KR20150101351A - Neuronal resonance magnetic resonance imaging method - Google Patents

Abstract

The present invention provides a novel imaging modality called neuronal resonance MRI for mapping a bandpass filtering phenomenon of a brain onto an image. The image modality is verified through a biometric experiment, and compared/analyzed from aspects of systems biology.

Description

[0001] The present invention relates to a neuron resonance magnetic resonance imaging method,

The present invention relates to a method of acquiring and processing MRI data of a neuron resonance magnetic resonance imaging.

It is obvious that communication is being carried out for the transmission of information between human brain regions. However, it is still not known how a particular brain region selectively transfers information to other brain regions to which it wishes to communicate. In recent system biology research, we hypothesized that communication between brain regions is selectively performed from the whole brain signal through a selective frequency band filter. Based on this hypothesis, the development of a new imaging technique that can measure the resonance frequency of neurons for communication between brain regions can reveal the mechanism of communication between brain regions that are not well known at present. It can have a tremendous impact on research.

The brain is made up of small areas that hold many functions and these brain areas are structurally connected to each other. Most brain functions, such as behavior, cognition, and perception, are accomplished by reassembling the brain's network quickly and flexibly. Recent studies have revealed some of the dynamic and flexible functional brain networks of cognition and perception, selective focus, and working memory. It is known that the vibrational characteristics and synchronization of neuron spiking are related to dynamic and flexible connectivity between brain regions. Recent studies have shown that the vibrations and synchronization of neurons occurring in such specific frequency bands control the flow of information between the regions of the brain that are structurally connected and enable flexible and selective communication between brain regions. However, in the dormant and active / stimulated states, it is not yet clear what mechanism is causing the selective and flexible communication of the brain to a specific frequency band. Furthermore, a number of clinical data suggest that patients with brain disorders such as autism, schizophrenia, epilepsy, dementia and Parkinson's have altered the synchronization characteristics of neurons over a broad frequency band and that such abnormal neuronal synapses are associated with symptoms of such diseases , Movement, etc.).

Therefore, by understanding the mechanism of selective inter-brain communication through synchronization, we can find clues that are very important for the treatment of biopathology and brain diseases. Furthermore, the development of a new imaging technique that can map the resonance characteristics of neurons to a high resolution for a wide frequency band used for communication between brain regions requires a tremendous ripple effect on academia, industry, and medical related to biomedical engineering Lt; / RTI >

On the other hand, functional magnetic resonance imaging (fMRI) indirectly measures the activity of the brain by the interaction of neurons and bloodstream rather than directly measuring the current through neurons. Twenty years ago, we first showed that functional magnetic resonance imaging (fMRI) can noninvasively map the activity of brain neurons. Since then, fMRI has shown tremendous impact on neuroscience, psychology, and psychopathology.

Unlike conventional MRI, fMRI is a technology that acquires images repeatedly during and after external stimuli and maps the brain regions that correlate with temporal patterns of external stimuli through statistical processing. fMRI is almost the only imaging technique that can map non-invasive and relatively high resolution brain activity.

fMRI is a special imaging technique that can non-invasively map the activity of neurons, but rather indirectly measures blood flow changes by neuron activity rather than directly measuring electrical signals of neurons. Efforts to directly measure the electrical signals of neurons using MRI have been made over the past decade, and the possibility of that has been controversial.

About a decade ago, it was first shown that fMRI can be performed without external stimuli. This new fMRI technology is called Resting-state fMRI. The basic assumption of resting-state fMRI is that if there is a functional association between two regions of the brain, temporal MRI signal changes will also correlate. Resting-state fMRI measures functional connectivity between brain regions. In the resting-state fMRI, the stimulation pattern used in the existing fMRI technique was replaced with the temporal change in the signal region of the specific brain region, Or statistical analysis. Some brain regions (ex: default mode network) are functionally linked and are measured consistently using a Resting-state fMRI. However, existing fMRI techniques, including Resting-state fMRI, indirectly measure neuronal activity through hemodynamic responses in the local area. The hemodynamic response is slow and there is a delay of about 4 seconds. Resting-state fMRIs show functional connectivity between brain regions, but they do not show which mechanisms selectively communicate between brain regions. This is a fundamental limitation of the existing fMRI technique and explains why "resonance of frequency-selective neurons" can not be confirmed by conventional methods.

On the other hand, the ability to measure in vivo neuron currents of existing MRI imaging techniques has been controversial for more than a decade. There have been several attempts to directly measure the current of a neuron directly with MRI. While promising results have been shown in phantom, cell culture, and theoretical computational studies, the in vivo measurement of neuronal currents using MRI has been controversial for more than a decade. Researchers in the field agree that the current in a neuron causes a change in the magnetic field around it and that the change can be measured by MRI. However, many researchers argue that the MRI signal due to changes in the magnetic field generated by currents in neurons is so small that it is very difficult to consistently measure with MRI in vivo.

There are two main approaches to MRI imaging for directly sensing the electrical signals of neurons. The first approach is to give a periodic stimulus with a certain time interval and to measure the electrical signal of the neuron by obtaining an MRI signal as soon as each stimulus is over (ie before the hemodynamic response occurs). The second approach is to increase the temporal resolution of the MRI image acquisition (≤100 ms) and find a component that is synchronized with the frequency of the external stimulus through Fourier transform. Both approaches depend on the frequency or frequency of the external stimulus and do not take into account the natural frequency of the neuron. Whether the neurons sense current signals for all existing MRI imaging techniques including the above two methods is still controversial.

In order to solve the above problems, the present invention provides a neuron resonance magnetic resonance imaging (NR-MRI).

In the present invention, the difference between the states of the dormant state and the external stimulus is measured for various frequency regions through the selective frequency band filter characteristic mapping of the neuron using the neuron resonance magnetic resonance imaging (NR-MRI) technique Technology.

And to provide a technique for revealing frequency selective inter-brain communication mechanisms based on systematic biological studies on frequency selective inter-brain communication mechanisms.

We also provide useful techniques to complete the proof-of-concept test and the frequency-selective brain-to-brain communication map.

According to one aspect of the present invention, a magnetic resonance signal processing method using a process of obtaining MRI data by an MRI imaging technique using a gradient magnetic field pattern to sense a magnetic field vibration signal can be provided. The method includes: obtaining a plurality of MRI data by repeating the process a plurality of times while changing a relative phase between the gradient magnetic field pattern and the magnetic field vibration signal; And calculating frequency characteristics from the obtained plurality of MRI data by a predetermined rule, wherein the plurality of MRI data is K-space data or MRI image data.

Here, the phase of the magnetic field vibration signal may not be known at the time of occurrence and at a specific point in time.

In this case, the magnetic field vibration signal may include a specific frequency component.

The predetermined rule may be a Fourier transform of a signal obtained by arranging the acquired plurality of MRI data on a time axis according to the relative phase.

Wherein the acquiring step uses an acquiring process for acquiring N MRI data by repeatedly performing the process for a predetermined N relative phases between the gradient magnetic field pattern signal and the magnetic field vibration signal, And repeating the process M times in sequence according to time to acquire a plurality of MRI data.

The predetermined rule may be to Fourier-transform the obtained plurality of MRI data on a signal arranged on the time axis according to the predetermined relative phase.

At this time, the plurality of MRI data may be temporally divided, and Fourier transform may be performed on each of the divided data to obtain time-base data for a signal change of the neuron.

At this time, the specific frequency may be the same as the frequency of the gradient magnetic field pattern.

The predetermined rule may be adapted to find a value of the frequency component of the obtained plurality of MRI data with respect to the specific frequency.

Wherein the obtaining step includes an obtaining process for obtaining N MRI data by repeatedly performing the process for a predetermined N relative phases between the gradient magnetic field pattern signal and the magnetic field vibration signal, The difference value between the two relative phases may be an integer multiple of 2 * pi / N with respect to the specific frequency.

At this time, the time to repeat (TR) of the MRI image parameters may be adjusted to make the difference value between the N relative phases an integral multiple of 2 *? / N with respect to the specific frequency.

Here, the magnetic field vibration signal may be a neuron signal.

In this case, the gradient magnetic field pattern may be a vibration gradient magnetic field pattern.

At this time, the vibration gradient magnetic field pattern may be a bipolar readout gradient pattern used in a multi-echo gradient echo imaging technique.

At this time, the spatial directionality of the gradient magnetic field pattern may be different for the two different processes that are repeatedly performed.

According to another aspect of the present invention, there is provided an MRI apparatus including: an MRI imaging technique that uses a gradient magnetic field pattern signal to resonate at a specific frequency and detect a magnetic field vibration signal whose phase is unknown at a specific time point, The process of obtaining is used. Obtaining a plurality of MRI data by repeating the process a plurality of times while changing a relative phase between the gradient magnetic field pattern signal and the magnetic field vibration signal; And a step of calculating a frequency characteristic from the obtained plurality of MRI data by a predetermined rule.

According to another aspect of the present invention, there is provided a computer-readable medium having a magnetic resonance imaging apparatus in which an MRI apparatus generates a gradient magnetic field pattern signal to resonate at a specific frequency and detect a magnetic field vibration signal, Obtaining a plurality of MRI data by performing the process repeatedly a plurality of times while changing a relative phase between the gradient magnetic field pattern signal and the magnetic field vibration signal using a process of obtaining MRI data using the MRI imaging technique to be used; And a step of calculating a frequency characteristic from the obtained plurality of MRI data by a predetermined rule.

According to still another aspect of the present invention, there is provided a magnetic resonance signal processing method comprising: detecting an MRI signal using a vibration gradient magnetic field pattern signal having a predetermined vibration frequency to detect a magnetic field vibration signal whose phase is unknown at a generation time point and a specific time point; Is a magnetic resonance signal processing method using a process of obtaining MRI data by an imaging technique. The method includes: obtaining a plurality of MRI data by repeating the process a plurality of times while changing a relative phase between the vibration gradient magnetic field pattern signal and the magnetic field vibration signal; And a step of calculating a value relating to the predetermined vibration frequency from the obtained plurality of MRI data by a predetermined rule. And obtaining P values for the predetermined vibration frequency for each of the first processes by performing the first process P times so that the predetermined vibration frequency has P different values.

According to another aspect of the present invention, there is provided a neuron resonance magnetic resonance imaging method, comprising: performing a first process of acquiring an image using an MRI imaging technique having a vibration gradient magnetic field pattern to sense a neuron signal resonating at a specific frequency ; And a second step of repeating the first process N times to acquire an MRI image for each of N different neuron resonance phases, wherein the second step is repeated M times to acquire a plurality of MRI images 2 performing a process; And performing a third process of analyzing all or a part of the plurality of multiphase MRI images by spectral analysis through Fourier transform and calculating the spectrum by a predetermined rule.

At this time, the specific frequency may be the same as the frequency of the gradient magnetic field pattern signal.

At this time, in the third process, the predetermined rule can find a neuron component that resonates at the specific frequency among the frequency components of the multi-phase data.

At this time, the difference of the N neuron resonance phases may satisfy an integral multiple of 2 * pi / N.

At this time, it is possible to adjust the time to repeat (TR) of the MRI image parameters in order to make the difference of the N neuron resonance phases an integral multiple of 2 * pi / N.

In this case, in the third process, the plurality of MRI data may be divided, and Fourier transform may be performed on each of the divided data to obtain time-base data for a signal change of the neuron.

At this time, the vibration gradient magnetic field pattern may be a bipolar readout gradient pattern used in a multi-echo gradient echo imaging technique.

At this time, the vibration gradient magnetic field pattern can be applied to various directions in space such as X, Y, and Z.

An MRI apparatus provided by another aspect of the present invention is an MRI apparatus including an MRI image acquisition unit and a processing unit configured to control the operation of the MRI image acquisition unit. In this case, the processing unit may include: performing a first process of obtaining an image using an MRI imaging technique having a vibration gradient magnetic field pattern to sense a neuron signal resonating at a specific frequency; And a second step of repeating the first process N times to acquire an MRI image for each of N different neuron resonance phases, wherein the second step is repeated M times to acquire a plurality of MRI images 2 performing a process; And performing a third process of analyzing all or a part of the plurality of multiphase MRI images by spectral analysis through Fourier transform and calculating the spectrum by a predetermined rule.

According to another aspect of the present invention, there is provided a computer-readable medium having a first process for obtaining an image by an MRI imaging technique having a vibration gradient magnetic field pattern in order to detect a neuron signal resonating at a specific frequency, ; And a second step of repeating the first process N times to acquire an MRI image for each of N different neuron resonance phases, wherein the second step is repeated M times to acquire a plurality of MRI images 2 performing a process; And performing a third process of analyzing all or a part of the plurality of multiphase MRI images by a spectrum through Fourier transform and calculating the spectrum by a predetermined rule, wherein the program is recorded in a computer-readable medium to be.

According to another aspect of the present invention, there is provided a method for processing MRI time series acquisition data comprising a first step of acquiring an MRI data set using a vibration gradient magnetic field pattern signal having a predetermined frequency, Performing a first process wherein the first process is adapted to obtain N sets of MRI data by performing N iterations during one period of the vibration gradient magnetic field pattern signal; For each of the N MRI data sets, energy of a signal mapped to a first image area of the total image represented by each of the MRI data sets is calculated so that an energy change Performing a second process of acquiring time-base data representing the time-axis data; And a third process of calculating a value relating to a specific frequency component of frequency components of the time-base data by a predetermined rule.

In this case, the third process may include Fourier transforming the time-base data.

At this time, the specific frequency may be the same as the frequency of the vibration gradient magnetic field pattern signal.

At this time, the N MRI data sets may be data obtained with respect to the neuron space.

At this time, the phase difference of the N vibration gradient magnetic pattern signals may be an integral multiple of 2 * pi / N.

At this time, the second process and the third process may be repeatedly performed on the first image area and the second image area of the entire image of the entire image, and the results thereof are stored.

At this time, the first process is repeated M times, and the second process and the third process are repeated M times using M result data obtained as a result of performing M times repeatedly Can be.

Wherein in the first process, the first process is repeated M times for M periods of the vibration gradient magnetic field pattern signal to obtain a plurality of MRI data sets, and in the second process, For each of the MRI data sets, calculating energy of a signal mapped to a first image area of the total image represented by the respective MRI data set, And in the third process, a value relating to a specific frequency component among frequency components of the time axis data acquired during the M periods may be calculated by a predetermined rule.

At this time, the vibration gradient magnetic field pattern signal may be a bipolar readout gradient pattern signal used in a multi-echo gradient echo imaging method.

According to the present invention, it is possible to use a technique of maximizing the signal of a neuron by applying an oblique magnetic field pattern oscillating in accordance with the resonance frequency of the neuron to the MRI imaging technique.

In addition, by repeating the multi-phase image acquisition and extracting the component corresponding to the resonance frequency of the neuron through the Fourier analysis method, the problem that the MRI signal due to the current of the neuron is very small, The phenomenon can solve the problem that the temporal interval and the phase occur randomly. The neuron signal can be frequency-selectively extracted by using the repetitive multi-phase image acquisition method and the Fourier analysis method, and signal-to-noise ratio (SNR) can be obtained through a temporal averaging effect. S / N ratio) can be greatly improved.

In addition, it is possible to provide a basic technique for identifying a specific frequency and a brain region related to brain function and brain disease by completing a communication channel map for each frequency region between brain regions which has not been previously possible.

1 is a diagram illustrating communication between frequency selective brain regions in a neuron resonance magnetic resonance imaging (NR-MRI) according to an embodiment of the present invention.
2A to 2C are schematic diagrams of a neuron resonance magnetic resonance imaging (NR-MRI) according to an embodiment of the present invention.
FIG. 3 shows a schematic diagram of an experimental study of a neuron resonance magnetic resonance imaging (NR-MRI) according to an embodiment of the present invention.
4A to 4F are graphs showing simulation results of a neuron resonance magnetic resonance imaging (NR-MRI) according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. However, the present invention is not limited to the embodiments described herein, but may be implemented in various other forms. The terminology used herein is for the purpose of understanding the embodiments and is not intended to limit the scope of the present invention. In addition, the singular forms used below include plural forms unless the phrases expressly have the opposite meaning.

Hereinafter, a neuron resonance MRI image according to an embodiment of the present invention will be described with reference to FIGS. 1 to 4F.

1 is a diagram illustrating communication between frequency selective brain regions in a neuron resonance magnetic resonance imaging (NR-MRI) according to an embodiment of the present invention.

The mammalian brain exhibits resonance in the population of neurons by interactions between neurons. As shown in Fig. 1, one neuron group NG1 selectively communicates with another neuron group NG2 that resonates at a specific frequency and resonates at the same frequency.

Accordingly, the present invention provides a neuron resonance magnetic resonance imaging (NR-MRI) capable of measuring frequency-selective inter-brain communication signals.

2A to 2C are schematic diagrams of a neuron resonance magnetic resonance imaging (NR-MRI) according to an embodiment of the present invention. FIG. 2A is a diagram of a pulse diagram of a neuron resonance magnetic resonance image. FIG. 2B is a graph showing the relationship between a multiphase measurement (left) of a neuron resonance magnetic resonance image (left) And FIG. 2C is a schematic diagram of phase-insensitive oscillation detection independent of the phase of the neuron.

(Ii) is a plot of a pulse train for a multi-echo gradient echo (ME-GE) that can be used for an NR-MRI, and (iii) Is a diagram showing an example of a K-region photograph and an image.

2A, the NR-MRI can include a superconducting magnet 10, a gradient coil 20, a gradient coil-Z 21, a gradient coil-Y 22, and a radio transmission / reception coil 30 have.

The graphs 91 to 96 in FIG. 2A represent the RF signal 91, the gradient Z signal pulse 92, the gradient Y pulse 93, and the gradient X pulse 94, an RX signal 95 received by the radio transmission / reception coil 30, and an example of a neuron resonance waveform 96 of an organism disposed in the NR-MRI.

(Iii) of FIG. 2A shows an example of a brain image generated by performing a K-region FFT using the RX signal 95 received in (ii) of FIG.

According to FIG. 2A, in ME-GE, the oscillating frequency of the bipolar reading gradient field can be adjusted to the resonance frequency of the neuron, and the later acquired echo signal resonates with the neuron for a longer time to reflect stronger neuronal signals, Lt; / RTI >

In the multi-phase measurement of the neuron resonance MRI image shown in FIG. 2B, the time to repeat (TR) is set such that the time phase of the neuron resonance versus acquisition of the MRI image is 2π / N (N: .

According to FIG. 2B, the multi-phase measurement method may be performed by repeating M cycles. In each cycle, the pulse train for ME-GE can be provided N times repeatedly. Hereinafter, a method of providing pulse trains for ME-GE in each cycle will be described.

While the multiphase measurement is being performed, it can be assumed that the neuron resonance waveform 96 has a period of T _NO . At this time, in one embodiment of the present invention, a pulse train may be provided so as to satisfy the following two conditions. First, the period of the pulse train provided repeated N times is set equal to the period T _NO of the assumed neuron resonance waveform 96. (Ii) the generation time of the (i) th pulse train among the pulse train provided N times repeatedly is delayed by K * T _NO + 2π / N from the generation time of the (i-1) ~ N). Here, K * T _NO means an integer multiple of the resonance period of the neuron resonance waveform 96. In this way, the phases of the pulse train repeated N times are different from each other in phase with respect to the neuron resonance waveform 96. In the present specification, this can be expressed by the term " multi-phase ". For example, in Fig. 2A, the generation timing of the (i) th pulse train is delayed by i * 2? / N from the zero crossing point of the neuron resonance waveform 96 toward the positive direction. That is, there exists a phase difference of? _I = i * 2? / N between the (i) th pulse train and the neuron resonance waveform 96.

2C is a diagram showing a case where the period of the pulse train according to the multi-phase measurement method is T _NO , the neuron resonance period in the regions A and B of the brain is T _NO , the resonance period of the neurons in a shows results that can be obtained in other cases, and T _NO. At this time, neurons in the brain region C may not resonate.

The graphs 201, 202, and 203 of (i), (ii), and (iii) in FIG. 2C are graphs of the brain regions A, B, And the amount of energy according to the acquired time. The graphs 201 and 202 may have periodicity with respect to regions A and B of the brain but may not have periodicity in the graph 203 with respect to region B. [

(Iv), (v) and (vi) of FIG. 2c are graphs 211, 212, and 213 obtained by performing FFT on the graphs 201, 202, and 203 of (i) 212, and 213, respectively. In the graphs 211 and 212, the resonance frequency is shown as f _NO , but the graph 213 shows that there is no special resonance frequency. As a result, by interpreting (iv), (v), and (vi) in FIG. 2C, it can be seen that the regions A and B of the brain are regions where the period of the pulse train resonates at a period equal to T _NO .

2A to 2C, a neuron resonance MRI image according to an exemplary embodiment of the present invention applies an MRI imaging technique to an oblique magnetic field that oscillates in accordance with a resonance frequency of a neuron, Can be maximized.

Conventional MRI techniques for directly detecting the electrical signals of neurons have two problems. The first is that the MRI signal changes are too small, and the second is that the resonance phenomenon of the neurons takes place almost randomly, without the rule of temporal phase and interval. Therefore, the neuron resonance MRI according to an embodiment of the present invention solves the above problems by repeatedly performing multiphase measurement corresponding to the resonance frequency of the neuron, and then applying the Fourier analysis method.

There has been an attempt to detect the resonance of a neuron by MRI, but the possibility of it has been controversial for the last decade. One of the main reasons is that the change in MRI signal due to resonance of the neurons is very low. Another important problem is that the time interval and phase of the resonance of the neurons are random and not known to us, making it almost impossible to acquire and synchronize MRI images. Another problem is that since the neurons periodically resonate, there is a high probability that they resonate more than once between sending and receiving RF energy for acquiring MRI images. In this case, when an RF signal is received by an actual MRI, Is almost completely canceled. It is thought that the possibility of existing methods is controversial because of these points.

The neuron resonance MRI according to an embodiment of the present invention is a new MRI imaging technique for directly detecting the resonance of a neuron with MRI, and is largely composed of three components. As the three elements of the new MRI imaging technique, (i) the gradient magnetic field pattern resonating to the resonance frequency of the neuron is applied to the MRI imaging technique, (ii) the multi-phase signal is repeatedly obtained according to the resonance frequency of the neuron , and (iii) applying the Fourier analysis method to separate the frequency components from the resonance frequency of the neuron. That is, according to the neuron resonance magnetic resonance imaging method according to the embodiment of the present invention, other undesired signals such as hemodynamic response, movement or flow due to neurons and blood flow, errors on the MRI system , Harmonic frequency components between the resonance frequency of the neuron and the gradient magnetic field current resonance frequency, noise components, and the like, and only the resonance frequency component of the desired neuron can be separated.

In order to realize a neuron resonance magnetic resonance imaging according to an embodiment of the present invention, a neuronal-resonance oscillating gradient (NROG) may be applied to several MRI imaging techniques. That is, it can be implemented in image techniques such as Multi-Echo Gradient Echo (ME-GE), Spin-Echo Echo Planar Imaging (SE-EPI), and Gradient Echo (GE). According to FIG. 2A, all MRI systems have three gradient magnetic coils. In MRI, three gradient magnetic coils linearly modulate the size of the magnetic field in three directions, thereby generating spatial image information, i.e., imaging ). SE-EPI (Spin-Echo Echo Planar Imaging) and GE (Gradient Echo) can be implemented using independent "Neuron Resonant Vibration Gradient Field (NROG)" to map neuron resonance frequency characteristics according to directionality. In an embodiment of the present invention, a bipolar readout gradient pattern originally existing in a multi-echo gradient echo (ME-GE) imaging technique instead of SE-EPI (Spin-Echo Echo Planar Imaging) or GE (Gradient Echo) Resonance vibration gradient magnetic field (NROG) ".

In the ME-GE (Multi-Echo Gradient Echo) imaging technique, the echo spacing time (ESP) between two adjacent echoes is calculated using the following Equation 1 As long as it is equal to half of the neuron resonance period.

[Formula 1]

T _NO (Neuron Resonance Cycle) = 2 * ESP = Period of bipolar readout gradient pattern

In multi-echo gradient echo (ME-GE), multiple echoes can be used to create a single MRI image, which can be used to reduce imaging time. According to FIG. 2A, as the resonance of the neuron and the gradient magnetic field vibration progress, the resulting change in the MRI signal becomes larger. Therefore, the echo data obtained later has a stronger neuron resonance signal. And conversely, it is advantageous to configure the echo data initially acquired to occupy the edge of the MRI K-space data.

However, since the magnitude of the echo signal decreases exponentially with a time constant called T2 ^* (usually <80ms), two conflicting effects (the signal with a cumulatively large neuron resonance signal and a T2 ^* A phenomenon of exponential decay). Finally, the number of multiple echoes in the ME-GE should be optimized through experimentation and simulation of several factors: the magnitude of the main magnetic field, the spatial resolution of the image, and the resonant frequency to be viewed.

The neuron resonance of the frequency f _NO can be expressed as [Equation 2] below.

[Formula 2]

If the resonance phase (phi) of the neuron is the same as the phase of the MRI vibration gradient magnetic field pattern, the change of the MRI signal by the neuron will be maximized as shown in Figs. 2A to 2C. However, since the resonance phase of a neuron is randomly generated, MRI data acquisition must be repeatedly obtained in multiple phases relative to the resonance phase of a neuron, and the result should be analyzed by the Fourier analysis. One of the easiest ways to obtain multi-phase data from multi-echo gradient echo (ME-GE) is to design the time-to-repeat (TR) A slight difference may be caused between the MRI gradient magnetic field vibration phases.

[Formula 3]

)와 공진 주기를 다중 위상의 수로 나눈 값(

)을 더한 값임을 의미한다. [식 3]에서, 뉴런의 공진 주기 정수배 term(

)은 뉴런의 위상과 진동 경사자계 위상 간의 차이를 바꾸지 않지만, 공진 주기를 다중 위상의 수로 나눈 값 ()은 둘 사이의 위상차를

만큼 증가시킨다. 두 위상의 차이는 궁극적으로 매 TR 마다

만큼 증가하게 되는데, 이것은 도 2b에서와 같이, 다중 위상 신호를 반복적으로 얻을 때 두 위상의 차이가 N_phases 를 주기로 반복적으로 나타남을 의미한다. 이러한 다중 위상 데이터 획득을 반복함으로써 (i) K-space 영역을 다 채워서 완성된 MRI 영상을 만들고 (ii) 다중 위상 MRI 영상을 반복적으로 얻어서 푸리에 해석법을 적용할 수 있다. Where N _Echoes and N _phases represent the number of multiple echoes and the number of multiple _phases , respectively. At this time, N _Echoes must be an even number. [Equation 3] indicates that TR is an integral multiple of the resonance period of the neuron (

) And the resonance period divided by the number of multi-phases (

) Is added. In Equation (3), the resonance period integral term of the neuron term

) Does not change the difference between the phase of the neuron and the phase of the vibration gradient magnetic field, but the value obtained by dividing the resonance period by the number of the multiple phases ) Is the phase difference between the two

. The difference between the two phases is ultimately

As shown in FIG. 2B, which means that when two or more phases are repeatedly obtained, the difference between the two _phases repeatedly appears at intervals of N _phases . By repeating the acquisition of the multiphase data, (i) the completed MRI image is obtained by filling the K-space region, and (ii) the multiphase MRI image is repeatedly obtained and the Fourier analysis method can be applied.

As shown in FIG. 2C, the signal of a neuron having a specific frequency is periodically changed when the repetitive multi-phase image acquisition mentioned above is performed. When the Fourier transform is applied to the signal, the corresponding frequency component can be found. As shown in FIG. At this time, the Fourier analysis method may be applied to all the repeated multi-phase data, but it can also be applied to the temporal change of the frequency characteristics of the neuron by dividing it by time. In addition, this sliding-time window analysis makes it possible to see brain frequency changes due to functional changes when the brain is in a resting-state or when there are external stimuli.

FIG. 3 shows a schematic diagram of an experimental study of a neuron resonance magnetic resonance imaging (NR-MRI) according to an embodiment of the present invention.

Since the difference between the neuron resonance phase and the image acquisition phase is increased by 2π / N every repetition time (TR), when the multi-phase acquisition is repeated, two phase differences appear repeatedly in cycles of N (ie, φ ₁ , φ ₂ ,. ..., φ _N , φ ₁ , φ ₂ , ..., φ _N , φ ₁ , φ ₂ , ..., φ _N (see FIG. This appears in resonant form in which the signal of the neuron is periodically cycled to N repeatedly obtained multi-phase data (see (ii), (iii) of FIG. 2C). There is repetitive multiphase data for each MRI image pixel present in space and a Fourier transform is performed in the repeated multiphase direction to extract a signal corresponding to the frequency of the neuron determined by N (Iv) and (v) in FIG. 2C).

3 (a) shows an example of a result obtained by mapping a resonance frequency distribution diagram of a neuron using a neuron resonance magnetic resonance image (NR-MRI) according to an embodiment of the present invention.

FIG. 3 (b) shows an example of a result obtained by analogy of a communication channel between brain regions through an algorithm for optimizing a population simulation such as the Wilson-Cowan model. Such a communication channel map can be applied to illuminate brain regions communicating with each other in a frequency selective manner. In FIG. 3 (b), reference numeral 601 denotes a Wilson-Cowan model, reference numeral 602 denotes an excitatory population, and reference numeral 603 denotes an inhibitory population.

FIG. 3 (c) shows an example of a result of creating a distribution diagram in which various feedback loops exist in each brain region, according to an embodiment of the present invention.

As shown in FIG. 3, the neuron frequency mapping technique described above can be applied to various neuron frequencies to study resonance characteristics of neurons according to frequency bands and communication mechanisms between brain regions. This mapping of brain-to-brain communication can be further developed and completed through system biology.

A vibration gradient magnetic field pattern can increase the signal of a neuron, but one image acquisition can still be low. In general, iterative image acquisition can increase the signal due to the averaging effect, but it is difficult to see the averaging effect in the conventional neuron imaging technique using MRI. This is because the temporal phase of the neuron resonance and the generation / extinction interval are random.

However, by applying the repetitive multiphase image acquisition and Fourier analysis according to an embodiment of the present invention, even if the temporal phase of the neuron resonance and the generation / extinction interval are unknown, the averaging effect can be observed temporally , Thereby increasing the neuron signal-to-noise ratio (S / N ratio). Accordingly, the neuron resonance magnetic resonance imaging (NR-MRI) according to one embodiment of the present invention has the following five advantages.

(1) It is possible to map the resonance of a neuron according to a frequency band which has not been previously possible.

(2) There is no need for prior knowledge about the resonance frequency, temporal phase, and interval of the neuron.

(3) Both dysfunctional and functional changes can be measured when external stimuli are present.

(4) It can be implemented in all existing MRI equipment.

(5) High frequency neuron resonance components are more advantageous because they can accelerate MRI image acquisition.

4A to 4F are graphs showing simulation results of a neuron resonance magnetic resonance imaging (NR-MRI) according to an embodiment of the present invention.

4A shows a result after repeating 40 cycles of a neuronal signal model, FIG. 4B shows a neuronal signal + noise, FIG. 4C shows a result of repeating 40 cycles of the neuron signal + noise, 4E shows a result after repeating 200 cycles of "neuron signal + noise". FIG. 4F shows a result of repeating 200 cycles of "neuron signal + noise" and "noise" And the data obtained by randomly crossing the other 200 cycles of the repeated data with each other. 4A to 4F, the upper portion of each graph represents the size, the middle represents the phase, and the lower represents the spectrum. At this time, the arrows 41 to 46 shown in FIG. 4 indicate the positions of the resonance frequencies of the neurons (33.3 Hz in the simulation according to the embodiment of the present invention).

At this time, in the simulation of a neuron resonance magnetic resonance imaging (NR-MRI) according to an embodiment of the present invention, the noise is five times stronger than the neuron signal.

To simulate the possibility of neuron current sensing in repetitive multiphase imaging and Fourier analysis, a simple simulation was performed. As shown in FIG. 4A, it is assumed that the MRI signal generated by the neuron represents a sinusoidal curve in time, and that the frequency is 33.3 Hz (assuming that there is only a temporal phase change and no change in size) ). As shown in FIG. 4B, the spectrum after the Fourier transform shows that the neuron signal corresponding to 33.3 Hz completely disappeared after adding the complex random noise which is 5 times higher than the MRI signal by the neuron. When the operation of FIG. 4B was repeatedly applied to 40 cycles, the electric signal of the neuron corresponding to 33.3 Hz began to appear again as shown in FIG. 4C. At this time, if the number of cycles is further increased, as shown in FIGS. 4D and 4E, the noise is further reduced. Even after the data ("noise") consisting only of random noise is randomly interleaved between the data ("signal plus noise") in FIG. 4e, the signal of the neuron is detected in the spectrum after the Fourier transform, (The total number of "signal + noise" cycles equals the total number of "noise" cycles).

That is, as shown in FIG. 4, both the magnitude and the phase signal are largely determined by the noise, and it can be seen that the 1/5 small signal by the neuron is hardly distinguished on the time axis. These characteristics are very similar to those of actual MRI signals. The simulation of a neuron resonance magnetic resonance imaging (NR-MRI) according to an embodiment of the present invention is based on the assumption that the MRI signal of the neuron changes only in phase and the size does not change, When the experiment was conducted, similar results were obtained (data not shown). These results show that the repetitive multi-phase measurement method and the Fourier analysis method can be applied to the present invention, and even if the resonance signal of the neuron has a time-wise random interval and phase and the noise of the MRI signal is 5 times or more, it means.

Referring to FIGS. 2A to 2C again, selective frequency band filter characteristic mapping of a neuron through a neuron resonance magnetic resonance imaging (NR-MRI) according to an embodiment of the present invention will be described.

According to a neuron resonance magnetic resonance imaging (NR-MRI) according to an embodiment of the present invention, as shown in FIG. 2A, a multi-echo gradient echo (ME-GE), a spin-echo echo planar imaging ), GE (Gradient Echo), and the above three MRI pulse train types (ME-GE, SE-EPI, GE). The SE-EPI and the GE sequences are implemented using a separate neuronal resonance oscillating gradient (NROG).

As shown in FIG. 2B, the Fourier analysis method is applied after repeatedly obtaining the multiphase data. In addition, time-division analysis is applied to repetitive multiphase data to observe the temporal change of the neuron's resonance frequency, and the resonance frequency of the target neuron, that is, the vibration frequency of the vibration gradient magnetic field, , And to test and optimize various MRI imaging parameters. The parameters to be optimized include the number of echoes, the number of oblique magnetic field pulses oscillating within one repetition time TR, the number of multiple phases, the magnitude of the vibration gradient magnetic field, the type of MRI pulse train to be imaged , SE-EPI, GE), and the magnitude of the MRI main magnetic field.

All of these studies are performed both in the dormant state and in the external stimulus and are to observe the frequency change trend between the dormant state and the stimulated state. The SE-EPI and the GE sequences are used to determine the difference in spatial direction of the resonance signals of the neurons As shown in Fig. In addition, NR-MRI is comparatively verified against the existing fMRI, resting-state fMRI, and neuron current measurement MRI techniques in various perspectives (temporal signal variation, spatial signal distribution, etc.).

According to a neuron resonance magnetic resonance imaging (NR-MRI) according to an embodiment of the present invention, a vibration gradient magnetic field pattern can be implemented in ME-GE, SE-EPI, and GE sequences have. At this time, the number of multi-phases can be changed from 4 to 16 by 4 increments. The resonance frequency of the target neurons is tested from 5 Hz to 500 Hz, varying at relatively fine intervals in the low frequency band and widely spaced in the high frequency band. In this case, the frequency band includes both theta waves (3.5-7.5 Hz), alpha waves (7.5-12.5 Hz), and beta waves (> 12.5 Hz). The number of echo in the ME-GE sequence and the number of oscillating gradient pulses in one TR of the SE-EPI or GE sequence can vary from 2 to 10. All other basic MRI imaging parameters are tailored accordingly. Optimization of these parameters is done in parallel with simulation and actual experiments.

MRI experiments are performed using an SD-rat (Sprague Dawley rats) weighing 200-500 g, and can be used for animal anesthesia using isoflurane.

In a neuron resonance magnetic resonance imaging (NR-MRI) according to an embodiment of the present invention, frequency characteristics of neurons according to the degree of anesthesia can be observed by testing at various isoflurane levels. At this time, electric stimulation can be applied in the following parameter ranges.

current = 0.5-1.5mA,

pulse duration = 1-3 ms,

repetition rate = 3-10 Hz,

stimulation duration = 5-30 sec,

inter-stimulation period ≥ 2 min.

The Fourier analysis method is applied to all data obtained once, and it can also be applied by dividing it by various time intervals (from 10 seconds to 10 minutes) to observe the temporal change of the resonance frequency of the neuron. The temporal change in the resonance frequency of these neurons can be evaluated throughout the brain.

Conventional fMRI, resting-state fMRI, and neuron current measurement MRI techniques follow previously published data acquisition and analysis techniques. In addition, the basic photographing conditions among the photographing parameters of the neuron resonance magnetic resonance imaging (NR-MRI) according to an embodiment of the present invention maintain the conditions as similar or similar to those of the existing techniques for a fair comparison. The frequency distribution of the brain measured by the NR-MRI in the dormant state is compared with the connectivity map obtained through the resting-state fMRI, and the neuron resonance magnetic resonance image (NR-MRI) The frequency distribution of the brain measured by resonance magnetic resonance imaging (NR-MRI) is compared with the conventional fMRI map. A comparison of the temporal change of the fMRI signal with that of the neuron resonance magnetic resonance image (NR-MRI) according to an embodiment of the present invention shows that the origin of the signal of the neuron resonance magnetic resonance image (NR-MRI) It is helpful to verify that it is due to resonance and is not a signal due to hemodynamics.

Meanwhile, the neuron resonance magnetic resonance imaging (NR-MRI) technique according to the present invention is a technique for directly and non-invasively measuring the electrical signals of a neuron using MRI at high resolution, . Successful implementation of neuron resonance magnetic resonance imaging (NR-MRI) will provide new information for studying our brain and can have enormous ripple effects on various brain-related studies including neuroscience, psychology, and pathology. For example, we can study how communication frequencies change between brain regions while our brain performs certain tasks. In addition, it is possible to study how the communication frequency between brain regions is different between normal and non-human patients, thereby observing changes in the communication frequency between specific brain regions that may cause brain diseases. It is possible to complete the communication map of the whole brain according to the frequency, which will greatly improve the understanding of the brain including the normal person and the mental patient.

The fact that many brain diseases are caused by communication between brain regions is still being revealed. Various methods, including a variety of electrical and magnetic stimuli, have been attempted to treat brain disorders such as schizophrenia and Parkinson's disease. However, such therapies are often accompanied by serious side effects. These side effects may be caused by amplifying the wrong brain communication frequency and stimulating unwanted areas of the brain.

Through the neuron resonance magnetic resonance imaging (NR-MRI) according to the present invention, it is possible to find a desirable frequency band and to reduce side effects through stimulation of such a wrong frequency band.

Meanwhile, brain research is considered to be very important in advanced countries, and the development of new image modality proposed by the neuron resonance magnetic resonance imaging (NR-MRI) according to the present invention is an absolute condition for leading the field .

The neuron resonance magnetic resonance imaging (NR-MRI) technique according to the present invention is expected to be able to present a new breakthrough in brain communication research and brain disease diagnosis.

Hereinafter, a magnetic resonance signal processing method according to an embodiment of the present invention will be described with reference to FIG. This processing method is a magnetic resonance signal processing method using a process of obtaining MRI data ((iii) in FIG. 2) using an MRI imaging technique using gradient magnetic field patterns 92 to 94 to sense the magnetic field vibration signal 96 Obtaining a plurality of MRI data by repeating the process a plurality of times while changing relative phases (? ₁ ,? ₂ , ...,? _N ) between the gradient magnetic field pattern and the magnetic field vibration signal; And calculating frequency characteristics from the obtained plurality of MRI data by a predetermined rule. Here, the plurality of MRI data may be K-space data or MRI image data. Here, the magnetic field vibration signal may mean a signal relating to a magnetic field including a specific frequency component. The inclined magnetic field pattern may refer to a signal relating to a magnetic field having a signal pattern including a certain frequency component. The MRI imaging technique using the gradient magnetic field pattern signal may refer to an MRI imaging technique that uses a gradient magnetic field pattern signal to generate a gradient magnetic field. The relative phase can be easily defined in the relative phase between the two signals when it is assumed that the gradient magnetic field pattern and the magnetic field vibration signal have the same frequency.

At this time, the magnetic field vibration signal may be one in which the phase at the time of occurrence and the phase at a specific time point are unknown.

At this time, the magnetic field vibration signal may include a specific frequency component (ex: _fNO ).

At this time, the predetermined rule may be to Fourier transform the signals (ex 201 and 202 in FIG. 2C) obtained by arranging the acquired plurality of MRI data on the time axis according to the relative phase.

At this time, the plurality of MRI data may be temporally divided, and Fourier transform may be performed on each of the divided data to obtain time-base data of a signal change of the neuron. For example, since the period of the magnetic field vibration signal may change with time, or since a magnetic field vibration signal may not be generated in some time periods, the above-described divided data can be used to grasp the changing pattern.

At this time, the predetermined rule may be configured to find a value related to the specific frequency (ex: _fNO ) among the frequency components of the acquired plurality of MRI data.

At this time, the acquiring step repeats the process for the predetermined N relative phases (? ₁ ,? ₂ , ...,? _N ) between the gradient magnetic field pattern signal and the magnetic field vibration signal, Wherein the difference value between the N relative phases may be an integer multiple of 2 * pi / N for the particular frequency.

At this time, the time to repeat (TR) of the MRI image parameters may be adjusted to make the difference value between the N relative phases an integral multiple of 2 *? / N with respect to the specific frequency.

At this time, the magnetic field vibration signal may be a neuron signal. The gradient magnetic field pattern may be a vibration gradient magnetic field pattern. The vibration gradient magnetic field pattern may be a bipolar readout gradient pattern used in a multi-echo gradient echo imaging technique.

At this time, the spatial directionality of the gradient magnetic field pattern may be different for the two different processes that are repeatedly performed.

An MRI apparatus adapted to perform the steps described above according to another embodiment of the present invention may be provided.

According to another embodiment of the present invention, a computer-readable medium having a program recorded thereon for causing the MRI apparatus to perform the steps described above may be provided.

According to another embodiment of the present invention, MRI data is obtained by an MRI imaging technique using a vibration gradient magnetic field pattern signal having a predetermined vibration frequency to detect a magnetic field vibration signal whose phase at an occurrence section and at a specific time point is unknown A magnetic resonance signal processing method using a process can be provided. The method includes: obtaining a plurality of MRI data by repeating the process a plurality of times while changing a relative phase between the vibration gradient magnetic field pattern signal and the magnetic field vibration signal; And a step of calculating a value relating to the predetermined vibration frequency from the obtained plurality of MRI data by a predetermined rule. And obtaining P values for the predetermined vibration frequency for each of the first processes by performing the first process P times so that the predetermined vibration frequency has P different values have.

According to the neuron resonance MRI technique of the present invention, the following three effects can be obtained.

Effects 1. Neuron resonant magnetic resonance imaging (NR-MRI) can map the selective frequency band filter characteristics of neurons. An MRI imaging technique that combines a resonant gradient magnetic field pattern is repeatedly applied to multiple phases of the neuron relative to the resonance frequency to obtain an image and a frequency selective neuron signal can be extracted by applying a Fourier analysis method. This technique can be applied to both resting-state and external stimuli for various frequency ranges and can be compared with conventional MRI imaging techniques.

Effect 2. Provide a technique for performing system biological studies on frequency selective communication mechanisms between brain regions. This allows spiking neural networks based neuron modeling to reveal frequency selective inter-brain communication mechanisms. And neuronal population modeling, such as the Wilson-Cowan model, can be used to study the interaction between excitatory and inhibitory neuronal populations.

Effect 3. Provide a technique to complete proof-of-concept experiments and a frequency-selective cross-domain communication map. The proof-of-concept experiment can be demonstrated by (i) using both NR-MRI and EEG and (ii) applying NR-MRI to genetically engineered mice. Combining experiments with NR-MRI and system biology-based simulations can provide additional data for proof-of-concept and contribute to the complete selection of frequency-selective brain-to-brain communications.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the essential characteristics thereof. The contents of each claim in the claims may be combined with other claims without departing from the scope of the claims.

Claims (18)

A magnetic resonance signal processing method using a process for obtaining MRI data by an MRI imaging technique using an oblique magnetic field pattern to sense a magnetic field vibration signal,
Obtaining a plurality of MRI data by repeating the process a plurality of times while changing a relative phase between the gradient magnetic field pattern and the magnetic field vibration signal; And
Calculating a frequency characteristic from the acquired plurality of MRI data by a predetermined rule;
/ RTI &gt;
Wherein the plurality of MRI data are K-space data or MRI image data,
Magnetic resonance signal processing method. The magnetic resonance signal processing method according to claim 1, wherein the phase of the magnetic field vibration signal at the time of occurrence and at a specific time point is unknown. 2. The method of claim 1, wherein the magnetic field vibration signal comprises a specific frequency component. 2. The magnetic resonance signal processing method according to claim 1, wherein the predetermined rule is to Fourier-transform a signal obtained by arranging the obtained plurality of MRI data on a time axis according to the relative phase. The method according to claim 1,
Wherein the acquiring comprises:
Using an acquisition process to obtain N MRI data by repeating the process for a predetermined N relative phases between the gradient magnetic field pattern signal and the magnetic field vibration signal,
Obtaining the plurality of MRI data by repeating the acquisition process M times in succession over time
/ RTI &gt;
Magnetic resonance signal processing method. 6. The magnetic resonance signal processing method according to claim 5, wherein the predetermined rule is to Fourier transform the obtained plurality of MRI data on a signal arranged on a time axis according to the predetermined relative phase. The magnetic resonance signal processing method according to claim 1, wherein the plurality of MRI data is temporally divided and Fourier transform is performed on each of the divided data to obtain time-axis data for a signal change of a neuron. The method of claim 1, wherein the specific frequency is the same as the frequency of the gradient magnetic field pattern. The method of claim 1, wherein the predetermined rule is adapted to find a value for the specific frequency among frequency components of the obtained plurality of MRI data. The method according to claim 1,
Wherein the acquiring step includes an acquiring process for acquiring N MRI data by repeating the process for a predetermined N relative phases between the gradient magnetic field pattern signal and the magnetic field vibration signal,
Wherein the difference value between the N relative phases is an integral multiple of 2 * pi / N for the particular frequency.
Magnetic resonance signal processing method. 11. The method of claim 10, further comprising: adjusting a time to repeat (TR) of the MRI image parameters to make the difference value between the N relative phases an integer multiple of 2 * [pi] / N for the particular frequency. Resonance signal processing method. The magnetic resonance signal processing method according to claim 1, wherein the magnetic field vibration signal is a neuron signal. The magnetic resonance signal processing method according to claim 1, wherein the gradient magnetic field pattern is a vibration gradient magnetic field pattern. 14. The neuron resonance magnetic resonance imaging method according to claim 13, wherein the vibration gradient magnetic field pattern is a bipolar readout gradient pattern used in a multi-echo gradient echo imaging method. 2. The method according to claim 1, wherein, in the two different processes repeatedly performed, the gradient directions of the gradient magnetic field patterns are different from each other. A process of obtaining MRI data by an MRI imaging technique using an oblique magnetic field pattern signal to resonate at a specific frequency and detect a magnetic field vibration signal whose phase is unknown at a specific time point and at a specific time point,
Obtaining a plurality of MRI data by repeating the process a plurality of times while changing a relative phase between the gradient magnetic field pattern signal and the magnetic field vibration signal; And
Calculating a frequency characteristic from the acquired plurality of MRI data by a predetermined rule;
Lt; / RTI &gt;
MRI apparatus. MRI apparatus,
A process of obtaining MRI data by an MRI imaging technique using an oblique magnetic field pattern signal to resonate at a specific frequency and detect a magnetic field vibration signal whose phase is unknown at a specific time point and at a specific time point,
Obtaining a plurality of MRI data by repeating the process a plurality of times while changing a relative phase between the gradient magnetic field pattern signal and the magnetic field vibration signal; And
Calculating a frequency characteristic from the acquired plurality of MRI data by a predetermined rule;
A program is recorded,
Computer-readable medium. A magnetic resonance signal processing method using a process for obtaining MRI data by an MRI imaging technique using a vibration gradient magnetic field pattern signal having a predetermined vibration frequency to detect a magnetic field vibration signal whose phase is unknown at a generation period and at a specific time point,
Obtaining a plurality of MRI data by repeating the process a plurality of times while changing a relative phase between the vibration gradient magnetic field pattern signal and the magnetic field vibration signal; And a step of calculating, by a predetermined rule, a value relating to the predetermined vibration frequency from the obtained plurality of MRI data,
Obtaining P values for the predetermined vibration frequency for each of the first processes by performing the first process P times so that the predetermined vibration frequency has P different values,
/ RTI &gt;
Magnetic resonance signal processing method.

Images