Joint dynamic causal and biophysical modeling enables modeling of multiscale brain network function

Abstract

Methods, systems, and devices, including computer programs encoded on computer storage media, are provided for combined dynamic causal and biophysical modeling of brain function. In particular, the disclosed brain function modeling methods can be used to integrate brain function measurements from two or more methods, such as functional neuroimaging and electrophysiology. Successive model adaptation is used to improve modeling accuracy to generate a more comprehensive model of brain neuronal circuits.

Description

[Background technology]

[0001] Brain function is studied using many different experimental techniques with different spatiotemporal sampling rates and contrasts. Some of the modalities utilized include functional magnetic resonance imaging (fMRI), electrophysiology, and optical imaging.

[0002] While different techniques offer unique advantages in their ability to probe brain function at different scales, none of them provide the whole picture, so it would be useful to find methods that allow different measurements to be utilized together to build a comprehensive model of brain function. Summary of the Invention

[0003] Methods, systems, and devices, including computer programs encoded on computer storage media, are provided for combined dynamic causal modeling and biophysical modeling of brain function. In particular, the methods can be used to integrate brain function measurements by two or more methods, such as functional neuroimaging and electrophysiology. Successive model adaptation is used to improve modeling accuracy to generate a more comprehensive model of brain neuronal circuits.

[0004] In one aspect, a method of modeling brain function is provided, the method including: a) performing functional neuroimaging of neural activity in one or more regions of interest in a subject's brain; b) obtaining experimental electrophysiology data for one or more neurons in the one or more regions of interest in the subject's brain; c) fitting the functional neuroimaging data of neural activity to a dynamic causal model (DCM); d) calculating neural inter- and intra-regional connectivity strength estimates from the dynamic causal model; e) generating a biophysical model that simulates synthetic electrophysiological data using the inter- and intra-regional connectivity strength estimates; f) comparing the synthetic electrophysiological data with the experimental electrophysiological data; and g) iteratively adjusting the biophysical model until model convergence with the experimental electrophysiological data is achieved.

[0005] Exemplary functional neuroimaging techniques that can be used to implement the subject methods include, but are not limited to, functional magnetic resonance imaging (fMRI), positron emission tomography (PET), functional near-infrared spectroscopy (fNIRS), single-photon emission computed tomography (SPECT), and functional ultrasound imaging (fUS). In some embodiments, fMRI is used to detect blood oxygen level-dependent (BOLD) signals.

[0006] Exemplary electrophysiological techniques that can be used in the practice of the subject methods include, but are not limited to, electroencephalography (EEG), magnetoencephalography (MEG), and patch clamp.

[0007] In certain embodiments, the method further includes using optogenetics to excite or inhibit one or more selected neurons of interest using light. In some embodiments, the fMRI experimental data is optogenetic fMRI (ofMRI) experimental data.

[0008] Brain function can be modeled by current methods for any region in the brain.In certain embodiments, one or more brain regions of interest are in the cerebrum, cerebellum or brainstem region of the brain.Brain regions of interest can include, but are not limited to, basal ganglia, striatum, medulla, pons, midbrain, medulla oblongata, hypothalamus, thalamus, superior thalamus, amygdala, superior colliculus, cerebral cortex, neocortex, ectocortex, hippocampus, claustrum, olfactory bulb, frontal lobe, temporal lobe, parietal lobe, occipital lobe, caudate putamen, external pallidum, internal pallidum, subthalamic nucleus, substantia nigra, thalamus and motor cortex region of the brain.

[0009] Functional neuroimaging and electrophysiology data may be obtained for any type of neuron, including, but not limited to, unipolar neurons, bipolar neurons, multipolar neurons, Golgi I neurons, Golgi II neurons, axonal neurons, pseudobipolar neurons, intermediate neurons, motor neurons, sensory neurons, efferent neurons, efferent neurons, cholinergic neurons, GABAergic neurons, glutamatergic neurons, dopaminergic neurons, serotonergic neurons, histaminergic neurons, Purkinje cells, spinous neurons, Renshaw cells, and granule cells, or combinations thereof.

[0010] In certain embodiments, the method further comprises performing optical imaging or multiphoton microscopy of one or more regions of interest in the subject's brain.

[0011] In certain embodiments, the subject has a neurological disease or disorder.

[0012] In certain embodiments, the method further comprises detecting an alteration in brain function in the subject as compared to brain function in a control subject based on modeling of brain function.

[0013] In certain embodiments, the method further comprises identifying a neural circuit involved in a pathological change associated with a neurological disease or disorder.

[0014] In certain embodiments, the method further includes detecting a change in brain function in the subject after administration of a treatment for a neurological disease or disorder based on modeling the subject's brain function before and after administration of the treatment, compared to the subject's brain function before the treatment.

[0015] In certain embodiments, the functional neuroimaging and / or electrophysiology data is acquired while the subject is performing a task or responding to a stimulus. In some embodiments, the method further comprises identifying the neural circuitry involved in performing the task or responding to the stimulus. In certain embodiments, the functional neuroimaging data and experimental electrophysiology data are acquired while the subject is in a resting state (e.g., to allow comparison with data acquired when the subject is performing a task or responding to a stimulus).

[0016] In certain embodiments, the DCM is classical DCM, stochastic DCM, spectral DCM, or dynamic DCM.

[0017] In another aspect, a computer-implemented method for modeling brain function is provided, comprising computer-implemented steps of: a) receiving functional neuroimaging data and experimental electrophysiology data for one or more brain regions of interest of a subject; b) fitting the functional neuroimaging data to a dynamic causal model; c) calculating inter- and intra-regional connectivity strength estimates from the dynamic causal model; d) generating a biophysical model that simulates synthetic electrophysiological data from the inter- and intra-regional connectivity strength estimates; e) comparing the synthetic electrophysiological data with the experimental electrophysiological data; f) iteratively adjusting the biophysical model until model convergence with the experimental electrophysiological data is achieved; and g) displaying information related to the modeling of brain function.

[0018] In certain embodiments, the functional neuroimaging data includes functional magnetic resonance imaging (fMRI) data, positron emission tomography (PET) data, functional near-infrared spectroscopy (fNIRS) data, single photon emission computed tomography (SPECT) data, or functional ultrasound imaging (fUS) data.

[0019] In certain embodiments, the experimental electrophysiology data includes electroencephalography (EEG) data, magnetoencephalography (MEG) data, or patch clamp data.

[0020] In certain embodiments, the subject has a neurological disease or disorder, and displaying the information includes enumerating one or more neural circuits involved in pathological changes associated with the neurological disease or disorder.

[0021] In certain embodiments, the functional neuroimaging data and the experimental electrophysiology data are acquired while the subject is performing a task or responding to a stimulus, and displaying the information includes enumerating one or more neural circuits involved in performing the task or responding to the stimulus. In certain embodiments, the functional neuroimaging data and the experimental electrophysiology data are acquired while the subject is in a resting state (e.g., to allow comparison with data acquired while the subject is performing a task or responding to a stimulus).

[0022] In another aspect, a system for modeling brain function using the computer-implemented methods described herein is provided, the system comprising: a) a storage component for storing data, the storage component having instructions for modeling brain function based on analysis of functional neuroimaging data and experimental electrophysiology data stored therein; b) a computer processor for processing the functional neuroimaging data and the experimental electrophysiology data using one or more algorithms, the computer processor coupled to the storage component and configured to execute the instructions stored in the storage component to receive input functional neuroimaging data and experimental electrophysiology data and analyze the data according to the computer-implemented methods described herein; and c) a display component for displaying information regarding the modeling of brain function.

[0023] In another aspect, a non-transitory computer-readable medium is provided that includes program instructions that, when executed by a processor in a computer, cause the processor to perform the computer-implemented methods described herein.

[0024] In another aspect, a kit is provided that includes a non-transitory computer-readable medium and instructions for modeling brain function from functional neuroimaging data and experimental electrophysiology data. [Brief description of the drawings]

[0025] [Figure 1] Modeling with fMRI BOLD data of seven regions of interest is shown. Experimental BOLD data is averaged across all subjects. Light stimulation was 20Hz 20s ON-20s OFF-20s ON, repeated six times per cycle. Regions of interest are caudate-putamen (CPu), globus pallidus external (GPe), globus pallidus internal (GPi), substantia nigra (STN), substantia nigra (SN), thalamus (THL), and motor cortex (MCX). [Diagram 2] Modeling with electrophysiological data for four regions of interest is shown. [Diagram 3] Predictive generative network model. Using the corresponding data for each subject, we first performed DCM analyses to reveal effective connectivity estimates between and within each region using the predictive generative network model. [Figure 4] The results of the analysis are shown. Connectivity strength estimates are given in Hz. The magnitude of the parameters in this figure reflects the influence of neuronal activity in one region on the rate of change of neuronal activity in another region. These estimates are then used as influences for neuronal population design in biophysical modeling. [Figure 5A] Electrophysiological data simulated using biophysical model are shown. In constructing the biophysical model, we used the same structural model of mouse basal ganglia-thalamocortical loop as in DCM analysis, and modified neuronal model (Hill and Tononi Journal of neurophysiology, 93(3):1671-1698, 2005, incorporated herein by reference). In DCM analysis, one node represents one region of interest, while in the biophysical model, we used 100 neuronal units per region of interest to better study the mechanism at the neuronal level. The biophysical model is constructed without making any assumptions, i.e., we derive the model directly from the connectivity parameters of DCM results. [Figure 5B]Electrophysiological data simulated using biophysical model are shown. In constructing the biophysical model, we used the same structural model of mouse basal ganglia-thalamocortical loop as in DCM analysis, and modified neuronal model (Hill and Tononi Journal of neurophysiology, 93(3):1671-1698, 2005, incorporated herein by reference). In DCM analysis, one node represents one region of interest, while in the biophysical model, we used 100 neuronal units per region of interest to better study the mechanism at the neuronal level. The biophysical model is constructed without making any assumptions, i.e., we derive the model directly from the connectivity parameters of DCM results. [Figure 5C] Electrophysiological data simulated using biophysical model are shown. In constructing the biophysical model, we used the same structural model of mouse basal ganglia-thalamocortical loop as in DCM analysis, and modified neuronal model (Hill and Tononi Journal of neurophysiology, 93(3):1671-1698, 2005, incorporated herein by reference). In DCM analysis, one node represents one region of interest, while in the biophysical model, we used 100 neuronal units per region of interest to better study the mechanism at the neuronal level. The biophysical model is constructed without making any assumptions, i.e., we derive the model directly from the connectivity parameters of DCM results. [Figure 5D]Electrophysiological data simulated using biophysical model are shown. In constructing the biophysical model, we used the same structural model of mouse basal ganglia-thalamocortical loop as in DCM analysis, and modified neuronal model (Hill and Tononi Journal of neurophysiology, 93(3):1671-1698, 2005, incorporated herein by reference). In DCM analysis, one node represents one region of interest, while in the biophysical model, we used 100 neuronal units per region of interest to better study the mechanism at the neuronal level. The biophysical model is constructed without making any assumptions, i.e., we derive the model directly from the connectivity parameters of DCM results. [Figure 6] 1 shows simulated activity in the thalamus. [Figure 7] This illustrates how a single thalamic neuron fires during GPi inhibitory input. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0026] Methods, systems, and devices including a computer program encoded on a computer storage medium are provided for combined dynamic causal modeling and biophysical modeling of brain function. In particular, the disclosed methods of modeling brain function can be used to integrate brain function measurements by two or more methods, such as functional magnetic resonance imaging (fMRI) and electrophysiology. Successive model adaptation is used to improve modeling accuracy to generate a more comprehensive model of brain neuron circuits.

[0027] Before the present compositions, methods, and kits are described, it is to be understood that this invention is not limited to the particular methods or compositions described, as such may, of course, vary. It is also to be understood that the terminology used herein is for the purpose of describing particular embodiments only, and is not intended to be limiting, since the scope of the present invention will be limited only by the appended claims.

[0028] Where a range of values ​​is provided, it is to be understood that each intervening value between the upper and lower limits of that range is also specifically disclosed to the tenth of the unit of the lower limit, unless the context dictates otherwise. Each smaller range between any stated value or intervening value within a stated range and any other stated value or intervening value within that stated range is encompassed within the invention. The upper and lower limits of these smaller ranges may be independently included or excluded within the range, and each range in which either, neither, or both limits are included in the smaller range is also encompassed within the invention, subject to any specifically excluded limits in the stated range. Where a stated range includes one or both of the limits, ranges excluding either or both of those included limits are likewise encompassed within the invention.

[0029] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention belongs.Any method and material similar or equivalent to the method and material described herein can be used to carry out or test this invention, but some potential and preferred methods and materials are described here.All publications mentioned herein are incorporated by reference to disclose and describe the method and / or material in connection with which the publication is cited.In case of discrepancy, it should be understood that the present disclosure takes precedence over any disclosure of the incorporated publication.

[0030] As will be apparent to those skilled in the art upon reading this disclosure, each of the individual embodiments described and illustrated herein has distinct components and features which may be readily separated from or combined with the features of any of the other various embodiments without departing from the scope or spirit of the invention. Any recited method may be carried out in the order of events recited or in any other order which is logically possible.

[0031] It should be noted that as used in this specification and the appended claims, the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise. Thus, for example, a reference to "a neuron" includes a plurality of such neurons, a reference to "the measurement" includes one or more measurements and equivalents thereof known to those skilled in the art, and so forth.

[0032] The publications discussed herein are provided solely for their disclosure prior to the filing date of the present application. Nothing herein should be construed as an admission that the present invention is not entitled to antedate such publications by virtue of prior invention. Further, the dates of publication provided may be different from the actual publication dates which may need to be independently confirmed.

[0033] definition The term "about" is meant to encompass a deviation of plus or minus 5%, particularly with respect to a given amount.

[0034] The term "connectivity" refers to causal, functional, or directional connections between brain regions and / or neuronal populations. Neural processing may include integrated networks of several brain regions and functional connectivity of spatially separated brain regions. Analysis of functional connectivity may include detecting inter- and intra-regional neural interactions involved in brain activity during specific cognitive or motor tasks or from spontaneous activity during rest. Inter- and intra-regional "connectivity strength" may be estimated from dynamic causal modeling of brain function.

[0035] The terms "individual," "subject," "host," and "patient" are used interchangeably herein and refer to any subject with a brain, including, but not limited to, mammals, including invertebrates and arthropods (e.g., insects, crustaceans, arachnids), cephalopods (e.g., octopus, squid), amphibians (e.g., frogs, salamanders, caesars), fish, reptiles (e.g., turtles, crocodiles, snakes, amphibians, lizards, tuatara), chimpanzees, other apes, and monkey species, including humans and non-human mammals such as non-primates, laboratory animals such as mice, rats, rabbits, hamsters, guinea pigs, and chinchillas, farm animals such as dogs and cats, domestic animals such as sheep, goats, pigs, and cows, and birds such as chickens, turkeys, and other galliformes, ducks, and geese, including domestic, wild, and game birds. In some cases, the methods of the invention are used in laboratory animals, in veterinary applications, and in the development of animal models for disease, including, but not limited to, rodents, including mice, rats, and hamsters, primates, and transgenic animals.

[0036] It will be apparent to those skilled in the art that various changes and modifications can be made in the present invention without departing from the spirit or scope of the present invention.

[0037] Modeling brain function using dynamic causal modeling combined with biophysical modeling A method is provided for generating improved models of brain function using dynamic causal modeling in combination with biophysical modeling. In particular, the subject method can be used to integrate brain function measurements by two or more techniques, such as functional neuroimaging and electrophysiology. For example, functional neuroimaging can be performed to image neural activity in one or more regions of interest in a subject's brain. In addition, electrophysiological measurements can be performed on one or more neurons in the same brain region of interest. First, the functional neuroimaging data of brain neural activity is fitted to a dynamic causal model (DCM), which is used to calculate neural inter- and intra-regional connectivity strength estimates. Next, a biophysical model is constructed using the inter- and intra-regional connectivity strength estimates calculated from the DCM analysis. This biophysical model is used to simulate synthetic electrophysiological data, which can be compared to experimental electrophysiological data acquired for neurons in the brain region of interest. The biophysical model can be iteratively adjusted until model convergence with the experimental electrophysiological data is achieved. In this method, successive model adaptation is used to improve modeling accuracy to generate more accurate and comprehensive models of brain neuronal circuits.

[0038] The methods described herein are applicable to modeling brain function in any subject that has a brain, including, but not limited to, invertebrates and vertebrates, as well as mammals, including invertebrates and arthropods (e.g., insects, crustaceans, arachnids), cephalopods (e.g., octopus, squid), amphibians (e.g., frogs, salamanders, caesars), fish, reptiles (e.g., turtles, crocodiles, snakes, amphibians, lizards, tuatara), chimpanzees, other apes, and monkey species, including humans and non-human mammals such as non-primates, laboratory animals such as mice, rats, rabbits, hamsters, guinea pigs, and chinchillas, domestic animals such as dogs and cats, domestic animals such as sheep, goats, pigs, and cows, and domestic animals such as wild birds, birds, and birds such as chickens, turkeys, and other galliformes, ducks, and geese, domestic, wild, and game birds. In some cases, the methods of the invention are used in laboratory animals, in veterinary applications, and in the development of animal models for disease, including, but not limited to, rodents, including mice, rats, and hamsters, primates, and transgenic animals.

[0039] Exemplary functional neuroimaging techniques that can be used to implement the subject methods include, but are not limited to, functional magnetic resonance imaging (fMRI), positron emission tomography (PET), functional near-infrared spectroscopy (fNIRS), single photon emission computed tomography (SPECT), and functional ultrasound imaging (fUS). These functional neuroimaging techniques measure local changes in cerebral blood flow and changes in the composition of blood associated with neural activity. Functional neuroimaging is useful for non-invasively detecting patterns of brain activity associated with specific stimuli or tasks.

[0040] In some embodiments, fMRI is used to monitor temporal changes in blood flow associated with changes in levels of brain activity. Blood flow increases upon neuronal activation when an area of ​​the brain is in use. Changes in brain activity can be imaged using fMRI by detection of the blood oxygen level dependent (BOLD) signal. For descriptions of fMRI and methods of using fMRI for imaging brain activity, see, for example, Ogawa, et al. (1990) Magnetic Resonance in Medicine 14(1):68-78, Kim et al. (2002) Curr Opin Neurobiol. 12(5):607-15, Kim et al. (2012) J Cereb Blood Flow Metab. 32(7):1188-206, Bandettini (2012) Neuroimage 62(2):575-88; Zarghami et al. (2020) Neuroimage 207:116453; Logothetis, NK (June 12, 2008), Logothetis et al. (2008) Nature 453(7197):869-78, and Logothetis et al. See, al. (2001) Nature. 412(6843):150-157.

[0041] In some embodiments, PET is used to monitor changes in blood flow associated with changes in neural activity. PET uses a radioactive tracer that emits positrons for imaging. Brain activity can be imaged using PET by detecting changes in blood flow, which can be measured indirectly, for example, using an oxygen-15 tracer. Areas with higher levels of radioactivity are associated with increased brain activity. For a description of PET and methods of using PET for imaging brain activity, see, for example, Hiura et al. (2014) J Cereb Blood Flow Metab. 34(3): 389-96, Baron et al. (2012) Neuroimage 61(2): 492-504, Law (2007) Dan Med Bull. 2007 Nov; 54(4): 289-305, Ramsey et al. (1996) J Cereb Blood Flow Metab. 16(5): 755-64, which are incorporated herein by reference.

[0042] In some embodiments, SPECT imaging is used in brain imaging. Like PET, SPECT also uses radioactive tracers to detect changes in blood flow, but instead uses tracers that emit gamma rays that can be detected by a gamma camera. Brain activity can be detected by, for example, technetium. 99m Tc-examethadime or 99mTc-D, L-hexamethylene-propyleneamine oxime can be used to detect changes in blood flow and image.For a description of SPECT and the method of using SPECT for imaging brain activity, see, for example, Cuocolo et al.(2018) Int Rev Neurobiol 141:77-96, Andersen(1989) Cerebrovasc Brain Metab Rev.1(4):288-318, Matsuda(2001) Ann Nucl Med 15(2):85-92, Gonul et al.(2009) Int Rev Psychiatry 21(4):323-35, which are incorporated herein by reference.

[0043] In some embodiments, fNIRS is used to monitor changes in the composition of blood near a neural event. In particular, fNIRS can be used to detect changes in the levels of oxyhemoglobin and deoxyhemoglobin using near-infrared light. Based on the difference in the absorption spectra of oxyhemoglobin and deoxyhemoglobin, the relative changes in hemoglobin concentration can be measured with fNIRS. Cerebral hemodynamic responses correlate with brain activation or deactivation. For a description of fNIRS and methods of using fNIRS for imaging brain activity, see, e.g., Tachtsidis et al. (2020) Ann NY Acad Sci. 1464(1):5-29, Ferrari et al. (2012) Neuroimage 63(2):921-35, Scholkmann et al. (2014) Neuroimage 85 Pt 1:6-27, Kim et al. (2017) Mol Cells 40(8):523-532, which are incorporated herein by reference.

[0044] In some embodiments, fUS is used to monitor local changes in cerebral blood volume that correlate with changes in neural activity. For a description of fUS and methods of using fUS for imaging brain activity, see, for example, Bleton et al. (2016) BMC Med Imaging 12;16:22, Raaij et al. (2012) Neuroimage 63(3):1030-7, which are incorporated herein by reference.

[0045] Electrophysiological techniques are used to measure electrical properties in the brain, typically voltage or current changes of neurons. Exemplary electrophysiological techniques that can be used to implement the subject methods include, but are not limited to, electroencephalography (EEG), magnetoencephalography (MEG), and patch clamp. These electrophysiological techniques are useful for identifying specific types of neurons involved in neural networks and measuring neuron-specific changes in activity associated with brain responses.

[0046] In some embodiments, EEG is used to record neuronal electrical activity in the brain. EEG can be performed non-invasively with electrodes placed on the scalp. EEG measurement has the advantage of having high time resolution, and can detect changes in electrical activity in the brain on the millisecond time scale. For a description of EEG and methods of using EEG to record electrical activity in the brain, see, e.g., Niedermeyer et al. (2004) Electroencephalography: Basic Principles, Clinical Applications, and Related Fields. Lippincott Williams & Wilkins; Jackson et al. (2014) Psychophysiology 51(11):1061-71, Khanna et al. (2015) Neurosci Biobehav Rev. 49:105-13, Feyissa et al. (2019) Handb Clin Neurol. 160:103-124, Beres et al. (2017) Appl Psychophysiol Biofeedback 42(4):247-255, which are incorporated herein by reference.

[0047] In some embodiments, MEG is used to record the magnetic field produced by the current generated in the brain. MEG detects the weak magnetic field produced by synchronized neuronal currents (i.e., the ionic currents that flow in the dendrites of neurons during synaptic transmission). These weak magnetic fields can be detected using magnetometers such as superconducting quantum unit interference devices (SQUIDs) or spin exchange relaxation free (SERF) magnetometers. For a description of MEG and methods of using MEG to record magnetic fields associated with neuronal activity in the brain, see, e.g., Hamalainen et al. (1993) Reviews of Modern Physics. 65 (2): 413-497, Baillet et al. (2017) Nat Neurosci. 20(3): 327-339, Gross et al. (2019) Neuron 104(2): 189-204, Stapleton-Kotloski et al. (2018) Brain Sci. 8(8): 157, which are incorporated herein by reference.

[0048] Dynamic causal modeling (DCM) is a Bayesian generative modeling paradigm that can model causal relationships in the brain, taking into account neuronal dynamics, modulating inputs, and observational data (e.g., BOLD fMRI, EEG, or MEG). DCM can be used to estimate connections between brain regions and quantify effective connectivity strengths between neuronal populations in one or more regions of the brain. Dynamic causal models of interacting neural populations can be used to estimate inter- and intra-neuronal connectivity strengths from functional neuroimaging data using Bayesian statistical methods. The models can include interactions between excitatory and inhibitory neural populations. Different types of DCM can be used in computing connectivity strength estimates. Classical DCM does not assume state noise, while stochastic DCM (sDCM) utilizes state noise and accounts for uncertainty due to neighboring regions. Spectral DCM is another paradigm that assumes state noise, but in this approach, model fitting is performed on the cross-spectrum of the system, which makes sDCM useful for resting-state modeling. Dynamic effective connectivity (or dynamic DCM) is another method that is useful for investigating changes in connectivity strength over time. For descriptions of different types of DCM techniques, see, for example, Friston et al. (2003) Neuroimage 19(4):1273-1302, Friston et al. (2014) NeuroImage 94:396-407, Park et al. (2018) Neuroimage 180:594-608, Li et al. (2011) NeuroImage 58(2):442-457, Stephan et al. (2010) NeuroImage 49(4):3099-3109, Stephan et al. (2008) NeuroImage.42(2):649-662, Marreiros et al. NeuroImage 39(1):269-278, Razi et al. (2015) NeuroImage 106:1-14, and Lee et al. See, al. (2006) NeuroImage. 30(4):1243-1254.

[0049] Biophysical modeling is based on biophysical attributes of neurons and brain regions. Biophysical neuron models may include various biophysical attributes, including but not limited to, firing patterns of individual neurons, neuron types, receptors, inputs and outputs, topological structures, and self-connectivity. The method utilizes a biophysical model to simulate synthetic electrophysiological data using inter- and intra-regional connectivity strength estimates from dynamic causal modeling. The synthetic electrophysiological data is then compared to experimental electrophysiological data acquired from the subject, and the model is improved by iteratively adjusting to reach convergence. For a description of biophysical modeling, see, for example, Example 5 of Hill and Tononi (2005) Journal of Neurophysiology 93(3):1671-1698, and Murray et al. (2018) Biol Psychiatry Cogn Neurosci Neuroimaging 3(9):777-787, which are incorporated herein by reference.

[0050] In some embodiments, functional neuroimaging and electrophysiological techniques are used in combination to detect brain responses when a subject is exposed to a stimulus or performs a task. In addition, functional neuroimaging and electrophysiological measurements of brain activity can be performed while the subject is in a resting state (e.g., in the absence of a stimulus or without a task), allowing brain activity to be compared with the subject's "baseline" brain state, i.e., to identify brain regions that show changes in neural activity associated with a particular stimulus or task.

[0051] In some embodiments, the methods described herein are used to assess changes in brain function in response to optogenetic perturbation of neural activity. In certain embodiments, optogenetics is used to induce cell-specific perturbations in the brain. For example, optogenetics can be used to excite or inhibit one or more selected neurons of interest using light. For descriptions of optogenetic techniques, see, for example, Abe et al., 2012; Desai et al., 2011; Duffy et al., 2015; Gerits et al., 2012; Kahn et al., 2013; Lee et al., 2010; Liu et al., 2015; Ohayon et al., 2013; Weitz et al., 2015; Weitz and Lee, 2013, which are incorporated herein by reference.

[0052] The methods described herein can also be used to assess changes in brain function in response to brain stimulation involving electrical current or magnetic fields applied to selected brain regions. For example, electrical brain stimulation (EBS) can be used to stimulate neurons or neural networks in the brain through direct or indirect excitation of their cell membranes by using electrical current. For descriptions of EBS techniques, see, e.g., Aum et al. (2018) Front Biosci (Landmark Ed) 23:162-182, Tellez-Zenteno et al. (2011) Neurosurg Clin N Am. 22(4):465-75, Padberg et al. (2009) Exp. Neurol. 219:2-13, Nahas et al. (2010) Biol. Psychiatry 67:101-109, Lefaucheur et al. (2010) Exp. Neurol. 223:609-614, Levy et al. (2008) J. Neurosurg. 108:707-714, Hanajima et al. (2002) Clin. Neurophysiol. 113:635-641, Picillo et al. al.(2015) Brain Stimul.8:840-842., Canavero(2014) Textbook of Cortical Brain Stimulation.Berlin:De Gruyter Open. Alternatively, transcranial magnetic stimulation (TMS) can be used to electrically stimulate the brain by electromagnetic induction and can be used to stimulate specific regions of the brain non-invasively. For a description of TMS techniques, see, for example, Klomjai et al.(2015) Ann Phys Rehabil Med.58(4):208-213, Lefaucheur(2019) Handb Clin Neurol.160:559-580, Burke et al.(2019) Handb Clin Neurol.163:73-92, which are incorporated herein by reference.

[0053] In some embodiments, the methods described herein are used to assess changes in brain activity in response to a subject performing a cognitive or motor task. Such tasks may include, but are not limited to, tests of memory, intelligence, speech / language, emotion, executive function (e.g., problem solving, planning, organizational skills, selective attention, inhibitory control), visuospatial function, balance, or physical activity / movement (e.g., walking, running, limb movement). The methods described herein can be used to analyze neural interactions within and between regions involved in brain activity during a particular cognitive or motor task, and can be compared to the subject's brain activity in a resting state.

[0054] Brain function can be modeled by current methods for any region in the brain.In certain embodiments, one or more brain regions of interest are in the cerebrum, cerebellum or brainstem region of the brain.Brain regions of interest can include, but are not limited to, basal ganglia, striatum, medulla, pons, midbrain, medulla oblongata, hypothalamus, thalamus, superior thalamus, amygdala, superior colliculus, cerebral cortex, neocortex, ectocortex, hippocampus, claustrum, olfactory bulb, frontal lobe, temporal lobe, parietal lobe, occipital lobe, caudate putamen, external pallidum, internal pallidum, subthalamic nucleus, substantia nigra, thalamus and motor cortex region of the brain.

[0055] Functional neuroimaging and electrophysiology data may be obtained for any type of neuron, including, but not limited to, unipolar neurons, bipolar neurons, multipolar neurons, Golgi I neurons, Golgi II neurons, axonal neurons, pseudobipolar neurons, intermediate neurons, motor neurons, sensory neurons, efferent neurons, efferent neurons, cholinergic neurons, GABAergic neurons, glutamatergic neurons, dopaminergic neurons, serotonergic neurons, histaminergic neurons, Purkinje cells, spinous neurons, Renshaw cells, and granule cells, or combinations thereof.

[0056] SYSTEM AND COMPUTER-IMPLEMENTED METHODS FOR MODELING BRAIN FUNCTION - Patent application In a further aspect, the present invention includes a computer-implemented method for modeling brain function using a combination of dynamic causal modeling and biophysical modeling as described herein. The computer performs the steps of: a) receiving functional neuroimaging data and experimental electrophysiology data for one or more brain regions of interest of a subject; b) fitting the functional neuroimaging data to a dynamic causal model; c) calculating neural inter- and intra-regional connectivity strength estimates from the dynamic causal model; d) generating a biophysical model that simulates synthetic electrophysiological data from the inter- and intra-regional connectivity strength estimates; e) comparing the synthetic electrophysiological data with the experimental electrophysiological data; f) iteratively adjusting the biophysical model until model convergence with the experimental electrophysiological data is achieved; and g) displaying information regarding the modeling of brain function.

[0057] In certain embodiments, the functional neuroimaging data includes functional magnetic resonance imaging (fMRI) data, positron emission tomography (PET) data, functional near-infrared spectroscopy (fNIRS) data, single photon emission computed tomography (SPECT) data, or functional ultrasound imaging (fUS) data.

[0058] In certain embodiments, the experimental electrophysiology data includes electroencephalography (EEG) data, magnetoencephalography (MEG) data, or patch clamp data.

[0059] The method can be implemented in digital electronic circuitry, or in computer software, firmware, or hardware. The disclosed and other embodiments can be implemented as one or more computer program products, i.e., one or more modules of computer program instructions encoded on a computer-readable medium for execution by or control of the operation of a data processing apparatus. The computer-readable medium can be a machine-readable storage device, a machine-readable storage substrate, a memory device, a composition of matter affecting a machine-readable propagated signal, or any combination thereof.

[0060] A computer program (also known as a program, software, software application, script, or code) can be written in any form of programming language, including compiled or interpreted languages, and it can be deployed in any form, including as a stand-alone program or as modules, components, subroutines, or other units suitable for use in a computing environment. A computer program does not necessarily correspond to a file in a file system. A program can be stored in part of a file that holds other programs or data (e.g., one or more scripts stored in a markup language document), in a single file dedicated to the program in question, or in multiple coordinated files (e.g., a file that stores one or more modules, subprograms, or portions of code). A computer program can be deployed to run on one computer or multiple computers located at one site or distributed across multiple sites and interconnected by a communication network.

[0061] In a further aspect, a system is provided for performing the computer-implemented method as described. Such a system includes a computer including a processor, a storage component (i.e., memory), a display component, and other components typically present in a general-purpose computer. The storage component stores information accessible by the processor, including instructions that can be executed by the processor and data that can be retrieved, manipulated, or stored by the processor.

[0062] The storage component includes instructions. For example, the storage component includes instructions for modeling brain function based on the analysis of the functional neuroimaging data and experimental electrophysiology data stored therein. The computer processor is coupled to the storage component and configured to execute the instructions stored therein to receive the functional neuroimaging data and experimental electrophysiology data and analyze the data according to one or more algorithms using dynamic causal modeling and biophysical modeling, as described herein. The display component displays information related to the modeling of brain function. In certain embodiments, the functional neuroimaging data and experimental electrophysiology data are obtained from a subject having a neurological disease or disorder, and the information displayed related to the modeling of brain function includes enumerating one or more neural circuits involved in pathological changes associated with the neurological disease or disorder. In certain embodiments, the functional neuroimaging data and experimental electrophysiology data are obtained while the subject is performing a task or responding to a stimulus, and the information displayed related to the modeling of brain function includes enumerating one or more neural circuits involved in performing the task or responding to the stimulus. In certain embodiments, the functional neuroimaging data and the experimental electrophysiology data are acquired while the subject is in a resting state (e.g., to allow comparison with data acquired while the subject is performing a task or responding to a stimulus).

[0063] The storage component may be of any type capable of storing information accessible by the processor, such as a hard drive, memory card, ROM, RAM, DVD, CD-ROM, USB flash drive, writable memory, and read-only memory. The processor may be any well-known processor, such as a processor from Intel Corporation. Alternatively, the processor may be a dedicated controller, such as an ASIC.

[0064] Instructions may be any set of instructions executed directly (such as machine code) or indirectly (such as a script) by a processor. In that regard, the terms "instructions," "steps," and "program" may be used interchangeably herein. Instructions may be stored in object code format for direct processing by a processor, or in any other computer language, including a script or collection of independent source code modules that are interpreted on demand or precompiled.

[0065] Data may be retrieved, stored, or modified by a processor according to instructions. For example, the system is not limited by any particular data structure, but data may be stored in a computer register, a relational database, as a table with multiple different fields and records, an XML document, or a flat file. Data may also be formatted in any computer-readable format, such as, but not limited to, binary values, ASCII, or Unicode. Furthermore, data may include any information sufficient to identify related information, such as numbers, descriptive text, unique codes, pointers, references to data stored in other memory (including other network locations), or information used by a function to calculate related data.

[0066] In certain embodiments, the processor and storage components may comprise multiple processors and storage components that may or may not be stored in the same physical housing. For example, some of the instructions and data may be stored on a removable CD-ROM and others in a read-only computer chip. Some or all of the instructions and data may be stored in a location that is physically separate from the processor but still accessible by the processor. Similarly, a processor may comprise a collection of processors that may or may not operate in parallel.

[0067] kit Also provided are kits for implementing the methods for modeling brain function from functional neuroimaging data and experimental electrophysiology data using a combination of dynamic causal modeling and biophysical modeling as described herein. In some embodiments, the kits include software for executing the computer-implemented methods for modeling brain function using a combination of dynamic causal modeling and biophysical modeling. In some embodiments, the kits include a system for modeling brain function as described herein.

[0068] In addition, the kit may further include instructions (in certain embodiments) for carrying out the subject method. These instructions may be present in the subject kit in various forms, one or more of which may be present in the kit. For example, the instructions may be present as printed information on a suitable medium or substrate, such as a sheet or sheets of paper on which the information is printed, in the kit packaging, in a package insert, etc. Another form of these instructions is a computer readable medium on which the information is recorded, such as a diskette, a compact disc (CD), a flash drive, etc. Yet another form of these instructions that may be present is a website address that may be used via the Internet to access the information at a remote location.

[0069] usefulness The disclosed method finds many applications in research and development in the field of neuroscience, particularly in the study of brain function and dysfunction, and in the development of therapeutics for the treatment of neurological diseases and disorders. The combined modeling approach of DCM and biophysical modeling described herein can relate different types of data with different spatial and temporal resolutions. In particular, the inventors have demonstrated that the combined use of DCM and biophysical modeling can generate a more comprehensive brain function model that reproduces both fMRI and electrophysiology data (see Examples). Thus, this modeling approach is useful for interpreting experimental data obtained by different techniques, and for testing neuroscience hypotheses. Applications of this modeling approach include fundamental research into neurological processes involved in cognitive and motor functions, including identifying neural circuits in various regions of the brain and determining their functions, identifying neural circuits and brain regions affected by pathology in subjects with neurological diseases and disorders, and monitoring brain responses to therapeutics for treating neurological diseases and disorders.

[0070] The subject methods are applicable to modeling brain function of any region of the brain, including the frontal, parietal, temporal, or occipital lobes, gyri (e.g., frontal, temporal, precentral, postcentral, or cingulate gyrus), sulci (e.g., central, cingulate, calcaneal, caulosal, or hippocampal sulcus), cerebrum or regions thereof, including but not limited to the cerebral cortex (neocortex or isocortex), and corpus callosum, the cerebellum or regions thereof, including but not limited to the anterior, posterior, and convoluted lobes, the midbrain, the cerebral cortex and medulla, or the basal ganglia or regions thereof, including but not limited to the striatum, including the dorsal striatum (caudate and putamen) and ventral striatum (nucleus accumbens and olfactory tubercle), the globus pallidus, ventral globus pallidus, substantia nigra, and subthalamic nucleus.

[0071] The subject methods are also applicable to modeling the function of unipolar neurons, bipolar neurons, multipolar neurons, Golgi I neurons, Golgi II neurons, axonal neurons, pseudobipolar neurons, interneurons (e.g., basket cells, Lugaro cells, unipolar brush cells, spindle cells), pyramidal neurons (e.g., RSad neurons, RSna serotonin neurons, IB neurons, and Betz cells), motor neurons (upper and lower motor neurons), sensory neurons, efferent neurons, efferent neurons, cholinergic neurons, GABAergic neurons, glutamatergic neurons, dopaminergic neurons, serotonergic neurons, histaminergic neurons, cerebellar Purkinje cells, striatal neurons (e.g., spinous neurons, cholinergic interneurons, and GABAergic interneurons), Renshaw cells, granule cells, as well as connected neuronal groups that form neural circuits, and inter- and intra-regional continuity between neurons in one or more selected regions of the brain.

[0072] The subject methods may also be used to treat neurological diseases and disorders, including brain injuries such as, but not limited to, frontal lobe injuries, parietal lobe injuries, temporal lobe injuries, and occipital lobe injuries; brain dysfunction such as, but not limited to, aphasia, optic ataxia, dysarthria, apraxia, agnosia, amnesia; cranial nerve disorders such as trigeminal neuralgia, hemifacial spasm, autonomic disorders (such as dysautonomia, multiple system atrophy), seizure disorders such as epilepsy, Parkinson's disease, essential tremor, amyotrophic lateral sclerosis, Tourette's syndrome, multiple sclerosis, and the like. The subject methods are applicable to modeling changes in brain function associated with central and peripheral nervous system movement disorders, various peripheral neuropathies, sleep disorders such as narcolepsy, and other neurological diseases and disorders associated with nervous system damage, underdevelopment, biochemical, anatomical or electrical dysfunction, and / or disease pathology such as attention deficit hyperactivity disorder, autism, Asperger's syndrome, obsessive-compulsive disorder, Huntington's disease, Alzheimer's disease, multiple sclerosis, and organic psychiatric disorders, among others. Additionally, the subject methods can be used to screen therapeutic agents for the treatment of such neurological diseases and disorders, as well as to monitor the effectiveness of selected treatments.

[0073] experiment The following examples are presented to provide those of ordinary skill in the art with a complete disclosure and description of how to make and use the present invention, and are not intended to limit the scope of what the inventors regard as their invention, nor are they intended to represent that the following experiments are all or the only experiments performed. Efforts have been made to ensure accuracy with respect to numbers used (e.g., amounts, temperatures, etc.), but some experimental error and deviation should be accounted for. Unless otherwise indicated, parts are parts by weight, molecular weight is weight average molecular weight, temperature is in degrees Celsius, and pressure is at or near atmospheric.

[0074] All publications and patent applications cited in this specification are herein incorporated by reference to the same extent as if each individual publication or patent application was specifically and individually indicated to be incorporated by reference.

[0075] The present invention has been described with respect to specific embodiments found or proposed by the inventor to include preferred modes for carrying out the invention. Those skilled in the art will appreciate that, in light of this disclosure, numerous modifications and changes can be made in the specific embodiments exemplified without departing from the intended scope of the invention. All such modifications are intended to be within the scope of the appended claims.

[0076] Example 1 Joint dynamic causal and biophysical modeling enables multiscale brain network functional modeling We designed a novel modeling approach that integrates brain function measurements from different modalities. Brain function is studied using many different experimental techniques with different spatiotemporal sampling rates and contrasts. Some of the modalities utilized include functional magnetic resonance imaging (fMRI), electrophysiology, and optical imaging. While different techniques offer unique advantages in terms of their ability to probe brain function at different scales, none of them provide the whole picture, so it is important to have an approach that allows us to understand how different measurements can be utilized to build a comprehensive model of brain function.

[0077] To build a comprehensive model of brain function, we required a strategy to model data from multimodal imaging / recording. Here, we provide a novel approach that allows comprehensive modeling of whole-brain scale function combined with the ability to model spiking-level activity. We first fit a model to data from one imaging / recording modality, e.g., functional magnetic resonance imaging. The outcome from this process can be, for example, connectivity estimates between selected regions. We then use the outcomes from the model fitting procedure to start modeling neural activity at the single unit scale. This modeling utilizes as a basis the findings from the initial dataset and is adapted under their influence. Using this influence from another paradigm to try to fit the experimental data, we obtain a powerful model.

[0078] More specifically, we provide an example that can be generalized to other measurement modalities and computational modeling approaches. We used DCM (Dynamic Causal Model) on fMRI data and a biophysical model on electrophysiology data in the following steps. DCM constituted a first model that influenced a second model, i.e., a biophysical model. The steps of our method are provided below. 1. fMRI data were fitted to a DCM model and effective inter- and intra-regional connectivity strength estimates were extracted. 2. Using these DCM estimates, we constructed a biophysical model to simulate synthetic electrophysiological data. One could include other biophysical attributes as hypotheses for testing, such as neuron type, receptors, and input and output topology structures. 3. The synthetic electrophysiology data was compared to the experimental electrophysiology data. If they matched, we terminated our pipeline. If not, we updated the biophysical model and iterated until convergence.

[0079] DCM is a data-driven approach, while biophysical modeling is highly hypothesis-dependent. This joint modeling approach integrates the advantages of DCM and biophysical modeling, and compensates for the main shortcomings of each. Using DCM estimates in biophysical modeling also reduces the parameter search time. This biophysical modeling approach has, in principle, the power to reproduce and interpret fMRI and electrophysiological data, but the parameter space grows exponentially with the number of regions. On the other hand, biophysical modeling extends DCM with many physiological details. For example, instead of modeling a brain region as a single node, biophysical modeling describes a region as a neuronal group with input and output mechanisms and topological structure.

[0080] In summary, by combining the two modeling approaches we are able to base our analyses directly on both fMRI and electrophysiological data and test hypotheses about the mechanisms underlying brain functions of interest.

[0081] Example 2 Modeling the basal ganglia in the mouse brain Using the basal ganglia in mouse brain as an example, we have shown that the modeling approach described in Example 1 is a powerful tool for building a comprehensive model of brain function. Here, we have utilized real fMRI and electrophysiology data. Details of dynamic causal modeling (DCM) modeling using optogenetic fMRI (ofMRI) data are described by Bernal-Casas et al. (Neuron, 93(3):522-532, 2017, incorporated herein by reference in its entirety), but for illustrative purposes, we provide a brief description here.

[0082] The basal ganglia is an essential region for motor control and many other important brain functions. This region has been difficult to study because of its anatomical mixture of different cell types and complex circuits. Bernal-Casas et al. (above) used optogenetic fMRI to separately stimulate D1 and D2 medium spiny neurons in the striatum and obtained high-resolution data on the pathways associated with the two types of cells in the basal ganglia. We performed modeling with fMRI BOLD data of seven regions of interest and electrophysiology data of four regions of interest (Figures 1 and 2). Figure 1 shows the experimental BOLD data averaged across all subjects. Optical stimulation was performed at 20 Hz 20 s ON-20 s OFF-20 s ON cycles and repeated six times. The regions of interest were the caudate-putamen (CPu), external pallidum (GPe), internal pallidum (GPi), substantia nigra (STN), substantia nigra (SN), thalamus (THL), and motor cortex (MCX).

[0083] Using the corresponding data for each subject, we first performed a DCM analysis to reveal effective connectivity estimates between and within each region. For this analysis, we employed the prior generative network model shown in Figure 3. The analysis results can be found in Figure 4, where the connectivity strength estimates are given in Hz. The magnitude of the parameters in this figure reflects the influence of neuronal activity in one region on the rate of change of neuronal activity in another region. These estimates are then used as influences for neuronal population design in biophysical modeling.

[0084] In constructing the biophysical model, we used the same structural model of the mouse basal ganglia-thalamocortical loop as in the DCM analysis, and a modified neuronal model from

[10] . The Hill and Tononi neuronal model is a variant of the Hodgkin-Huxley model that features biologically accurate details. In the DCM analysis, one node represents one region of interest, whereas in the biophysical model, we used 100 neuronal units per region of interest to better study the neuronal-level mechanisms. The number of units is set arbitrarily.

[0085] We first constructed a biophysical model without making any hypotheses, i.e., we derived the model directly from the connectivity parameters of the DCM results. Figure 5 shows simulated electrophysiology data using this model that are consistent with the experimental data shown in Figure 2. However, because DCM is a region-level analysis that may not reflect true neuronal-level mechanisms, testing hypotheses of neuronal-level mechanisms is an important advantage of our biophysical modeling. For example, anatomically, the projections from GPi to the thalamus are GABAergic, a type of inhibitory connection. In principle, this connection should lead to opposite polarities of activity in GPi and thalamus, yet our experiments found that both showed increased activity during D1-MSN stimulation.

[0086] We hypothesized that the thalamus can be activated during GPi inhibitory signals, and that synchronization between GPi neurons is the key to this seemingly paradoxical phenomenon. This hypothesis is based on the observation that in our D1-MSN stimulation experiments, the synchronization level increased among neurons in multiple focal areas. One thalamic neuron can receive inputs from multiple GPi neurons. Before D1 stimulation occurs, the synchronization level between GPi neurons is relatively low, such that one thalamic neuron receives asynchronous inhibitory inputs from the intrinsic firing of GPi. During stimulation, GPi firing and inhibition to the thalamus become more synchronized, and thus intrinsically active thalamic neuron firing can occur during synchronized GPi inhibitory signals.

[0087] To test this hypothesis, we generated a new model by forcibly inhibiting the connections from GPi to the thalamus. The remainder of the model remains the same as the first model. We succeeded in increasing thalamic activity during stimulation despite GPi inhibition. Figure 6 shows the simulated activity in the thalamus, and Figure 7 illustrates how one thalamic neuron fires between GPi inhibitory inputs. The results demonstrate that our joint DCM and biophysical modeling approach can serve as a powerful tool in representing both fMRI and electrophysiological data and testing hypotheses of underlying mechanisms.

[0088] Example 3 Continuous Model Fitting Our method can be applied to perform successive model fitting, using information from one model to improve another model, when data from three or more different modalities are acquired at different scales. Thus, all existing applications of the modeling paradigm we use can benefit from our method. This enlarges the number of possible applications, which include, but are not limited to, the following scenarios: 1. Optogenetic experiments in selective excitation of cells can be performed during fMRI scanning and electrophysiological data acquisition. Our method allows for comprehensive modeling of brain circuits. 2. Resting-state experiments with no direct external input to the brain. Collection of resting-state fMRI (rsfMRI) and electrophysiology data is common and often follows connectivity modeling. Our method allows for comprehensive modeling of brain circuits.

[0089] The experimental methods described above can be used in the study of diseases including Parkinson's disease and epilepsy. As an example, in [2] MRI and DCM are used to investigate neural circuits that play an important role in Parkinson's disease. Similarly, in [4] fMRI is used for the analysis of epilepsy networks. Our modeling approach can play an important role in the advancement of these fields. Our joint DCM and biophysical modeling can uniquely generate functional brain models that can reproduce both fMRI and electrophysiology data. Importantly, this approach can relate two types of data with different spatial and temporal resolutions. This modeling approach can be a powerful tool to interpret experimental data from different techniques and test neuroscience hypotheses.

[0090] Example 4 BOLD fMRI and electrophysiology data Functional magnetic resonance imaging (fMRI) is a neuroimaging method that allows non-invasive and large-scale monitoring of temporal changes in brain activity. It has shown great success as a functional neuroimaging method in the last 30 years and has found use in both clinical and cognitive neuroscience [9]. In a 1992 paper [1], the authors proposed an MRI method that allows imaging temporal changes in brain activity by detecting blood oxygenation levels in MRI. This paper marks the birth of fMRI through the detection of blood oxygen level-dependent (BOLD) signals. The underlying phenomenon of these signals is the oxygen demand of neurons.

[0091] With images taken periodically, we can extract the activity of each voxel throughout the course of the scan. Once the fMRI scan is complete, what we have is a 4D signal, where the fourth dimension is time. We then anatomically mask this tensor to select regions of interest and perform dimensionality reduction on the selected regions. We calculate the average time series across each region to represent it in a single time series. In our method, the BOLD fMRI signal is one of the sources of data we use for modeling.

[0092] Electrophysiological techniques study changes in the electrical properties of neurons, usually voltage or current. Popular non-invasive electrophysiological techniques include electroencephalography (EEG) and magnetoencephalography (MEG), as well as invasive techniques such as intracellular, patch clamp and extracellular recordings, which have proven successful in collecting high temporal resolution data [5]

[18]

[11] . In our current study, we used spike trains collected from extracellular recordings. One advantage of electrophysiology is that it has recently been combined with cutting-edge techniques such as multiphoton microscopy and optogenetics. In our study, optogenetic fMRI data was used to allow researchers to excite or inhibit specific neurons using light, giving researchers unprecedented high spatial and temporal resolution

[13] . By recording the electrical activity surrounding a stimulated or inhibited neuron, it is possible to establish a direct link between neuronal excitability and responses in downstream / associated cells.

[0093] In summary, our method uses the non-invasive imaging technique fMRI to monitor temporal changes in brain regions of interest through a proxy called the BOLD signal, and uses electrophysiological data to study the behavior of specific neurons and the associated mechanisms.

[0094] Example 5 Effective Connectivity and Dynamic Causal Modeling (DCM) DCM is a Bayesian generative modeling paradigm that can comprehensively model causal relationships in the brain [6], [7]. It considers neuronal dynamics, modulatory and / or excitatory inputs, and observation methods (BOLD fMRI, EEG / MEG, etc.) at once. It is therefore considered one of the most biologically plausible and technologically advanced modeling methods

[17] . The goal of DCM is to estimate the connections between brain regions and their dependence on the experimental context, as well as to find model evidence to compare different hypotheses. The papers mentioned in the results section also make use of this methodology and can be found in [2].

[0095] DCM uses a state-space system to model brain dynamics. There are two equations governing this system: the evolution equation and the observation equation given in (1).

[0096]

number

[0097] The time series x(t) is the neuronal activity, u(t) is the input stimulus, and y(t) is the observed signal. w(t) is the state noise, ∈(t) is the observation noise. The function g(.) is the observation model and varies with respect to the imaging method, but the function f(.) is given as follows:

[0098]

number

[0099] Here, matrix A represents the effective connectivity strength since it defines the change in neuronal activity in a region in terms of the current activity in all regions. This matrix is ​​generally the main goal of the model fitting task. Matrices B and C are in the model to reflect the effect of the inputs on the dynamics. Matrix B governs the connectivity modulation effect of the inputs, whereas matrix C models the direct contribution of the inputs to the dynamics.

[0100] There are several different types of DCM in the literature. Classical DCM [6] does not assume state noise, whereas stochastic DCM (sDCM)

[14] uses state noise to account for uncertainty due to neighboring regions. Spectral DCM [7] is another paradigm that assumes state noise, but in this approach, model fitting is performed on the cross-spectrum of the system, making it a useful model for resting-state modeling. Dynamic Effective Connectivity (or Dynamic DCM)

[16] is also one of the techniques we use in our method, where the temporal changes in connectivity strength are investigated. In our method, we can use any type of DCM for one of the joint modeling.

[0101] DCM Inversion DCM is a Bayesian modeling technique, meaning that its parameters are treated as distributions to be estimated, i.e. random variables, rather than as constants to be found. What we are interested in is therefore p(θ|y,m), i.e. the posterior, where θ is the sum of all the parameters to be estimated and m is the model hypothesis we use. A model hypothesis in the DCM context can be seen as a particular choice of f, g and E and their respective priors. To illustrate, there is DCM

[19] , which uses a nonlinear evolution function, and several other models

[15] with two neuronal states. Hence the choice is called a model in this context.

[0102] During inversion, the Bayesian nature of the method requires us to use a prior. The choice of prior is in fact an important consideration for proper modeling. Some of the parameters allow the use of information from previous neuroscience studies (e.g., hemodynamic prior), while for other parameters we can only assume a support set (e.g., A). For the latter type, we only make a binary decision about the existence of the parameter (i.e., whether its posterior is nonzero or not). When set to zero, the posterior has zero mean and variance.

[0103] Inversion of DCM models is based on calculating the model evidence p(y|m), which is the probability that the model observes the data given. The analytical calculation of model evidence is often difficult. Therefore, a common method used for inversion is the variational Bayesian approach.

[0104] In this variation method, our task is to iteratively update a distribution called the variation distribution (denoted by q(θ)). The log model evidence can be written as ln p(y|m)=KL(q(θ)||p(θ|y,m))+F(q(θ),y)(3)

[0105] Since we do not know p(θ|y,m), the first term cannot be computed. However, due to the properties of KL divergence, we know that it is non-negative and therefore the second term constitutes a lower bound on the model evidence. Thus, instead of maximizing the log-model evidence, the best estimate is to maximize what is called the negative free energy.

[0106]

number

[0107] where q(θ) is a variance distribution that is also an approximation of the posterior distribution, and KL(.) is the operation of Kullback-Leibler divergence. The first term represents the precision, and the second term has a complexity interpretation (to measure the deviation from the prior distribution). The above lower bound holds in equality if q*(θ)=p(θ|y,m). In DCM, we also make assumptions about the distribution of the parameters, which are assumed to be normal. This assumption is called the Laplace approximation, and we reflect it in the prior distribution p(θ|m).

[0108] In summary, our problem can be formulated as follows. max(ln p(y|θ,m))qK L(q(θ)||p(θ|m))(5) q(θ)

[0109] The solution to this problem ends with an iterative method called Expectation Maximization (EM). EM is a steepest descent algorithm based on Jacobian calculations and is a common tool in variance analysis. When the algorithm converges, it returns the resulting model evidence and the variance distribution that led to its values. As mentioned before, this distribution is an approximation for the posterior distribution of the parameters. Maximum a posteriori (MAP) estimates of these distributions give us the parameter estimates we are looking for.

[0110] Biophysical modelling methods Modeling based on the biophysical attributes of neurons and brain regions is another major neural modeling method. The main advantage of such an approach is that the models can reflect brain behavior at different scales and resolutions.

[0111] At least one prior model should be constructed based on the experimental results and hypotheses for testing. Prior models usually consist of two parts, a structural model and a neuronal model. The structural model determines the brain regions and regional connections hypothetically involved in the process of interest, and the neuronal model expands each region into a group of neuronal units to reflect the detailed biophysical mechanism. In our joint DCM and biophysical modeling process, the structural model is also the prior model for DCM analysis. The neuronal model first describes a single neuron firing pattern, typically by a differential equation

[22]

[12] . Other biophysical attributes, such as self-connectivity, neuronal type, receptors, inputs and outputs, and topological structure, have also been described from time to time

[10]

[21] .

[0112] In our study example, we used the neuron model described in Hill and Tononi (Modeling sleep and wakefulness in the thalamocortical system). See Journal of Neurophysiology, 93(3):1671-1698, 2005, which is incorporated herein by reference in its entirety. Other types of biophysical models may be selected. The simulated neuron emits a single spike if it is not refractory and the membrane potential passes a threshold. A neuron enters a refractory period after a certain time of spike emission, during which no new spikes can be generated. The membrane potential V is governed by the following equation:

[0113]

number

[0114] g Na and g K are the Na and K constant leak conductances. E Na and E K is the reversal potential, which is assumed to be constant. I syn is the synaptic current from other neurons. I int is the characteristic current. Ire is the repolarizing current that is active only during the refractory period. m is the time constant.

[0115] The output of the biophysical model is electrophysiological data, either in the form of EEG / MEG-style time series or single neuron recording-style spike trains, depending on the type of neuron model chosen. In the example of our study, the output is a spike train from a single neuron. The simulated electrophysiological data can then be converted into fMRI BOLD time series using hemodynamic functions. A comprehensive model describing this process, named the "balloon model", is provided by Buxton et al. [3]. In this model, the complex relationship between blood oxygen levels and neuronal activity is proposed as a system of differential equations. The exact set of equations can be found in (7), where x is the neuronal activity, s is the vasodilatory signal, f is the blood inflow, κ is the signal decay constant, γ is a feedback regulation constant, v is the blood volume of the brain region, f is the blood flow rate, κ is the signal decay constant, γ is the feedback regulation constant, v is the blood volume of the brain region, f is the blood flow rate, κ is the signal decay constant, γ is the feedback regulation constant, v is the blood flow rate ... out is the blood outflow, τ is the mean transit time of blood, α is the vascular stiffness, q is the deoxyhemoglobin content, and E0 is the oxygen extraction fraction at rest.

[0116]

number

[0117] Under these underlying dynamics, the observed signal y(t), i.e., the time course of the voxel intensity, is proposed to be: where the coefficients k1, k2 and k3 have been considered in

[20] and are a combination of scanning and biological parameters:

[0118]

number

[0119] In this model, neuronal activity is seen to dilate blood vessels, resulting in increased blood inflow. The net blood flow directly affects blood volume, and the relationship between the final two shows a tight relationship between blood flow, volume, and deoxyhemoglobin content. The model transfers neuronal activity to fMRI BOLD signal changes. Thus, by comparing the fMRI time series and DCM results from the simulation with those from the experiment, one can directly compare the simulation and experimental results and modify the prior model accordingly or select a different prior model.

[0120] software Professor Karl Friston and his colleagues at the Wellcome Centre for Human Neuroimaging, University College London, have published a MATLAB toolbox named SPM (Statistical Parametric Mapping) that can be used for many functional neuroimaging applications [8]. DCM inversion and analysis is one of the features of this toolbox and is commonly used in the DCM literature. Our method is based on their toolbox.

[0121] References [1] Peter A Bandettini, Eric C Wong, R Scott Hinks, Ronald S Tikofsky, and James S Hyde. Time course epi of human brain function during task activation. Magnetic resonance in medicine, 25(2):390-397, 1992. [2]David Bernal-Casas,Hyun Joo Lee,Andrew J Weitz,and Jin Hyung Lee.Studying brain circuit function with dynamic causal modeling for optogenetic fmri.Neuron,93(3):522-532,2017. [3]Richard B Buxton,Eric C Wong,and Lawrence R Frank.Dynamics of blood flow and oxygenation changes during brain activation:the balloon model.Magnetic resonance in medicine,39(6):855-864,1998. [4]ManKin Choy,Ben A Duffy,and Jin Hyung Lee.Optogenetic study of networks in epilepsy.Journal of neuroscience research,95(12):2325-2335,2017. [5]Fernando Lopes da Silva.EEG and MEG:relevance to neuroscience.Neuron,80(5):1112-1128,2013.

[0122] [6]Karl J Friston,Lee Harrison,and Will Penny.Dynamic causal modelling.Neuroimage,19(4):1273-1302,2003. [7]Karl J.Friston,Joshua Kahan,Bharat Biswal,and Adeel Razi.A DCM for resting state fMRI.NeuroImage,94:396-407,2014. [8]K.J.Friston,J.Ashburner,S.J.Kiebel,T.E.Nichols,and W.D.Penny,editors.Statistical Parametric Mapping:The Analysis of Functional Brain Images.Academic Press,2007. [9]Gary H Glover.Overview of functional magnetic resonance imaging.Neurosurgery Clinics,22(2):133-139,2011.

[10] Sean Hill and Giulio Tononi.Modeling sleep and wakefulness in the thalamocortical system.Journal of neurophysiology,93(3):1671-1698,2005.

[0123]

[11] Donald R Humphrey and Edward M Schmidt.Extracellular single-unit recording methods.In Neuro-physiological techniques,pages 1-64.Springer,1990.

[12] Eugene M Izhikevich.Simple model of spiking neurons.IEEE Transactions on neural networks,14(6):1569-1572,2003.

[13] Hyun Joo Lee,Andrew J Weitz,David Bernal-Casas,Ben A Duffy,ManKin Choy,Alexxai V Kravitz,Anatol C Kreitzer,and Jin Hyung Lee.Activation of direct and indirect pathway medium spiny neurons drives distinct brain-wide responses.Neuron,91(2):412-424,2016.

[14] Baojuan Li,Jean Daunizeau,Klaas E.Stephan,Will Penny,Dewen Hu,and Karl Friston.Generalised filtering and stochastic DCM for fMRI.NeuroImage,58(2):442-457,2011.

[15] Andr´e C Marreiros,Stefan J Kiebel,and Karl J Friston.Dynamic causal modelling for fmri:a two-state model.Neuroimage,39(1):269-278,2008.

[0124]

[16] Hae-Jeong Park,Karl J Friston,Chongwon Pae,Bumhee Park,and Adeel Razi.Dynamic effective connectivity in resting state fmri.Neuroimage,180:594-608,2018.

[17] Stephen M Smith.The future of fmri connectivity.Neuroimage,62(2):1257-1266,2012.

[18] Harald Sontheimer and Christopher B Ransom.Whole-cell patch-clamp recordings.In Patch-Clamp Analysis,pages 35-67.Springer,2002.

[19] Klaas Enno Stephan,Lars Kasper,Lee M Harrison,Jean Daunizeau,Hanneke EM den Ouden,Michael Breakspear,and Karl J Friston.Nonlinear dynamic causal models for fmri.Neuroimage,42(2):649-662,2008.

[20] Klaas Enno Stephan,Nikolaus Weiskopf,Peter M Drysdale,Peter A Robinson,and Karl J Friston.Comparing hemodynamic models with dcm.Neuroimage,38(3):387-401,2007.

[0125]

[21] MA Tagamets and B Horwitz.Integrating electrophysiological and anatomical experimental data to cre-ate a large-scale model that simulates a delayed match-to-sample human brain imaging study.Cerebral cortex(New York,NY:1991),8(4):310-320,1998.

[22] Hugh R Wilson and Jack D Cowan.Excitatory and inhibitory interactions in localized populations of model neurons.Biophysical journal,12(1):1-24,1972.

[0126] STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT This invention was made with Government support under Contracts MH114227 and NS091461 awarded by the National Institutes of Health. The United States Government has certain rights in the invention.

Claims

1. 1. A method for modeling brain function, the method comprising: a) performing functional neuroimaging of neural activity in one or more regions of interest in a subject's brain; b) obtaining experimental electrophysiology data for one or more neurons in the one or more regions of interest in the brain of the subject; c) fitting the functional neuroimaging data of neural activity to a dynamic causal model (DCM); d) calculating inter- and intra-regional connectivity strength estimates from said DCM; and e) generating a biophysical model that simulates synthetic electrophysiological data using the inter- and intra-regional connectivity strength estimates; and f) comparing the synthetic electrophysiological data with the experimental electrophysiological data; g) iteratively adjusting the biophysical model until model convergence with the experimental electrophysiological data is achieved; A method comprising:

2. 2. The method of claim 1, wherein performing the functional neuroimaging comprises performing functional magnetic resonance imaging (fMRI), positron emission tomography (PET), functional near-infrared spectroscopy (fNIRS), single photon emission computed tomography (SPECT), or functional ultrasound imaging (fUS).

3. The method of claim 2 , wherein the fMRI is used to detect blood oxygen level dependent (BOLD) signals.

4. The method of any one of claims 1 to 3, wherein obtaining the experimental electrophysiological data comprises performing electroencephalography (EEG), magnetoencephalography (MEG), or patch clamp.

5. The method of claim 2, wherein the fMRI experimental data is optogenetic fMRI (ofMRI) experimental data.

6. The method of any one of claims 1 to 5, wherein the one or more regions of interest are in the cerebral, cerebellar, or brainstem regions of the brain.

7. 7. The method of claim 6, wherein the one or more regions of interest are in the basal ganglia, striatum, medulla, pons, hypothalamus, thalamus, superior colliculus, cerebral cortex, neocortex, isocortex, hippocampus, or olfactory bulb regions of the brain.

8. 8. The method of claim 7, wherein the one or more regions of interest are in the frontal, temporal, parietal, or occipital lobe regions of the brain.

9. 8. The method of claim 7, wherein the one or more regions of interest are in the caudate putamen, external pallidum, internal pallidum, subthalamic nucleus, substantia nigra, thalamus, or motor cortex regions of the brain.

10. 10. The method of any one of claims 1 to 9, wherein the one or more neurons are unipolar neurons, bipolar neurons, multipolar neurons, Golgi I neurons, Golgi II neurons, axonal neurons, pseudobipolar neurons, intermediate neurons, motor neurons, sensory neurons, efferent neurons, efferent neurons, cholinergic neurons, GABAergic neurons, glutamatergic neurons, dopaminergic neurons, serotonergic neurons, histaminergic neurons, Purkinje cells, spinous neurons, Renshaw cells, or granule cells, or combinations thereof.

11. The method of any one of claims 1 to 10, wherein the subject has a neurological disease or disorder.

12. 12. The method of claim 11, wherein the neurological disease or disorder is Parkinson's disease, epilepsy, or Alzheimer's disease.

13. 13. The method of claim 11 or 12, further comprising detecting changes in brain function in the subject based on said modeling of brain function compared to brain function in a control subject.

14. The method of any one of claims 11 to 13, further comprising identifying neural circuits involved in pathological changes associated with said neurological disease or disorder.

15. 14. The method of any one of claims 11 to 13, further comprising detecting a change in brain function in the subject after administration of a treatment for the neurological disease or disorder based on the modeling of the subject's brain function before and after administration of the treatment compared to the subject's brain function before the treatment.

16. The method of any one of claims 1 to 15, wherein the DCM is a classical DCM, a stochastic DCM, or a spectral DCM.

17. The method of any one of claims 1 to 16, wherein the functional neuroimaging or experimental electrophysiology data is acquired while the subject is in a resting state.

18. 17. The method of any one of claims 1 to 16, wherein the functional neuroimaging or experimental electrophysiology data is acquired while the subject is performing a task or responding to a stimulus.

19. 20. The method of claim 18, further comprising identifying neural circuitry involved in performing the task or responding to the stimulus.

20. A method for causing a computer to execute a process for modeling brain function, the computer executing steps comprising: a) receiving functional neuroimaging data and experimental electrophysiology data for one or more brain regions of interest in a subject; b) fitting the functional neuroimaging data to a dynamic causal model; and c) calculating inter- and intra-regional connectivity strength estimates from said dynamic causal model; and d) generating a biophysical model that simulates synthetic electrophysiological data from the inter- and intra-regional connectivity strength estimates; and e) comparing the synthetic electrophysiological data with the experimental electrophysiological data; f) iteratively adjusting the biophysical model until model convergence with the experimental electrophysiological data is achieved; and g) displaying information about the dynamic causal model, the biophysical model, or a combination thereof; A method comprising:

21. 21. The method of claim 20, wherein the functional neuroimaging data comprises functional magnetic resonance imaging (fMRI) data, positron emission tomography (PET) data, functional near-infrared spectroscopy (fNIRS) data, single photon emission computed tomography (SPECT) data, or functional ultrasound imaging (fUS) data.

22. 22. The method of claim 20 or 21, wherein the experimental electrophysiology data comprises electroencephalography (EEG) data, magnetoencephalography (MEG) data, or patch clamp data.

23. The method of any one of claims 20 to 22, wherein the subject has a neurological disease or disorder.

24. 24. The method of claim 23, wherein displaying the information comprises listing one or more neural circuits involved in pathological changes associated with the neurological disease or disorder.

25. 25. The method of any one of claims 20-24, wherein the functional neuroimaging data and the experimental electrophysiology data are acquired while the subject is performing a task or responding to a stimulus, and wherein displaying the information comprises enumerating one or more neural circuits involved in performing the task or responding to the stimulus.

26. A system for modelling brain function using a method according to any one of claims 20 to 25, said system comprising: a) a storage component for storing data, the storage component having instructions for modeling brain function based on the functional neuroimaging data and experimental electrophysiology data stored therein; b) a computer processor for processing the functional neuroimaging data and the experimental electrophysiology data using one or more algorithms, the computer processor being coupled to the storage component and configured to execute the instructions stored in the storage component in order to read the functional neuroimaging data and the experimental electrophysiology data from the storage component and to process the data according to the method of any one of claims 20 to 25; c) a display component for displaying the information regarding the dynamic causal model, the biophysical model, or a combination thereof; A system comprising:

27. A non-transitory computer readable medium comprising program instructions which, when executed by a processor in a computer, cause the processor to perform the method of any one of claims 20 to 25.

28. A program for causing a computer to execute a process for modeling brain function, a) receiving functional neuroimaging data and experimental electrophysiology data for one or more brain regions of interest in a subject; b) fitting the functional neuroimaging data to a dynamic causal model; and c) calculating inter- and intra-regional connectivity strength estimates from said dynamic causal model; and d) generating a biophysical model that simulates synthetic electrophysiological data from the inter- and intra-regional connectivity strength estimates; and e) comparing the synthetic electrophysiological data with the experimental electrophysiological data; f) iteratively adjusting the biophysical model until model convergence with the experimental electrophysiological data is achieved; and g) displaying information about the dynamic causal model, the biophysical model, or a combination thereof; Including, the program.

Images