EP1906822A2 - Method and device for representing a dynamic functional image of the brain, by locating and discriminating intracerebral neuroelectric generators and uses thereof - Google Patents

Abstract

The invention concerns a method for representing a dynamic functional image of the brain. It consists in acquiring (A) for a specific duration a plurality of electrophysiological signals {eSi}<SUP>N</SUP> <SUB>1</SUB>

Description

 METHOD AND DEVICE FOR REPRESENTING AN IMAGE

FUNCTIONAL DYNAMIC BRAIN. BY LOCATION AND

DISCRIMINATION OF NEUROELECTRIOUS GENERATORS

INTRACEREBRAUX AND THEIR APPLICATIONS

The present invention relates to a method and a device for representing a dynamic functional image of the brain, by locating and discriminating intracerebral neuroelectric generators and their applications. Every cerebral act, in an individual, emerges from a cooperation between several neural networks spatially distributed in the intracerebral space into a functional network.

At present, despite their recent progress, the main brain imaging techniques such as EEG for Electro Encephalography, MEG for Magneto Encephalography, fMRI for Functional Magnetic Resonance Imaging and PET for Spinal Emission Tomography Positons can provide only a map of brain activation zones, but can not directly account for the interactions between these zones and these activators.

Indeed, the characterization of these functional networks requires, at the same time, the identification of the involved brain zones, the understanding of the mechanisms of interaction between these and the precise quantification of these interactions.

The observation of the operation of the above-mentioned neural networks is not possible from the sole mapping of brain activities provided by the aforementioned imaging techniques. Indeed, the discrimination, among all simultaneously active cerebral zones, of those participating in the same functional network can not be carried out, because the simple observation that zones are simultaneously active is not sufficient to conclude that these zones are engaged in a same functional, pathological or cognitive process. The set of similar goal-oriented approaches, currently known, are based on the concept that the existence of a coupling between two intracerebral zones must result in a correlation between their neuroelectric activity.

The activity of a group of neurons, such as for example a cortical column, can be characterized by two types of physiological measurements:

1) a temporal coding with the rate of neuronal discharges per second; or

2) a coding of the synchronization of the oscillatory activities of the brain zones involved in the same functional network. The second type of physiological measurements mentioned above has been the subject of application work.

A first application has been the subject of US Pat. Nos. 6,442,421 and 6,507,754 issued in the name of LE VAN QUYEN, J. MARTINERIE, F. VARELA and M. BAULAC. The above application essentially concerns a method and a device for anticipating epileptic seizures, from a surface electroencephalogram.

A second application was the subject of a French patent application FR 2 845 883 in the name of CNRS (National Center for Scientific Research).

This second application concerns the characterization of cognitive states from surface encephalograms.

The aforementioned application gives satisfaction. However, it appears limited, since it is essentially based on a process of statistical validation of a period analyzed in real time.

The present invention relates to the implementation of a method and a device for establishing a true representation of a dynamic functional image of the brain, by location and discrimination of intracerebral neuroelectric generators, in the whole of the world. intracerebral space.

Another object of the present invention is, in addition, the implementation of a method and a device for establishing a plurality of functional images. dynamics of the brain, one or more of these dynamic functional images that can be associated with the same functional, pathological or cognitive process.

Another object of the present invention is, in addition, the implementation of a method and a device for establishing one or more dynamic functional images of the brain for characterizing both temporal information and spatial information. on the neuroelectric activity of areas of brain activity forming a functional network.

Another object of the present invention is, in particular, the implementation of a method and a device for representing a dynamic functional image of the brain allowing the execution of true non-invasive imaging of the integration. functional and functional connectivity of brain areas in a state of functional, pathological or cognitive processes.

Another object of the present invention is in particular the implementation of a tool, from the method and the device of the invention, allowing the execution of a characterization of the signatures of substances, drugs or drugs generating d one or more states of functional abnormality of the brain caused, represented as dynamic functional images.

Another object of the present invention is finally the implementation of a tool from the method and device objects of the invention, allowing the execution of a characterization of the signatures of specific cognitive states such as vigilance, attention diffuse, sleepy or otherwise, represented as dynamic functional images.

The method of representing a dynamic functional image of the brain, by location and discrimination of the intracerebral neuroelectric generators, object of the invention, is remarkable in that it consists at least in acquiring, during a determined recording duration, a plurality of electrophysiological signals emitted and / or induced by cerebral activity from a plurality of electrodes substantially distributed on the scalp of the skull containing the brain and digitizing these electrophysiological signals to form a database of analysis of the brain. brain activity, locate all the generators - AT -

neuroelectrics in the cerebral volume from an acquisition of an electronic map of the position of the electrodes, a three-dimensional image of this brain by successive sections, and the recording of electrophysiological signals on these electrodes. Starting from a segmentation of the cerebral cortex obtained from the three-dimensional image of this brain, electrophysiological signals and the electronic mapping of the localization, the application of the inverse problem allows the spatial localization of the intracerebral neuroelectric generators, and to calculate among the active zones comprising neuroelectric generators, the synchronism existing between all the pairs of these neuroelectric generators. This quantization is performed for a plurality of frequency bands, to detect groups of discriminating neural networks and to constitute a database of reference states representative of this dynamic functional image.

The method which is the subject of the invention is, moreover, remarkable in that, for a dynamic functional image acquired during a determined recording duration, it consists in identifying the functional image with a class of functional images among a plurality of functional images, each class of functional images of this plurality of functional image classes characterizing a brain state of the subject's brain. The dynamic functional image of the brain, object of the invention, is remarkable in that it comprises at least one three-dimensional image of this brain in successive sections each representing an elementary image of this brain, and on at least one elementary image. at least one neuroelectric generator of neuroelectric signals represented by a marker, each neuroelectric generator being characterized in position in this elementary image, in electrical current density and in the direction of emission of neuroelectric signals, each neuroelectric generator of a current elementary image next to a neuroelectric generator of a previous and / or subsequent elementary image and having substantially the same direction of emission of neuroelectric signals and a synchronism over a determined duration of coherence constituting a group of discriminant neural networks of functional state representative of this dynamic functional image of the brain.

The device for representing a dynamic functional image of the brain, object of the invention, is remarkable in that it comprises at least one acquisition circuit for a determined recording duration of a plurality of electrophysiological signals emitted and / or induced by the brain activity from a plurality of electrodes substantially distributed on the scalp of the cranial box housing the brain, this acquisition circuit for storing and saving these electrophysiological signals to form a database of cerebral activity analysis, a circuit for acquiring a three-dimensional image of the brain by successive sections, a module for calculating the location of the set of neuroelectric generators of the intracerebral neuroelectric signals in the cerebral volume from the position electrodes, the three-dimensional image of this brain, segmentation of the cerebral cortex, and the application of the inverse problem, a discrimination module, among the active zones comprising neuroelectric generators, of the synchronism existing between the pairs of neuroelectric generators in a plurality of frequency bands, for detecting groups of discriminating neural networks and constituting a database of reference states representative of this dynamic functional image. The method and the device, objects of the invention, find application to the non-invasive functional study of the human brain in the most diverse situations such as, in particular, the study of functional abnormalities caused or not by the ingestion of drugs, drugs, the categorization of functional and / or clinical states and their rationalized relationship to specific pathological or cognitive states.

They will be better understood from reading the description and from the following drawings in which:

FIG. 1a illustrates, in an illustrative manner, a sectional view, along a plane of vertical symmetry, of the entire head of a subject whose scalp is provided with an array of electrodes, in order to allow the implementation of the method which is the subject of the invention;

FIG. 1b represents a flow diagram of the essential steps for implementing the method that is the subject of the invention under the conditions illustrated in FIG. FIG. 1c represents, in an illustrative manner, a succession of elementary dynamic functional images constituting a dynamic functional image, in accordance with the object of the present invention, making it possible to highlight groups of discriminating neural networks constituting an area of brain activity constituting a functional network; FIG. 2 represents, in the form of a temporal diagram, the implementation of a window for recording and analyzing electrophysiological signals, the recording duration and the duration of the window being parameterizable according to the class. functional image chosen to characterize the brain state of the subject's brain; FIG. 3 represents, by way of illustration, a detail of implementation of the step of locating all the neuroelectric generators in the cerebral volume represented in FIG. 1b;

FIG. 4a represents a timing diagram of EEG type raw signals delivered by a pair of electrodes placed on the scalp of a subject for a determined recording duration;

FIG. 4b represents a timing diagram of the signals of FIG. 4a obtained after filtering;

FIG. 4c represents the phase difference obtained by spectral analysis of the signals represented in FIG. 4b; FIG. 4d shows the signal of variation of the phase difference of the signals represented in FIG. 4c on the recording duration making it possible to highlight the synchronism of these signals over certain ranges of the recording duration; FIG. 5 represents a specific dynamic functional image of a brain in which the neuroelectric generators associated with fingers of the right hand of a normal subject have been represented;

FIG. 6a represents, by way of illustration, a functional block diagram of a device for representing a dynamic functional image of the brain that is the subject of the invention;

FIG. 6b shows a flexible helmet provided with electrodes enabling the acquisition of electrophysiological signals.

The method of representing a dynamic functional image of the brain, object of the invention will now be described in connection with Figures la, Ib and the following figures.

In the figure there is shown a sectional view along a vertical plane of symmetry of an entire head of a subject, for which the method object of the invention is applied. The choice of the cutting plane is given by way of non-limiting example, any cutting plane distinct from the latter may be used.

As shown in FIG. 1a, C _k denotes the section of the brain C and the entire head along the abovementioned cutting plane, this section being accordingly represented in the plane of FIG. The head of the subject and, in particular, the scalp S of the latter is equipped with a plurality of electrodes distributed on the scalp S of the cranial box housing the brain.

C. The plurality of electrodes is denoted {E}; the plurality of electrodes being deemed to have N electrodes, for example distributed substantially uniformly over the scalp of the subject. As represented in FIG. 1a, O denotes an arbitrary reference point located for example in the plane of section C _k and Oxyz a given referential for locating any point P of the brain C in polar coordinates, for example r, θ, φ, vis-à-vis this repository.

It is thus understood that, for the implementation of the method that is the subject of the invention, each electrode Ej makes it possible to collect an electrophysiological signal noted es; EEG type and / or MEG to allow the implementation of the method object of the present invention.

With reference to FIG. 1b, and taking into account the indications given in connection with FIG. 1a, the method which is the subject of the invention is remarkable in that it consists at least in acquiring, in a step A, during a period D determined record, a plurality of electrophysiological signals these electrophysiological signals being noted .

The aforementioned electrophysiological signals are emitted and / or induced by brain activity C and are collected from the plurality of electrodes (Ei) ι. The aforementioned electrophysiological signals are digitized to form a cerebral activity analysis database, this database being denoted by DBe and the whole of the recording being noted M (t).

Regarding the nature of the electrophysiological signals es; In addition to the signals directly generated by the cerebral activity, as mentioned above, additional signals can be acquired simultaneously, these additional signals being able to consist of signals generated by the movement of the subject's eyes, signals of cardiac activity and finally any electrophysiological signal that may be recorded during the recording period. The set of aforementioned signals is then organized as mentioned above to form the DBe database.

As represented in FIG. 1b, step A is then followed by a step B consisting of locating all the neuroelectric generators corresponding to the brain activity of the subject, this location being carried out in the cerebral volume. The aforementioned location is advantageously performed from an acquisition of the electronic map of the position of the electrodes placed on the patient's scalp, as shown in FIG. 1a, and, in addition, a three-dimensional image of the patient. brain C represented by a set of successive cuts, this set of successive cuts being noted {Ck} f. It is understood in particular that, given the known position of the acquisition electrodes E ₁ , and, of course, the three-dimensional image of the brain

C formed by the set of sections {θt} f we obtain a segmentation of the cerebral cortex, as will be described later in the description, with the location of the positions of the electrodes on this modeling.

The location of all the neuroelectric generators in the brain volume is then obtained by applying the inverse problem, the inverse problem being defined as that relating to obtaining the local density of currents in the cerebral cortex and, in particular, in the segmentation of the latter from the voltage measurements M (t) obtained from the electrophysiological signals esi delivered by the set of electrodes {Eι} f.

It is thus understood that the application of the inverse problem makes it possible to determine, from all the electrophysiological signals {esι} f, the electronic mapping of the location of the electrodes, and, of course, of the three-dimensional image of the brain formed by the set of successive sections, making it possible to obtain a segmentation of the cerebral cortex, the spatial location of the neuroelectric generators of the intracerebral neuroelectric signals.

In FIG. 1b, in step B thereof, all the generators of the intra-cerebral neuroelectric signals are noted. It is understood, in particular, that each neuroelectric generator of the intracerebral neuroelectric signals is defined not only in amplitude, that is to say in local current density, but also in orientation at each point P (r, θ, φ) of the brain C as described previously in the description.

When step B has been carried out, in accordance with the method that is the subject of the invention, the set of the aforementioned intracerebral generators is available at each instant t in each successive section of rank k and thus ultimately in the entire volume. intracerebral.

Step B is then followed, as shown in FIG. 1a, of a step C consisting of discriminating, among the active zones of the brain and in particular of each cut C _k comprising neuroelectric generators, the synchronism existing between the pairs of neuroelectric generators in a plurality of frequency bands for detecting groups of discriminating neural networks constituting functional networks born from the brain activity of the subject. In step C of FIG. 1b, this synchronism discrimination operation is noted {gft} jf -> RNdk symbolically.

In the preceding relation, RN _dk designates the discriminant neural network groups corresponding to a functional network as mentioned above, for a section C _k for example. Following the execution of the above-mentioned step C, at the end of the execution of the process which is the subject of the invention, that is to say in step D, a functional image which can be formed by elementary dynamic functional images that may each correspond to one of the sections C _k to which at least one active neuroelectric generator gjk and a group or part of a group of discriminating neural networks RNd _{k are associated} , the elementary functional image being noted for this reason

It is furthermore indicated that each functional image may correspond to a projection or intersection of a set of elementary dynamic functional images, each corresponding for example to one of the sections C _k5 by a representation plane of any orientation with respect to the direction of the cuts.

Referring to FIG. 1c, there is shown a plurality of functional images formed by successive sections Ck _-1 , Ck and C _{k +} t on which different neuroelectric generators g / k have been represented, each generator being of course positioned relative to each other. to the Oxyz repository as mentioned above and each neuroelectric generator being defined in amplitude, that is to say in current density, and in orientation with respect to a reference Px'y'z ¹ linked to the original reference frame. Referring to FIG. 1c, it is furthermore indicated that a group of discriminant neural networks is then constituted by a group of neuroelectric generators present on elementary images and thus on successive sections Ck _-1 , Ck and Ck _{+ 15} the generators above having a similar orientation and meeting the criterion of synchrony, as defined in connection with step C of Figure Ib.

The method, object of the invention, further allows, for each functional image acquired during a recording duration D, to identify the functional image to a functional image class among a plurality of functional image classes, each class of functional images of this plurality of functional image classes for characterizing a brain state of the subject, as will be described later in the description.

Thus, with reference to FIG. 2, it is indicated that the step of acquiring and processing a plurality of electrophysiological signals {is} f is performed in real time with a maximum recording delay of less than 100 milliseconds.

With reference to the above-mentioned FIG. 2, it is indicated that the recording duration can be set over a duration range, the recording duration D being able to be between a minimum recording time of the order of 20 minutes for the duration of recording. recording and representation of a functional image of the brain relating to one or more cognitive states and a recording duration D of several days denoted D = x days in FIG. 2, for the recording and representation of a functional image brain related to one or more states of functional abnormality of the brain caused or not provoked. The states of abnormality provoked can be caused by the ingestion of drugs, drugs or finally any substance by accidental ingestion for example.

It is understood, in particular, that in view of the maximum recording delay of less than 100 milliseconds, the recording of electrophysiological signals is performed by sampling with a sufficient sampling frequency for this purpose. With regard to the exploitation of the above-mentioned registration data, that is to say data M (t) previously mentioned in the description constituting the database DBe, these recorded data can be exploited in this manner. after discriminating among the active zones of the synchrony existing between the electrophysiological signal generating neuroelectric pairs during the recording duration as represented in FIG. 2.

The exploitation of the aforementioned signals then consists in making this discrimination on a sliding time window whose duration f is between 50 milliseconds and 2 seconds for the representation of a functional image of the brain relating to one or more cognitive states and on a window slippery temporal duration whose duration is between 5 s and 20 s for a representation of a functional image of the brain relating to one or more states of functional brain abnormality caused or not caused as well as further represented in FIG. 2.

A more detailed description of the locating step B of the neuroelectric generators will now be given in connection with FIG.

A justification of the procedure described in connection with the aforementioned figure will be introduced beforehand. The discretization of the integral equations governing the calculation of electric scalp potentials makes it possible to establish at each moment a linear relationship between the measurements M (t) and the amplitudes, that is to say the current densities of the distributed neuroelectric generators. in the brain volume. In the presence of an additive noise, the problem therefore consists in estimating the distribution of the cortical currents or current densities J at the origin of the recorded signals M (t) and thus consists in solving an inverse problem like many other other applications of image reconstruction in medical imaging for example.

The source estimation problem, that is to say the neuroelectric generators of an electromagnetic field measured on the outer surface of a conducting volume, does not admit of a single solution. In the sense of J. Hadamard this is a fundamentally ill-posed problem. The method, which is the subject of the invention, therefore proposes using an estimator which makes it possible to impose controlled anatomical and electrophysiological constraints and guaranteeing the uniqueness of the estimate obtained. The corresponding estimator is now described with reference to FIG.

With reference to the above-mentioned FIG. 3, the entire recording of the measurements denoted M (t) = G (r, θ, φ) J (t) is available.

In the previous relationship, we recall that:

M (t) denotes the set of recordings obtained, that is to say the values of the electrophysiological signals XeSi) ₁ in the form of an electric potential value, for example on the surface of the scalp,

- G (r, θ, φ) denotes the transfer matrix between the surface electrophysiological signals present on the scalp at each local point of the intracerebral volume and the estimated corresponding local current density, denoted by J (t).

The locating step B as shown in FIG. 3 then consists of executing in a step B ₁ consisting in applying the stresses resulting from the individual anatomy introduced by segmentation and surface mesh of the parenchyma.

This operation is performed from the set of successive cuts {Ck} f to obtain the aforementioned mesh noted m _u .

Step B ₁ is then followed by a step B ₂ for calculating local current densities by solving the inverse problem according to the relation: + λ I. λ is the regularization term and I the identity matrix. Following step B _2, local current densities are obtained at a given instant at any point with coordinates r, θ, φ of the intracerebral volume.

In the previous relationship, we indicate that:

J (t) denotes the estimate of the local current density, G ¹ denotes the transposed transfer matrix of the matrix G representing the transfer matrix G (r, θ, φ),

- (G G) denotes the pseudo inverse of the transfer matrix G.

Due to the constraints of speed of estimation of the aforementioned current density values sought for the applications of the method and the device, objects of the invention, and under the assumption of an independent and identically distributed Gaussian noise, a satisfactory solution in terms of the compromise between spatial resolution and computational time is that which minimizes the energy of the residuals and the norm of the neural currents, the estimator thus constituted being an unbiased estimator with minimal norm in the sense of least squares.

Step B ₂ is then followed by a step B ₃ of calculating in position the functional parameters, that is to say the amplitude and orientation of the neuroelectric generators gjk, in the form of a source of elementary electric current on the mesh of the cortical surface. In step B ₂ of FIG. 3, this operation represented by the symbolic relation:

There are thus active zones, each active zone comprising at least one neuroelectric generator. With regard to the implementation of step B _{2, it} is indicated that the physical models involving the measurements M (t) use the resolution of Ohm's law in three dimensions. Indeed, it is justified to neglect the phenomena of electromagnetic field propagation at physiological frequencies implemented. The corresponding modeling can then be performed either analytically in the context of the spherical geometry vis-à-vis the aforementioned original reference, or numerically considering the specific geometry of the envelopes of bone tissue and scalp S.

A more detailed description of a specific mode of implementation of step C of synchrony discrimination, previously described in connection with FIG. 1b, will now be described with reference to FIGS. 4a to 4d. In general, with reference to the above figures, it is pointed out that the step of synchronism discrimination existing between the pairs of neuroelectric generators among the active zones comprising neuroelectric generators in a frequency band consists at least in evaluating the statistical level of the PLS synchronization between two signals of a pair of neuroelectric generators, by means of the circular variance of the phase difference between these signals or the standardized Shannon entropy, of this phase difference.

Prior to a description proper in connection with Figures 4a to 4d, a theoretical justification will be given below. In general, the instantaneous phase of a signal can be calculated using an analytical signal, the concept of analytic signal having been introduced by Gabor in 1946 and recently applied to experimental data.

Thus, with reference to the aforementioned concept, for an arbitrary signal s (t), that is to say finally for any recorded electrophysiological signal M (t), the analytic signal z is a complex time-dependent function and defined according to the following relation: ζ {t) = s {t) + Js (t) = A (t) e ^m) (1).

In the previous relation, the function Js (t) is the Hubert transform of s (t) of the form:

In the aforementioned Hubert transform, P.V. indicates that the integral is calculated in the sense of the main value of Cauchy. The instantaneous amplitude A (t) and the instantaneous phase F (t) of the signal S (t) are only defined by the preceding relation 1. With reference to the relation 2,? (T) is considered as the product of convolution of the signal s (t) and of 1 / π.

Consequently, this means that applying the Hubert transform to the signal s (t) is equivalent to applying a filter whose amplitude response is unitary and the phase response shifted by π / 2 for all the frequencies. While the aforementioned transform process can theoretically be applied to broadband signals, the notion of phase in this case is not very explicit and, in practice, only narrow band signals obtained by filtering are used. As a result, filtering is applied in a specific frequency band. However, several frequency bands can then be retained, but the same frequency band is used for two signals in case of 1: 1 synchrony. Other frequency bands can be used for studying synchrony n: m. The statistical level of PLS synchronism between two signals is evaluated by two indices: the circular variance and the phase difference between the signals or Shannon's normalized entropy of the phase difference.

We recall that the circular variance satisfies the relation:

and that Shannon's normalized entropy satisfies the relation: = (H _msk -H) _{lH ^} .

In the aforementioned relation, entropy is defined by the relation:

.

In the previous relationship:

M denotes the number of classes of phase value, H _max = ln (M) the maximum entropy,

- p _m the relative frequency of the phase difference in the mth phase value class,

- In the natural logarithm.

The optimal number of phase value classes is M = exp [0.626 + 0.41n (P-I)] where P is the number of phase difference to be classified.

Given the introduction of the aforementioned normalization, the values of γ are between 0 uniform distribution and no synchronization and 1 perfect synchronization. The above calculation is done for all the estimated source pairs or possibly by random draws or oriented to reduce the calculation time.

For a number of sources, that is to say neuroelectric generators, equal to 27, the number of distinct pairs is 325 and for 64 generators, it amounts to 1953. For a few hundred sources this procedure is impossible to establish.

In practice, and in accordance with the implementation of the method, object of the invention, the real-time calculation of the synchronies can advantageously be limited to 100 generators. A choice of regions of interest for real-time processing is then performed and is a function of the experimental protocol adopted and the use of statistical techniques for reducing information (discriminant analysis, spatial filters, etc.).

Thus, with reference to FIGS. 4a to 4d above, step C of synchrony discrimination can consist, for example, from the raw signals represented in FIG. 4a, for two signals constituting a pair recorded over the duration D d. recording, filter on a plurality of frequency bands to obtain the filtered signals as shown in Figure 4b, and then perform the spectral analysis process previously mentioned in the description, to obtain the instantaneous phase differences between the aforementioned signals as shown in Figure 4c.

The statistical study as previously mentioned from the circular variance indices of the phase difference between the signals where the standardized Shannon entropy of this phase difference can then be executed to obtain a quantization of the phase differences represented in FIG. 4b, where the synchronies SyI and Sy2 can be highlighted, for a substantially minimal phase difference and constant in relative value vis-à-vis other areas of the recording time.

Finally, with reference to FIG. 4d, the synchronism between pairs of neuroelectric generators can advantageously be established by synchronous time ranges. This allows to temporally represent the activity of the pairs of neuroelectric generators and to obtain a true dynamic functional image of the brain.

Since the neuroelectric generators have been placed on the elementary functional dynamic images and, in particular, on a succession of them as shown in FIG. 1c, the method which is the subject of the invention then makes it possible to obtain any dynamic functional image of the brain as shown in Figure 5.

Such an image comprises at least one three-dimensional image of the brain in successive sections, each section representing an elementary image of the brain as mentioned in connection with FIG. In Figure 5, the successive sections are not shown, so as not to overload the drawing.

In addition, as shown in FIG. 5 above, the dynamic functional image comprises, on at least one elementary image, if necessary on several, at least one neuroelectric generator of intracerebral neuroelectric signals represented by a marker. In FIG. 5 the marker is constituted by an oriented arrow whose amplitude represents in fact the local current density at the positioning point of the corresponding neuroelectric generator and whose orientation corresponds exactly to the orientation in the reference frame of origin of the electrical current generated by the neuroelectric generator.

With reference to FIG. 5 it is indicated that each neuroelectric generator is characterized in position in the elementary image and, finally, in the resultant dynamic functional image in electrical current density and in the emission direction of the corresponding neuroelectric signals.

It is thus conceivable that each neuroelectric generator of a current elementary image, neighboring a neuroelectric generator of a previous and / or subsequent elementary image, as represented in FIG. 1c and having substantially the same direction of emission of electrical signals and a synchronism over a determined duration of coherence constitutes a group of neural networks discriminating functional states representative of the dynamic functional image of the brain. More particularly, it is pointed out that FIG. 5 advantageously represents the neuroelectric generators associated with the finger of the right hand of a normal subject, that is to say which has no functional abnormality in the fingers of the and, consequently, no functional abnormality of the corresponding brain.

In FIG. 5 it can be observed that each finger is represented by a neuroelectric generator constituting an equivalent dipole. The aforementioned generators are oriented and are perpendicular to the cortical surface and tangential to the surface of the head and correspond to the activity of macro column of neurons located in the central groove shown in Figure 5. It can be recognized, at the observation In Figure 5, the thumb Th is represented by the arrow oriented, the index I by a particular arrow, the middle finger M by another parallel arrow and the little finger A by a different parallel arrow.

It can be seen that the neuroelectric generators associated with the aforementioned fingers are represented in the anatomical order with great precision.

It is understood, in particular, that the functional images obtained, thanks to the implementation of the method that is the subject of the present invention, then make it possible to immediately detect any functional abnormality of cortical representation of the human body in the brain, the functional images mentioned above being able to of course, be subdivided into representative classes either of the state of absence of functional anomaly, or conversely, representative of a class of functional anomalies and subclasses corresponding to an anomaly of one of the fingers considered.

The distribution of the dynamic functional images obtained, thanks to the implementation of the method that is the subject of the invention, according to a category of classes, makes it possible to implement the method, object of the invention, for the purpose of discrimination with a decisional purpose. .

This is the case in particular in the example given above in connection with FIG. Thus, for a given recording time D of one or more seconds, for example, and for a given synchrony of the electrophysiological signals es, it is then possible to assign the corresponding dynamic functional image obtained to a specific class characterizing a state. brain among many.

The corresponding problem is that of a classification and assumes, of course, the definition a priori of a set of classes as mentioned above in relation with FIG.

This decision-making procedure must take into account all electrode pairs Ej. Under these conditions, for an N number of electrodes equal to 100 and for 14 frequency bands determined by the filtering performed in the case of the processing shown in FIGS. 4a to 4b, a classification variable space of dimension p = 70700 is obtained.

In an important dimension space as previously indicated, obtaining stable predictions is conditioned by operating constraints of working in several small contiguous spaces and using a multi-file strategy to take into account interactions between these spaces.

Thus, a first sorting of the variables is performed for all the frequency bands by a Fisher discrimination test for example between the classes selected to retain only the best 300 for example.

Then, for all these last variables and for each frequency band, an LDA or SVM analysis is carried out and the class boundaries are retained. Such a procedure for a binary discrimination, that is to say between two classes, as mentioned previously in the description in connection with FIG. 5 for example, made it possible to go from a space of 70700 to 300 and then to 14 dimensions, that is to say one dimension per frequency band used. The final classification is performed by a combination of multiclassers such as LDA, NN or SVM on the reduced space mentioned above. A more detailed description of a device for representing a dynamic functional image of the brain, in accordance with the subject of the present invention will now be given in connection with FIGS. 6a and 6b.

With reference to FIG. 6a, it is pointed out that the device which is the subject of the invention comprises acquisition resources 1 during a determined recording duration of a plurality of electrophysiological signals emitted and / or induced by brain activity. , the signals {E <} f previously described in the description. These signals are acquired from a plurality of helmet electrodes ₁₀ which are placed in operation on the scalp of the subject so as to distribute the electrodes Ej evenly over the skull containing the C-brain.

Advantageously, as shown in FIG. 6a, the electrodes Ej and the aforementioned headset can advantageously be connected by a WIFI-type link, for example, to an acquisition computer 1 ₁ for storing and saving the electrophysiological signals to constitute a base of brain activity analysis data. Preferably, the database DBe mentioned above may be offset with respect to the acquisition computer I ₁ as will be described below.

As further represented in FIG. 6a, the device that is the subject of the invention further comprises a resource 2 for acquiring a three-dimensional image of the brain by successive sections, that is to say by the set of sections. {Ck} f. Figure 6a is shown above, advantageously, the resource acquisition 2 as formed by a reader or electronic file receiver networked to the acquisition computer 1 _\ and, on the other hand, an auxiliary processing unit 3, which performs the functions of calculating the location of all the neuroelectric generators and discriminating among the active zones comprising the aforementioned neuroelectric generators of the synchronism existing between the pairs of neuroelectric generators, as described previously in the description.

It is thus understood that the resources of the acquisition of the three-dimensional image make it possible either to access an external database managed by a clinical treatment entity of the subject, or to access the latter by a reader very large capacity optical disc type double layer DVD for example or other.

With regard to the processing unit 3, it is indicated that the latter is also connected in network to the acquisition computer I ₁ and can therefore be relocated with respect to the latter, which makes it possible to make the acquisition system for a specific subject.

In particular, it is conceivable that during the implementation of the method and device objects of the invention, for testing and for obtaining dynamic functional images for recording times of several days, the headset I ₀ can be made independent of the acquisition computer I ₁ via the indicated Wifi link, while the acquisition computer 1 ₁ may be constituted by a laptop interconnected network with the processing unit 3 .

Thus, the device object of the invention allows the implementation of the corresponding method by imposing a minimum of constraints on the subject, which can, of course, remain free of movement and in a near-normal situation, home for example.

The processing unit 3, as shown in FIG. 6a, comprises, in addition to an input I / O input device enabling the connection of this network processing unit via the Internet network, for example or otherwise, to a unit CPU processing unit, a RAM working memory and a hard disk type storage unit for storing the DBe data database does not analyze brain activity.

In addition, the central processing unit 3 comprises a module for calculating the location of the set of neuroelectric generators formed for example by the program storage modules M ₀ and M ₁ shown in FIG. 6a from the position of the electrodes and three-dimensional image of the brain acquired from the resources 2.

The above-mentioned calculation module can be formed by the modules M ₀ and M ₁ , the module M ₀ being for example dedicated to calculating the inverse problem to execute step B ₀ of Figure 3 for example, and the module M ₁ being dedicated to the execution of the mesh operation, that is to say of the step B ₁ shown in Figure 3 for example to from the successive sections {Ck} f obtained from the acquisition resource of three-dimensional images 2. In addition, a module M ₂ is provided which makes it possible to effectively locate all the neuroelectric generators of the intracerebral neuroelectric signals in accordance with FIG. step B ₂ shown in FIG. 3 and taking into account the indications given previously in the description.

Finally, the processing resource 3 advantageously comprises a calculation module denoted M ₃ making it possible to execute the discrimination calculation processing, in the active zones comprising neuroelectric generators, of the synchronism existing between the pairs of signals in a plurality of bands. frequency, that is to say in accordance with Figures 4a, 4b, 4c and 4d shown in the drawings.

It is understood, in particular, that the calculation modules M ₀ , M ₁ , M ₂ and M ₃ may advantageously consist of program modules stored in read-only memory and called by the central processing unit CPU in working memory RAM for execution of the corresponding operations.

The database of reference states representative of the dynamic functional image may, where appropriate, be stored on the hard disk drive already containing the database DBe, but, preferably, be transmitted for storage, storage and use on a particular resource connected in a network and preferably located at the level of the entity already holding the three-dimensional image of the brain in successive sections.

Finally, the device which is the subject of the invention may advantageously comprise a resource 4 for stimulating the subject, the resource 4 comprising a computer 4o for stimulating a subject to transmit to the subject either an audio stimulus via headphones 4 or ₂ , on the contrary, a visual stimulus, by means of display on display screens A _\ successive images to change the state of consciousness of the subject, psychological or other test images for example. The clinical and / or diagnostic applications of the method and of the device, objects of the invention, are numerous.

Indeed, the method and device of the invention allow a better location of the underlying neuroelectric generators located in the brain volume or on the surface of the latter.

The process implemented has the advantage of accessing an excellent temporal resolution of the reconstructed functional images. In addition, while the surface electrodes measure an instantaneous mixture of multiple distributed brain activations, the imaging works performs a spatial deconvolution of the information by providing access to a reconstructed time-course estimate in each position of interest. of the brain. Thanks to the implementation of the method and device objects of the invention, a more refined characterization of the brain states can be implemented in real time, given the synchronies demonstrated between the neuroelectric generators detected. In particular, some diagnostic results have been highlighted.

Thus, it has been observed that certain pairs of intracerebral electrodes placed in the vicinity of the periphery of the epileptogenic zone systematically exhibit, before any crisis, a significant change in their degree of synchrony, especially in the band of fast frequencies α 8 at 12 Hz, β 15 at 30 Hz and γ 30 at 70 Hz.

In addition, the aforementioned synchronizations have recently received a great deal of attention because of their possible involvement in large-scale integration phenomena during the cognition process. The corresponding results thus suggest that the neuronal populations underlying the epileptogenic zone modify their relationship with larger scale dynamics before the crisis.

The aforementioned changes in the synchronizations can then lead to a dynamic isolation of the epileptogenic focus and are then likely to provide recurrently, a neuronal population easily mobilized by epileptic processes. The method and the device that are the subject of the invention make it possible to quantify very precisely the precritical brain activity. This possibility of anticipating the onset of crises opens up diagnostic and, if necessary, very broad therapeutic perspectives, by characterizing the neurobiological changes that occur during the precritical phase.

In addition, in clinical matters, the possibility of preventing the subject and attempting to abort a crisis in preparation by a therapeutic intervention may be considered. In particular, electrical neurostimulation has emerged recently as a promising therapeutic solution for other pathologies, such as in particular Parkinson's disease.

In this view, the mechanical destruction of a predefined brain region can be replaced by conservative electrical stimulation treatment to enhance or inhibit neuronal activity. The possibility of crisis anticipation through the implementation of the method and device objects of the invention is decisive because it gives an answer to the question when to stimulate. Indeed, the aforementioned stimuli can be applied when a preictal state is detected and will then have the object of destabilizing the epileptogenic processes before they become irreversible at the time of the crisis.

In addition, the method and the device that are the subject of the invention can enable cognitive intervention developments. Indeed, some subjects describe the faculty they have of interrupting their debilitating crisis by specific cognitive activities or motor activities. These phenomena are probably based on a destabilization of the epileptic process by the appearance of new electrical activities within the cerebral cortex. Thus, thanks to the implementation of the method and device of the invention, the modulation of epileptic activity by cognitive synchronizations has also been demonstrated.

Finally, other interventions may be envisaged, such as, for example, the pharmacological intervention consisting in the administration of fast-acting anti-epileptic drugs, such as benzodiazepines. These alert and intervention possibilities offered by crisis anticipation necessarily imply anticipation in real time, that is to say that the results of calculations and corresponding detections are obtained instantly and not deferred.

The ability to anticipate seizures also makes it possible to improve the performance of examinations carried out during the pre-surgical assessment of partial drug-resistant epilepsies. Notably, the realization of the precritical brain scintigraphy, designated SPEC-ictale, is facilitated by the alerting of the personnel treating for the injection of the radioactive tracer at the very beginning of the crisis, or just before, which makes it possible to better localize the epileptogenic focus. Hospitalization times can be significantly reduced and the occupancy time of the imaging systems optimized.

Finally, the aforementioned example of application to the clinical study of epilepsy, can easily be transposed to cognitive activities such as the measurement of alertness, mental load and drug / cognition interactions including modifying, by download for example, the learning base constituted by the functional images characterizing a brain state of the subject's brain.

Consequently, it will be understood that the limitations of the prior art resulting from the fact that they implied a linear relationship between the recorded signals could be lifted, thanks to the phase synchronization process as described previously in the description, which appears particularly well adapted to measure the degree of interdependence between the activities of various brain regions in one or more specific frequency bands.

Thus, remarkably, neuronal synchronizations in the fast 30 to 50 Hz frequency band have recently received a lot of attention for their possible role in large-scale integration phenomena during cognition and in the case of certain pathologies.

In summary, the device developed and the method of the invention make it possible to locate and quantify, in real time, from electroencephalographic (EEG) signals collected in humans, the interactions between different intracerebral activities for the purpose of Characterize by a signature: 1) alertness, attention, stress, effort, fatigue etc. ; 2) the very short-term evolution of certain pathological states such as epileptic seizures; and

3) the action of drugs and / or drugs, acting specifically on the central nervous system (CNS). They also make it possible to visualize, classify, compare these different cerebral states. More specifically, they make it possible to test whether or not a new type of drug or pharmacopoeia is approaching, by its signature, such a drug or drug already known. In this sense, the method and the device which are the subject of the invention appear extremely useful for clarifying in man the potential action field of a new molecule before it is placed on the market.

The invention finally covers a computer program product recorded on a storage medium for execution by a computer, remarkable in that, during this execution, it allows the implementation of the method, object of the invention, such that described in Figures Ib to 4d and a device for representing a dynamic functional image of the brain as described in connection with Figure 6a.

Claims

1. A method for representing a dynamic functional image of the brain, by location and discrimination of intracerebral neuroelectric generators, characterized in that it consists at least of:

acquiring, during a determined recording duration, a plurality of electrophysiological signals emitted and / or induced by the brain activity from a plurality of electrodes substantially distributed on the scalp of the cranial box housing the brain and digitizing said signals electrophysiological to form a database of analysis of brain activity;

locating the set of neuroelectric generators in the cerebral volume from an acquisition of an electronic map of the position of said electrodes, of a three-dimensional image of this brain by successive sections, from a segmentation of the cerebral cortex obtained from said three-dimensional image, and from the application of the inverse problem to determine, from at least one of the electrophysiological signals, the electronic mapping of the location of the electrodes and of said three-dimensional image of this brain, the spatial localization neuroelectric generators of said intracerebral neuroelectric signals; discriminating among the active zones comprising neuroelectric generators the synchronism existing between the pairs of neuroelectric generators in a plurality of frequency bands, for detecting groups of discriminating neural networks and constituting a database of reference states representative of said image dynamic functional.

Method according to claim 1, characterized in that, for a dynamic functional image acquired during said determined recording duration, said method further comprises identifying said functional image with a class of functional images from among a plurality of classes. functional images, each class of functional images of said plurality of functional image classes characterizing a brain state of the subject's brain.

3. Method according to one of claims 1 or 2, characterized in that the step of acquiring a plurality of electrophysiological signals is performed in real time, with a maximum recording delay of less than 100 milliseconds.

4. Method according to one of claims 1 to 3, characterized in that the recording duration is parameterizable over a range of time, said range of time being between a minimum recording time of about 20 minutes for recording and representing a functional image of the brain relating to one or more cognitive states and a duration of several days for recording and representing a functional image of the brain relating to one or more states of anomaly functional brain function caused or not provoked.

5. Method according to one of the preceding claims, characterized in that the step of discriminating, among the active areas, the synchronism existing between the pairs of neuroelectric generators is performed during said recording time on a sliding time window of which the duration is between 50 milliseconds and 2 seconds for the representation of a functional image of the brain relative to one or more cognitive states respectively on a sliding temporal window of temporal coherence whose duration is between 5 seconds and 20 seconds for the representation a functional image of the brain relating to one or more states of functional abnormality of the brain caused or not provoked.

6. Method according to one of claims 1 to 5, characterized in that the application of the inverse problem for determining, from at least one of the electrophysiological signals, the electronic map of the location of the electrodes and the image. three-dimensional of this brain, the spatial location of the neuroelectric generators of intracerebral neuroelectric signals consists of at least:

- apply constraints from the individual anatomy introduced by segmentation and surface mesh of the parenchyma; - to estimate the cortical electric currents from a multimodal treatment of the electrophysiological signals;

calculating in position and functional parameters said neuroelectric generators in the form of sources of elementary electric currents on the mesh of the cortical surface, an active zone comprising at least one neuroelectric generator.

7. Method according to one of claims 1 to 6, characterized in that the step of discriminating, from said active areas comprising neuroelectric generators, the synchrony existing between the pairs of neuroelectric generators in a frequency band consists of at least evaluating the statistical level of the PLS synchronization between two signals of a pair of neuroelectric generators by means of the circular variance of the phase difference between these signals or the standardized Shannon entropy of this phase difference.

8. Method according to claim 7, characterized in that the synchronie is established by synchronous time ranges, which allows to temporally represent the activity of said pairs of neuroelectric generators.

9. Dynamic functional image of the brain, characterized in that said dynamic functional image comprises at least:

a three-dimensional image of this brain by successive sections, each representing an elementary image of this brain; and, on at least one elementary image,

at least one neuroelectric generator of intracerebral neuroelectric signals represented by a marker, each neuroelectric generator being characterized in position in said elementary image, in electrical current density and in the direction of emission of neuroelectric signals, each neuroelectric generator of an elementary image current neighbor of a neuroelectric generator of a previous and / or subsequent elementary image and having substantially the same direction of emission of neuroelectric signals and a synchronism over a determined duration of coherence constituting a network group functional state discriminant neurons representative of said dynamic functional image of the brain.

10. Functional image according to claim 9, characterized in that for a functional image relating to several cognitive states the duration of temporal coherence is between 50 milliseconds and 2 seconds.

11. Functional image according to claim 9, characterized in that for a functional image relating to one or more functional abnormality states of the brain caused or not caused the duration of temporal coherence is between 5 and 20 seconds.

12. Device for representing a dynamic functional image of the brain, characterized in that it comprises at least:

Acquisition means during a determined recording duration of a plurality of electrophysiological signals emitted and / or induced by cerebral activity from a plurality of electrodes substantially distributed on the scalp of the cranial box housing the brain, said acquisition means for storing and saving said electrophysiological signals to form a brain activity analysis database;

means for acquiring a three-dimensional image of the brain by successive sections; means for calculating the location of the set of neuroelectric generators of the intracerebral neuroelectric signals in the cerebral volume from the position of said electrodes and of said three-dimensional image of this brain, so as to obtain a segmentation of the cerebral cortex, and the application of the opposite problem; means for discriminating, among the active zones comprising neuroelectric generators, the synchronism existing between the electrophysiological signal-generator pairs in a plurality of frequency bands, for detecting groups of discriminant neural networks and for constituting a database of neuroelectric generators; reference states representative of said dynamic functional image.

13. Device according to claim 12, characterized in that said means for acquiring a plurality of electrophysiological signals comprise at least one flexible helmet provided with a set of electromagnetic sensors constituting said electrodes, said electromagnetic sensors constituting an array of sensors. Plated on the scalp of the cranial box of the subject.

14. Device according to one of claims 12 or 13, characterized in that it further comprises means for video and / or audio stimulation of the subject.

15. Computer program product recorded on a storage medium for execution by a computer, characterized in that, when executed by a computer, it allows the implementation of the method of representation of a functional image. brain dynamics according to one of claims 1 to 8.

16. Use of the method according to one of claims 1 to 8, of at least one dynamic functional image of the brain according to one of claims 9 to 11, of the device for representing a dynamic functional image of the brain according to the invention. one of claims 12 to 14 and a computer program product according to claim 15 for characterizing, by a signature, different brain states among the groups of states relating to alertness, attention, stress, stress, fatigue, either the evolution of certain short-term disease states, or the action of drugs and / or drugs acting on the central nervous system.