319227 2000-07-21 ~-f ~'-', -';' C
f ~ta;~_ On*vanpsrtt 22/~2/9fl 22:53; our oo~ -.......
Y2/22/99 ' WED 13: SO FAZ 804 881 4081 OPEN WIGGS GREEN MUTALA I~ O1 _ 8 -mate of the RMS source current density at any designated location in the head. The present invention also provides, in the example of evaluating brain activity, a method for displaying the brain activity that differs between two or more states of brain activity. This latter process is referred to herein as Differential Current Density Mapping . (DCDM). In applying DCDM, individual SAM images are derived from MEG and/or EEG data which has been partitioned into discrete time segments. The time segments correspond to at Ieast two mental states under examination. The SAM image derived for each mental state are then combined using DCDM to display the locations and intensities of the brain that differ between the at least two mental states. Since the common mode brain activity is attenuated by the subtraction process, the locations and interactions of the brain activity that differ between any two brain states is readily identified.
Accordingly, in one aspect of the present invention, there is provided a method of performing synthetic aperture magnetometery on the signals from a target organ using an array of biomagnetic sensors positioned in a predetermined manner around the target organ (e.g., the brain), each sensor in the array having a position vector and an orientation vector relative to a common coordinate system encompass-ing the target organ (e.g., the brain), the method comprising the steps of:
(i) simultaneously measuring EM signals from each sensor positioned in the array for a selected time interval;
(ii) computing a covariance matrix of the measured EM sig-nals over a user-selected time sub-interval within the selected time interval;
CA 02319227 2000-07-21 ANiLi'~7~D S~-'IFE~
IPcA/~P
OntYanqen ~ 22/12/9~ C~:a~, ,... . ,..... _ _ 12/22/99 WED 13:51 FAg B04 681 4081 OPEN WIGGS GREEN !(LiTALA 1001 (iii) selecting a set of coordinates for a region of interest to be imaged and a distance between vowels to form a grid of voxels;
(iv) computing a forward solution for a current element at each of the voxels for each of the sensors in the array sensors;
(v) computing an RMS current density estimate for each vowel from the covariance matrix and the forward solution for that vowel; and (vi) displaying the voxels estimating RMS current density as a first image. This method is typical of SAM.
Thus, in the present method, as will be illustrated below, when reference is made to positioning of biomagnetic sensors in a predeter-mined manner around a target organ, those of skill in the art will appreciate that this means that the sensors are placed in the vicinity of the target organ, usually external to the surface of the body. As is known in the art, the sensitivity of biomagnetic sensors declines rapidly with distance (inverse cube law for simple magnetometers, inverse 4th power for first-order gradiometers, etc.). Also, the fine (i.e., ."higher-order") spatial features of the biomagnetic field, necess-wry for distinguishing different sources, also decline with distance.
This means that the biomagnetic sensors must be placed as close to the body as is practically feasible. Ideally, biomagetic sensors of a num-ber sufficient to obtain as many different "perspective" measurements as is possible are placed around or in the vicinity of the target organ.
Further, the common coordinate system encompassing the target organ is related to the position vector and orientation vector of each sensor in the array. The magnetic field of the target organ (e.g., the brain, as CA 02319227 2000-07-21 ,~~,n,L;y JCS Se..~~~-~
iP=t~l~ P
Ontvanp~nc 22/12/00 22:e4; ....~ ..... .__ 12/22/99 ' WED 13:51 FAX B04 881 4081 OYEN WIGGS GREEN MUTALA f~01 well as other organs), is a vector quantity. The features of such a field convey information as to the location and intensity of each of the cortical generators (sources)- The field should be sampled, spatially, at small enough intervals, surrounding as much of the target organ as possible, to convey information needed for localization and imaging.
As used throughout this specification, the term "EM signals" is intended to mean the signals generated from ion currents in electrically active organs. Generally, these signals will be bioelectric signals, biomagnetic signals or a combination of these. Thus, if the electrically active organ is a brain the EM signals can be magnetoencephalogram (MEG) signals, electroencephalogram (EEG) signals or a combination of these. Alternatively, if the electrically active organ is a heart the EM signals can be electrocardiogram (ECG) signals, magnetocardio-gram (MCG) signals or a combination of these. Further alternatively, if the electrically active organ is an eye, the EMG signals can be electrooculogram (EOG) signals, magnetooculogram (MOG) signals or a combination of these. Persons of skill in the art will recognize that the precise nomenclature of the EM signals useful in the present method will depend on the particular target organ.
Further, as used throughout this specification, the term "forward solution" is intended to mean a computation of the magnetic field or electrical potential response of a mathematically modelled sensor or electrode to a mathematically modeled current distribution within a mathematically modeled conducting volume. As will be understood by those of skill in the art, closed mathematical solutions east for a unique forward solution of signal from source. By contrast, there is no closed and unique mathematical solution for the bioelectromagnetic IP_~~; ~,:p Ont~iangen 22/12/A9 22:54; oua oo~ ...,o~ _ _ 12/22/99 WED 13:52 FAX 604 B81 4081 OPEN WIGGS GREEN MUTALA ~01~
inverse. Hence, all inverse solutions rely upon forward solutions.
More information on forward solutions may be found in Basic Math-ematical and Electromagnetic Concepts of the Biomagnetic Inverse Problem by J. Sarvas (Phys. Med. Biol. 32:11-22 (198'n).
Preferably, the method includes repeating Steps (ii) through (vi) over a second time sub-interval within the selected time interval, to produce a second RMS current density image and, the additional step of subtracting the second RMS current density image, voxel by vowel, from the first image to form a third RMS current density image representing the difference source activity in the brain between the first and second time windows. This preferred embodiment is typical of DCDM.
Figure 1 illustrates a multi-channel biomagaetometer system useful to carry out the present method;
Figure 2 illustrates a cross-section of a configuration of a hel-met, array of sensors and reference system used in the biomagneto-meter system of Figure 1;
Figure 3 is a flowchart representation of the method of one embodiment of the present invention; and, Figure 4 is a schematic representation of the relationship between the coordinate, vector and sensor systems in the present invention.
p ~' T
., Orltvanpsn: 22/12/99 22:65; ou.. oo~ -...... -12/22/89' WED 13:52 FAa B04 881 4081 OYEN WIGGS GREEN MUTALA ~ O1 r~escrirtion The method of the present invention, conveniently referred to herein as Synthetic Aperture Magnetometery (S~) methodology, generally resembles techniques previously employed for forming synthetic aperture antenna arrays, as used in radio. In synthetic aperture methods, an array of antennas can be combined into a much more directional antenna than any of its individual elements taken alone. The present inventor notes that, while the method of the present invention may be used advantageously m magnetometery and magnetoencephalographic (MEG) data measurement, the method may also be used in electroencephalographic (EEG) ~~ m~surements or even in a combination of MEG and EEG data measurement. While the bulk of this specification is directed to describing how SAM
methodology can be usod to measure magnetic signals from the brain, it will be readily understood by those of skill in the art that the method of the present invention is applicable to any electrical and/or magnetic measurements from any electrically active organ of the body, including but not limited to: the heart, nerves, smooth muscle, skeletal muscle and secretory organs. While MEG and EEG data measurement is specifically suited to electrophysiological data measurement of the brain, signal data from other electrically active organs may also be used in the present method. As discussed above, the nomenclature for signals underlying such data will depend on the particular target organ.
For much of this specification, and to facilitate an understanding of the present method, the context of this disclosure is directed, but not limited to, MEG data measurement.
n, r r ~~. r ANi c~ v;.t ~''~.~ ~'"~ ~-.
(~_:
OntVanpen: 22/12/AA 22:55; ou~v oo~ ........, _ _ 12/22/99 VVED 1.3-53 FAZ 804 681 4081 OPEN WIGGS GREEN MLiTALA
When SAM is applied to measuremeats from a sensor array formed from a plurality of MEG andlor EEG sensors, the spatial selectivity of the sensor array to source activity can be much greater than that of the individual sensors. The results of the method of the present invention displays an estimate of the source activity in the form of images of the intensity versus location of the brain's electrical activity, as known to those of skill in the art as being encompassed by "Functional Neuroimaging" . Functional neuroimaging conventionally has been used primarily with PET, fMRI and single photon emission computed tomography (SPECTj - see, for e~cample, "Functional Neuroimaging" (subtitle: Technical Foundations) ISBN:0126858454, Editors: Thatcher et al, Academic Press (1994). Accordingly, these images showing the distribution of brain activity in both healthy and diseased states may be employed directly as a medical diagnostic tool.
However, since the normal brain is active to varying degrees through-out its entire volume, it is often difficult to discern the activity of centers relating to specific faculties such as speech, emotion, and other higher cognitive functions. Therefore, by having the subject presented with stimuli or perform tasks that preferably differ by only one qual-ity, an image of each activity state can be produced. The images are then subtracted from each other to remove the common-mode brain activity. The result of this method reveals images of the brain activity representative of the selected cognitive state.
As an example, the MEG of a subject might be measured while reading a story containing additional unreadable words, presented one word at a time or tachistoscopically. Two SAM images can then be created from the measurements of the brain activity: one image illus---, ., ,;!~~3 'i~: ~ ~ ~:. ~.~ J!_! ~~T
~"'~;~D
!tr~r~: -cm wangonc czi~eisrei t~:ao; ....-. ..... ,__ i2/22/99 'WED 13:53 FAZ B04 B81 4081 OPEN WIGGS GREEN MUTALA
trating brain activity during perception of the words of the story; and the second illustrating brain activity during perception of the unread-able words. By subtracting these two images of brain activity, a third "difference image" is derivod for which the activity common to both states or the "common-mode" brain activity is removed. By repre-senting the brain activity by a difference image, one can then localize where in the brain reading activity takes place.
Accordingly, the SAM method, which permits tomographic imaging of brain activity, represents a radical departure from the conventional previously described prior art methods for analysing MEG and/or EEG signals. In contrast to the conventional methods, SAM converts the data measurements of the sensor array over a user selected time window rather than at a single instant in dme, into an estimate of the RMS source current density at any designated location in the brain.
With reference to Figure 1 there is illustrated a schematic of a 143 whole-head biomagaetometer system 100 useful for carrying out the present method. System 100 uses SQUID (acronym for "supercon-ducting quantum interference devices") detectors and supercondueting gradiometer coils for obtaining the highest sensitivity. System 100 comprises a dewar 104 which is supported by a gantry 108, the dewar having a head-shaped helmet 116 at is lower end, and a patient support 120 which supports a patient while the patient's head is inserted into helmet 116.
2,5 Dewar 104 further includes an array 112 of magnetic sensors about helmet 116 and a magnetic reference sensor system 124 above helmet 116. Array 112 of magnetic sensors may comprises gradio-CA 02319227 2000-07-21 A~~Pc.~',!~~~L ~"~'T
I~~r,~'-'~
OAT\aHrlOer'1: ZZ/7Z/El6r GL:DO~ ..'. . ._ 12/22/99 WED 13:53 FA% B04 B81 4081 OYEN WIGGS GREEN MUTALA
meter sensors, magnetometer sensors, MEG sensors and the like. The SQUIDs (not illustrated) associated with array 112 of magnetic sensors and reference sensor system 124 are located above reference sensor system 124 within dewar 104 which is filled with a cryogen (for example, liquid helium for low temperature superconductor or liquid nitrogen for high temperature superconductors) in operation, or may be refrigerated by mechanical means. As the measurements obtained from both array 112 of magnetic sensors and reference sensor system 124 are susceptible to errors due to mechanical vibration, gantry 108 is designed to minimize vibrations and to have a relatively high char-acteristic frequency, preferably outside the range of frequencies characteristic of brain signals. Of course, as will be understood by those of skill in the art, gantry 108 and dewar 104, with its array 112 of magnetic sensors and reference sensor system 124, can be designed with different characteristics for different sensor array and biomagnedc sources. For example, for cardiac measurements, instead of helmet 116, a different gantry and dewar that are arranged to bring the sensors proumal to the chest wall typically would be used. The design of such a system is not particularly limited and various tech-piques and design will occur to those of skill in the art.
In use, a set of outputs 118 from the SQUIDS associated with array 112 of magnetic sensors and reference sensor system 124 are amplified to a desired level by a SQUID pre-amplifier 132 and the resulting signals 136 are processed by System Electronics 140.
System Electronics 140 comprises a plurality of SQUm control-lers and analog to digital (A/D) conversion means, to convert signals 136 to digital values of the magnetic fields and gradients measured by ., CA 02319227 2000-07-21 '',~-~HN"~i-Ontvanp~n: 22/12/~g 22:57; 604 69i 4odi -~ .rd~~ o.. ...__ , _ 12/22/99 WED 13:54 FAI 804 881 4081 OPEN WIGGS GREEN ~IUTALA 102 array 112 of magnetic sensors and reference sensor system 124, and a plurality of digital signal processors , (DSPs) to perform desired pro-cessing of these digital values. The choice of DSP used in system 100 is not particularly restricted and is within the purview of a person skilled in the art.
Once processed, the resulting signals are transmitted to an Acquisition computer 144 via a suitable communications link, such as a SCSI interface and, thereafter, to a Processing computer 148 via another communications link, such as an Ethernet interface. Preferab-ly, Acquisition compute= 144 and Processing computer 148 are differ-ent computer systems, but in some circumstances they may be com-biped in a single computer system.
Acquisition computer 144 and Processing computer 148 can be any suitable computer systems with graphical workstation capabilities such as, for e~cample, a suitably equipped UNIX based workstation or a member of the Macintosh family of microcomputers manufactured by Apple. Acquisition computer 144 performs several tasks, including tuning of SQUIDS, data collection and storage and control of optional peripheral components, such as stimulus and EEG systems. Process-ing computer 148 performs off line data processing of stored data and display of real time or stored data, and the results of analysis of the data. As will be apparent to those of skill in the art, processing computer 148 may also combine data from the biomagnetometer with other data, such as MRI or CAT scans, to produce graphical displays which can be interpreted in a more intuitive manner.
Additionally, EEG or other data of interest may be collected simultaneously with the measurements made by array 112 of magnetic ~li~~at:..'C.r? w..i~~~
CA 02319227 2000-07-21 ~t''=.:'-'.!~_-i~
OtltVeFtplfl: 22/12/99 22:57; ou~. uu~
12/22/99 WED 13:54 FAX B04 B81 4081 OYEN WIGGS GREEN MUTALA
sensors and reference sensor system 124 and, for example, system electronics 140 may include 64 additional channels to which such inputs may be applied.
Figure ~ 2 illustrates helmet 116, array 112 of magnetic sensors, reference sensor system 124 and a portion of dewar 104 in more detail. As shown, helmet 116 is formed of two spaced and generally parallel walls 149, I50. A series of individual sensor gradiometers 152, which make up array 112 of magnetic sensors, is disposed in a liquid helium bath adjacent parallel walls 149,150. Wall 149 is shaped to receive a human head. Helmet 116 is shaped such that each sensor 152 is located in close proumity to the surface of the human head received therein. In an attempt to ensure a good fit of helmet 116 to a variety of human heads, it has been found that different helmets can be constructed, along anthropometric lines, for biomagnetometer systems intended for use with different races.
Helmet 116 is used in combination with some means of deter-mining the head position and head frame coordinate system, relative to the sensor array. For example, a number (e.g., three) of tiny coils, also known in the art as "head coils" , are affixed to a respective number of fiduciary points on the head (for eacample, the nasion (NAS), the left pre-auricular (LPA) point and the right pre-auricular (RPA) point as illustrated in Figure 4), using double-sided adhesive tape. When the head coordinates are to be measured, the head coils are simultaneously energized with small AC currents, each at a different frequency, generated by digital-to-analog converters in System Electronics 140. The MEG and reference sensors are used to measure the AC magnetic fields emitted by the coils. A precise A ~i~i~l~~~~ ~~~~T
-, fP~~~~_.
O!7t'Ya 11p1I1 ! 22 / 1 2 / A9 22 : SB ; cau-. uu , ....,.. , 12/22/98 '9VED 13:55 FAZ 604 881 4081 OPEN VVIGGS GREEN MUTALA gyp;
inverse solution for Ehe positions of each of the head coils, relative to the sensor array coordinates, can then be calculated. A further head coil (i.e., a fourth if three head coils are used originally) can also be energized so that any other coordinate on the head can be determined -for e~cample, the coordinates of EEG electrodes.
Further details on a system useful for carrying out the present method is described in published International patent application number WO 96/41209 [Vrba et al.].
The present method may conveniently be carried out in process-ing computer 148 (Figure 1) or a supplementary computer (not shown). Referring now to Figure 3, a flowchart representing the method of the present invention is indicated generally at 10. As indicated in Figure 3, the first step in the method is to obtain EM
signals from the organ under observation (block 14). As previously mentioned, data representing the electrical activity of the organ are measured by either biomagnetic sensors (e.g., MEG, MCG, MOG, etc.), bioelectric sensors (e.g., EEG, ECG EOG, etc.) or by both types of sensors. As will be apparent, the sensors may also include a means of detecting ambient biomagnetic and bioelectrical signals such as reference sensors, and a means of measuring the coordinates of the head relative to the sensors known as head localization (e.g. , when the target organ is the brain). It is noted that for the purposes of the method of the present invention, data is measured simultaneously by the plurality of sensors in the array of sensors, optionally in combina-lion with reference sensors. Furthermore, data is measured at discrete and regular time intervals determined by the measurement sample frequency. Further, the SAM method can ~ be applied to any set of ~i''~IEiJ~~'~ ~ ~-;~~ a 1 S"~" ~' y,l ~.' Ontvanpsn: 22/12/99 22:55; c",~ ,.,., ..._ 12/22/99 WED 13:55 FA7C B04 B81 4081 OPEN WIGGS GREEN MUTALA ~0;
time samples such as for multiple intervals of time. The time interval is generally dictated by the conditions of the test and the operator, and should be of a sufficient length to enable to adequately measure at least one state of the target organ.
The precise conditions of the test will depend, at least in part, on the target organ being evahiated. For example, if the target organ is the brain, the conditions of the test may include "Object-Naming"
language paradigm. In an "Object-Naming" language paradigm, a prescribed number of images (e.g., 120) of objects are presented to the subject, one image at a time. The subject is instructed to state the name of each object, as soon as it presented. For each image of object presented to the subject, the "active-state covariance" is com-puted (as described in detail hereinbelow) from the MEG data for a prescribed time interval (e.g., one second) prior to vocalization. For each image of object presented to the subject, the "control-state"
covariance is computed (as described in detail hereinbelow) from the MEG data for a prescribed time interval (e.g., one second) after vocalization is complete. The data may therefore be divided in any logical manner to distinguish intervals of brain. activity, including multiple intervals for a common state of brain activity.
In one situation, the present inventor has employed a CTF
Systems, Inc.T"'' 143-channel whole cortex MEG system; a Hewlett Packard model C-100 UNIX workstation was employed as a host computer for data measurement. However, as will be understood by those of skill in the art, other MEG and computing systems may be employed with similar success.
;_ ._.. T
iVW~Vw~~._L viv 1..~~
CA 02319227 2000-07-21 'r~~t r I ~ ~~=,! , Ontvanpene 22/12/AA 22:6A; ou~. oo. -......
12/22/99 ~ WED 13:56 FA1C B04 B81 4081 OYEN WIGGS GREEN MUT~1..A ~I~c~
At block 18 of Figure 3, the position of each sensor in the array of sensors is determined relative to the organ under observation. For this, both the sensors and the organ positions are referenced to a common coordinate system. As an example of the implementation of block 18, reference is made to Figure 4, in which the organ under eacamination is a human brain. In the present example, wherein the above-identified whole cortex MEO system is employed, the coordinate ~ system is established in a "head ".~ ~ the content typical of an MEC3'r system, the head frame is a coordinate system based upon three fiducial marks on the head (as referred to above).
These are: the nasion (NAS; i.e., the indentation point where the bridge of the nose meets the forehead, and the right and left pre-auricular points (I-PA and RPA, ~ respectively; i.e., points that are forward of the ear canals). These landmarks are used as follows to define a coordinate system:
(i) the Y-axis is defined by the line connecting the two pre-auricular points; .
(ii) the X-axis is defined as the line connecting the nasion with ~o Y-axis that is perpendicular to the Y-aus (i.e., not necessarily at the midpoint of the two pre-auricular points); and (iii) the Z-aus is defined as perpendicular to the XY-plane, at its origin.
The origin of this head frame is at the intersection of the three princi pal axes. Other coordinate systems, bash upon other fiducial points (either internal or external) exist for the head, and may also be used, provided that they are consistent with one another. Initially, the sensor i~~~i=a Ontiianpen: 22/12/A9 22:5~; ....-. .....
12/22/99 WED 13:56 FAZ 604 681 4081 OYEN WIGGS GREEN MUTALA ~ u~t coordinates are only known within the "helmet" frame of reference.
The fiducial points are measured relative to this "helmet" frame of reference. Once the positions of the fiducial points are determined, the head frame is constructed, and the sensors transformed into the head reference frame. Having a common frame of reference for the brain and the sensors is important, as the signals detected by MEG or EEG sensors depend upon. the relative positions and vectors of (i) the sensors, (ii) the source generators, and (iii) the intervening conducting medium (e.g., brain, skull, scalp, etc.).
As indicated in Figure 4, a plurality of sensors 152a through 152n are positioned in a predetermined manner about the head. As discussed above with reference to Figures 1 and 2; the plurality of sensors are fixed in a helmet worn by a subject such that the geometric relationship and orientation of each sensor 152 is known. A
coordinate system is shown in Figure 4, as indicated by the X, Y and Z axes, and an origin is selected as appropriate proximal a region desired for measurement. As will be understood by those of skill in the art, the coordinate system does not have to be Cartesian in nature but may be any other system suitable to the particular application. As indicated in Figure 4, the position of each of the plurality of sensors 152a...n can be located by a vector r, relative to the origin. The vector describing the orientation of each sensor is denoted u~.
Returning now to Figure 3, for the purpose of mapping the measured data to a voxel image array within a desired region of interest (ROI), an intcrvoxel spacing is selected and a grid of vowels is formed within the ROI referenced to the coordinate system (block 22).
As shown in Figure 4, any voxel position can be located by a vector ro CA 02319227 2000-07-21 ~MCi~t"~'w~ ~f"'~~~~
~~~/;!~~
On*vanp~n: 22/~2/AA 23:00 (7iU4 Obi rwca~ -- ...._, ,.....
12/22/99' WED 13:57 FAZ B04 B81 4081 OPEN WIGGS GREEN MUTALA (mQ
having both a magnitude and direction from the origin of the coordinate system. Accordingly, the current vector at ro is jo, corre-sponding to unit length current vector uo. This step corresponds to block 26 in Figure 3. The current vector at each vowel is best deter-s mined using anatomical constraints. For example, source current due to cortical activity in the brain is known to flow in a direction normal to the local cortical surface, at each point on that surface. This vector data could be obtained from an anatomical image, such as a magnetic resonance image (11~R~ or computed tomography (C'17 of the head.
This is provided that the image is translated into the common coordinate system determined for application the SAM method. In the absence of anatomical images, the current vector may be estimated by searching in a for a maximum in current deasity at each voxel as would be apparent to those of skill in the art..
Returning once again to Figure 3, at block 30, the forward solution per unit source current is computed for the potential data measured by the bioelectric sensors and/or tho magnetic field data measured by the biomagnetic sensors. The data measured by each sensor in the array of sensors at time sample k, neglecting sensor noise, is given by:
mi (k) = j.T(r, k) ~Gi (r) dv Vii) n where J(r) is the source current distribution throughout the conductive volume of the head ~ , and G,(r) is the sensitivity pattern of the ~
sensor, also known as the Green's function. This function applies to CA 02319227 2000-07-21 ~~!~~i~4!~ ~ jH
Ontvanpan: 22/12/A9 23:00; 604 681 4Ub~ -- --- -12./22/99 RWED 13:57 FAa B04 681 4081 OYEN ViIGGS GREEN MUTALA X030 either magnetic sensors such as a SQUID (superconducting quantum interference device) magnetometer or gradiometers, or to electrodes applied to the scalp. Equation 1 is a general expression which repre-sents forward solutions but is not readily solved. For practical pur-poses, the Green's function should account for factors such as: the size and extent of the current density; the conductive medium and bound-aries within which the currents flow; and, the geometry and positions of the sensors.
Practical forward solutions are known in the art. E~Ples of some of these published works include: a current dipole in free space (Law of Biot-Savart); a current dipole in a homogeneously conducting sphere (see, for example, Model Studies of the Magnetocardiogrom by F. Grynzpan and D.B. Geselowitz (Biophysics Journal 13:911-925 (1973), and Basic Mathematical and Electromagnetic Concepts of the BiomagneHc Inverse Problem by J. Sarvas (Phys. Med. Biol. 32:11-22 (1987)); a current dipole in a homogenously conducting volume bounded by a finite mesh (see, for example, Sarvas); and a current dipole in multiple concentric spheres, each sphere having differing conductivity (see, for example, Eccentric Dipole in a Spherical Medium: Generalized Expression For Surface Potentials by D.A.
Brody, F.H. Terry and E. Ideker (IEEE Transactions on Biomedical Engineering, BME-20:141,143). The current dipole model generally simplifies the need for knowing the current density everywhere in space, J(r), to simply assuming it is zero everywhere except for a single point in space. With the exception of the current dipole in free space, the various previously-mentioned forward solution models make use of bounded homogeneously conducting spaces, which may have ~~d~Ct~L.l,r__~ ..~.'iC~..l ~~L~~~ri Ontv~npsn: 22/12/~~ 23:01; s04 ee~ 4uo~ -- ..~...... ..._ 12/22/99 WED 13:58 FAa B04 881 4081 OPEN WIGGS GREEN MUTALA ~ 0;
-multiple compartments. Furthermore, they are a simplification of the actual medium which consists of various organs of the body and various tissue structures. At a finer level, conductivity varies rapidly, for example, between the interior of a blood vessel and the surround-s ing tissue. However,, as the method of the . present invention measures 'data at a macroscopic level, these effects tend to average out. There-fore, the assumption of homogeneous conductivity in each tissue compartment leads to accurate forward solutions.
Returning once again to Figure 3, at block 34, the operator optionally filters the measured bioelectric and/or biomagnetic data into selected frequency bands of interest. Specifically, since the distribu-tion of brain activity sometimes differs with frequency, it is useful to examine images of brain activity in specific frequency bands. Thus, for example, the collected data can ~be divided into distinct frequency bands, using a second-order digital bandpass filter. In some situations, it is preferable to filter the data for the purposes of separating artifacts due to, for example, mouth and eye movements. This filtering is also justified by the observation that somewhat different sources may appear in each frequency band.
The operator then selects at least one time window within the predetermined time interval within which the data is to be analyzed (block 38) and for which an image is to be constructed. At block 42 the optionally filtered and time segmented data is employed .to compute the covariance of the measured data. P.ach element of the covariance matrix is given by:
C,~ _ . K~ Imi (k) -mtl (mj (k) -m jl CA 02319227 2000-07-21 _ AI~CI~inJCJ ji"1.~.WT
. ,-, - ~, ~--, Ontvanpsn: 22/12/A9 23:01 ~ u..-. .."., "._ 12/22/99 WED 13:58 FA% 604 B81 4081 OYEN 1VIGGS GREEN MUTALA ~ U3 -where i, j =1,2,. . . ,M and the mean signal is computed in the conven-tional manner:
tits ~ K~ mi l k) ( 3 ) The above Equation is used for calculating the covariance of the data m and is used for measured data which is both bioelectric or biomagaetic in aature. For M channels, the covariance matrix will be MxM in size. The covariance may be computed as a time integral over any time window or multiple time windows in the time interval.
Covariance, for which the mean signal is removed is preferred, instead of the correlation matrix R, as real measurements may contain a DC
offset which originates in the sensors and are therefore not part of the electrophysiological measurements.
At block 46 of Figure 3, the results of blocks 30 and 42 are then combine to compute the Voxel RMS current density estimate. The mean RMS current density at any coordinate ro can then be repre-sented by the following linear algebraic relationship:
i J(r~, n~) G (Zo ~ noI ~ ~~ _~no) where ro is the selected three dimensional coordinate, no is the unit current vector at ro as previously described with respect to Figure 4.
C is the covariance matrix of the measurements, S is the uncorrelated sensor noise variance of all the measurements, and ~c is a regulariz-ation parameter which is selected to control the trade off between noise and spatial selectivity of the currant density estimate. In this matrix notation, G (without subscript) refers to the Green's functions of all CA 02319227 2000-07-21 f~l~JiE'lWl:-:-~ :J:~~~-~
OntJanpen: 22/l2/A9 23:02; uu-. uo. -,.",..
12/22/99 WED 13:59 FAR B04 B81 4081 OYEN WIGGS GREEN MUTALA ~ 03 sensors (i.e., G=G,,Gz,G3,...GM). The superscript "T" refers to the matrix transpose. Matrix and vector variables are designated by bold letters. For any given coordinate ro, the vector no specifying the direction of the current, can be selected by using the vector normal to that point on the surface of the cerebral cortex, for example, extracted from a segmented MRI or CT . image of the subject's brain. The results of Equation 3 is the SAM solution for the RMS current density for a voxel located at ro and having a current. (unit) vector direction no.
An alternative method which is only a useful approximation is to search for the angle no that maximizes the RMS current density at each position ro.
The direct conversion of MEG and/or EEG measurements, over some designated time period, into RMS currents by Equation 3 can be 1 S seen by suiistituting the measurements from Equation 1 into the covariance Equation 2, the result of which is approximated by:
csj = f ~ tr) IGs (r) ~Gj (r) l dv (5) The quantity in the square brackets is conventionally referred to as the "Gram~chmidt matrix", appearing frequently in the linear algebra of inverse problems. It is normally computed using an a priori model. of all possible sources. The covariance of the measure-ment can be seen here to be equivalent to a Gram-Schmidt matrix that has been weighted by the mean square current density that originally gave rise to the measurement. Thus, use of the covariance obviates ~IV~i~.~ ~~c~~ ~';='~
~ C_ = '~ ~ : . =.
OntVanp~n: 22/12/fl9 23:02; o~.. .,~~ ..,.,.. _ 12/22/89 WED 13:59 FAQ B04 881 4081 OPEN WIGGS GREEN MUTALA
the need of having an a priori model of the sources to perform inverse calculations.
Normally, and preferably, spontaneous brain signals are col-lected, without signal averaging. The correlation or covariance matrix is computed from the signals over some time interval or intervals selected from some or all of the measured signals. Alternatively, prior to computing covariance, signals from the array may be averaged synchronously to a stimulus presented to a subject, or be triggered by an external event such as a subject pressing a switch. The stimulus may be either visual, auditory, or somatosensory.
The vowels estimating the RMS current density are then dis-played, for example on a monitor, as a first image or a second image.
~MEi~DE~ S. iEE i
As an example, a single ECD may be described using five free para-meters for magnetic measurements - three for position, one describing the tangential dipole-moment vector (radial currents are "silent" in magnetic measurement) and one describing the dipole-moment magni-tude. Further, for ECD representation to be operable, a relatively high signal-to-noise (SNR) ratio is needed. However, for certain organs such as the brain, the spontaneous signals generated by small functional regions do not provide an adequate SNR ratio (in the case of the brain, this is due to the fact that brain functions are also being carried out in areas which are not of interest). Accordingly, it is necessary to use signal averaging techniques to improve the SN'R ratio.
For signal averaging techniques to be useful, it is necessary that the regions of interest in, for example, the brain are in precise synchrony with external events resulting in poor vision of associative areas.
Accordingly, practical imaging of high brain function from MEG and EEG signals has heretofore been at least difficult (if not possible) to achieve. In the Minimum Norm solutions, there are many times more free parameters. When there are more parameters than sensors (measurements) the problem becomes underdetermined.
Certain other prior art approaches (e.g., the LFS method dis-cussed above) rely generally involve the following steps (i) the signals are initially observed;
l't~CAI~P
Ontvanpen: 22/~2/aa cc :~.., lE/22/99 WED 13:50 FAZ B04 B81 4081 OPEN wIGGS GREEN MUTALA ~ u14 (ii) observed signals are weighted by some coefficient; and (iii) derivation of an additional signal which is an estimate of activity (e.g., brain activity).
This approach has limited value in evaluating higher cognitive fune-lions due to the rapid fluctuations (liability) of the activity of certain organs (e.g., the brain).
It would be desirable to have a system and method for measur-ing, estimating and displaying root mean square (RMS) current density maps which obviates or mitigates the above-mentioned limitations of the prior art.
It is an object of the present invention to provide a novel appar-atus and method for measuring, estimating, and displaying RMS
current density maps of brain activity which obviates or mitigates at least one of the disadvantages of the prior art methods for calculating estimates of the distribution and intensity of source activity from electrophysiological measurements.
The system and method of the present invention, referred to herein as Synthetic Aperture Magnetometery (SAM) methodology, permits tomographic imaging of brain activity and represents a radical departure from the previously described prior art methods for analysing MEG and/or EEG data. In contrast to the conventional methods (e.g., ECD, minimum norm and LFS taught in Robinson et al. patents), SAM converts the measured data from, for example when brain activity is being evaluated, an MEG and/or EEG sensor array over a segment of time (rather than at a single instant), into an esti-Af~m~a,~~~~ . : ~~~T
CA 02319227 2000-07-21 ~-f ~'-', -';' C
f ~ta;~_ On*vanpsrtt 22/~2/9fl 22:53; our oo~ -.......
Y2/22/99 ' WED 13: SO FAZ 804 881 4081 OPEN WIGGS GREEN MUTALA I~ O1 _ 8 -mate of the RMS source current density at any designated location in the head. The present invention also provides, in the example of evaluating brain activity, a method for displaying the brain activity that differs between two or more states of brain activity. This latter process is referred to herein as Differential Current Density Mapping . (DCDM). In applying DCDM, individual SAM images are derived from MEG and/or EEG data which has been partitioned into discrete time segments. The time segments correspond to at Ieast two mental states under examination. The SAM image derived for each mental state are then combined using DCDM to display the locations and intensities of the brain that differ between the at least two mental states. Since the common mode brain activity is attenuated by the subtraction process, the locations and interactions of the brain activity that differ between any two brain states is readily identified.
Accordingly, in one aspect of the present invention, there is provided a method of performing synthetic aperture magnetometery on the signals from a target organ using an array of biomagnetic sensors positioned in a predetermined manner around the target organ (e.g., the brain), each sensor in the array having a position vector and an orientation vector relative to a common coordinate system encompass-ing the target organ (e.g., the brain), the method comprising the steps of:
(i) simultaneously measuring EM signals from each sensor positioned in the array for a selected time interval;
(ii) computing a covariance matrix of the measured EM sig-nals over a user-selected time sub-interval within the selected time interval;
CA 02319227 2000-07-21 ANiLi'~7~D S~-'IFE~
IPcA/~P
OntYanqen ~ 22/12/9~ C~:a~, ,... . ,..... _ _ 12/22/99 WED 13:51 FAg B04 681 4081 OPEN WIGGS GREEN !(LiTALA 1001 (iii) selecting a set of coordinates for a region of interest to be imaged and a distance between vowels to form a grid of voxels;
(iv) computing a forward solution for a current element at each of the voxels for each of the sensors in the array sensors;
(v) computing an RMS current density estimate for each vowel from the covariance matrix and the forward solution for that vowel; and (vi) displaying the voxels estimating RMS current density as a first image. This method is typical of SAM.
Thus, in the present method, as will be illustrated below, when reference is made to positioning of biomagnetic sensors in a predeter-mined manner around a target organ, those of skill in the art will appreciate that this means that the sensors are placed in the vicinity of the target organ, usually external to the surface of the body. As is known in the art, the sensitivity of biomagnetic sensors declines rapidly with distance (inverse cube law for simple magnetometers, inverse 4th power for first-order gradiometers, etc.). Also, the fine (i.e., ."higher-order") spatial features of the biomagnetic field, necess-wry for distinguishing different sources, also decline with distance.
This means that the biomagnetic sensors must be placed as close to the body as is practically feasible. Ideally, biomagetic sensors of a num-ber sufficient to obtain as many different "perspective" measurements as is possible are placed around or in the vicinity of the target organ.
Further, the common coordinate system encompassing the target organ is related to the position vector and orientation vector of each sensor in the array. The magnetic field of the target organ (e.g., the brain, as CA 02319227 2000-07-21 ,~~,n,L;y JCS Se..~~~-~
iP=t~l~ P
Ontvanp~nc 22/12/00 22:e4; ....~ ..... .__ 12/22/99 ' WED 13:51 FAX B04 881 4081 OYEN WIGGS GREEN MUTALA f~01 well as other organs), is a vector quantity. The features of such a field convey information as to the location and intensity of each of the cortical generators (sources)- The field should be sampled, spatially, at small enough intervals, surrounding as much of the target organ as possible, to convey information needed for localization and imaging.
As used throughout this specification, the term "EM signals" is intended to mean the signals generated from ion currents in electrically active organs. Generally, these signals will be bioelectric signals, biomagnetic signals or a combination of these. Thus, if the electrically active organ is a brain the EM signals can be magnetoencephalogram (MEG) signals, electroencephalogram (EEG) signals or a combination of these. Alternatively, if the electrically active organ is a heart the EM signals can be electrocardiogram (ECG) signals, magnetocardio-gram (MCG) signals or a combination of these. Further alternatively, if the electrically active organ is an eye, the EMG signals can be electrooculogram (EOG) signals, magnetooculogram (MOG) signals or a combination of these. Persons of skill in the art will recognize that the precise nomenclature of the EM signals useful in the present method will depend on the particular target organ.
Further, as used throughout this specification, the term "forward solution" is intended to mean a computation of the magnetic field or electrical potential response of a mathematically modelled sensor or electrode to a mathematically modeled current distribution within a mathematically modeled conducting volume. As will be understood by those of skill in the art, closed mathematical solutions east for a unique forward solution of signal from source. By contrast, there is no closed and unique mathematical solution for the bioelectromagnetic IP_~~; ~,:p Ont~iangen 22/12/A9 22:54; oua oo~ ...,o~ _ _ 12/22/99 WED 13:52 FAX 604 B81 4081 OPEN WIGGS GREEN MUTALA ~01~
inverse. Hence, all inverse solutions rely upon forward solutions.
More information on forward solutions may be found in Basic Math-ematical and Electromagnetic Concepts of the Biomagnetic Inverse Problem by J. Sarvas (Phys. Med. Biol. 32:11-22 (198'n).
Preferably, the method includes repeating Steps (ii) through (vi) over a second time sub-interval within the selected time interval, to produce a second RMS current density image and, the additional step of subtracting the second RMS current density image, voxel by vowel, from the first image to form a third RMS current density image representing the difference source activity in the brain between the first and second time windows. This preferred embodiment is typical of DCDM.
Figure 1 illustrates a multi-channel biomagaetometer system useful to carry out the present method;
Figure 2 illustrates a cross-section of a configuration of a hel-met, array of sensors and reference system used in the biomagneto-meter system of Figure 1;
Figure 3 is a flowchart representation of the method of one embodiment of the present invention; and, Figure 4 is a schematic representation of the relationship between the coordinate, vector and sensor systems in the present invention.
p ~' T
., Orltvanpsn: 22/12/99 22:65; ou.. oo~ -...... -12/22/89' WED 13:52 FAa B04 881 4081 OYEN WIGGS GREEN MUTALA ~ O1 r~escrirtion The method of the present invention, conveniently referred to herein as Synthetic Aperture Magnetometery (S~) methodology, generally resembles techniques previously employed for forming synthetic aperture antenna arrays, as used in radio. In synthetic aperture methods, an array of antennas can be combined into a much more directional antenna than any of its individual elements taken alone. The present inventor notes that, while the method of the present invention may be used advantageously m magnetometery and magnetoencephalographic (MEG) data measurement, the method may also be used in electroencephalographic (EEG) ~~ m~surements or even in a combination of MEG and EEG data measurement. While the bulk of this specification is directed to describing how SAM
methodology can be usod to measure magnetic signals from the brain, it will be readily understood by those of skill in the art that the method of the present invention is applicable to any electrical and/or magnetic measurements from any electrically active organ of the body, including but not limited to: the heart, nerves, smooth muscle, skeletal muscle and secretory organs. While MEG and EEG data measurement is specifically suited to electrophysiological data measurement of the brain, signal data from other electrically active organs may also be used in the present method. As discussed above, the nomenclature for signals underlying such data will depend on the particular target organ.
For much of this specification, and to facilitate an understanding of the present method, the context of this disclosure is directed, but not limited to, MEG data measurement.
n, r r ~~. r ANi c~ v;.t ~''~.~ ~'"~ ~-.
(~_:
OntVanpen: 22/12/AA 22:55; ou~v oo~ ........, _ _ 12/22/99 VVED 1.3-53 FAZ 804 681 4081 OPEN WIGGS GREEN MLiTALA
When SAM is applied to measuremeats from a sensor array formed from a plurality of MEG andlor EEG sensors, the spatial selectivity of the sensor array to source activity can be much greater than that of the individual sensors. The results of the method of the present invention displays an estimate of the source activity in the form of images of the intensity versus location of the brain's electrical activity, as known to those of skill in the art as being encompassed by "Functional Neuroimaging" . Functional neuroimaging conventionally has been used primarily with PET, fMRI and single photon emission computed tomography (SPECTj - see, for e~cample, "Functional Neuroimaging" (subtitle: Technical Foundations) ISBN:0126858454, Editors: Thatcher et al, Academic Press (1994). Accordingly, these images showing the distribution of brain activity in both healthy and diseased states may be employed directly as a medical diagnostic tool.
However, since the normal brain is active to varying degrees through-out its entire volume, it is often difficult to discern the activity of centers relating to specific faculties such as speech, emotion, and other higher cognitive functions. Therefore, by having the subject presented with stimuli or perform tasks that preferably differ by only one qual-ity, an image of each activity state can be produced. The images are then subtracted from each other to remove the common-mode brain activity. The result of this method reveals images of the brain activity representative of the selected cognitive state.
As an example, the MEG of a subject might be measured while reading a story containing additional unreadable words, presented one word at a time or tachistoscopically. Two SAM images can then be created from the measurements of the brain activity: one image illus---, ., ,;!~~3 'i~: ~ ~ ~:. ~.~ J!_! ~~T
~"'~;~D
!tr~r~: -cm wangonc czi~eisrei t~:ao; ....-. ..... ,__ i2/22/99 'WED 13:53 FAZ B04 B81 4081 OPEN WIGGS GREEN MUTALA
trating brain activity during perception of the words of the story; and the second illustrating brain activity during perception of the unread-able words. By subtracting these two images of brain activity, a third "difference image" is derivod for which the activity common to both states or the "common-mode" brain activity is removed. By repre-senting the brain activity by a difference image, one can then localize where in the brain reading activity takes place.
Accordingly, the SAM method, which permits tomographic imaging of brain activity, represents a radical departure from the conventional previously described prior art methods for analysing MEG and/or EEG signals. In contrast to the conventional methods, SAM converts the data measurements of the sensor array over a user selected time window rather than at a single instant in dme, into an estimate of the RMS source current density at any designated location in the brain.
With reference to Figure 1 there is illustrated a schematic of a 143 whole-head biomagaetometer system 100 useful for carrying out the present method. System 100 uses SQUID (acronym for "supercon-ducting quantum interference devices") detectors and supercondueting gradiometer coils for obtaining the highest sensitivity. System 100 comprises a dewar 104 which is supported by a gantry 108, the dewar having a head-shaped helmet 116 at is lower end, and a patient support 120 which supports a patient while the patient's head is inserted into helmet 116.
2,5 Dewar 104 further includes an array 112 of magnetic sensors about helmet 116 and a magnetic reference sensor system 124 above helmet 116. Array 112 of magnetic sensors may comprises gradio-CA 02319227 2000-07-21 A~~Pc.~',!~~~L ~"~'T
I~~r,~'-'~
OAT\aHrlOer'1: ZZ/7Z/El6r GL:DO~ ..'. . ._ 12/22/99 WED 13:53 FA% B04 B81 4081 OYEN WIGGS GREEN MUTALA
meter sensors, magnetometer sensors, MEG sensors and the like. The SQUIDs (not illustrated) associated with array 112 of magnetic sensors and reference sensor system 124 are located above reference sensor system 124 within dewar 104 which is filled with a cryogen (for example, liquid helium for low temperature superconductor or liquid nitrogen for high temperature superconductors) in operation, or may be refrigerated by mechanical means. As the measurements obtained from both array 112 of magnetic sensors and reference sensor system 124 are susceptible to errors due to mechanical vibration, gantry 108 is designed to minimize vibrations and to have a relatively high char-acteristic frequency, preferably outside the range of frequencies characteristic of brain signals. Of course, as will be understood by those of skill in the art, gantry 108 and dewar 104, with its array 112 of magnetic sensors and reference sensor system 124, can be designed with different characteristics for different sensor array and biomagnedc sources. For example, for cardiac measurements, instead of helmet 116, a different gantry and dewar that are arranged to bring the sensors proumal to the chest wall typically would be used. The design of such a system is not particularly limited and various tech-piques and design will occur to those of skill in the art.
In use, a set of outputs 118 from the SQUIDS associated with array 112 of magnetic sensors and reference sensor system 124 are amplified to a desired level by a SQUID pre-amplifier 132 and the resulting signals 136 are processed by System Electronics 140.
System Electronics 140 comprises a plurality of SQUm control-lers and analog to digital (A/D) conversion means, to convert signals 136 to digital values of the magnetic fields and gradients measured by ., CA 02319227 2000-07-21 '',~-~HN"~i-Ontvanp~n: 22/12/~g 22:57; 604 69i 4odi -~ .rd~~ o.. ...__ , _ 12/22/99 WED 13:54 FAI 804 881 4081 OPEN WIGGS GREEN ~IUTALA 102 array 112 of magnetic sensors and reference sensor system 124, and a plurality of digital signal processors , (DSPs) to perform desired pro-cessing of these digital values. The choice of DSP used in system 100 is not particularly restricted and is within the purview of a person skilled in the art.
Once processed, the resulting signals are transmitted to an Acquisition computer 144 via a suitable communications link, such as a SCSI interface and, thereafter, to a Processing computer 148 via another communications link, such as an Ethernet interface. Preferab-ly, Acquisition compute= 144 and Processing computer 148 are differ-ent computer systems, but in some circumstances they may be com-biped in a single computer system.
Acquisition computer 144 and Processing computer 148 can be any suitable computer systems with graphical workstation capabilities such as, for e~cample, a suitably equipped UNIX based workstation or a member of the Macintosh family of microcomputers manufactured by Apple. Acquisition computer 144 performs several tasks, including tuning of SQUIDS, data collection and storage and control of optional peripheral components, such as stimulus and EEG systems. Process-ing computer 148 performs off line data processing of stored data and display of real time or stored data, and the results of analysis of the data. As will be apparent to those of skill in the art, processing computer 148 may also combine data from the biomagnetometer with other data, such as MRI or CAT scans, to produce graphical displays which can be interpreted in a more intuitive manner.
Additionally, EEG or other data of interest may be collected simultaneously with the measurements made by array 112 of magnetic ~li~~at:..'C.r? w..i~~~
CA 02319227 2000-07-21 ~t''=.:'-'.!~_-i~
OtltVeFtplfl: 22/12/99 22:57; ou~. uu~
12/22/99 WED 13:54 FAX B04 B81 4081 OYEN WIGGS GREEN MUTALA
sensors and reference sensor system 124 and, for example, system electronics 140 may include 64 additional channels to which such inputs may be applied.
Figure ~ 2 illustrates helmet 116, array 112 of magnetic sensors, reference sensor system 124 and a portion of dewar 104 in more detail. As shown, helmet 116 is formed of two spaced and generally parallel walls 149, I50. A series of individual sensor gradiometers 152, which make up array 112 of magnetic sensors, is disposed in a liquid helium bath adjacent parallel walls 149,150. Wall 149 is shaped to receive a human head. Helmet 116 is shaped such that each sensor 152 is located in close proumity to the surface of the human head received therein. In an attempt to ensure a good fit of helmet 116 to a variety of human heads, it has been found that different helmets can be constructed, along anthropometric lines, for biomagnetometer systems intended for use with different races.
Helmet 116 is used in combination with some means of deter-mining the head position and head frame coordinate system, relative to the sensor array. For example, a number (e.g., three) of tiny coils, also known in the art as "head coils" , are affixed to a respective number of fiduciary points on the head (for eacample, the nasion (NAS), the left pre-auricular (LPA) point and the right pre-auricular (RPA) point as illustrated in Figure 4), using double-sided adhesive tape. When the head coordinates are to be measured, the head coils are simultaneously energized with small AC currents, each at a different frequency, generated by digital-to-analog converters in System Electronics 140. The MEG and reference sensors are used to measure the AC magnetic fields emitted by the coils. A precise A ~i~i~l~~~~ ~~~~T
-, fP~~~~_.
O!7t'Ya 11p1I1 ! 22 / 1 2 / A9 22 : SB ; cau-. uu , ....,.. , 12/22/98 '9VED 13:55 FAZ 604 881 4081 OPEN VVIGGS GREEN MUTALA gyp;
inverse solution for Ehe positions of each of the head coils, relative to the sensor array coordinates, can then be calculated. A further head coil (i.e., a fourth if three head coils are used originally) can also be energized so that any other coordinate on the head can be determined -for e~cample, the coordinates of EEG electrodes.
Further details on a system useful for carrying out the present method is described in published International patent application number WO 96/41209 [Vrba et al.].
The present method may conveniently be carried out in process-ing computer 148 (Figure 1) or a supplementary computer (not shown). Referring now to Figure 3, a flowchart representing the method of the present invention is indicated generally at 10. As indicated in Figure 3, the first step in the method is to obtain EM
signals from the organ under observation (block 14). As previously mentioned, data representing the electrical activity of the organ are measured by either biomagnetic sensors (e.g., MEG, MCG, MOG, etc.), bioelectric sensors (e.g., EEG, ECG EOG, etc.) or by both types of sensors. As will be apparent, the sensors may also include a means of detecting ambient biomagnetic and bioelectrical signals such as reference sensors, and a means of measuring the coordinates of the head relative to the sensors known as head localization (e.g. , when the target organ is the brain). It is noted that for the purposes of the method of the present invention, data is measured simultaneously by the plurality of sensors in the array of sensors, optionally in combina-lion with reference sensors. Furthermore, data is measured at discrete and regular time intervals determined by the measurement sample frequency. Further, the SAM method can ~ be applied to any set of ~i''~IEiJ~~'~ ~ ~-;~~ a 1 S"~" ~' y,l ~.' Ontvanpsn: 22/12/99 22:55; c",~ ,.,., ..._ 12/22/99 WED 13:55 FA7C B04 B81 4081 OPEN WIGGS GREEN MUTALA ~0;
time samples such as for multiple intervals of time. The time interval is generally dictated by the conditions of the test and the operator, and should be of a sufficient length to enable to adequately measure at least one state of the target organ.
The precise conditions of the test will depend, at least in part, on the target organ being evahiated. For example, if the target organ is the brain, the conditions of the test may include "Object-Naming"
language paradigm. In an "Object-Naming" language paradigm, a prescribed number of images (e.g., 120) of objects are presented to the subject, one image at a time. The subject is instructed to state the name of each object, as soon as it presented. For each image of object presented to the subject, the "active-state covariance" is com-puted (as described in detail hereinbelow) from the MEG data for a prescribed time interval (e.g., one second) prior to vocalization. For each image of object presented to the subject, the "control-state"
covariance is computed (as described in detail hereinbelow) from the MEG data for a prescribed time interval (e.g., one second) after vocalization is complete. The data may therefore be divided in any logical manner to distinguish intervals of brain. activity, including multiple intervals for a common state of brain activity.
In one situation, the present inventor has employed a CTF
Systems, Inc.T"'' 143-channel whole cortex MEG system; a Hewlett Packard model C-100 UNIX workstation was employed as a host computer for data measurement. However, as will be understood by those of skill in the art, other MEG and computing systems may be employed with similar success.
;_ ._.. T
iVW~Vw~~._L viv 1..~~
CA 02319227 2000-07-21 'r~~t r I ~ ~~=,! , Ontvanpene 22/12/AA 22:6A; ou~. oo. -......
12/22/99 ~ WED 13:56 FA1C B04 B81 4081 OYEN WIGGS GREEN MUT~1..A ~I~c~
At block 18 of Figure 3, the position of each sensor in the array of sensors is determined relative to the organ under observation. For this, both the sensors and the organ positions are referenced to a common coordinate system. As an example of the implementation of block 18, reference is made to Figure 4, in which the organ under eacamination is a human brain. In the present example, wherein the above-identified whole cortex MEO system is employed, the coordinate ~ system is established in a "head ".~ ~ the content typical of an MEC3'r system, the head frame is a coordinate system based upon three fiducial marks on the head (as referred to above).
These are: the nasion (NAS; i.e., the indentation point where the bridge of the nose meets the forehead, and the right and left pre-auricular points (I-PA and RPA, ~ respectively; i.e., points that are forward of the ear canals). These landmarks are used as follows to define a coordinate system:
(i) the Y-axis is defined by the line connecting the two pre-auricular points; .
(ii) the X-axis is defined as the line connecting the nasion with ~o Y-axis that is perpendicular to the Y-aus (i.e., not necessarily at the midpoint of the two pre-auricular points); and (iii) the Z-aus is defined as perpendicular to the XY-plane, at its origin.
The origin of this head frame is at the intersection of the three princi pal axes. Other coordinate systems, bash upon other fiducial points (either internal or external) exist for the head, and may also be used, provided that they are consistent with one another. Initially, the sensor i~~~i=a Ontiianpen: 22/12/A9 22:5~; ....-. .....
12/22/99 WED 13:56 FAZ 604 681 4081 OYEN WIGGS GREEN MUTALA ~ u~t coordinates are only known within the "helmet" frame of reference.
The fiducial points are measured relative to this "helmet" frame of reference. Once the positions of the fiducial points are determined, the head frame is constructed, and the sensors transformed into the head reference frame. Having a common frame of reference for the brain and the sensors is important, as the signals detected by MEG or EEG sensors depend upon. the relative positions and vectors of (i) the sensors, (ii) the source generators, and (iii) the intervening conducting medium (e.g., brain, skull, scalp, etc.).
As indicated in Figure 4, a plurality of sensors 152a through 152n are positioned in a predetermined manner about the head. As discussed above with reference to Figures 1 and 2; the plurality of sensors are fixed in a helmet worn by a subject such that the geometric relationship and orientation of each sensor 152 is known. A
coordinate system is shown in Figure 4, as indicated by the X, Y and Z axes, and an origin is selected as appropriate proximal a region desired for measurement. As will be understood by those of skill in the art, the coordinate system does not have to be Cartesian in nature but may be any other system suitable to the particular application. As indicated in Figure 4, the position of each of the plurality of sensors 152a...n can be located by a vector r, relative to the origin. The vector describing the orientation of each sensor is denoted u~.
Returning now to Figure 3, for the purpose of mapping the measured data to a voxel image array within a desired region of interest (ROI), an intcrvoxel spacing is selected and a grid of vowels is formed within the ROI referenced to the coordinate system (block 22).
As shown in Figure 4, any voxel position can be located by a vector ro CA 02319227 2000-07-21 ~MCi~t"~'w~ ~f"'~~~~
~~~/;!~~
On*vanp~n: 22/~2/AA 23:00 (7iU4 Obi rwca~ -- ...._, ,.....
12/22/99' WED 13:57 FAZ B04 B81 4081 OPEN WIGGS GREEN MUTALA (mQ
having both a magnitude and direction from the origin of the coordinate system. Accordingly, the current vector at ro is jo, corre-sponding to unit length current vector uo. This step corresponds to block 26 in Figure 3. The current vector at each vowel is best deter-s mined using anatomical constraints. For example, source current due to cortical activity in the brain is known to flow in a direction normal to the local cortical surface, at each point on that surface. This vector data could be obtained from an anatomical image, such as a magnetic resonance image (11~R~ or computed tomography (C'17 of the head.
This is provided that the image is translated into the common coordinate system determined for application the SAM method. In the absence of anatomical images, the current vector may be estimated by searching in a for a maximum in current deasity at each voxel as would be apparent to those of skill in the art..
Returning once again to Figure 3, at block 30, the forward solution per unit source current is computed for the potential data measured by the bioelectric sensors and/or tho magnetic field data measured by the biomagnetic sensors. The data measured by each sensor in the array of sensors at time sample k, neglecting sensor noise, is given by:
mi (k) = j.T(r, k) ~Gi (r) dv Vii) n where J(r) is the source current distribution throughout the conductive volume of the head ~ , and G,(r) is the sensitivity pattern of the ~
sensor, also known as the Green's function. This function applies to CA 02319227 2000-07-21 ~~!~~i~4!~ ~ jH
Ontvanpan: 22/12/A9 23:00; 604 681 4Ub~ -- --- -12./22/99 RWED 13:57 FAa B04 681 4081 OYEN ViIGGS GREEN MUTALA X030 either magnetic sensors such as a SQUID (superconducting quantum interference device) magnetometer or gradiometers, or to electrodes applied to the scalp. Equation 1 is a general expression which repre-sents forward solutions but is not readily solved. For practical pur-poses, the Green's function should account for factors such as: the size and extent of the current density; the conductive medium and bound-aries within which the currents flow; and, the geometry and positions of the sensors.
Practical forward solutions are known in the art. E~Ples of some of these published works include: a current dipole in free space (Law of Biot-Savart); a current dipole in a homogeneously conducting sphere (see, for example, Model Studies of the Magnetocardiogrom by F. Grynzpan and D.B. Geselowitz (Biophysics Journal 13:911-925 (1973), and Basic Mathematical and Electromagnetic Concepts of the BiomagneHc Inverse Problem by J. Sarvas (Phys. Med. Biol. 32:11-22 (1987)); a current dipole in a homogenously conducting volume bounded by a finite mesh (see, for example, Sarvas); and a current dipole in multiple concentric spheres, each sphere having differing conductivity (see, for example, Eccentric Dipole in a Spherical Medium: Generalized Expression For Surface Potentials by D.A.
Brody, F.H. Terry and E. Ideker (IEEE Transactions on Biomedical Engineering, BME-20:141,143). The current dipole model generally simplifies the need for knowing the current density everywhere in space, J(r), to simply assuming it is zero everywhere except for a single point in space. With the exception of the current dipole in free space, the various previously-mentioned forward solution models make use of bounded homogeneously conducting spaces, which may have ~~d~Ct~L.l,r__~ ..~.'iC~..l ~~L~~~ri Ontv~npsn: 22/12/~~ 23:01; s04 ee~ 4uo~ -- ..~...... ..._ 12/22/99 WED 13:58 FAa B04 881 4081 OPEN WIGGS GREEN MUTALA ~ 0;
-multiple compartments. Furthermore, they are a simplification of the actual medium which consists of various organs of the body and various tissue structures. At a finer level, conductivity varies rapidly, for example, between the interior of a blood vessel and the surround-s ing tissue. However,, as the method of the . present invention measures 'data at a macroscopic level, these effects tend to average out. There-fore, the assumption of homogeneous conductivity in each tissue compartment leads to accurate forward solutions.
Returning once again to Figure 3, at block 34, the operator optionally filters the measured bioelectric and/or biomagnetic data into selected frequency bands of interest. Specifically, since the distribu-tion of brain activity sometimes differs with frequency, it is useful to examine images of brain activity in specific frequency bands. Thus, for example, the collected data can ~be divided into distinct frequency bands, using a second-order digital bandpass filter. In some situations, it is preferable to filter the data for the purposes of separating artifacts due to, for example, mouth and eye movements. This filtering is also justified by the observation that somewhat different sources may appear in each frequency band.
The operator then selects at least one time window within the predetermined time interval within which the data is to be analyzed (block 38) and for which an image is to be constructed. At block 42 the optionally filtered and time segmented data is employed .to compute the covariance of the measured data. P.ach element of the covariance matrix is given by:
C,~ _ . K~ Imi (k) -mtl (mj (k) -m jl CA 02319227 2000-07-21 _ AI~CI~inJCJ ji"1.~.WT
. ,-, - ~, ~--, Ontvanpsn: 22/12/A9 23:01 ~ u..-. .."., "._ 12/22/99 WED 13:58 FA% 604 B81 4081 OYEN 1VIGGS GREEN MUTALA ~ U3 -where i, j =1,2,. . . ,M and the mean signal is computed in the conven-tional manner:
tits ~ K~ mi l k) ( 3 ) The above Equation is used for calculating the covariance of the data m and is used for measured data which is both bioelectric or biomagaetic in aature. For M channels, the covariance matrix will be MxM in size. The covariance may be computed as a time integral over any time window or multiple time windows in the time interval.
Covariance, for which the mean signal is removed is preferred, instead of the correlation matrix R, as real measurements may contain a DC
offset which originates in the sensors and are therefore not part of the electrophysiological measurements.
At block 46 of Figure 3, the results of blocks 30 and 42 are then combine to compute the Voxel RMS current density estimate. The mean RMS current density at any coordinate ro can then be repre-sented by the following linear algebraic relationship:
i J(r~, n~) G (Zo ~ noI ~ ~~ _~no) where ro is the selected three dimensional coordinate, no is the unit current vector at ro as previously described with respect to Figure 4.
C is the covariance matrix of the measurements, S is the uncorrelated sensor noise variance of all the measurements, and ~c is a regulariz-ation parameter which is selected to control the trade off between noise and spatial selectivity of the currant density estimate. In this matrix notation, G (without subscript) refers to the Green's functions of all CA 02319227 2000-07-21 f~l~JiE'lWl:-:-~ :J:~~~-~
OntJanpen: 22/l2/A9 23:02; uu-. uo. -,.",..
12/22/99 WED 13:59 FAR B04 B81 4081 OYEN WIGGS GREEN MUTALA ~ 03 sensors (i.e., G=G,,Gz,G3,...GM). The superscript "T" refers to the matrix transpose. Matrix and vector variables are designated by bold letters. For any given coordinate ro, the vector no specifying the direction of the current, can be selected by using the vector normal to that point on the surface of the cerebral cortex, for example, extracted from a segmented MRI or CT . image of the subject's brain. The results of Equation 3 is the SAM solution for the RMS current density for a voxel located at ro and having a current. (unit) vector direction no.
An alternative method which is only a useful approximation is to search for the angle no that maximizes the RMS current density at each position ro.
The direct conversion of MEG and/or EEG measurements, over some designated time period, into RMS currents by Equation 3 can be 1 S seen by suiistituting the measurements from Equation 1 into the covariance Equation 2, the result of which is approximated by:
csj = f ~ tr) IGs (r) ~Gj (r) l dv (5) The quantity in the square brackets is conventionally referred to as the "Gram~chmidt matrix", appearing frequently in the linear algebra of inverse problems. It is normally computed using an a priori model. of all possible sources. The covariance of the measure-ment can be seen here to be equivalent to a Gram-Schmidt matrix that has been weighted by the mean square current density that originally gave rise to the measurement. Thus, use of the covariance obviates ~IV~i~.~ ~~c~~ ~';='~
~ C_ = '~ ~ : . =.
OntVanp~n: 22/12/fl9 23:02; o~.. .,~~ ..,.,.. _ 12/22/89 WED 13:59 FAQ B04 881 4081 OPEN WIGGS GREEN MUTALA
the need of having an a priori model of the sources to perform inverse calculations.
Normally, and preferably, spontaneous brain signals are col-lected, without signal averaging. The correlation or covariance matrix is computed from the signals over some time interval or intervals selected from some or all of the measured signals. Alternatively, prior to computing covariance, signals from the array may be averaged synchronously to a stimulus presented to a subject, or be triggered by an external event such as a subject pressing a switch. The stimulus may be either visual, auditory, or somatosensory.
The vowels estimating the RMS current density are then dis-played, for example on a monitor, as a first image or a second image.
~MEi~DE~ S. iEE i
Claims (14)
1. A method of performing synthetic aperture magnetometery on the signals from a target organ using an array of biomagnetic sensors positioned in a predetermined manner around a target organ, each sensor in the array having a position vector and an orientation vector relative to a common coordinate system encompassing the target organ, the method comprising the steps of:
(i) simultaneously measuring EM signals from each sensor positioned in the array for a selected time interval;
(ii) computing a covariance matrix of the measured EM
signals over a user-selected time sub-interval within the selected time interval;
(iii) selecting a set of coordinates for a region of interest to be imaged and a distance between voxels to form a grid of voxels;
(iv) computing a forward solution for a current element at each of the voxels for each of the sensors in the array sensors;
(v) computing an RMS current density estimate for each voxel from the covariance matrix and the forward solution for that voxel; and (vi) displaying the voxels estimating RMS current density as a first image.
(i) simultaneously measuring EM signals from each sensor positioned in the array for a selected time interval;
(ii) computing a covariance matrix of the measured EM
signals over a user-selected time sub-interval within the selected time interval;
(iii) selecting a set of coordinates for a region of interest to be imaged and a distance between voxels to form a grid of voxels;
(iv) computing a forward solution for a current element at each of the voxels for each of the sensors in the array sensors;
(v) computing an RMS current density estimate for each voxel from the covariance matrix and the forward solution for that voxel; and (vi) displaying the voxels estimating RMS current density as a first image.
2. The method defined in claim 1, wherein the organ is a brain.
3. The method defined in claim 1, wherein the voxels are spaced to form Cartesian coordinate grid.
4. The method defined in claim 1, wherein the EM signals are magnetoencephalagraphic (MEG) signals.
5. The method defined in claim 1, wherein the EM signals are electroencephalographic (EEG) signals.
6. The method defined in claim 1, wherein the EM signals are a combination of MEG and EEG signals.
7. The method defined in claim 1, comprising, prior to Step (ii), the further Step of filtering the EM signals to a specified frequency band of interest.
8. The method defined in claim 1, wherein the further steps of repeating Steps (ii) through (vi) thereby computing a second RMS
current density image over a second selected time interval and subtracting the second RMS current density image, voxel by voxel, from the first image to form a third difference image.
current density image over a second selected time interval and subtracting the second RMS current density image, voxel by voxel, from the first image to form a third difference image.
9. The method defined in claim 1, wherein the first RMS current density image is displayed relative to the anatomy of the organ.
10. The method defined in claim 1, wherein the second RMS current density image is displayed relative to the anatomy of the organ.
11. The method defined in claim 1, wherein the third difference image is displayed relative to the anatomy of the organ.
12. A method of performing synthetic aperture magnecometery on an electrically active organ comprising the steps of:
(i) measuring a plurality of EM data simultaneously from an array of sensors positioned around the organ;
(ii) measuring the position of an organ in a common frame of reference with the array of sensors;
(iii) computing a covariance matrix of the signals over a first selected interval of time;
(iv) selecting a set of coordinates of a region of interest to be imaged and a distance between voxels to form a grid of vowels;
(v) computing a forward solution for. each sensor in the array of sensors at each of the vowels with each of the sensors current vector;
(vi) computing an RMS current density estimate for each voxel from said covariance matrix and from said forward solution for each said voxel; and (vii) displaying the voxels estimating RMS current density as a first image.
(i) measuring a plurality of EM data simultaneously from an array of sensors positioned around the organ;
(ii) measuring the position of an organ in a common frame of reference with the array of sensors;
(iii) computing a covariance matrix of the signals over a first selected interval of time;
(iv) selecting a set of coordinates of a region of interest to be imaged and a distance between voxels to form a grid of vowels;
(v) computing a forward solution for. each sensor in the array of sensors at each of the vowels with each of the sensors current vector;
(vi) computing an RMS current density estimate for each voxel from said covariance matrix and from said forward solution for each said voxel; and (vii) displaying the voxels estimating RMS current density as a first image.
13. The method defined in claim 12, wherein the organ is a brain.
14. The method defined in claim 12, wherein the vowels are spaced to form Cartesian coordinate grid.
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US7234098P | 1998-01-23 | 1998-01-23 | |
| US60/072,340 | 1998-01-23 | ||
| PCT/CA1999/000044 WO1999037206A1 (en) | 1998-01-23 | 1999-01-20 | System and method for measuring, estimating and displaying rms current density maps |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CA2319227A1 CA2319227A1 (en) | 1999-07-29 |
| CA2319227C true CA2319227C (en) | 2003-09-30 |
Family
ID=22106978
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA002319227A Expired - Fee Related CA2319227C (en) | 1998-01-23 | 1999-01-20 | System and method for measuring, estimating and displaying rms current density maps |
Country Status (8)
| Country | Link |
|---|---|
| US (1) | US6370414B1 (en) |
| EP (1) | EP1049402B1 (en) |
| JP (1) | JP2002500909A (en) |
| AT (1) | ATE209465T1 (en) |
| AU (1) | AU2043699A (en) |
| CA (1) | CA2319227C (en) |
| DE (1) | DE69900494D1 (en) |
| WO (1) | WO1999037206A1 (en) |
Families Citing this family (47)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3712348B2 (en) | 2000-07-18 | 2005-11-02 | 独立行政法人科学技術振興機構 | Cardiac magnetic field diagnostic apparatus for atrial flutter and atrial fibrillation and method for operating the same |
| US20050283053A1 (en) * | 2002-01-30 | 2005-12-22 | Decharms Richard C | Methods for physiological monitoring, training, exercise and regulation |
| US6996261B2 (en) * | 2001-01-30 | 2006-02-07 | Decharms R Christopher | Methods for physiological monitoring, training, exercise and regulation |
| JP3582495B2 (en) * | 2001-02-21 | 2004-10-27 | 株式会社日立製作所 | Biomagnetic field measurement device |
| JP3757815B2 (en) * | 2001-04-27 | 2006-03-22 | 株式会社日立製作所 | Biomagnetic field measurement device |
| JP2002355229A (en) * | 2001-05-31 | 2002-12-10 | Japan Science & Technology Corp | Method to analyze magnetic field and instrument to visualize current distribution |
| AU2003231992A1 (en) * | 2002-04-17 | 2003-11-03 | Board Of Trustees Of The University Of Arkansas | Uterine magnetomyography |
| US20040092809A1 (en) * | 2002-07-26 | 2004-05-13 | Neurion Inc. | Methods for measurement and analysis of brain activity |
| WO2004109300A2 (en) * | 2003-06-03 | 2004-12-16 | Decharms R Christopher | Methods for magnetic resonance signal perturbations measurement |
| US7672707B2 (en) * | 2003-06-11 | 2010-03-02 | Japan Science And Technology Agency | Sensor for magnetoencephalography meter and supermultichannel magnetoencephalography meter system using the same |
| FI115737B (en) * | 2003-09-26 | 2005-06-30 | Elekta Neuromag Oy | A method for using a multichannel measurement signal in source modeling |
| EP1814630A4 (en) * | 2004-11-15 | 2008-05-07 | Christopher Decharms | Applications of the stimulation of neural tissue using light |
| US7844211B2 (en) * | 2004-12-01 | 2010-11-30 | Kaonetics Technologies, Inc. | Ionizing communication disruptor unit |
| US8244342B2 (en) * | 2004-12-03 | 2012-08-14 | The Invention Science Fund I, Llc | Method and system for adaptive vision modification |
| US7656569B2 (en) * | 2004-12-03 | 2010-02-02 | Searete Llc | Vision modification with reflected image |
| US9155483B2 (en) | 2004-12-03 | 2015-10-13 | The Invention Science Fund I, Llc | Vision modification with reflected image |
| US7390088B2 (en) * | 2004-12-03 | 2008-06-24 | Searete Llc | Adjustable lens system with neural-based control |
| US7350919B2 (en) * | 2004-12-03 | 2008-04-01 | Searete Llc | Vision modification with reflected image |
| US7470027B2 (en) * | 2004-12-03 | 2008-12-30 | Searete Llc | Temporal vision modification |
| US7931373B2 (en) * | 2004-12-03 | 2011-04-26 | The Invention Science Fund I, Llc | Vision modification with reflected image |
| US7334892B2 (en) * | 2004-12-03 | 2008-02-26 | Searete Llc | Method and system for vision enhancement |
| US7594727B2 (en) * | 2004-12-03 | 2009-09-29 | Searete Llc | Vision modification with reflected image |
| US7334894B2 (en) * | 2004-12-03 | 2008-02-26 | Searete, Llc | Temporal vision modification |
| US8104892B2 (en) * | 2004-12-03 | 2012-01-31 | The Invention Science Fund I, Llc | Vision modification with reflected image |
| US7486988B2 (en) * | 2004-12-03 | 2009-02-03 | Searete Llc | Method and system for adaptive vision modification |
| US7344244B2 (en) * | 2004-12-03 | 2008-03-18 | Searete, Llc | Adjustable lens system with neural-based control |
| WO2006122398A1 (en) * | 2005-05-16 | 2006-11-23 | Cerebral Diagnostics Canada Incorporated | Near-real time three-dimensional localization, display , recording , and analysis of electrical activity in the cerebral cortex |
| US7698946B2 (en) | 2006-02-24 | 2010-04-20 | Caterpillar Inc. | System and method for ultrasonic detection and imaging |
| US20070239059A1 (en) * | 2006-03-21 | 2007-10-11 | Mciver Christopher R | Neurophysiology testing system |
| US7231723B1 (en) | 2006-08-08 | 2007-06-19 | The Boeing Company | Device for neural sensor placement and reference system measurements |
| JP4893213B2 (en) * | 2006-10-05 | 2012-03-07 | 株式会社日立製作所 | Magnetic field measuring device |
| US8170304B2 (en) * | 2007-04-03 | 2012-05-01 | Siemens Aktiengesellschaft | Modeling cerebral aneurysms in medical images |
| WO2008127720A2 (en) * | 2007-04-13 | 2008-10-23 | University Of Florida Research Foundation, Inc. | Atomic magnetometer sensor array magnetoencephalogram systems and methods |
| US20090306532A1 (en) * | 2008-06-06 | 2009-12-10 | Tucker Don M | Method for locating tracts of electrical brain activity |
| US8593141B1 (en) | 2009-11-24 | 2013-11-26 | Hypres, Inc. | Magnetic resonance system and method employing a digital squid |
| US8970217B1 (en) | 2010-04-14 | 2015-03-03 | Hypres, Inc. | System and method for noise reduction in magnetic resonance imaging |
| US8463571B2 (en) | 2010-11-12 | 2013-06-11 | International Business Machines Corporation | Performing reliability analysis of signal wires |
| WO2018047083A2 (en) | 2016-09-07 | 2018-03-15 | Texas Tech University System | Electric current imaging system |
| US11766204B2 (en) * | 2017-05-12 | 2023-09-26 | The Korea Research Institute of Standards and Science (“KRISS”) | Multi-sensor magneto-monitoring-imaging system |
| WO2019060298A1 (en) | 2017-09-19 | 2019-03-28 | Neuroenhancement Lab, LLC | Method and apparatus for neuroenhancement |
| US11717686B2 (en) | 2017-12-04 | 2023-08-08 | Neuroenhancement Lab, LLC | Method and apparatus for neuroenhancement to facilitate learning and performance |
| WO2019133997A1 (en) | 2017-12-31 | 2019-07-04 | Neuroenhancement Lab, LLC | System and method for neuroenhancement to enhance emotional response |
| US11364361B2 (en) | 2018-04-20 | 2022-06-21 | Neuroenhancement Lab, LLC | System and method for inducing sleep by transplanting mental states |
| EP3849410A4 (en) | 2018-09-14 | 2022-11-02 | Neuroenhancement Lab, LLC | System and method of improving sleep |
| JP7293814B2 (en) * | 2019-04-01 | 2023-06-20 | 株式会社リコー | Biological information measuring device, biological information measuring method and program |
| US11786694B2 (en) | 2019-05-24 | 2023-10-17 | NeuroLight, Inc. | Device, method, and app for facilitating sleep |
| CN111973180B (en) * | 2020-09-03 | 2021-09-17 | 北京航空航天大学 | Brain structure imaging system and method based on MEG and EEG fusion |
Family Cites Families (31)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4949725A (en) | 1988-07-01 | 1990-08-21 | Bio-Logic Systems Corporation | Apparatus and method for displaying electrical activity generated within a living body |
| US4947480A (en) | 1988-10-31 | 1990-08-07 | The United States Of America As Represented By The United States Department Of Energy | Multichannel signal enhancement |
| JPH02246926A (en) | 1989-03-17 | 1990-10-02 | Fujitsu Ltd | System for displaying estimated current source |
| US5269325A (en) | 1989-05-26 | 1993-12-14 | Biomagnetic Technologies, Inc. | Analysis of biological signals using data from arrays of sensors |
| US4977896A (en) | 1989-05-26 | 1990-12-18 | Biomagnetic Technologies, Inc. | Analysis of biological signals using data from arrays of sensors |
| DE3922150A1 (en) | 1989-07-06 | 1991-01-17 | Philips Patentverwaltung | METHOD FOR RECONSTRUCTING THE SPATIAL POWER DISTRIBUTION IN A BIOLOGICAL OBJECT AND ARRANGEMENT FOR IMPLEMENTING THE METHOD |
| EP0443069A1 (en) | 1990-02-22 | 1991-08-28 | Siemens Aktiengesellschaft | Method of measuring the electric or magnetic field pattern with a sensor arrangement |
| EP0449231B1 (en) | 1990-03-28 | 1995-11-29 | Hitachi, Ltd. | Process and apparatus for measuring the biocurrent distribution |
| EP0504027A3 (en) | 1991-03-15 | 1993-04-21 | Centro De Neurociencias De Cuba | Method and system for three-dimensional tomography of activity and connectivity of brain and heart electromagnetic waves generators |
| JPH0515504A (en) | 1991-07-09 | 1993-01-26 | Hitachi Ltd | Estimation method of magnetic field generation source |
| US5594849A (en) | 1991-08-09 | 1997-01-14 | Yale University | Biomedical magnetism imaging apparatus and method |
| DE4126949A1 (en) | 1991-08-16 | 1993-02-18 | Siemens Ag | METHOD FOR SIGNAL EVALUATION IN BIOMAGNETIC MEASURING DEVICES |
| JPH05220123A (en) | 1992-02-12 | 1993-08-31 | Hitachi Ltd | Method for estimating magnetic field generating source |
| US5228443A (en) | 1992-03-27 | 1993-07-20 | General Electric Company | Method for estimation and display of current source distribution from electric and magnetic measurements and 3D anatomical data |
| EP0582885A3 (en) | 1992-08-05 | 1997-07-02 | Siemens Ag | Procedure to classify field patterns |
| US5263488A (en) | 1992-10-05 | 1993-11-23 | Nicolet Instrument Corporation | Method and apparatus for localization of intracerebral sources of electrical activity |
| DE4326043C2 (en) | 1992-10-19 | 1995-08-17 | Siemens Ag | Method for localization of electrophysological activities overlaid with high noise |
| DE4326041C2 (en) | 1992-10-19 | 1995-11-16 | Siemens Ag | Method for localization of electrophysiological activities overlaid with high noise |
| DE4326044C2 (en) | 1992-10-19 | 1995-08-17 | Siemens Ag | Method for the localization of electrophysiological activities overlaid with strong noise |
| US5315308A (en) | 1992-11-04 | 1994-05-24 | Yale University | Method for eletromagnetic source localization |
| FI98337C (en) * | 1992-11-30 | 1997-06-10 | Risto Juhani Ilmoniemi | Method and apparatus for distinguishing between brain excitation responses and spontaneous function and different components of signals measured from the heart |
| FR2699347B1 (en) | 1992-12-14 | 1995-02-10 | Commissariat Energie Atomique | Method and device for extracting a useful signal of spatial extension finite at each instant and variable over time. |
| JPH06245916A (en) | 1993-02-25 | 1994-09-06 | Hitachi Ltd | Estimating method for magnetic field generation source in living body magnetic field measurement |
| JP2739804B2 (en) | 1993-05-14 | 1998-04-15 | 日本電気株式会社 | Dipole estimator |
| US5601081A (en) | 1993-06-04 | 1997-02-11 | Shimdaszu Corporation | Method and apparatus for deducing bioelectric current sources |
| US5526811A (en) | 1993-06-15 | 1996-06-18 | Biomagnetic Technologies, Inc. | Apparatus and process for determining the sources of biomagnetic activity |
| US5687724A (en) | 1993-10-26 | 1997-11-18 | Abratech Corporation | Apparatus and method for determining variations in measured physical parameters of signal-generators |
| JP3519141B2 (en) | 1994-10-31 | 2004-04-12 | 株式会社東芝 | In vivo current source estimation method |
| US5701909A (en) | 1994-11-01 | 1997-12-30 | Abratech Corporation | Machine and method for the determination of nervous-system-generator parameters using lead fields |
| US5657756A (en) | 1995-06-07 | 1997-08-19 | Ctf Systems Inc. | Method and systems for obtaining higher order gradiometer measurements with lower order gradiometers |
| JP3473210B2 (en) | 1995-08-31 | 2003-12-02 | 株式会社島津製作所 | Biomagnetic measurement device |
-
1999
- 1999-01-20 US US09/600,730 patent/US6370414B1/en not_active Expired - Lifetime
- 1999-01-20 AT AT99900857T patent/ATE209465T1/en not_active IP Right Cessation
- 1999-01-20 CA CA002319227A patent/CA2319227C/en not_active Expired - Fee Related
- 1999-01-20 EP EP99900857A patent/EP1049402B1/en not_active Expired - Lifetime
- 1999-01-20 WO PCT/CA1999/000044 patent/WO1999037206A1/en active IP Right Grant
- 1999-01-20 DE DE69900494T patent/DE69900494D1/en not_active Expired - Lifetime
- 1999-01-20 AU AU20436/99A patent/AU2043699A/en not_active Abandoned
- 1999-01-20 JP JP2000528204A patent/JP2002500909A/en not_active Ceased
Also Published As
| Publication number | Publication date |
|---|---|
| JP2002500909A (en) | 2002-01-15 |
| AU2043699A (en) | 1999-08-09 |
| WO1999037206A1 (en) | 1999-07-29 |
| DE69900494D1 (en) | 2002-01-10 |
| EP1049402B1 (en) | 2001-11-28 |
| CA2319227A1 (en) | 1999-07-29 |
| US6370414B1 (en) | 2002-04-09 |
| EP1049402A1 (en) | 2000-11-08 |
| ATE209465T1 (en) | 2001-12-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA2319227C (en) | System and method for measuring, estimating and displaying rms current density maps | |
| US6539245B2 (en) | Biomagnetic field measuring method and apparatus | |
| Hämäläinen et al. | Magnetoencephalographic (MEG) characterization of dynamic brain activation | |
| US5307807A (en) | Method and system for three dimensional tomography of activity and connectivity of brain and heart electromagnetic waves generators | |
| US6681131B2 (en) | Apparatus for measuring bio-magnetic fields | |
| US5269325A (en) | Analysis of biological signals using data from arrays of sensors | |
| US5885215A (en) | Method of reconstructing the spatial current distribution in a biological object, and device for performing the method | |
| Abraham-Fuchs et al. | Pattern recognition in biomagnetic signals by spatio-temporal correlation and application to the localisation of propagating neuronal activity | |
| Leder et al. | Noninvasive biomagnetic imaging in coronary artery disease based on individual current density maps of the heart | |
| George et al. | Anatomical constraints for neuromagnetic source models | |
| US6941165B2 (en) | Cardiac magnetic field diagnosing apparatus by late ventricular potential and method of locating intramyocardial excitement uneven propagation portion | |
| Ustinin et al. | Functional Tomography of Complex Systems Using Spectral Analysis of Multichannel Measurement Data | |
| Erné et al. | Magnetocardiography under clinical conditions | |
| US20240122549A1 (en) | Localizing physiological signals | |
| Ziolkowski et al. | Postprocessing of 3-D current density reconstruction results with equivalent ellipsoids | |
| JP2003038455A (en) | Organism activity analyzer | |
| Ioannides | Magnetic field tomography: theory, applications and examples from the visual system | |
| Aburidi et al. | Normalized Cumulative Periodogram Method For Tuning A Regularization Parameter of The EEG Inverse Problem | |
| Knösche et al. | Electro-and Magnetoencephalographic Measurements | |
| KR20230150303A (en) | Localization of physiological signals | |
| König et al. | Tiny signals from the human brain: acquisition and processing of biomagnetic fields in magnetoencephalography | |
| Rodriguez-Rivera | Biomagnetic and Bioelectric Imaging | |
| Lewis et al. | Magnetoencephalography | |
| CN116801791A (en) | Diagnostic support device, diagnostic support method, and diagnostic support program | |
| Zhou et al. | Imaging of cardiac electrical excitation conduction |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| EEER | Examination request | ||
| MKLA | Lapsed |