WSGR Docket No. 53242—701301
`
`METHODS AND SYSTEMS FOR IDENTIFYING BRAIN DISORDERS
`
`CROSS-REFERENCE
`
`[0001]
`
`This application claims the benefit of US. Provisional Patent Application Serial
`
`Number 62/430,351, filed December 6, 2016, and US. Provisional Patent Application Serial
`
`Number 62/481,839, filed April 5, 2017, each of which is entirely incorporated herein by
`
`reference for all purposes.
`
`BACKGROUND
`
`[0002]
`
`Neurodegenerative diseases leading to dementia are a tremendous societal burden,
`
`currently devastating 9 million people domestically and 47 million people worldwide. Current
`
`inability to effectively prevent, diagnose and combat neurodegeneration results in staggering
`
`direct and indirect costs. Alzheimer's disease (AD), the most common cause of dementia, alone
`
`afflicts over 5 million Americans and accounts for the 6th leading cause of death in the USA. AD
`
`requires an estimated 18 billion hours of unpaid caretaking and well over $250 billion of medical
`
`costs annually. Prevalence of the disease is projected to escalate to nearly 14 million people
`
`domestically and 135 million worldwide by 2050, with no potential cure in immediate sight.
`
`There is a dire need for technological advancements toward diagnostics, prevention, therapeutics
`
`and eventual cures that will each have profound beneficial impacts on the population.
`
`[0003]
`
`Current clinical evaluation typically includes non-invasive brain imaging with
`
`magnetic resonance imaging (MRI), positron emission tomography (PET), or other advanced
`
`imaging strategies which provide insight into tissue volume changes, chemical composition,
`
`cortical metabolic rate, alterations associated with tissue cellularity and disease biomarkers, and
`
`structural abnormalities attributed to neurodegenerative disease. To aid in the diagnosis of AD
`
`and differential diagnosis from non-Alzheimer dementias, fluorodeoxyglucose (FDG) PET and
`
`amyloid PET reveal AD-associated patterns of cerebral cortical metabolism and beta amyloid
`
`deposits in the gray matter, respectively. Similarly, tau PET reveals neurofibrillary tangles in the
`
`brain. However, due to a lack of advancement in analysis technologies, meaningful use of these
`
`imaging techniques for neurodegenerative disease is restricted to late stages when considerable
`
`tissue damage and cognitive or other clinical abnormalities are present. As we deepen our
`
`understanding of the multiplicity of abnormalities associated with AD, there is increasing
`
`evidence that the continual targeting of these amyloid plaques and neurofibrillary tangles may
`
`merely be treating late stage symptoms rather than the underlying causes. The inability to
`
`

`

`WSGR Docket No. 53242—701301
`
`effectively detect early stages of AD precludes pre-symptomatic intervention and conceals the
`
`potential beneficial effects of drug candidates.
`
`[0004]
`
`Implicit to the neurodegenerative process is the death of the signaling nerve cells in
`
`the brain, though this can merely be the ultimate consequence in a cascade of degeneration within
`
`brain tissue. The structural integrity of tissue is necessary for neuron support and survival and
`
`clearance of molecular waste that must be removed from the brain for maintenance of neural
`
`tissue homeostasis and efficient function. Alterations in non-cellular components of the brain are
`
`complicit in the degenerative process and may be a precursor of lost nerve cell function. It has
`
`been shown that proper regulation of neural tissue homeostasis is necessary for eliminating toxic
`
`residue buildup, a process that can be altered in the AD brain. Yet, there remains limited
`
`understanding of brain structural content and its impact on transport of molecules in the brain
`
`interstitium. Currently, the clinical use and the diagnostic capacity of brain MRI remains limited
`
`to differential diagnosis, only after symptomatic presentation, principally due to the inherently
`
`low spatial resolution — MRI image voxels are in mm dimensions, whereas structural changes
`
`contributing to tissue degeneration originate at the sub-micron scale. FDG, amyloid, and tau PET
`
`scans suffer from similar limitations.
`
`SUMMARY
`
`[0005]
`
`Recognized herein is a need for tools that allow early detection of Alzheimer’s
`
`disease and other neurodegenerative disorders, including tools that may utilize approaches that
`
`detect microscopic changes in brain tissue from low-resolution magnetic resonance imaging
`
`(MRI) scans. Such approaches may leverage a deeper understanding of brain tissue
`
`microstructure to more reliably predict and interpret the health of the brain from MRI scans well
`
`before severe tissue damage irreversibly impedes healthy cognitive function.
`
`[0006]
`
`Provided herein is an image analysis platform that can detect and quantify brain tissue
`
`abnormality (such as neurodegeneration) in every voxel of standard clinical brain MRI. The
`
`platform may provide detailed information about the brain tissue health at the microscopic level
`
`and the resulting observed patterns of pathologic involvement that is currently missing in the
`
`neuroimaging/ brain diagnostics field. As a result, complex brain diseases, such as Alzheimer's
`
`disease, which are currently diagnosed very late (i.e., at the late symptomatic stages), may be
`
`diagnosed or otherwise identified prior to the onset of advanced symptoms. The platform may
`
`allow early-stage testing of novel drug candidates for clinical trials, which have previously failed
`
`due to poor patient selection, late intervention, and very high trial costs. All of these factors may
`
`be significantly improved using the platform.
`
`

`

`WSGR Docket No. 53242—701301
`
`[0007]
`
`Provided herein are methods and systems for determining whether brain tissue is
`
`indicative of a disorder, such as a neurodegenerative disorder. The methods and systems may
`
`allow the early diagnosis of a brain disorder much earlier than would be possible using prior
`
`methods and systems, such as many years before the development of symptoms associated with
`
`the disorder that are detectable using prior methods and systems. The methods and systems may
`
`provide high accuracy in diagnosing a brain disorder (such as greater than 90% accuracy), as
`
`measured by a variety of criteria described herein.
`
`[0008]
`
`The methods and systems of the present disclosure may utilize data processing
`
`techniques to assess a level of congruence between measured parameters obtained from magnetic
`
`resonance imaging (MRI) data and simulated parameters obtained from computational modeling
`
`of brain tissues. The methods and systems generally operate by determining a level of
`
`congruence between the one or more measured parameters and the one or more simulated
`
`parameters for one or more voxels of the MRI data. The simulated parameters are obtained from
`
`a plurality of microstructural models. Each microstructural model of the plurality of
`
`microstructural models is obtained by subjecting a microstructural model that is not indicative of
`
`a disorder to a series of microstructural perturbations. After assessing the level of congruence
`
`between the one or more measured parameters and the one or more simulated parameters for a
`
`number of microstructural models of the plurality of microstructural models, a diagnostic
`
`microstructural model that meets a threshold congruence is selected. The diagnostic
`
`microstructural model is used to determine the disorder state of the brain tissue associated with
`
`the voxel.
`
`[0009]
`
`The methods and systems may be applied to a plurality of voxels of the MRI data,
`
`such that a level of congruence is determined for each voxel of the plurality of voxels. In this
`
`manner, a diagnostic model and a disorder state may be determined for each voxel of the
`
`plurality of voxels. The methods and systems may be applied to determine a diagnostic model
`
`and a disorder state for a plurality of voxels located within a particular region of a brain, within a
`
`whole brain, or across a plurality of brains from a plurality of subjects.
`
`[0010]
`
`In an aspect, a method for determining a disorder state of brain tissue in a brain of a
`
`subject may comprise: (a) obtaining magnetic resonance imaging (MRI) data comprising at least
`
`one MRI image of the brain, the MRI image comprising a plurality of voxels, a voxel of the
`
`plurality of voxels being associated with the brain tissue of the brain of the subject and
`
`comprising one or more measured MRI parameters in the MRI data; (b) for the voxel of the
`
`plurality of voxels, using one or more computer processors to process the one or more measured
`
`MRI parameters with one or more simulated MRI parameters for the voxel, the one or more
`
`

`

`WSGR Docket No. 53242—701301
`
`simulated MRI parameters being generated from one or more microstructural models at the
`
`voxel; (c) for the voxel of the plurality of voxels, selecting a diagnostic model from the one or
`
`more microstructural models, the diagnostic model meeting a threshold congruence between the
`
`one or more measured MRI parameters and the one or more simulated MRI parameters
`
`associated with the diagnostic model; and (d) for the voxel of the plurality of voxels, using the
`
`diagnostic model to determine the disorder state of the brain tissue associated with the voxel.
`
`[0011]
`
`Each voxel may comprise a plurality of measured MRI parameters. The one or more
`
`measured MRI parameters may be a plurality of measured MRI parameters. The one or more
`
`simulated MRI parameter may be a plurality of simulated MRI parameters.
`
`[0012]
`
`The method may further comprise repeating (b)-(d) one or more times for additional
`
`voxels of the plurality of voxels. The method may further comprise repeating (b)-(d) for all other
`
`voxels of the plurality of voxels. The method may further comprise repeating (b)-(d) for all
`
`voxels associated with a specified region of the brain. The method may further comprise
`
`repeating (b)-(d) for all voxels associated with an entirety of the brain. The method may further
`
`comprise repeating (a)-(d) for a plurality of MRI images, each MRI image of the plurality of
`
`MRI images associated with a brain selected from a plurality of brains, each brain of the plurality
`
`of brains associated with a subject selected from a plurality of subjects.
`
`[0013]
`
`The MRI image may be selected from the group consisting of: a longitudinal
`
`relaxation time (Tl)-weighted MRI image, a transverse relaxation time (T2)-weighted MRI
`
`image, and a diffusion-weighted MRI image. The measured MRI parameter may be selected
`
`from the group consisting of: a longitudinal relaxation time (T1), a transverse relaxation time
`
`(T2), and a diffusion coefficient. The simulated MRI parameter may be selected from the group
`
`consisting of: a longitudinal relaxation time (T1), a transverse relaxation time (T2), and a
`
`diffusion coefficient.
`
`[0014]
`
`The one or more microstructural models may comprise information regarding a
`
`parameter selected from the group consisting of: intracellular content, extracellular content,
`
`distribution of extracellular content within interstitial space, distribution of intracellular content
`
`within intracellular space, and tissue geometry. The one or more microstructural models may
`
`comprise measured or predicted values of a parameter selected from the group consisting of: cell
`
`density, cell shape, cell geometry, cell size, cell distribution, intercellular spacing, extracellular
`
`matrix homogeneity, interstitial tortuosity, water to protein ratio, water to lipid ratio, water to
`
`carbohydrate ratio, protein to lipid ratio, protein to carbohydrate ratio, and lipid to carbohydrate
`
`ratio. The one or more microstructural models may be selected from a microstructural model
`
`library. The microstructural model library may comprise at least 100 microstructural models.
`
`4
`
`

`

`WSGR Docket No. 53242—701301
`
`[0015]
`
`The microstructural model library may be constructed by: (a) creating a first
`
`microstructural model corresponding to a brain state that is not associated with a disorder; and
`
`(b) iteratively subjecting the first microstructural model to a perturbation, each iteration
`
`producing an additional perturbed microstructural model. (b) may comprise subjecting the first
`
`microstructural model to at least 100 iterations to generate at least 100 perturbed microstructural
`
`models. The first microstructural model may be selected based on knowledge of the brain region
`
`associated with the voxel. The perturbation may comprise an operation selected from the group
`
`consisting of: depleting cells, altering cellular morphology or distribution, altering intracellular or
`
`interstitial physico-chemical composition or distribution, altering extracellular matrix
`
`composition or distribution, and altering intercellular spacing. The perturbation may comprise a
`
`stochastic procedure.
`
`[0016]
`
`The threshold congruence may be determined by computing an objective function
`
`between the one or more measured MRI parameters and the one or more simulated MRI
`
`parameters. The objective function may comprise an L1 norm or an L2 norm.
`
`[0017]
`
`Determining the disorder state of the brain tissue associated with the voxel may be
`
`achieved at an accuracy of at least 90%. Determining the disorder state across the brain tissue
`
`associated with the specified region of the brain may be achieved at an accuracy of at least 90%.
`
`Determining the disorder state of the brain tissue associated with the whole brain of the subject
`
`may be achieved at an accuracy of at least 90%. Determining the disorder state of the brain tissue
`
`associated with the plurality of subjects may be achieved at an accuracy of at least 90%.
`
`[0018]
`
`The disorder may be a non-neurodegenerative disorder. The disorder may be selected
`
`from the group consisting of: a primary neoplasm, a metastatic neoplasm, a seizure disorder, a
`
`seizure disorder with focal cortical dysplasia, a demyelinating disorder, a non-neurodegenerative
`
`encephalopathy, a cerebrovascular disease, and a psychological disorder. The disorder may be a
`
`neurodegenerative disorder. The disorder may be selected from the group consisting of:
`
`Alzheimer’s disease, a non-Alzheimer’s dementia disorder, Parkinson’s disease, a Parkinsonism
`
`disorder, a motor neuron disease, Huntington’s disease, a Huntington’s disease-like syndrome,
`
`transmissible spongiform encephalopathy, chronic traumatic encephalopathy, and a tauopathy.
`
`[0019]
`
`The method may enable diagnosis of a neurodegenerative disorder more than 5 years
`
`prior to the development of symptoms associated with the neurodegenerative disorder. The
`
`method may enable monitoring of the neurodegenerative disorder at a plurality of time points, the
`
`plurality of time points separated by a plurality of time intervals.
`
`[0020]
`
`The method may further comprise constructing a brain map that, for each voxel of the
`
`plurality of voxels, indicates the disorder state of the brain tissue associated with the voxel. The
`
`

`

`WSGR Docket No. 53242—701301
`
`method may further comprise displaying the brain map on a graphical user interface of an
`
`electronic device of a user. The brain map may comprise a qualitative abnormality map. The
`
`brain map may comprise a binary abnormality map. The brain map may comprise a quantitative
`
`abnormality map. The brain map may comprise a percent abnormality map.
`
`[0021]
`
`In an aspect, a method for determining a disorder state of a tissue in a portion of a
`
`body of a subject may comprise: obtaining magnetic resonance imaging (MRI) data comprising
`
`at least one MRI image of the tissue, the MRI image comprising a plurality of voxels, a voxel of
`
`the plurality of voxels being associated with the tissue of the subject and comprising one or more
`
`measured MRI parameters in the MRI data; (b) for the voxel of the plurality of voxels, using one
`
`or more computer processors to process the one or more measured MRI parameters with one or
`
`more simulated MRI parameters for the voxel, the one or more simulated MRI parameters being
`
`generated from one or more microstructural models at the voxel; (c) for the voxel of the plurality
`
`of voxels, selecting a diagnostic model from the one or more microstructural models, the
`
`diagnostic model meeting a threshold congruence between the one or more measured MRI
`
`parameters and the one or more simulated MRI parameters associated with the diagnostic model;
`
`and (d) for the voxel of the plurality of voxels, using the diagnostic model to determine the
`
`disorder state of the tissue associated with the voxel.
`
`[0022]
`
`The tissue may be selected from the group consisting of: spinal cord tissue, heart
`
`tissue, vascular tissue, lung tissue, liver tissue, kidney tissue, esophageal tissue, stomach tissue,
`
`intestinal tissue, pancreatic tissue, thyroid tissue, adrenal tissue, spleen tissue, lymphatic tissue,
`
`appendix tissue, breast tissue, bladder tissue, vaginal tissue, ovarian tissue, uterine tissue, penile
`
`tissue, testicular tissue, prostatic tissue, skeletal muscle tissue, skin, and non-brain tissue of the
`
`head and neck.
`
`[0023]
`
`In an aspect, a non-transitory computer-readable medium may comprise machine-
`
`executable code that, upon execution by one or more computer processors, implements a method
`
`for detecting a disorder state of brain tissue in a brain of a subject, the method comprising: (a)
`
`obtaining magnetic resonance imaging (MRI) data comprising at least one MRI image of the
`
`brain, the MRI image comprising a plurality of voxels, a voxel of the plurality of voxels being
`
`associated with the brain tissue of the brain of the subject and comprising one or more measured
`
`MRI parameters in the MRI data; (b) for the voxel of the plurality of voxels, using one or more
`
`computer processors to process the one or more measured MRI parameters with one or more
`
`simulated MRI parameters for the voxel, the one or more simulated MRI parameters being
`
`generated from one or more microstructural models at the voxel; (c) for the voxel of the plurality
`
`of voxels, selecting a diagnostic model from the one or more microstructural models, the
`
`6
`
`

`

`WSGR Docket No. 53242—701301
`
`diagnostic model meeting a threshold congruence between the one or more measured MRI
`
`parameters and the one or more simulated MRI parameters associated with the diagnostic model;
`
`and (d) for the voxel of the plurality of voxels, using the diagnostic model to determine the
`
`disorder state of the brain tissue associated with the voxel.
`
`[0024]
`
`Each voxel may comprise a plurality of measured MRI parameters. The one or more
`
`measured MRI parameters may be a plurality of measured MRI parameters. The one or more
`
`simulated MRI parameter may be a plurality of simulated MRI parameters.
`
`[0025]
`
`The method may further comprise repeating (b)-(d) one or more times for additional
`
`voxels of the plurality of voxels. The method may further comprise repeating (b)-(d) for all other
`
`voxels of the plurality of voxels. The method may further comprise repeating (b)-(d) for all
`
`voxels associated with a specified region of the brain. The method may further comprise
`
`repeating (b)-(d) for all voxels associated with an entirety of the brain. The method may further
`
`comprise repeating (a)-(d) for a plurality of MRI images, each MRI image of the plurality of
`
`MRI images associated with a brain selected from a plurality of brains, each brain of the plurality
`
`of brains associated with a subject selected from a plurality of subjects.
`
`[0026]
`
`The MRI image may be selected from the group consisting of: a longitudinal
`
`relaxation time (Tl)-weighted MRI image, a transverse relaxation time (T2)-weighted MRI
`
`image, and a diffusion-weighted MRI image. The measured MRI parameter may be selected
`
`from the group consisting of: a longitudinal relaxation time (T1), a transverse relaxation time
`
`(T2), and a diffusion coefficient. The simulated MRI parameter may be selected from the group
`
`consisting of: a longitudinal relaxation time (T1), a transverse relaxation time (T2), and a
`
`diffusion coefficient.
`
`[0027]
`
`The one or more microstructural models may comprise information regarding a
`
`parameter selected from the group consisting of: intracellular content, extracellular content,
`
`distribution of extracellular content within interstitial space, distribution of intracellular content
`
`within intracellular space, and tissue geometry. The one or more microstructural models may
`
`comprise measured or predicted values of a parameter selected from the group consisting of: cell
`
`density, cell shape, cell geometry, cell size, cell distribution, intercellular spacing, extracellular
`
`matrix homogeneity, interstitial tortuosity, water to protein ratio, water to lipid ratio, water to
`
`carbohydrate ratio, protein to lipid ratio, protein to carbohydrate ratio, and lipid to carbohydrate
`
`ratio. The one or more microstructural models may be selected from a microstructural model
`
`library. The microstructural model library may comprise at least 100 microstructural models.
`
`[0028]
`
`The microstructural model library may be constructed by: (a) creating a first
`
`microstructural model corresponding to a brain state that is not associated with a disorder; and
`
`7
`
`

`

`WSGR Docket No. 53242—701301
`
`(b) iteratively subjecting the first microstructural model to a perturbation, each iteration
`
`producing an additional perturbed microstructural model. (b) may comprise subjecting the first
`
`microstructural model to at least 100 iterations to generate at least 100 perturbed microstructural
`
`models. The first microstructural model may be selected based on knowledge of the brain region
`
`associated with the voxel. The perturbation may comprise an operation selected from the group
`
`consisting of: depleting cells, altering cellular morphology or distribution, altering intracellular or
`
`interstitial physico-chemical composition or distribution, altering extracellular matrix
`
`composition or distribution, and altering intercellular spacing. The perturbation may comprise a
`
`stochastic procedure.
`
`[0029]
`
`The threshold congruence may be determined by computing an objective function
`
`between the one or more measured MRI parameters and the one or more simulated MRI
`
`parameters. The objective function may comprise an L1 norm or an L2 norm.
`
`[0030]
`
`Determining the disorder state of the brain tissue associated with the voxel may be
`
`achieved at an accuracy of at least 90%. Determining the disorder state across the brain tissue
`
`associated with the specified region of the brain may be achieved at an accuracy of at least 90%.
`
`Determining the disorder state of the brain tissue associated with the whole brain of the subject
`
`may be achieved at an accuracy of at least 90%. Determining the disorder state of the brain tissue
`
`associated the plurality of subjects may be achieved at an accuracy of at least 90%.
`
`[0031]
`
`The disorder may be a non-neurodegenerative disorder. The disorder may be selected
`
`from the group consisting of: a primary neoplasm, a metastatic neoplasm, a seizure disorder, a
`
`seizure disorder with focal cortical dysplasia, a demyelinating disorder, a non-neurodegenerative
`
`encephalopathy, a cerebrovascular disease, and a psychological disorder. The disorder may be a
`
`neurodegenerative disorder. The disorder may be selected from the group consisting of:
`
`Alzheimer’s disease, a non-Alzheimer’s dementia disorder, Parkinson’s disease, a Parkinsonism
`
`disorder, a motor neuron disease, Huntington’s disease, a Huntington’s disease-like syndrome,
`
`transmissible spongiform encephalopathy, chronic traumatic encephalopathy, and a tauopathy.
`
`[0032]
`
`The method may enable diagnosis of a neurodegenerative disorder more than 5 years
`
`prior to the development of symptoms associated with the neurodegenerative disorder. The
`
`method may enable monitoring of the neurodegenerative disorder at a plurality of time points, the
`
`plurality of time points separated by a plurality of time intervals.
`
`[0033]
`
`The method may further comprise constructing a brain map that, for each voxel of the
`
`plurality of voxels, indicates the disorder state of the brain tissue associated with the voxel. The
`
`method may further comprise displaying the brain map on a graphical user interface of an
`
`electronic device of a user. The brain map may comprise a qualitative abnormality map. The
`
`

`

`WSGR Docket No. 53242—701301
`
`brain map may comprise a binary abnormality map. The brain map may comprise a quantitative
`
`abnormality map. The brain map may comprise a percent abnormality map.
`
`[0034]
`
`In an aspect, a non-transitory computer-readable medium may comprise machine-
`
`executable code that, upon execution by one or more computer processors, implements a method
`
`for detecting a disorder state of brain tissue in a brain of a subject, the method comprising:
`
`obtaining magnetic resonance imaging (MRI) data comprising at least one MRI image of the
`
`tissue, the MRI image comprising a plurality of voxels, a voxel of the plurality of voxels being
`
`associated with the tissue of the subject and comprising one or more measured MRI parameters
`
`in the MRI data; (b) for the voxel of the plurality of voxels, using one or more computer
`
`processors to process the one or more measured MRI parameters with one or more simulated
`
`MRI parameters for the voxel, the one or more simulated MRI parameters being generated from
`
`one or more microstructural models at the voxel; (c) for the voxel of the plurality of voxels,
`
`selecting a diagnostic model from the one or more microstructural models, the diagnostic model
`
`meeting a threshold congruence between the one or more measured MRI parameters and the one
`
`or more simulated MRI parameters associated with the diagnostic model; and (d) for the voxel of
`
`the plurality of voxels, using the diagnostic model to determine the disorder state of the tissue
`
`associated with the voxel.
`
`[0035]
`
`The tissue may be selected from the group consisting of: spinal cord tissue, heart
`
`tissue, vascular tissue, lung tissue, liver tissue, kidney tissue, esophageal tissue, stomach tissue,
`
`intestinal tissue, pancreatic tissue, thyroid tissue, adrenal tissue, spleen tissue, lymphatic tissue,
`
`appendix tissue, breast tissue, bladder tissue, vaginal tissue, ovarian tissue, uterine tissue, penile
`
`tissue, testicular tissue, prostatic tissue, skeletal muscle tissue, skin, and non-brain tissue of the
`
`head and neck.
`
`[0036]
`
`In an aspect, a system for determining a disorder state of brain tissue in a brain of a
`
`subject may comprise: (a) a database comprising magnetic resonance imaging (MRI) data
`
`comprising at least one MRI image of the brain, the MRI image comprising a plurality of voxels,
`
`a voxel of the plurality of voxels being associated with the brain tissue of the brain of the subject
`
`and comprising a measured MRI parameter in the MRI data; and (b) one or more computer
`
`processors operatively coupled to the database, wherein the one or more computer processors are
`
`individually or collectively programmed to: (i) for the voxel of the plurality of voxels, use one or
`
`more computer processors to process the one or more measured MRI parameters with one or
`
`more simulated MRI parameters for the voxel, the one or more simulated MRI parameters being
`
`generated from one or more microstructural models at the voxel; (ii) for the voxel of the plurality
`
`of voxels, select a diagnostic model from the one or more microstructural models, the diagnostic
`
`9
`
`

`

`WSGR Docket No. 53242—701301
`
`model meeting a threshold congruence between the one or more measured MRI parameters and
`
`the one or more simulated MRI parameters associated with the diagnostic model; and (iii) for the
`
`voxel of the plurality of voxels, use the diagnostic model to determine the disorder state of the
`
`brain tissue associated with the voxel.
`
`[0037]
`
`Each voxel may comprise a plurality of measured MRI parameters. The one or more
`
`measured MRI parameters may be a plurality of measured MRI parameters. The one or more
`
`simulated MRI parameter may be a plurality of simulated MRI parameters.
`
`[0038]
`
`The one or more computer processors may be further individually or collectively
`
`programmed to repeat (i)-(iii) one or more times for additional voxels of the plurality of voxels.
`
`The one or more computer processors may be further individually or collectively programmed to
`
`repeat (i)-(iii) for all other voxels of the plurality of voxels. The one or more computer processors
`
`may be further individually or collectively programmed to repeat (i)-(iii) for all voxels associated
`
`with a specified region of the brain. The one or more computer processors may be further
`
`individually or collectively programmed to repeat (i)-(iii) for all voxels associated with an
`
`entirety of the brain. The one or more computer processors may be further individually or
`
`collectively programmed to repeat (i)-(iii) for a plurality of MRI images, each MRI image of the
`
`plurality of MRI images associated with a brain selected from a plurality of brains, each brain of
`
`the plurality of brains associated with a subject selected from a plurality of subjects.
`
`[0039]
`
`The MRI image may be selected from the group consisting of: a longitudinal
`
`relaxation time (Tl)-weighted MRI image, a transverse relaxation time (T2)-weighted MRI
`
`image, and a diffusion-weighted MRI image. The measured MRI parameter may be selected
`
`from the group consisting of: a longitudinal relaxation time (T1), a transverse relaxation time
`
`(T2), and a diffusion coefficient. The simulated MRI parameter may be selected from the group
`
`consisting of: a longitudinal relaxation time (T1), a transverse relaxation time (T2), and a
`
`diffusion coefficient.
`
`[0040]
`
`The one or more microstructural models may comprise information regarding a
`
`parameter selected from the group consisting of: intracellular content, extracellular content,
`
`distribution of extracellular content within interstitial space, distribution of intracellular content
`
`within intracellular space, and tissue geometry. The one or more microstructural models may
`
`comprise measured or predicted values of a parameter selected from the group consisting of: cell
`
`density, cell shape, cell geometry, cell size, cell distribution, intercellular spacing, extracellular
`
`matrix homogeneity, interstitial tortuosity, water to protein ratio, water to lipid ratio, water to
`
`carbohydrate ratio, protein to lipid ratio, protein to carbohydrate ratio, and lipid to carbohydrate
`
`10
`
`

`

`WSGR Docket No. 53242—701301
`
`ratio. The one or more microstructural models may be selected from a microstructural model
`
`library. The microstructural model library may comprise at least 100 microstructural models.
`
`[0041]
`
`The microstructural model library may be constructed by: (a) creating a first
`
`microstructural model corresponding to a brain state that is not associated with a disorder; and
`
`(b) iteratively subjecting the first microstructural model to a perturbation, each iteration
`
`producing an additional perturbed microstructural model. (b) may comprise subjecting the first
`
`microstructural model to at least 100 iterations to generate at least 100 perturbed microstructural
`
`models. The first microstructural model may be selected based on knowledge of the brain region
`
`associated with the voxel. The perturbation may comprise an operation selected from the group
`
`consisting of: depleting cells, altering cellular morphology or distribution, altering intracellular or
`
`interstitial physico-chemical composition or distribution, altering extracellular matrix
`
`composition or distribution, and altering intercellular spacing. The perturbation may comprise a
`
`stochastic procedure.
`
`[0042]
`
`The threshold congruence may be determined by computing an objective function
`
`between the one or more measured MRI parameters and the one or more simulated MRI
`
`parameters. The objective function may comprise an L1 norm or an L2 norm.
`
`[0043]
`
`Determining the disorder state of the brain tissue associated with the voxel may be
`
`achieved at an accuracy of at least 90%. Determining the disorder state across the brain tissue
`
`associated with the specified region of the brain may be achieved at an accuracy of at least 90%.
`
`Determining the disorder state of the brain tissue associated with the whole brain of the subject
`
`may be achieved at an accuracy of at least 90%. Determining the disorder state of the brain tissue
`
`associated the plurality of subjects may be achieved at an accuracy of at least 90%.
`
`[0044]
`
`The disorder may be a non-neurodegenerative disorder. The disorder may be selected
`
`from the group consisting of: a primary neoplasm, a metastatic neoplasm, a seizure disorder, a
`
`seizure disorder with focal cortical dysplasia, a demyelinating disorder, a non-neurodegenerative
`
`encephalopathy, a cerebr