Method and apparatus for dynamic imaging of vascular activity on a microscopic scale - Patents.com

Abstract

1. A method for dynamic imaging of vascular activity at a microscopic scale in a human or animal vascular network, the method comprising: (a) performing time series ultrasound localization microscopy images of a region of the vascular network to obtain values ​​of a vascular dynamics parameter in a region of interest within the region, wherein recording periods of the time series of ULM images correspond to dynamic events of a cause other than cardiac pulsatily that activate the vascular network in the region; and (b) calculating a measure of evolution of the vascular dynamics parameter in response to the dynamic event based on the values ​​of the vascular dynamics parameter.

Description

[Technical field]

[0001] The present disclosure relates to methods and apparatus for dynamic imaging of vascular activity on a microscopic scale in humans or animals. [Background technology]

[0002] Most whole brain functional imaging modalities utilize neurovascular coupling to map brain activity with millimeter resolution. More specifically, activated regions of the nervous system require more oxygen and therefore locally increase blood flow in the vascular network of the nervous system, particularly in the capillaries, venules, and arterioles of the vascular network. Thus, neural activity in the nervous system can be estimated on a millimeter scale based on the activity of the vascular network of the nervous system.

[0003] Nevertheless, recording functional responses at the microscopic scale in mammals remains the prerogative of electrophysiological or optical techniques, which are, however, limited to either the cortical surface or the immediate vicinity of the implanted sensors.

[0004] That is, functional imaging is limited by spatial resolution for neurovascular coupling-based methods and by penetration depth for electrophysiological and optical methods.

[0005] One type of imaging of particular interest in terms of spatial resolution and penetration depth is Ultrasound Localization Microscopy (ULM), described in particular by Errico et al. "Errico, C. et al., Ultrafast ultrasound localization microscopy for deep super-resolution vascular imaging. Nature 527, 499-+ (2015)" and Demene et al. "Demene, C. et al., Transcranial ultrafast ultrasound localization microscopy of brain vasculature in patients. Nat. Biomed. Eng 5, 219-228 (2021)". Ultrasound localization microscopy achieves transcranial imaging of cerebral vascular flow down to the micron scale by localizing millions of microbubbles (MB) injected intravenously. However, as explained by Higot et al., "Hingot, V. et al., Microvascular flow dictates the compromise between sensitivity to tiny vessels and acquisition time in Ultrasound Localization Microscopy. Scientific Reports 9, (2019)," the long acquisition times required to detect single microbubble signatures within small vessels have thus far limited the application of ULM primarily to imaging the structure of the microvasculature. Thus, there is a need to dynamically image vascular activity at the microscopic scale, especially throughout the brain. [Prior art documents] [Non-patent literature]

[0006] [Non-Patent Document 1] Errico, C. et al., Ultrafast ultrasound localization microscopy for deep super-resolution vascular imaging. Nature 527, 499-+ (2015) [Non-Patent Document 2] Demene, C. et al., Transcranial ultrafast ultrasound localization microscopy of brain vasculature in patients. Nat. Biomed. Eng 5, 219–228 (2021) [Non-Patent Document 3] Hingot, V. et al., Microvascular flow dictates the compromise between sensitivity to tiny vessels and acquisition time in Ultrasound Localization Microscopy. Scientific Reports 9, (2019) [Non-Patent Document 4] Hingot, V. et al., Microvascular flow dictates the compromise between spatial resolution and acquisition time in Ultrasound Localization Microscopy. Scientific Reports 9, (2019) [Non-Patent Document 5] Dayton, PA et al., Molecular ultrasound imaging using microbubble contrast agent. Frontiers in Bioscience 12, 5124–5142 (2007) [Non-Patent Document 6] Demene, C. et al., Spatiotemporal Clutter Filtering of Ultrafast Ultrasound Data Highly Increases Doppler and fUltrasound Sensitivity. IEEE Transactions on Medical Imaging 34, 2271~2285 (2015) [Non-Patent Document 7] Hingot, V. et al., Subwavelength motion-correction for ultrafast ultrasound localization microscopy. Ultrasonics 77, 17–21 (2017) [Non-Patent Document 8] Jerman, T. et al., Enhancement of Vascular Structures in 3D and 2D Angiographic Images. IEEE Transactions on Medical Imaging 35, 2107~2118 (2016) Summary of the Invention [Means for solving the problem]

[0007] To this end, the present disclosure proposes a method for dynamic imaging of vascular activity on a microscopic scale in a human or animal vascular network, preferably in the vascular network of the nervous system, the method comprising: (a) performing time-series ultrasound localization microscopy (ULM) imaging of at least one region of a vascular network to obtain values ​​of at least one vascular dynamic parameter in at least one region of interest within the at least one region during a recording period, the time-series ULM images being constructed from the time-series images of the region, each ULM image being constructed from a stack of images of the time-series images, at least some of the images of the stack of images corresponding to a (k+1)th ULM image having a time position greater than a time position of an image of the stack of images corresponding to a kth ULM image immediately preceding the (k+1)th ULM image, the recording period of the time-series ULM images including at least one dynamic event due to a cause other than cardiac beats, the dynamic event activating a change in hemodynamics and / or a change in structural conformation of the vascular network in the at least one region; (b) calculating a measure of evolution of the at least one vascular dynamic parameter in response to the dynamic event based on values ​​of the at least one vascular dynamic parameter from the time series of ULM images; Includes.

[0008] The proposed method advantageously allows for obtaining a dynamic understanding of vascular activity at a microscopic scale, depending on the embodiment. The acquisition of a time series of individual ULM images may aggregate information from the individual ULM images. Aggregating information from the individual ULM images may reduce the acquisition period of the individual ULM images while maintaining high spatial resolution. Thus, in some embodiments, the acquisition period of the time series may be shorter than the theoretical limit acquisition period disclosed by Higot et al. "Hingot, V. et al., Microvascular flow dictates the compromise between spatial resolution and acquisition time in Ultrasound Localization Microscopy. Scientific Reports 9, (2019)". Three-dimensional (3D) information from the time series of individual ULM images allows for the derivation of a measure of the evolution of at least one vascular dynamic parameter in response to a dynamic event.

[0009] In some embodiments, the inventors have achieved a dynamic study of functional hyperemia, an increase in blood flow to various tissues in the body during brain activation. Notably, in some embodiments, the inventors have achieved a spatiotemporal resolution of (6.5 μm, 1 s) in deep cortical and subcortical regions of the rat brain, surpassing the "classical" ULM performance limited by long acquisition periods. That is, quantitative information on both function and dysfunction of living organs can be obtained.

[0010] In one embodiment, each image in the time series of images may be constructed from successive composite images. Each composite image of a region may be obtained from a set of planar or divergent ultrasound waves transmitted from the region.

[0011] In one embodiment, the vascular network is pre-injected with an ultrasound contrast agent prior to acquiring the images.

[0012] The following features may be optionally implemented individually or in combination with other features.

[0013] In one or more embodiments, the vascular network can be a vascular network of the nervous system and the at least one dynamic event can be a stimulus delivered to the nervous system. (c) delivering at least one stimulus to the nervous system. It may further include.

[0014] A measure of the evolution of at least one kinetic parameter calculated from the time series of ULM images can reflect the response of at least one region of the nervous system to at least one stimulus. In some embodiments, therefore, vascular activity throughout the brain can be measured during the task-induced activity. Additionally, in some embodiments, the spatial extent and influence of at least one region of the nervous system can be deciphered during the task-induced activity.

[0015] In one or more embodiments, the vascular network can be a vascular network of the nervous system, and the at least one dynamic event can be spontaneous activity of the nervous system.

[0016] A measure of the evolution of at least one dynamic parameter calculated from the time series of ULM images may reflect the functional connectivity of at least one region of the nervous system. Thus, in some embodiments, it may be possible to measure interactions occurring in the brain during a resting or task-negative state.

[0017] In one or more embodiments, the recording period of a time series of ULM images may be longer than the duration of a cardiac cycle and shorter than one minute.

[0018] Thus, the recording period of a ULM image may be shortened relative to a "classical" ULM image while still being long enough to observe the ULM contrast agent.

[0019] In one or more embodiments, steps (a) and (c) may be repeated for N trials, and the evolution of at least one vascular dynamic parameter may be calculated based on a time series of the n trials, where n is an integer greater than 1.

[0020] In some embodiments, the information for calculating the measure of the evolution of at least one vascular dynamics parameter can be enhanced by repeating the pattern of stimulation during n trials and combining equivalent time points in n trials. Thus, the sparsity of the information for calculating the measure of the evolution of at least one vascular dynamics parameter can be reduced. The sensitivity of the information for calculating the measure of the evolution of at least one vascular dynamics parameter can be increased.

[0021] In one or more embodiments, calculating the measure of evolution of at least one vascular dynamic parameter in response to a dynamic event may include a correlation analysis between the stimulus and a time series of ULM images.

[0022] Correlation analysis between stimulation and time series of ULM images can provide a quantitative map of functional hyperemia.

[0023] In one or more embodiments, calculation of a measure of the evolution of at least one vascular dynamic parameter in response to a dynamic event may involve more advanced processing analysis, including, but not limited to, Singular Value Decomposition (SVD) or Principal Component Analysis (PCA) of the time series of ULM images.

[0024] Processing analysis of the time series of ULM images may allow to isolate a measure of the evolution of at least one vascular dynamics parameter in response to a dynamic event from the profile of the ultrasound agent injection. Thus, the evolution of at least one vascular dynamics parameter in response to a dynamic event may be observed independently of the profile of the ultrasound agent injection. For example, SVD of the time series of ULM images may allow to isolate the functional hyperemia in one singular vector (and the profile of the ultrasound agent injection in another singular vector). The local changes of the contrast agent of the ULM method may be represented in the singular vectors of the SVD. Thus, a quantitative map of the functional hyperemia may be obtained.

[0025] In one or more embodiments, the calculation of the measure of evolution of at least one vascular dynamic parameter in response to a dynamic event may include a sliding average of a time series of ULM images.

[0026] Performing a sliding average of a time series of ULM images may allow observing a smoothed evolution of brain activity. In particular, for diffuse or propagating events such as epileptic seizures or spreading depression, an analysis of the spread or propagation of the seizure or spreading depression may be obtained.

[0027] In one or more embodiments, successive ULM images in a time series of ULM images may be obtained from a stack of at least partially overlapping images.

[0028] Computing successive ULM images of a time series of ULM images based on a stack of images that at least partially overlap in the time dimension may increase the temporal resolution of a measure of at least one vascular dynamic parameter.

[0029] In one or more embodiments, successive ULM images of a time series of ULM images may be obtained from respective stacks of successive images in the time dimension.

[0030] In some embodiments, each stack of images may be immediately adjacent in the time dimension.

[0031] In one or more embodiments, the at least one vascular dynamic parameter calculated may be selected from the group including blood flow, blood flow velocity, blood volume, blood pressure, blood vessel diameter, and any combination thereof.

[0032] Calculating kinetic parameters from among blood flow rate, blood flow velocity, blood volume, blood pressure, vessel diameter, and any combination thereof may allow to derive the behavior of at least one region of the vascular network and to quantify the activity of at least one region without requiring a large amount of information from different types. The diameter may be obtained by various segmentation techniques applied to the ULM images. In its simplest form, the diameter may be defined as the width of the vessel at a threshold set by the half-maximum of the resting profile. The blood flow rate may be derived from the flow rate of the ULM contrast agent during the recording of the time series. The blood flow velocity may be derived from the velocity of the ULM contrast agent during the recording of the time series.

[0033] In one or more embodiments, the region may be further segmented into one or more vascular compartments, such as perforator arterioles, pial vessels, intraparenchymal vessels, or venules, and at least one region of interest may correspond to at least one of the vascular compartments.

[0034] The present disclosure also relates to an apparatus for dynamic imaging of vascular activity on a microscopic scale in a human or animal vascular network, preferably in the vascular network of the nervous system, the apparatus comprising: (a) an ultrasound contrast agent injection device and an ultrasound measurement device adapted to perform time-series Ultrasound Localization Microscopy (ULM) imaging of at least one region of a vascular network to obtain values ​​of at least one vascular dynamic parameter in at least one region of interest within the at least one region during a recording period, the time-series ULM images being constructed from a time-series image of the region, each ULM image being constructed from a stack of images of the time-series images, at least some of the images of the stack of images corresponding to a (k+1)th ULM image having a time position greater than a time position of an image of the stack of images corresponding to a kth ULM image immediately preceding the (k+1)th ULM image, the recording period of the time-series ULM images including at least one dynamic event due to a cause other than cardiac beats, the dynamic event activating a change in hemodynamics and / or a change in structural conformation of the vascular network in the at least one region; (b) a computing module adapted to calculate a measure of evolution of the at least one vascular dynamic parameter in response to a dynamic event based on values ​​of the at least one vascular dynamic parameter from the time series of ULM images; Includes.

[0035] In embodiments of the device, the following features may be used alone or in combination.

[0036] In one or more embodiments, the dynamic event is an epileptic seizure or a spreading depression, and the computing module may be adapted to estimate a location of the seizure focus or an origin of the spreading depression from a measure of the evolution of the at least one vascular dynamic parameter in response to the epileptic seizure or the spreading depression.

[0037] The measure of the evolution of the vascular dynamic parameters of the area of ​​the vascular network allows to track the dynamics of brain activity. In some embodiments, the location of the focus of epileptic seizures occurring during the time series of recordings can be identified by examining the evolution of brain activity. For example, in some embodiments, the epileptic focus can be identified by analyzing the direction and / or speed of flow. Thus, the object of the present specification is to enable accurate and rapid identification of the epileptic focus by the spatiotemporal resolution. Similarly, the origin of spreading depression can be identified accurately and quickly.

[0038] In one or more embodiments, the computing module may be adapted to diagnose whether a measure of the evolution of at least one vascular dynamic parameter in response to a dynamic event corresponds to a given disease, in particular a neurodegenerative disease.

[0039] In some embodiments, the object of this specification can map the functional activity of the whole brain, providing insight into neural brain circuits.The measure of the evolution of at least one vascular dynamics parameter in response to dynamic events can be compared with a predetermined threshold value that characterizes whether the measure of the evolution of the vascular dynamics parameter is normal and / or whether the vascular dynamics parameter corresponds to a predetermined disease (in the case of multiple vascular dynamics parameters, each response parameter is compared with the corresponding predetermined threshold value, respectively).

[0040] In one or more embodiments, the computing device is adapted to monitor the efficiency of a medical treatment for a given disease, in particular one of epilepsy, spreading depression, and neurodegenerative diseases, based on a measure of the evolution of at least one vascular dynamic parameter in response to a dynamic event.

[0041] The evolution of the vascular dynamic parameters in response to a medical treatment may allow the effect of the medical treatment to be effectively monitored, from which adaptations of the medical treatment may be derived.

[0042] In another aspect, proposed is computer software comprising instructions for performing at least a part of the methods defined herein when the software is executed by a processor.In another aspect, proposed is a computer readable non-transitory recording medium registered with software for performing the methods defined herein when the software is executed by a processor.

[0043] Other features, details, and advantages are set forth in the following detailed description and figures. [Brief description of the drawings]

[0044] [Figure 1] 1 is a block diagram illustrating an embodiment of an apparatus according to the present disclosure. [Diagram 2] FIG. 2 is a block diagram showing a possible method for acquiring a ULM image using the device of FIG. 1. [Diagram 3] FIG. 2 illustrates a possible method for acquiring a ULM image using the device of FIG. 1. [Figure 4] FIG. 2 illustrates a possible method for calculating at least one measure of the evolution of at least one vascular dynamic parameter in response to a dynamic event according to an embodiment of the present invention. [Diagram 5] FIG. 2 illustrates a possible method for calculating at least one measure of the evolution of at least one vascular dynamic parameter in response to a dynamic event according to an embodiment of the present invention. [Figure 6] 1A and 1B show examples of measures (activation maps) of the evolution of at least one vascular dynamic parameter in response to a dynamic event in a rat, obtained using a method according to the present disclosure (a) and (b) using functional ultrasound (fUS) (b). [Figure 7A] FIG. 1 shows an example of a functional Ulm map illustrating segmentation of activated barrel cortex (S1BF) into four vascular compartments including perforator arterioles, venules, pial vessels, and intraparenchymal vessels. [Figure 7B]FIG. 7B shows a dynamic histogram of microbubble velocity distribution in the section defined in FIG. 7A. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0045] In the figures, the same reference numbers indicate identical or similar elements.

[0046] The present disclosure proposes a method and apparatus for dynamic imaging of vascular activity at a microscopic scale in a human or animal vascular network, preferably a vascular network of the nervous system, by performing time-series ultrasound localization microscopy images of at least one region of the vascular network. The recording period of the time-series ULM images may include at least one dynamic event due to a cause other than cardiac beats. The at least one dynamic event may be due to neuronal or other cellular activity. The at least one dynamic event may be induced, such as a task-induced stimulus, or may be uninduced, such as spontaneous activity, seizures, or spreading depression. The time-series ULM images allow obtaining values ​​of at least one vascular dynamic parameter in at least one region of interest in the at least one region. Based on the values ​​of the at least one vascular dynamic parameter from the time-series ULM images, a measure of the evolution of the at least one vascular dynamic parameter in response to the dynamic event may be calculated.

[0047] The region of the vascular network may be a slice of the brain in the case of 2D imaging, or a 3D region of the brain in the case of 3D imaging.

[0048] In some embodiments, ULM images can be obtained based on methods already known in the art and described in the above-mentioned papers by Errico et al. in 2015 and Demene et al. in 2021.

[0049] An example of an apparatus 1 for imaging vascular activity (VA APP) that can be used in carrying out the method according to the present disclosure is shown in FIG.

[0050] The apparatus 1 may include a processor 2 (PROC), for example a specialized signal processing device controlled by a computer or a group of computers, possibly including a server.

[0051] The processor 2 may include a computing module 3 (COMP), the operation of which will be described later.

[0052] The processor 2 may control an ultrasound contrast agent injection device 4 (INJ).

[0053] The ultrasound contrast agent injection device 4 may include an ultrasound contrast agent. The ultrasound contrast agent may be, for example, a microbubble or equivalent ultrasound contrast agent as described by Dayton et al., "Dayton, PA et al., Molecular ultrasound imaging using microbubble contrast agent. Frontiers in Bioscience 12, 5124-5142 (2007)". In some embodiments, the ultrasound contrast agent may be based on SonoVue®.

[0054] The ultrasound agent injection device 4 may be a push syringe in the example considered here.

[0055] The ultrasound agent injection device 4 may include a magnet to mix the solution containing the ultrasound contrast agent.

[0056] The processor 2 may control the probe 5 (PRB).

[0057] The probe 5, in the example considered here, may for example be an ultrasound probe.

[0058] The probe 5 may include an array 7 (ARR) of ultrasound transducers. The array may be a linear array adapted to generate a 2D image of a slice of the area to be imaged, or a 2D array adapted to generate a 3D image of the area. If the array is a 2D array, it may be a sparse matrix of transducers, as is known in the art.

[0059] A typical array of transducers may include hundreds to thousands of transducers. The array may also be limited in some instances to a single transducer adapted to image only one line of the region in the direction of depth from the transducer, or several transducers adapted to image each line of the region in the direction of depth from the transducer.

[0060] The following detailed description is given for the case of a linear array or a 2D array, so that the device generates an ultrasonic measurement having pixels. If the array includes only one transducer or several transducers, the device generates an image limited to one line (ultrasonic measurement) or several lines in the depth direction, the line having pixels, and the process is similar, except that the generation of ultrasonic measurements does not require tilted plane waves with different tilt angles.

[0061] The transducer may be adapted to transmit and receive ultrasound waves having a center frequency comprised, for example, between 0.5 MHz and 100 MHz, or between 1 MHz and 20 MHz. An example of a usable center frequency is 15 MHz.

[0062] In certain embodiments, the probe 5 may further comprise a motorization 6 (MOT) adapted to position the array 7 .

[0063] In some embodiments, processor 2 may additionally control stimulation devices not shown in FIG.

[0064] An example of a method for ultrasound localization microscopy imaging that is already known in the art and described, for example, in the above-mentioned paper by Demene et al. in 2021 will now be described with reference to Figures 2 and 3.

[0065] An ultrasound contrast agent injection device 4 may be controlled by the processor 2 to intravenously inject (S1(MB_INJ)) an ultrasound contrast agent into a region of the vascular network.

[0066] The ultrasound contrast agent injection may be a bolus injection with a maximum injection volume of 40 mL / kg for rats (corresponding to 12 mL for a 300 g rat and approximately 3 mL for a 70 g mouse), typically an injection volume of 6 mL for rats and 2 mL for mice. For humans, a maximum volume of 10 mL may be injected. For example, two boluses of 2 mL and 4 mL may be injected in humans.

[0067] The ultrasound contrast agent injection can be a continuous infusion. For example, the flow rate of the continuous infusion can be comprised between 3 mL / h / kg and 60 mL / h / kg, preferably about 10 mL / h / kg (or typically 3.5 mL / h for a 300 g rat and 1.0 mL / h for a 70 g mouse).

[0068] The array of transducers 7 may be controlled by the processor 2 to acquire a composite image of the area during ultrasound contrast injection (S2(DATA_ACQ)).

[0069] In one embodiment, an array of transducers may acquire a composite image of a region of the vascular network that has previously been administered an ultrasound contrast agent.

[0070] Computing device 3 may process the acquired composite image of the area to obtain a filtered image of the area (S3(DATA_PROC)).

[0071] Computing module 3 may calculate a ULM image of the region of the vascular network based on the filtered image of the region (S4(ULM_COMP)).

[0072] In the example of FIG. 3, the array 7 of transducers may be controlled by the processor 2 to obtain a composite image of the region during ultrasound contrast agent injection (S2) by transmitting planar ultrasound waves in the region to be imaged and receiving the resulting backscattered ultrasound waves at a rate of, for example, 5 kHz (pulse repetition frequency PRF), i.e., every 0.2 milliseconds. More generally, the pulse repetition frequency PRF may be greater than 500 Hz. The received signals may be registered as a set of raw data for each transmitted planar ultrasound wave. The successively transmitted plane waves may have a propagation direction inclined at a continuously varying angle with respect to the direction of depth in the region to be imaged, i.e., with respect to the direction perpendicular to the array 7. For each image of the region, a number Np of planar ultrasound waves may be transmitted successively at different angles, and the N sets of raw data may be coherently added to synthesize said image of the region, thus resulting in a composite image. For example, Np may be 5, with the angles taking the values ​​of -5 degrees, -2 degrees, 0 degrees, +2 degrees, +5 degrees. Thus, for Np=5 and PRF=5kHz, the rate of the composite image of the region (frame rate) is 1KHz. Np may be different from 5, in which case the frame rate of the composite image will be different. For example, Np=11 may be used. The array of transducers 7 may be placed on the skull of a human or animal for transcranial experiments, or directly on the brain of a human or animal for experiments involving craniotomy.

[0073] Then, based on successive composite images of a region, a filtered image of the vascular network in said region can be calculated by the computing module 3 (S3). In the example of FIG. 3, for each filtered image, N C N consecutive composite images are used. C The consecutive composite images may be consecutive in time or may overlap partially in the time dimension. For example,C may be between 10 and 1000, for example between 200 and 600. An example number of consecutive composite images available for one filtered image is 400. A time position may be associated with the filtered image. The time position of a given filtered image is determined by the number of N composite images forming the filtered image. C This may correspond to the average acquisition time of the four composite images.

[0074] The filtered images may be calculated, for example, by singular value decomposition (SVD). More specifically, the SVD spatiotemporal clutter filter described, for example, by Demene et al., "Demene, C. et al., Spatiotemporal Clutter Filtering of Ultrafast Ultrasound Data Highly Increases Doppler and fUltrasound Sensitivity. IEEE Transactions on Medical Imaging 34, 2271-2285 (2015)" may be applied. In the example of FIG. 3, each filtered image has N C The singular value or sum of the singular values ​​of the SVD applied to the composite images. In some embodiments, as described by Hingot et al. "Hingot, V. et al., Subwavelength motion-correction for ultrafast ultrasound localization microscopy. Ultrasonics 77, 17-21 (2017)", the ultrasound signatures of the individual ultrasound contrast agents can be differentiated from the tissue signal of the region due to cardiac and respiratory pulsatility by performing a speckle tracking correlation analysis of the composite image before applying the SVD. Alternatively, the ultrasound signatures of the individual ultrasound contrast agents can be differentiated from the tissue signal of the region for each filtered image by the SVD of the composite image. The tissue signal can be filtered by discarding some singular values ​​of the SVD. For example, the first 10 singular values ​​of the SVD, which mainly reflect the tissue signal of the region, can be discarded.

[0075] Each of the filtered images may be interpolated. For example, each of the filtered images may be interpolated by a Lanczos interpolation kernel. Each of the filtered images may be interpolated to achieve a sampling rate on the order of (α / 6×ε / 6), where α is the spatial pitch of the probe and ε is the wavelength of the ultrasound.

[0076] The stack of filtered images may be filtered based on vesselness filtering, for example, as described by Jerman et al., "Jerman, T. et al., Enhancement of Vascular Structures in 3D and 2D Angiographic Images. IEEE Transactions on Medical Imaging 35, 2107-2118 (2016)."

[0077] The ultrasound contrast agent may be detected in the filtered image. For example, the ultrasound contrast agent may be detected as the brightest local maximum that has a high correlation with the point spread function. High correlation may be defined as a correlation above a threshold. The threshold may be a value comprised between 0.5 and 1. For example, the threshold may be equal to 0.7. The point spread function is the imaging response of an isolated ultrasound contrast agent and may be modeled as a Gaussian spot with axial and lateral dimensions ε. Localization of the sub-pixel maximum may be performed using a fast local quadratic polynomial fit. The neighborhood of the fast local quadratic polynomial fit may be a 5×5 pixel neighborhood. The coordinates of the localized sub-pixel maximum may be rounded to a selected pixel size. The selected pixel size may be smaller than the pixel size of the filtered image. For example, the selected pixel size may be a divisor of the pixel size of the filtered image. The divisor may be a multiple of 2. For example, the divisor may be equal to 16.

[0078] The ultrasound contrast agent may be tracked in the filtered image. The tracking of the ultrasound contrast agent may be performed using a particle tracking algorithm known in the art. A trajectory may be calculated based on the tracking of the ultrasound contrast agent in the filtered image. The trajectory may correspond to a position of the tracked ultrasound contrast agent in the filtered image. Each position of the tracked ultrasound contrast agent in the filtered image may be associated with a time position corresponding to the time position of the filtered image in which the position is located. A trajectory with ultrasound contrast agents detected in a predefined number of consecutive filtered images may be calculated. The predefined number of consecutive filtered images may be comprised between 1 and 100. For example, the predefined number of consecutive filtered images may be 10. A spatial interpolation may be calculated for each trajectory to obtain one ultrasound contrast agent at each pixel located on a path between two consecutive pixels of the trajectory.

[0079] Successive positions of the ultrasound contrast agent within the trajectory can be used to calculate velocity parameters. For example, the frame-to-frame ultrasound contrast agent velocity vector components (along the probe x-axis and depth z-axis for 2D imaging, and along the probe x-axis, y-axis, and depth z-axis for 3D imaging) can be calculated. Absolute velocity magnitude can be calculated.

[0080] Based on the filtered images, the ULM images may be calculated by the calculation module 3 (S4(ULM_COMP)). The ULM images may be calculated based on the trajectory based on the filtered images. In the example of FIG. 3, for each ULM image, N F Each ULM image is filtered by selecting a pixel size and dividing each pixel by N F In some embodiments, N FThe values ​​of pixels accumulating values ​​of the detected ultrasound contrast agent in the filtered images below the threshold are set to zero. The threshold may be comprised between 1 and 10, and is preferably 5.

[0081] A ULM image of ultrasound contrast agent counts may be calculated based on the detected ultrasound contrast agents. For example, the ULM count image may be calculated by counting, for each pixel, the number of ultrasound contrast agents detected in the corresponding pixel of the filtered image. In the example of FIG. 3, for each pixel, N F The number of ultrasound contrast agents detected in the corresponding pixels of the filtered images may be used.

[0082] A ULM velocity image of the ultrasound contrast agent velocity may be calculated based on the detected ultrasound contrast agent, e.g., a velocity map may be calculated based on the frame-to-frame ultrasound contrast agent velocity vector components and / or absolute velocity magnitude.

[0083] Slow motion drift may occur in ULM images due to recording periods that are sometimes longer than the periods of cardiac and respiratory cycles. Slow motion drift may also result from anesthesia in anesthetized humans or animals. When a craniotomy is performed, potential brain swelling of the brain may contribute to slow motion drift. Correction of slow motion drift may be performed via intensity-based spatial registration. The intensity-based spatial registration may be one of a translation transformation, a translation and rotation transformation, or a more complex non-rigid transformation. The intensity-based spatial registration may be performed based on ULM images of a period shorter than the recording period. For example, a 10-second ULM count image may be used to correct drift that occurs in ULM images of a longer recording period. Correction of slow motion drift may allow observation of dynamic events of interest of the present disclosure, which may be due to causes other than cardiac beats and may be slower than cardiac beats (e.g., the dynamic events may correspond to neural activity and / or inflammatory responses).

[0084] FIG. 4 shows how a time series of ULM images of a region can be created.

[0085] As known in the art, N F Instead of calculating a unique ULM count image or a unique velocity image using the entire acquisition data set, which is a single filtered image, a time series of consecutive ULM images of the region may be derived from the acquisition data set. The recording period of the acquisition data set includes at least one dynamic event. The dynamic event activates a vascular network in at least one region being imaged. Successive composite images N transmitted from the region may be derived from the entire acquisition data set. C From N F Each filtered image is constructed. The entire acquisition data set forms a time series of filtered images. Each filtered image is associated with a time position within the recording period. The time series of filtered images can be used to form a time series of stacks of filtered images. Each ULM image of successive ULM images in the time series is associated with a time position within the recording period. Each ULM image can be constructed from the stacks of filtered images of the time series of images, where at least some of the images of the stack of images corresponding to the (k+1)th ULM image have a time position greater than the time position of the image of the stack of images corresponding to the kth ULM image immediately preceding the (k+1)th ULM image.

[0086] In one embodiment, the kth ULM image and the (k+1)th ULM image are immediately adjacent in the time dimension. The respective stacks of images are immediately adjacent in the time dimension. In other words, all images of the stack of images corresponding to the (k+1)th ULM image have time positions that are greater than the time positions of the images of the stack of images corresponding to the kth ULM image.

[0087] In another example shown in FIG. 4, successive ULM images in a time series are t Contains N ULM images. tEach ULM image is associated with a time position, i.e., 1≦k≦N t The k-th ULM image in a series of consecutive ULM images is k The kth image in a series of consecutive ULM images is hereafter referred to as M(t k ) and label it as M(t k ) is N F A set S containing the filtered images F A subset of S Fk The subset S Fk is N F Of the N filtered images, Fk may contain N filtered images, Fk <N F For example, the subset S Fk As shown in FIG. 4, the time position is W and the time is t k Set S contained within a time window centered on F Further, 2≦k≦N t As two consecutive time positions t k-1 and k The time step between W and t may be constant. In this case, successive ULM images of a time series of ULM images may be obtained from a stack of images that at least partially overlap in the time dimension depending on the time step between two successive time positions. k -t k-1 In the case of M(t k-1 The final filtered image used to calculate (t k-1 +t k ) / 2 k-1 M(t k The first filtered image used to calculate (t k-1 +t k ) / 2(≦t k-1 ) less than or equal to the time position t k Therefore, M(t k ) and M(t k-1) can be calculated from a stack of images that at least partially overlap in the time dimension. In other words, in the example of FIG. 4, a time series of successive ULM images is acquired based on a time sliding window of filtered images. In this case, the number of ULM images N t can be described by [Math.1], where Acq t is the acquisition duration of the entire data set, and Step t is the sliding step of the time sliding window, i.e., Step t =t k -t k-1 The ULM image M(t k The length of the time sliding window W, which may correspond to a recording period of 100 Hz, may be longer than the duration of a cardiac cycle and shorter than 1 minute. For example, the length of the time window W may be comprised between 1 second and 20 seconds, e.g., between 1 second and 5 seconds. The sliding step of the time sliding window t can be comprised between 500 milliseconds and 1 second. For example, a time window W of 5 seconds and a sliding step Step of 1 second t A sliding window having

[0088]

number

[0089] Therefore, 1 ≤ k ≤ N t As the ULM image M(t k ) may have a value corresponding to a kinetic parameter, which may be the number of ultrasound contrast agents detected at the pixel during the time window W, or the average velocity of ultrasound contrast agents detected at the pixel during the time window W.

[0090] Based on the time series of ULM images, a measure of the evolution of the value of at least one vascular dynamic parameter in response to a dynamic event may be calculated.

[0091] A 3D ULM time matrix, commonly denoted as M(x,z,t), can be calculated based on a time series of consecutive ULM images. The 3D ULM time matrix has dimensions (N x ,N z ,N t ), where (N x ,N z ) may correspond to the spatial dimension of each ULM image. k ) and MB(t k Note that, k,th ULM image, k,th,i,∈K ...

[0092] The at least one vascular dynamic parameter may be selected from the group comprising blood flow, blood flow velocity, blood volume, blood pressure, blood vessel diameter, and any combination thereof.

[0093] The first type of 3D ULM time matrix M(x,z,t) may be a 3D count matrix, specifically denoted as MB(x,z,t), which may be calculated based on successive ULM count images in a time series. Therefore, 1≦x≦N x , 1≦x≦N z , and 1 ≤ k ≤ N t For each pixel MB(x,z,t k ) may have a value corresponding to the number of ultrasound contrast agents detected at pixel (x,z) during the time window W.

[0094] The second type of 3D ULM time matrix M(x,z,t) is specifically MB F It can be a 3D flow matrix, denoted as (x,z,t). The 3D flow matrix MB F (x,z,t) may correspond to a 3D count matrix divided by the length of the sliding window to obtain the ultrasound contrast agent flux value, i.e.

[0095]

number

[0096] Based on the 3D flow matrix, the evolution of the flow rate of the ULM contrast agent during the time series recording can be measured. Since the flow rate of the ULM contrast agent is driven by the blood flow rate, the blood flow rate can be derived. In particular, the flow rate of the ULM contrast agent can be equal to the blood flow rate.

[0097] The third type of 3D ULM time matrix M(x,z,t) can be specifically a 3D velocity matrix denoted as V(x,z,t), which can be calculated based on a time series of consecutive ULM velocity images. Therefore, 1≦x≦N x , 1≦x≦N z , and 1 ≤ k ≤ N t For each pixel V(x,z,t k ) may have a value corresponding to the average velocity of the ultrasound contrast agent detected at pixel (x,z) during the time window W. Based on the 3D velocity matrix, the evolution of the velocity of the ULM contrast agent during the recording of the time series may be measured. Since the velocity of the ULM contrast agent is driven by the blood flow velocity, the blood flow velocity may be derived. In particular, the velocity of the ULM contrast agent may be equal to the blood flow velocity.

[0098] Similarly, a 3D blood volume matrix can be calculated, based on which the evolution of the blood volume during the time series of recordings can be measured, and a 3D blood pressure matrix can also be calculated, based on which the evolution of the blood pressure during the time series of recordings can be measured.

[0099] The 3D ULM time matrix M(x,z,t) provides a 3D measure of the evolution of at least one vascular dynamic parameter in response to a dynamic event. Additionally, the 3D ULM time matrix M(x,z,t) allows for the calculation of two-dimensional (2D) and one-dimensional (1D) measures of the evolution of at least one vascular dynamic parameter in response to a dynamic event. For example, for a given pixel (x p ,y p ), the time function t→M(x p ,z p ,t) is (x p ,y p ) may provide a measure of the evolution over time of the number of ultrasound contrast agents passing through.

[0100] In the example of FIG. 4, the measure of the evolution of the vascular dynamics parameter may be the temporal ULM response P(t) in the region. The temporal ULM response P(t) may be calculated based on the 3D ULM time matrix M(x,z,t). k ) at each time position t k For a region, the pixel values ​​M(x,z,t) along the two spatial dimensions of the region are expressed as k ).

[0101]

number

[0102] Alternatively, the temporal ULM response P(t k ) at each time position t k For a region, the pixel values ​​M(x,z,t) along the two spatial dimensions of the region are expressed as k ) may be an average of

[0103]

number

[0104] In either case, the temporal ULM response P(t) can be normalized by a baseline value to obtain the relative variation, which can be the average value of the kinetic parameter when no kinetic events occur.

[0105] In some embodiments, the calculation of the measure of evolution of at least one vascular dynamic parameter in response to a dynamic event comprises a temporal sliding average of the time series of ULM images. Advantageously, calculating the measure of evolution of at least one vascular dynamic parameter in response to a dynamic event using a temporal sliding average of the time series of ULM images does not require knowledge of the temporal pattern A(t) of the dynamic event that can be used to analyze epileptic seizures, spreading depression, or spontaneous activity.

[0106] In some embodiments, the measure of the evolution of at least one vascular dynamic parameter in response to a dynamic event can be an activation map. The activation map, denoted AM(x,z), can be calculated using correlation analysis if the time pattern A(t) of the dynamic event is known, or using SVD as described. Other process analyses, such as PCA, can also be used by adapting the described methods to said process analysis.

[0107] For 1≦x≦Nx, 1≦x≦Nz, the value of a pixel in the activation map AM(x,z) may be equal to the Pearson product-moment correlation coefficient c(x,z) between the temporal pattern of the dynamic event, denoted as A(t), and the value of M(x,z,t), as described in [Math.4].

[0108]

number

[0109] Alternatively, SVD may be applied to the 3D ULM time matrix M(x,z,t) to obtain a measure of the evolution of at least one vascular dynamics parameter associated with the 3D ULM time matrix.

[0110] More specifically, a size of (Nx*Nz,Nt)

[0111]

number

[0112] SVD may be applied to the reshaped ULM Casorati matrix, denoted as ∑ m = ...

[0113]

number

[0114] λ i are the singular values ​​of U i is λ i is a spatial singular vector associated with t *N t )V i is λ i is the time singular vector associated with

[0115] Space singular vector U i and the time singular vector V i is the covariance matrix

[0116]

number

[0117] and

[0118]

number

[0119] This decomposition corresponds to the eigenvectors of the time signal V i Images modulated independently by i (corresponding to the image U i Every pixel from V i That is, each spatial singular vector U i is the time singular vector V i Furthermore, we can transform the spatial response of a region to a dynamic event described by the spatial activation singular vector U i,a is the time activation singular vector V i,a corresponds to the regional activation for a dynamic event modeled by i,a We determined that can be considered as a measure of the evolution of at least one vascular dynamic parameter in response to a dynamic event.

[0120] For a known temporal pattern of dynamic events A(t), we define the corresponding temporal activation singular vector V i,a can be determined. For each time singular vector V i For the time pattern A(t) of dynamic events and the time singular vector V i The scalar product between p i can be calculated as described in [Math.6].

[0121]

number

[0122] Temporal activation singular vector V i,a can be determined as the singular vector with the highest value of the scalar product, i.e., V i,a can be described by [Math.7] (equation 11).

[0123]

number

[0124] In some embodiments, M AOIA 3D ULM time matrix of at least one area of ​​interest (AOI), denoted (x,z,t), contained within at least one region, may be generated by extracting pixels contained within the AOI from the 3D time matrix M(x,z,t). The AOI may be equal to the region. Alternatively, the AOI may be smaller than the region. The AOI may be determined manually or automatically. For example, the AOI may be determined automatically based on the values ​​of the 3D time matrix M(x,z,t). The AOI may be an area of ​​the region where the values ​​M(x,z,t) show high temporal variability in response to dynamic events compared to other areas of the region. In particular, the AOI may be selected among the group of arterioles, venules, intraparenchymal vessels, and pial vessels. In case of spontaneous activity of the nervous system, the AOI may be selected depending on the functional connectivity to be studied. In particular, the AOI may be the whole brain slice (in case of 2D imaging) or the whole brain (in case of 3D imaging). In case of dynamic events corresponding to a stimulus, the AOI may be selected depending on the stimulus. For example, for whisker stimulation in rodents, the AOI may be the barrel field (S1BF) and / or the ventral-posterior-median (VPM) thalamic nucleus. For visual stimulation, the AOI may be the superior colliculus (SC). For each AOI, the corresponding 3D ULM time matrix of the AOI is AOIx ,N AOIy ,N t ), where (N AOIx ,N AOIy ) corresponds to the spatial dimension of the AOI, where N AOIx is N AOIy A measure of the evolution of the at least one vascular dynamics parameter in at least one area of ​​interest (AOI) of the region may be calculated by adapting the method described above for measuring the evolution of the at least one vascular dynamics parameter in the region.

[0125] In the example of Fig. 5, the stimulation pattern may be repeated periodically during the recording of the time series of ULM images. The number of trials N may be comprised between 2 and 100, for example between 5 and 30, and in particular 20.

[0126] The stimulation pattern may include an initial period without stimulation (i.e., a baseline period), for example, 30 seconds, and a final period without stimulation, for example, 10 seconds, in the example of FIG. 4. The stimulation pattern may include a stimulation period with stimulation, for example, 30 seconds, in the example of FIG. 4. Other stimulation patterns may also be used, such as a pattern including a stimulation period of 5 seconds. The stimulation may be a task-eliciting stimulation, such as a visual stimulation. In rodents, the stimulation may be a whisker stimulation.

[0127] As shown in Figure 4, equivalent time points of N trials can be combined to obtain an average pattern. In other words, ULM images corresponding to equivalent images can be summed. Spatial registration can be used to correct for possible motion drift. In this case, the 3D pattern average time matrix is ​​N s t where N s t is the number of time points in the pattern.

[0128]

number

[0129] It is.

[0130] A measure of the evolution of at least one vascular dynamic parameter in response to said dynamic event may be applied.

[0131] In particular, we developed an activation map AM that quantifies the variation of ultrasound contrast agents during stimulation patterns based on SVD. SVD (x,z) can be calculated as shown in [Math.8] (equation 13).

[0132]

number

[0133] Furthermore, the relative increase map AM R SVD(x,z) can be computed as described in [Math.9].

[0134]

number

[0135] Alternatively, the N trials are not averaged, in which case the 3D time matrix is ​​N*N s t may have a time dimension of

[0136] In particular, we developed activation maps of non-averaged data that quantify the variability of ultrasound contrast agents during stimulation patterns based on SVD. SVD (x,z) can be calculated as shown in [Math.10].

[0137]

number

[0138] As shown in FIG. 6(a), the method of the present disclosure based on the Pearson correlation between the pattern-averaged temporal ULM matrix and the stimulation paradigm can obtain maps of functional hyperemia in cortical and subcortical regions at a resolution of 6.5 μm. In particular, in the example of FIG. 6(a), vascular activity in the pial vessel β and the smaller vessel σ, which is the primary branch after descending arteriole, can be obtained. A functional ultrasound (fUS) imaging experiment shown in FIG. 6(b) confirms brain activation in the same region. In this example, the spatial resolution obtained with the method of the present disclosure, which may be named functional ultrasound localization microscopy (fULM), is 16 times better than the spatial resolution achieved in the fUS imaging experiment.

[0139] Thanks to the microscopic resolution, the disclosed functional ultrasound localization microscopy can be used to measure dynamic parameters such as microbubble flow rate, velocity, and vessel diameter in different vascular compartments. Figure 7A shows the segmentation of the barrel cortex (S1BF) into four vascular compartments, namely perforating arterioles, venules, pial vessels, and intraparenchymal vessels, based on the functional ULM map. Figure 7B shows the dynamic histogram of microbubble velocity, as well as the time course of microbubble flow rate and velocity. These results revealed the highest relative increase in microbubble flow rate during activation in intraparenchymal vessels compared to larger vessels such as perforating arterioles or pial arterioles, confirming that these intraparenchymal vessels are the most important contributors to neurovascular coupling. Thus, functional ULM images allow quantification of the contributions of different vascular compartments.

[0140] By carrying out the method of the present disclosure, the computing module 3 of the device 1 of Fig. 1 may be adapted to estimate a seizure focus based on a measure of the evolution of at least one vascular dynamic parameter in response to a dynamic event obtained by the method of the present disclosure. For example, based on values ​​of the flow and / or velocity evolution associated with pixels of areas during epileptic seizures, spreading depression, respectively, the seizure focus, which is the source of activity, the origin of the spreading depression, respectively, may be localized.

[0141] Additionally, the computing module 3 may be adapted to diagnose whether a measure of the evolution of at least one vascular dynamic parameter in response to a dynamic event corresponds to a given disease, in particular a neurodegenerative disease.

[0142] To determine whether a measure of the evolution of at least one vascular dynamic parameter in response to a dynamic event is normal and / or whether a measure of the evolution of at least one vascular dynamic parameter in response to a dynamic event corresponds to a predetermined disease, the computing module 3 may compare said parameters with predetermined thresholds.

[0143] In a variant, to determine whether the measure of evolution of at least one vascular dynamic parameter in response to a dynamic event is normal and / or whether the measure of evolution of at least one vascular dynamic parameter in response to a dynamic event corresponds to a predetermined disease, the computing module 3 may use a neural network trained to determine whether the measure of evolution of at least one vascular dynamic parameter in response to a dynamic event is normal and / or whether the measure of evolution of at least one vascular dynamic parameter in response to a dynamic event corresponds to a predetermined disease.

[0144] The device 1 may also be used to monitor the efficiency of a medical treatment for a given disease, in particular for diseases among epilepsy, spreading depression and neurodegenerative diseases, based on a measure of the evolution of at least one vascular dynamic parameter in response to a dynamic event. For this purpose, a measure of the evolution of at least one vascular dynamic parameter in response to a dynamic event may be measured at different time points, including at least before and after treatment, and possibly during the medical treatment, in order to determine whether the medical treatment reduces seizures and / or improves neurovascular coupling.

[0145] In this example, the vascular network described is the vascular network of the nervous system, nevertheless, the methods and devices described may be applied to any vascular network. [Explanation of symbols]

[0146] 1 device 2 processors 3 Computing Module, Computing Device 4. Ultrasound contrast agent injection device, ultrasound agent injection device 5. Probe 6. Motorization 7. Array

Claims

1. A method for dynamic imaging of vascular activity on a microscopic scale in a human or animal vascular network, preferably in the vascular network of the nervous system, comprising: (a) performing time series Ultrasound Localization Microscopy (ULM) imaging of the at least one region of the vascular network to obtain values ​​of at least one vascular dynamic parameter in at least one region of interest within the at least one region during a recording period, the time series ULM images being constructed from a time series of images of the region, each ULM image being constructed from a stack of images of the time series of images, at least some of the images of the stack of images corresponding to a (k+1)th ULM image having a time position greater than a time position of an image of a stack of images corresponding to a kth ULM image immediately preceding the (k+1)th ULM image, the recording period of the time series of ULM images including at least one dynamic event due to a cause other than cardiac beats, the dynamic event activating a change in local hemodynamics and / or a change in structural conformation of the vascular network in the at least one region; (b) calculating a measure of evolution of the at least one vascular dynamic parameter in response to the dynamic event based on the values ​​of the at least one vascular dynamic parameter from the time series of ULM images. method.

2. The method of claim 1 , wherein each image in the time series of images is constructed from successive composite images.

3. 3. The method according to claim 1 or 2, wherein the recording period of the ULM images is longer than the duration of a cardiac cycle and shorter than 1 minute.

4. the vascular network is a vascular network of a nervous system, the at least one dynamic event is a stimulus delivered to the nervous system, and the method comprises: (c) delivering the at least one stimulus to the nervous system.

4. The method of claim 1, further comprising:

5. 4. The method of claim 1, wherein the vascular network is a vascular network of the nervous system and the at least one dynamic event is a spontaneous activity of the nervous system.

6. 5. The method of claim 4, wherein steps (a) and (c) are repeated for N trials, and the measure of the evolution of the at least one vascular dynamic parameter is calculated based on a time series of the N trials, where N is an integer greater than 1.

7. 7. The method of any one of claims 4 to 6, wherein the calculation of the measure of the evolution of the at least one vascular dynamic parameter in response to the dynamic event comprises a correlation between the at least one stimulus and the time series of ULM images.

8. 8. The method of any one of claims 1 to 7, wherein the calculation of the measure of the evolution of the at least one vascular dynamic parameter in response to the dynamic event comprises singular value decomposition (SVD) and / or principal component analysis (PCA) of the time series of ULM images.

9. 9. The method of claim 1, wherein the calculation of the measure of the evolution of the at least one vascular dynamic parameter in response to the dynamic event comprises a sliding average of the time series of ULM images.

10. 10. The method of claim 1, wherein successive ULM images of the time series of ULM images are obtained from a stack of images that at least partially overlap in time.

11. 11. The method according to any one of claims 1 to 10, wherein the at least one vascular dynamic parameter is selected from the group comprising blood flow, blood flow velocity, blood volume, blood pressure, blood vessel diameter, and any combination thereof.

12. 12. The method of claim 1, wherein the region is segmented into at least one vascular compartment.

13. The method of claim 12 , wherein the at least one region of interest of the region corresponds to one of the vascular compartments.

14. 14. The method of claim 1, wherein successive ULM images of the time series of ULM images are obtained from respective stacks of successive images in the time dimension.

15. The method of claim 14 , wherein each stack of images is immediately adjacent in the time dimension.

16. 1. An apparatus for dynamic imaging of vascular activity on a microscopic scale in a human or animal vascular network, preferably in the vascular network of the nervous system, comprising: (a) an ultrasound contrast agent injection device and an ultrasound measurement device adapted to perform time series Ultrasound Localization Microscopy (ULM) images of at least one region of the vascular network to obtain values ​​of at least one vascular dynamic parameter in at least one region of interest within the at least one region during a recording period, wherein the time series ULM images are constructed from a time series of images of the region, each ULM image being constructed from a stack of images of the time series of images, at least some of the images of the stack of images corresponding to a (k+1)th ULM image having a time position greater than a time position of an image of a stack of images corresponding to a kth ULM image immediately preceding the (k+1)th ULM image, and wherein the recording period of the time series of ULM images includes at least one dynamic event due to a cause other than cardiac beats, the dynamic event activating a change in local hemodynamics and / or a change in structural conformation of the vascular network in the at least one region; (b) a computing module adapted to calculate a measure of evolution of the at least one vascular dynamic parameter in response to the dynamic event based on the values ​​of the at least one vascular dynamic parameter from the time series of ULM images.

17. 17. The apparatus of claim 16, wherein the dynamic event is an epileptic seizure or a spreading depression, and the computing module is adapted to estimate a location of a seizure focus or an origin of a spreading depression from the measure of the evolution of the at least one vascular dynamic parameter in response to the epileptic seizure or the spreading depression.

18. 18. The apparatus according to claim 16 or 17, wherein the computing module is adapted to diagnose whether the measure of the evolution of the at least one vascular dynamic parameter in response to the dynamic event corresponds to a given disease, in particular a neurodegenerative disease.

19. 19. The apparatus of claim 16, wherein the computing device is adapted to monitor the efficiency of a medical treatment for a predetermined disease, in particular one of epilepsy, spreading depression and neurodegenerative diseases, based on the measure of the evolution of the at least one vascular dynamic parameter in response to the dynamic event.

20. Computer software comprising instructions for carrying out the method of any one of claims 1 to 15 when the software is executed by a processor.

21. 16. A computer readable, non-transitory storage medium having registered thereon software for performing the method of any one of claims 1 to 15 when the software is executed by a processor.

Images