System, method and computer program product for co-mapping multi-modal brain imaging techniques using artificial intelligence (AI)

Abstract

Described herein are methods, systems and computer program products for individualized, non-invasive brain stimulation for enhanced learning. In some instances, high spatial-resolution and / or high temporal-resolution brain data is predicted based on new low spatial-resolution and / or new low temporal-resolution brain data using trained artificial intelligence (AI).

Description

SYSTEM, METHOD AND COMPUTER PROGRAM PRODUCT FOR COMAPPING MULTI-MODAL BRAIN IMAGING TECHNIQUES USING ARTIFICIAL INTELLIGENCE (Al)CROSS REFERENCE TO RELATED APPLICATION

[0001] This application claims priority to and benefit of U.S. provisional patent application number 63 / 536,616 filed September 5, 2023, which is fully incorporated by reference and made a part hereof.GOVERNMENT SUPPORT CLAUSE

[0002] This invention was made with government support under Grant No. OD031843 awarded by the National Institutes of Health. The government has certain rights in the invention.BACKGROUND

[0003] Currently, brain stimulation approaches are largely non-individualized, despite that fact that there are large, dynamic differences in how peoples' brains process information. Individualized approaches to date have used single brain imaging modalities (e.g. magnetic resonance imaging (MRI) alone) to generate targets or other parameters. Furthermore, individualized approaches are difficult to employ in clinical settings for nonscientist use.

[0004] Also, a common problem in brain imaging is the inability to integrate high spatial resolution with high temporal resolution. While modalities such as electroencephalography (EEG) have the ability to identify brain signals at the millisecond level, they are unable to localize that signal to the source brain area. EEG systems are relatively inexpensive and mobile, making them the most applicable brain imaging approach in real- world applications. On the other hand, imaging modalities like magnetic resonance imaging (MRI) can identify where brain processes are occurring with a high degree of spatial accuracy (millimeter), but are unable to measure rapid brain processes due to methodological and physiological restrictions of the BOLD signal. Additionally, while MRI systems have the highest spatial resolution of any non-invasive brain imaging approach, MRI's are expensive and immobile.

[0005] Therefore, what are needed are systems, methods and computer program products that overcome challenges in the art, some of which are described above. Specifically, what is desired is a system, method and computer program product that provides an accessible platform for non-scientists to get brain signatures of a task, and translate them into stimulation application parameters to increase or suppress neuronal activation in order to enhance cognition. Brain imaging approaches that increase the spatial and temporal-resolution additionally allow for individualized, dynamic brain stimulation, informed by brain data at a preferred resolution of milliseconds and millimeters. This can be applied in any clinical group or cognitive domain that justifies non-invasive brain stimulation. This can also specifically be applied in environments that don't have direct scientific support. There is also a need to identify methods that merge information from different brain imaging modalities in a unified and mobile system.SUMMARY

[0006] One aspect described and disclosed herein are embodiments of systems, methods and computer program products for individualized, non-invasive brain stimulation for enhanced learning. One embodiment of the method comprises obtaining brain data of a person; identifying one or more areas of a brain of the person for stimulation based on the brain data; and stimulating the identified one or more areas of the brain, wherein the stimulation results in enhanced learning. In some instances, the brain data are obtained while the person is performing a learning or other cognitive activities such as emotional regulation, semantic and episodic memory, and other memory processes, executive functions, including but not limited to working memory, cognitive control, inferential processes, prediction error, reading and language comprehension / production, social cognition, vision and auditory, motor processing, and the like. The learning or cognitive activity may be a specific learning activity such as factual learning or skill learning. In some instances, the person may not be performing any task, and / or be in a resting state. In some instances, the brain data comprise high spatial- resolution brain data and / or high temporal-resolution brain data. In some instances, the high spatial -resolution brain data and / or the high temporal-resolution brain data is predicted from low-spatial-resolution brain data and / or low temporal-resolution brain data using a trained artificial intelligence (Al) model.

[0007] Another aspect described herein are embodiments of systems, methods and computer program products for predicting high spatial-resolution brain data and / or hightemporal-resolution brain data from low-spatial-resolution brain data and / or low temporalresolution brain data using a trained Al model. One embodiment of the method comprises training the Al model by co-mapping existing low spatial-resolution brain data and / or low temporal-resolution brain data and existing high spatial-resolution brain data and / or existing high temporal-resolution brain data and using transformers and / or other Al architectures to identify patterns of co-occurrence between the existing low-spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data; obtaining new low spatial- resolution brain data and / or new low temporal-resolution brain data for the person; applying the trained Al model to the new low spatial-resolution brain data and / or the new low temporalresolution brain data; and predicting new high spatial-resolution brain data and / or high temporal-resolution brain data from the new low spatial-resolution brain data and / or the new low temporal -resolution brain data using the trained Al model. Other Al architectures that can be used include neural architectures such as neural networks, recurrent neural networks (RNN), long short-term memory networks (LSTMs), convolutional neural networks (CNN), hybrid combinations of CNNs, RNNs and others, state space machines, autoencoders, generative adversarial networks (GANs), graph neural networks, multi-modal fusion networks, and the like.

[0008] These systems, methods and computer program products for predicting high spatial-resolution brain data and / or high temporal-resolution brain data from low-spatial- resolution brain data and / or low temporal-resolution brain data using a trained Al model merge information from different brain imaging modalities in a unified and mobile system. This approach will have broad impact on the any industry that uses brain imaging. Currently, consumer use of brain imaging devices (including at-home devices, clinical devices, military devices) are largely restricted to low-resolution EEG solutions. This incomplete picture of the brain severely limits the efficacy of this market. Regardless of cognitive domain, brain processes are 1.) rapid, 2.) widespread, and 3.) highly individualized. High spatial resolution is necessary for effective brain-based interventions, including non-invasive brain stimulation and biofeedback applications. Additionally, embodiments disclosed herein would provide an extreme cost-cutting measure for hospitals, allowing for in-depth examination of brain-based medical issues without the need for MRI.

[0009] It should be understood that the above-described subject matter may also be implemented as a computer-controlled apparatus, a computing system, or an article ofmanufacture, such as a computer program product stored on a non-transient computer-readable storage medium.

[0010] Other systems, methods, features and / or advantages will be or may become apparent to one with skill in the art upon examination of the following drawings and detailed description. It is intended that all such additional systems, methods, features and / or advantages be included within this description and be protected by the accompanying claims.BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments and together with the description, serve to explain the principles of the methods and systems:FIG. 1 illustrates an example of fused MRI / EEG analysis, implemented to track brain networks over the milliseconds after a word is displayed;FIGS. 2A and 2B illustrate an example of how fused MRI / EEG can pinpoint where and when information processing can break down during reading, resulting in lower comprehension; FIG. 3A is an example flowchart illustrating steps performed in performing individualized, non-invasive brain stimulation for enhanced learning;FIG. 3B is an example flowchart illustrating a method of predicting high spatial- resolution brain data and / or the high temporal-resolution brain data is predicted from low- spatial-resolution brain data and / or low temporal-resolution brain data using a trained artificial intelligence (Al) model; andFIG. 4 is a block diagram of an example computing device upon which embodiments of the invention may be implemented.DETAILED DESCRIPTION

[0012] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art. Methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present disclosure.

[0013] As used in the specification and the appended claims, the singular forms “a,” “an,” “the,” and “data” include plural referents unless the context clearly dictates otherwise. Ranges may be expressed herein as from “about” one particular value, and / or to “about” another particular value. When such a range is expressed, another embodiment includes fromthe one particular value and / or to the other particular value. Similarly, when values are expressed as approximations, by use of the antecedent “about,” it will be understood that the particular value forms another embodiment. It will be further understood that the endpoints of each of the ranges are significant both in relation to the other endpoint, and independently of the other endpoint.

[0014] “Optional” or “optionally” means that the subsequently described event or circumstance may or may not occur, and that the description includes instances where said event or circumstance occurs and instances where it does not.

[0015] Throughout the description and claims of this specification, the word “comprise” and variations of the word, such as “comprising” and “comprises,” means “including but not limited to,” and is not intended to exclude, for example, other additives, components, integers or steps. “Exemplary” means “an example of’ and is not intended to convey an indication of a preferred or ideal embodiment. “Such as” is not used in a restrictive sense, but for explanatory purposes.

[0016] Furthermore, as used herein, the terms “word” or “words” includes a complete word, only a portion of a word, or a word segment.

[0017] Disclosed are components that can be used to perform the disclosed methods and systems. These and other components are disclosed herein, and it is understood that when combinations, subsets, interactions, groups, etc. of these components are disclosed that while specific reference of each various individual and collective combinations and permutation of these may not be explicitly disclosed, each is specifically contemplated and described herein, for all methods and systems. This applies to all aspects of this application including, but not limited to, steps in disclosed methods. Thus, if there are a variety of additional steps that can be performed it is understood that each of these additional steps can be performed with any specific embodiment or combination of embodiments of the disclosed methods.

[0018] The present methods and systems may be understood more readily by reference to the following detailed description of preferred embodiments and the Examples included therein and to the Figures and their previous and following description.

[0019] Disclosed herein is a system, method and computer-program product that converts an individual's high-resolution brain patterns into an individualized stimulation protocol. Generally, a user of the disclosed a system, method and computer-program product is naive to neuroimaging processing, and only needs to input a subject identifier to get individualized stimulation parameters for optimal stimulation gains.

[0020] The disclosed system, method and computer-program product may be configured to automatically obtain a subject's functional and / or structural brain imaging data (e.g. from one or more modalities including for example structural and / or functional magnetic resonance imaging (MRI), electroencephalogram (EEG), functional near-infrared spectroscopy (fNIRS), magnetoencephalography (MEG), or other devices or combinations of devices that measure brain structure or activation) as they view a standardized learning and / or other cognitive paradigm, along with the subject's behavioral learning and / or cognitive patterns (e.g. what they recall from the paradigm), and / or as structural / functional brain data is collected. The disclosed system, method and computer-program product then applies automated preprocessing and modelling algorithms in individual or combined brain imaging modalities (in Matlab, for instance using SPM and EEGLab packages). One non-limiting example of this would be the use of joint independent component analysis (jlCA) of EEG / MRI data to identify where and when learning happens while an individual is reading new medical facts for a test. In a current use-case, individuals study material for the Army’s Expert Field Medical Badge Test while they are in the MRI and while they are wearing the EEG. An example is provided below of using use joint independent component analysis of fMRI and EEG to track the rapid (millisecond) interactions of widespread cortical networks necessary for semantic cognition beyond the single word, and identify which networks and network interactions are most critical for language comprehension (LC) ability. This example is further described in Katherine S. Aboud, Tin Q. Nguyen, Stephanie N. Del Tufo, Catie Chang, David H. Zald, Alexandra P. Key, Gavin R. Price, Bennett A. Landman and Laurie E. Cutting; Journal of Neuroscience 16 November 2022, JN-RM-0529-21; DOI: https: / / doi.org / 10.1523 / JNEURQSCI.0529-21.2022, which is fully incorporated by reference and made a part hereof. jlCA co-estimates MRI and EEG components. In so doing, it identifies detailed MRI spatial maps and corresponding time courses, leading to an understanding of where and when brain processes of learning occur in that individual, for instance where and what parts of the brain are most active or most suppressed when a person learns a new medical fact. FIG. 1 illustrates an example of fused MRIZEEG analysis, implemented to track brain networks over the milliseconds after a word is displayed. The peak of the EEG waveform indicates when the peak of the network activation occurs. Here, it can be seen that after a word is displayed, temporal brain regions activate -300 milliseconds after the word (Joint Component 1), followed by fronto-temporal language areas at -400 milliseconds (Joint Component 2), proceeding across other cortical networks until finally arriving at higher order comprehension areas at -800 ms (Joint Component 5). Thisdegree of spatial and temporal-resolution cannot be gained through other imaging modalities / analyses. The disclosed system, method and computer-program product allows for both single- and multi-modality identification of learning in the brain, depending on available technology of the user. The disclosed system, method and computer-program product then identifies which fused MRI / EEG output is most predictive of the subject's successful online learning of information.

[0021] FIGS. 2A and 2B illustrate an example of how fused MRI / EEG can pinpoint where and when information processing can break down during reading, resulting in lower comprehension. If multiple brain areas are identified as important to learning, filtering algorithms are used to identify the top areas that contribute to learning (e.g. quantifying the brain areas based on their global connectivity during task, using measures of the brain regions’ contribution to the spatial map, or using a priori information that points to the most important areas in the spatial map for the task, among other approaches). FIG. 2A illustrates a comparison of single subject grand averages of event-related potential (ERP) data for two representative subjects with high (red) and low (blue) LC scores. The group spatial components and related time windows (indicated by box length) for P300 (light blue), N400 (orange), and P600c (green) are overlaid onto the subject ERP responses, demonstrating the dynamic relationship between the significant components, and how those dynamics contribute to LC ability (the P300, N400, and P600). FIG. 2B illustrates mediation analysis revealed that default mode network (DMN) comprehension areas activated -750 ms post-stimulus (JC5; green box) are partially mediated the relationship between early semantic-phonological areas activated at -300 ms post-stimulus (JC1; blue box) and LC ability. The disclosed system, method and computer-program product then identifies the non-invasive brain stimulation electrode array that most closely targets the brain areas identified from the fused MRI / EEG analysis. Stimulation parameters include intensity, frequency and / or duration. Intensity ranges from -1- 4 mA in total across multiple electrodes. Frequency can be individualized within a predetermined frequency band depending on task (e.g. 5-7 Hz for memory paradigms). Stimulation sends current through the scalp from an anode to a cathode to either facilitate or inhibit neuronal activity. Some stimulation approaches, such as tDCS, directly act on the neurons’ action potentials. Stimulation can also be used to elicit groups of neurons to fire together at a certain frequency (using tACS), which impacts communication across brain areas. In some instances, the stimulation may vary over time, while in others the stimulation is consistent over the time applied. In some instances, this may be done by interfacing thedisclosed system, method and computer-program product with existing software such as, for example, HD-Targets and HD-Explore by Soterix Medical. To identify the best fit electrode array, the disclosed system, method and computer-program product contribute an algorithm to HD-Targets / HD-Explore that iteratively tests electric field maps (e.g. the simulated impact of stimulation from an electrode array) to find the array with the closest spatial match to the jICA targets. The user consequently gets an electrode array after inputting the subject identifier. The disclosed system, method and computer-program product also provides a subject's specific peak frequency value within each frequency band (from EEG), which will be used as another parameter in non-invasive brain stimulation. For all processes, cloud computing may be used to allow for rapid outputs.

[0022] The disclosed system, method and computer-program product comprise a pipeline that can be applied to many brain domains, including motor learning, attention, visual processing, etc. In some instances, the disclosed system, method and computer-program product include time-variant targets for non-invasive brain stimulation. jICA and Al comapping allow for detailed information into the online dynamics of brain activations related to a task. Currently, the majority of non-invasive brain stimulation techniques consistently target the same brain areas over the course of a simulation session, despite the fact that brain activations dynamically and flexibly change during tasks in a way that is specific to an individual. The disclosed system, method and computer-program product can, in some instances, be used to input time-variant stimulation protocols.

[0023] In some instances, the disclosed system, method and computer-program product can be deployed to whole classrooms. Once a person's EEG-MRI alignment is established, classroom applications can employ real-time adjustments to non-invasive brain stimulation based on the person's online EEG patterns alone. For instance, it can be identified that EEG pattern X is related to MRI pattern Y, while EEG pattern A is related to MRI pattern B. If it is desired to enhance areas in Y when they occur, but suppress areas in B when they occur, then online EEG measurements of X and A (from the subject's previous MRI-EEG mapping) can be used to make online stimulation target adjustments.

[0024] FIG. 3A illustrates an exemplary flowchart for performing a method of individualized, non-invasive brain stimulation for enhanced learning. At 302, brain data of a person is obtained. In some instances, brain data are obtained while the person is performing a learning or other cognitive activity, not performing a task, or at rest. For example, the learning activity may be a specific learning activity such as medical learning, marksmanship, and thelike. Generally, though not required, the brain data comprise high-resolution brain data (defined as data with spatial and temporal information at a preferred resolution on the millisecond and millimeter scale). For example, the high resolution brain imaging may comprise functional and / or structural magnetic resonance imaging (MRI) and / or electroencephalogram (EEG) data, among other modalities. However, a challenge in brain imaging is the inability to integrate high spatial resolution with high temporal resolution. While modalities such as electroencephalography (EEG) have the ability to identify brain signals at the millisecond level, they are unable to localize that signal to the source brain area. EEG systems are relatively inexpensive and mobile, making them the most applicable brain imaging approach in real- world applications. On the other hand, imaging modalities like MRI can identify where brain processes are occurring with a high degree of spatial accuracy (millimeter), but are unable to measure rapid brain processes due to methodological and physiological restrictions of the blood-oxygen-level-dependent (BOLD) signal. Additionally, while MRI systems have the highest spatial resolution of any non-invasive brain imaging approach, MRI's are expensive and immobile. There is consequently a need to identify methods that merge information from different brain imaging modalities in a unified and mobile system.

[0025] Therefore, it would be advantageous and useful to be able to predict high spatial-resolution brain data (such as functional and / or structural MRI) and / or the high temporal-resolution brain data from low-spatial-resolution brain data and / or low temporalresolution brain data using a trained artificial intelligence (Al) model. For example, using Al to co-map EEG and MRI data, results in an algorithm that predicts high-resolution, full-brain MRI signals from real-time EEG / fNIRS data alone to identify patterns of co-occurrence between existing EEG and MRI data. In some instances, the Al may use transformers to identify patterns of co-occurrence between existing EEG and MRI data. The resulting model can then be applied to new EEG / fNIRS data and predict functional MRI signals and / or structural MRI from the low-spatial resolution data, without actually collecting MRI data.

[0026] As such, FIG. 3B illustrates a flowchart for a method of predicting high spatial -resolution brain data and / or the high temporal-resolution brain data is predicted from low-spatial-resolution brain data and / or low temporal-resolution brain data using a trained artificial intelligence (Al) model. As shown in FIG. 3B, the method includes 308, training the Al model by co-mapping existing low spatial-resolution brain data and / or low temporalresolution brain data and existing high spatial-resolution brain data and / or existing high temporal-resolution brain data and using transformers and / or other Al architectures to identifypatterns of co-occurrence between the existing low-spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal -resolution brain data. At 310, new low spatial -resolution brain data and / or new low temporal-resolution brain data is obtained for the person. At 312, the trained Al model is applied to the new low spatial-resolution brain data and / or the new low temporal-resolution brain data. And, at 314, new high spatial-resolution brain data and / or high temporal-resolution brain data is predicted from the new low spatial-resolution brain data and / or the new low temporal-resolution brain data using the trained Al model.

[0027] In one example, training the Al model by co-mapping the existing low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data and using transformers and / or other Al architectures to identify patterns of co-occurrence between the low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporalresolution brain data comprises training the Al model by co-mapping existing functional magnetic resonance imaging (fMRI) data and / or structural MRI data, and existing electroencephalogram (EEG) data and functional near-infrared spectroscopy (fNIRS) data and using the transformers and / or other Al architectures to identify the patterns of co-occurrence between the existing fMRI data and / or structural MRI data and the existing EEG and fNIRS data. In such an example, obtaining the new low-resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining new EEG and fNIRS data for the person. Applying the trained Al model to the new low-spatial-resolution brain data and / or the new low temporal-resolution brain data comprises applying the trained Al model to the new EEG and fNIRS data. Then, predicting the new high spatial-resolution brain data and / or the new high temporal-resolution brain data from the new low spatial-resolution brain data and / or the new low temporal-resolution brain data using the trained Al model comprises predicting new fMRI data and / or structural MRI data of the person from the new EEG and fNIRS data using the trained Al model.

[0028] In another example, training the Al model by co-mapping the existing low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data and using transformers or other Al architecture to identify patterns of co-occurrence between the low spatial-resolution brain data and / or the existing low temporal-resolution braindata and the existing high spatial-resolution brain data and / or the existing high temporalresolution brain data comprises training the Al model by co-mapping existing functional magnetic resonance imaging (fMRI) data and / or structural MRI data, and existing electroencephalogram (EEG) data and using the trained Al model to identify the patterns of co-occurrence between the existing fMRI data and the existing EEG data. In some instances, the trained Al model may employ transformers to identify the patterns of co-occurrence. Other Al architectures that can be used include neural architectures such as neural networks, recurrent neural networks (RNN), long short-term memory networks (LSTMs), convolutional neural networks (CNN), hybrid combinations of CNNs, RNNs and others, state space machines, autoencoders, generative adversarial networks (GANs), graph neural networks, multi-modal fusion networks, and the like. In such an example, obtaining the new low-resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining new EEG data for the person. Applying the trained Al model to the new low-spatial-resolution brain data and / or the new low temporal-resolution brain data comprises applying the trained Al model to the new EEG data. Then, predicting the new high spatial-resolution brain data and / or the new high temporal-resolution brain data from the new low spatial-resolution brain data and / or the new low temporal-resolution brain data using the trained Al model comprises predicting new fMRI data and / or structural MRI data of the person from the new EEG data using the trained Al model.

[0029] In yet another example, training the Al model by co-mapping the existing low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data and using transformers or other Al architectures to identify patterns of co-occurrence between the low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporalresolution brain data comprises training the Al model by co-mapping existing functional magnetic resonance imaging (fMRI) data and / or structural MRI data and existing functional near-infrared spectroscopy (fNIRS) data and using the transformers to identify the patterns of co-occurrence between the existing fMRI data and / or structural MRI data and the existing fNIRS data. In such an example, obtaining the new low-resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining new fNIRS data for the person. Applying the trained Al model to the new low-spatial-resolution brain data and / or the new low temporal-resolution brain data comprises applying the trained Al model to the newfNIRS data. Then, predicting the new high spatial-resolution brain data and / or the new high temporal-resolution brain data from the new low spatial-resolution brain data and / or the new low temporal-resolution brain data using the trained Al model comprises predicting new fMRI data and / or structural MRI data of the person from the new fNIRS data using the trained Al model.

[0030] In another example, training the Al model by co-mapping the existing low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data and using transformers or other Al architectures to identify patterns of co-occurrence between the low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporalresolution brain data comprises training the Al model by co-mapping existing functional magnetic resonance imaging (fMRI) data and / or structural MRI data and existing electroencephalogram (EEG) and / or functional near-infrared spectroscopy (fNIRS) data and using the transformers or other Al architecture to identify the patterns of co-occurrence between the existing fMRI data and / or structural MRI data and the existing EEG and / or fNIRS data. In such an example, obtaining the new low-resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining new fMRI data and / or structural MRI data for the person. Applying the trained Al model to the new low-spatial- resolution brain data and / or the new low temporal-resolution brain data comprises applying the trained Al model to the new fMRI data and / or structural MRI data. Then, predicting the new high spatial-resolution brain data and / or the new high temporal-resolution brain data from the new low spatial-resolution brain data and / or the new low temporal-resolution brain data using the trained Al model comprises predicting new EEG and / or fNIRS data of the person from the new fMRI data and / or structural MRI data using the trained Al model.

[0031] In another example, training the Al model by co-mapping the existing low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data and using transformers or other Al architectures to identify patterns of co-occurrence between the low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporalresolution brain data comprises training the Al model by co-mapping existing electroencephalogram (EEG) and functional near-infrared spectroscopy (fNIRS) data andusing the transformers or other Al architectures to identify the patterns of co-occurrence between the existing EEG and the existing fNIRS data. In such an example, obtaining the new low-resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining new EEG data for the person. Applying the trained Al model to the new low-spatial-resolution brain data and / or the new low temporal-resolution brain data comprises applying the trained Al model to the new EEG data. Then, predicting the new high spatial- resolution brain data and / or the new high temporal -re solution brain data from the new low spatial-resolution brain data and / or the new low temporal-resolution brain data using the trained Al model comprises predicting new fNIRS data of the person from the new EEG data using the trained Al model.

[0032] In any of the above examples, obtaining the existing low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial- resolution brain data and / or the existing high temporal-resolution brain data may comprise obtaining the existing low spatial-resolution brain data and / or the existing low temporalresolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal -resolution brain data simultaneously or in parallel. Similarly, obtaining the new low spatial-resolution brain data and / or the new low temporal-resolution brain data for the person may comprise obtaining the new low spatial-resolution brain data and / or the new low temporalresolution brain data for the person simultaneously or in parallel.

[0033] Furthermore, in any of the above examples, using the transformers to identify patterns of co-occurrence between the existing low-spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data comprises using transformer model architecture to perform cross-modal training in brain imaging data from different imaging modalities (e.g. EEG, fNIRS, and fMRI) to make predictions of how these imaging modalities correspond with one another. The cross-modal training comprises combining various pairings of imaging modality (e.g., MRIZEEG / fNIRS) and single- and / or cross-modal training approaches within the transformer model architecture. The single- and / or cross-modal training approaches within the transformer model architecture include masked signal modelling, causal- signal modelling, cross-attention, multi-modal fusion, multi-modal pre-training, late fusion, and early fusion, among others.

[0034] In any of the above examples, obtaining the new low spatial-resolution brain data and / or the new low temporal-resolution brain data for the person comprises usingcontinuous or discrete temporal and / or spatial signals from the new low spatial-resolution brain data and / or the new low temporal-resolution brain data. The new low spatial-resolution brain data and / or the new low temporal-resolution brain data is obtained using brain-imaging modalities and the continuous or discrete temporal and / or spatial signals are restricted based on the brain imaging modality used (for instance, only inputting spatial signals from fNIRS and temporal signals from EEG).

[0035] In any of the above examples, obtaining the new low spatial-resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining the new low spatial-resolution brain data and / or the new low temporal-resolution brain data for the person using whole-brain measurements (e.g. voxel), or region of interest measurements (e.g. atlas region in the brain).

[0036] One example of how this transformer model architecture work comprises the use of encoder / decoder transformer architecture, in which each brain imaging modality is first trained separately using modality-specific modeling (e.g. training an EEG model using masked signal modeling; training an fMRI model using causal signal modeling). These models train on time course data (including both spatial and temporal information), to develop a prediction of the next time point of signal data in space for each modality. After this training, attention mechanisms (e.g. cross-attention) are used to develop predictions across modalities (e.g. using EEG data to predict MRI data). Additional unlabeled data can then be input to generate cross-modal predictions, with input only including one modality from an individual (e.g. EEG data being used to predict fMRI patterns and / or structural MRI data without fMRI data and / or structural MRI data being collected). Labelled data can be used to continually optimize the model. This process can be done on the whole brain scale, or dimensionality can be reduced using predefined regions of interest. This can be accomplished within an individual subject or across multiple subjects, during tasks or at rest.

[0037] Returning to FIG. 3 A, at 304, one or more areas of a brain of the person are identified for stimulation based on the brain data. In instances where the brain data comprise MRI and / or EEG data, identifying the one or more areas of the brain of the person for stimulation comprises performing joint independent component analysis (jlCA) of the MRI and EEG data. The jlCA analysis identifies one or more detailed MRI spatial maps and corresponding time courses, leading to a high resolution understanding of brain processes. At least one of the one or more detailed MRI spatial maps and corresponding time courses is identified as most predictive of the subject's successful online learning of information. Theidentified one detailed MRI spatial map and corresponding time course is masked to contain only one or more greatest contributing brain areas (based either on weighting value of the area, or a priori characterization of the region). A non-invasive brain stimulation electrode array that most closely targets the one or more greatest contributing brain areas is identified from the fused MRIZEEG analysis. For example, identifying the non-invasive brain stimulation electrode array that most closely targets the one or more greatest contributing brain areas identified from the fused MRIZEEG analysis may be done by using HD-Targets and HD- Explore by Soterix Medical. Using HD-Targets and HD-Explore by Soterix Medical to identify a best fit electrode array generally comprises contributing an algorithm to HD- Targets / HD-Explore that iteratively tests electric field maps (e.g. the simulated impact of stimulation from an electrode array) to find the electrode array with the closest spatial match to thejICA targets.

[0038] At 306, the identified one or more areas of the brain are stimulated using the identified electrode array wherein the stimulation results in enhanced learning. In some instances, the person’s specific peak frequency value within each frequency band (from EEG) may be used as another parameter in the non-invasive brain stimulation. Furthermore, in some instances, the stimulation for stimulating the identified one or more areas of the brain varies over time. Such variance may occur in the form of varying peak voltage, frequencies, duration, and the like.

[0039] It is to be appreciated that the above described steps can be performed by computer-readable instructions executed by a processor or one or more processors. As used herein, a processor is a physical, tangible device used to execute computer-readable instructions. Furthermore, “processor,” as used herein, may be used to refer to a single processor, or it may be used to refer to one or more processors.

[0040] When the logical operations described herein are implemented in software, the process may be executed on any type of computing architecture or platform. For example, referring to FIG. 4, an example computing device upon which embodiments of the invention may be implemented is illustrated. In particular, at least one processing device described above may be a computing device, such as computing device 1000 shown in FIG. 4. For example, computing device 1000 may be a component of the cloud computing and storage system. Computing device 1000 may comprise all or a portion of server. The computing device 1000 may include a bus or other communication mechanism for communicating information among various components of the computing device 1000. In its most basic configuration, computingdevice 1000 typically includes at least one processing unit 1006 and system memory 1004. Depending on the exact configuration and type of computing device, system memory 1004 may be volatile (such as random access memory (RAM)), non-volatile (such as read-only memory (ROM), flash memory, etc.), or some combination of the two. This most basic configuration is illustrated in FIG. 4 by dashed line 1002. The processing unit 1006 may be a standard programmable processor that performs arithmetic and logic operations necessary for operation of the computing device 1000.

[0041] Computing device 1000 may have additional features / functionality. For example, computing device 1000 may include additional storage such as removable storage 1008 and non-removable storage 1010 including, but not limited to, magnetic or optical disks or tapes. Computing device 1000 may also contain network connect! on(s) 1016 that allow the device to communicate with other devices. Computing device 1000 may also have input device(s) 1014 such as a keyboard, mouse, touch screen, scanner, etc. Output device(s) 1012 such as a display, speakers, printer, etc. may also be included. The additional devices may be connected to the bus in order to facilitate communication of data among the components of the computing device 1000. All these devices are well known in the art and need not be discussed at length here. Though not shown in FIG. 4, in some instances computing device 1000 includes an interface. The interface may include one or more components configured to transmit and receive data via a communication network, such as the Internet, Ethernet, a local area network, a wide-area network, a workstation peer-to-peer network, a direct link network, a wireless network, or any other suitable communication platform. For example, interface may include one or more modulators, demodulators, multiplexers, demultiplexers, network communication devices, wireless devices, antennas, modems, and any other type of device configured to enable data communication via a communication network. Interface may also allow the computing device to connect with and communicate with an input or an output peripheral device such as a scanner, printer, and the like.

[0042] The processing unit 1006 may be configured to execute program code encoded in tangible, computer-readable media. Computer-readable media refers to any media that is capable of providing data that causes the computing device 1000 (i.e., a machine) to operate in a particular fashion. Various computer-readable media may be utilized to provide instructions to the processing unit 1006 for execution. Common forms of computer-readable media include, for example, magnetic media, optical media, physical media, memory chips or cartridges, a carrier wave, or any other medium from which a computer can read. Examplecomputer-readable media may include, but is not limited to, volatile media, non-volatile media and transmission media. Volatile and non-volatile media may be implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data and common forms are discussed in detail below. Transmission media may include coaxial cables, copper wires and / or fiber optic cables, as well as acoustic or light waves, such as those generated during radio-wave and infra-red data communication. Example tangible, computer-readable recording media include, but are not limited to, an integrated circuit (e.g., field-programmable gate array or application-specific IC), a hard disk, an optical disk, a magneto-optical disk, a floppy disk, a magnetic tape, a holographic storage medium, a solid-state device, RAM, ROM, electrically erasable program read-only memory (EEPROM), flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices.

[0043] In an example implementation, the processing unit 1006 may execute program code stored in the system memory 1004. For example, the bus may carry data to the system memory 1004, from which the processing unit 1006 receives and executes instructions. The data received by the system memory 1004 may optionally be stored on the removable storage 1008 or the non-removable storage 1010 before or after execution by the processing unit 1006.

[0044] Computing device 1000 typically includes a variety of computer-readable media. Computer-readable media can be any available media that can be accessed by device 1000 and includes both volatile and non-volatile media, removable and non-removable media. Computer storage media include volatile and non-volatile, and removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. System memory 1004, removable storage 1008, and non-removable storage 1010 are all examples of computer storage media. Computer storage media include, but are not limited to, RAM, ROM, electrically erasable program read-only memory (EEPROM), flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store the desired information and which can be accessed by computing device 1000. Any such computer storage media may be part of computing device 1000.

[0045] It should be understood that the various techniques described herein may be implemented in connection with hardware or software or, where appropriate, with a combination thereof. Thus, the methods and apparatuses of the presently disclosed subject matter, or certain aspects or portions thereof, may take the form of program code (i.e., instructions) embodied in tangible media, such as floppy diskettes, CD-ROMs, hard drives, or any other machine-readable storage medium wherein, when the program code is loaded into and executed by a machine, such as a computing device, the machine becomes an apparatus for practicing the presently disclosed subject matter. In the case of program code execution on programmable computers, the computing device generally includes a processor, a storage medium readable by the processor (including volatile and non-volatile memory and / or storage elements), at least one input device, and at least one output device. One or more programs may implement or utilize the processes described in connection with the presently disclosed subject matter, e.g., through the use of an application programming interface (API), reusable controls, or the like. Such programs may be implemented in a high level procedural or object-oriented programming language to communicate with a computer system. However, the program(s) can be implemented in assembly or machine language, if desired. In any case, the language may be a compiled or interpreted language and it may be combined with hardware implementations.

[0046] References

[0047] Unless otherwise noted, each of the below references are fully incorporated by reference and made a part hereof:

[0048] Aboud, K. S., Bailey, S. K., Petrill, S. A., & Cutting, L. E. (2016). Comprehending text versus reading words in young readers with varying reading ability: Distinct patterns of functional connectivity from common processing hubs. Developmental Science, 1-25. https: / / doi.org / 10. l l l l / desc.12422

[0049] Ager, R. & De Boeck, P (2017). On the Interpretation and Use of Mediation: Multiple Perspectives on Mediation Analysis. Front. Psych. https: / / doi.org / 10.3389 / fpsyg.2017.01984

[0050] Artoni, F., Menicucci, D., Delorme, A., Makeig, S., & Micera, S. (2014). Full Length Articles RELICA: A method for estimating the reliability of 3 independent components 4Q1. https: / / doi.Org / 10.1016 / j.neuroimage.2014.09.010

[0051] Azizian, A., Freitas, A. L., Watson, T. D., & Squires, N. K. (2005). Electrophysiological correlates of categorization: P300 amplitude as index of target similarity. https: / / doi.Org / 10.1016 / j.biopsycho.2005.05.002

[0052] Bain, S. K., & Jaspers, K. E. (2010). Test Review: Review of Kaufman Brief Intelligence Test, Second Edition: Kaufman, A. S., & Kaufman, N. L. (2004). Kaufman Brief Intelligence Test, Second Edition. Bloomington, MN: Pearson, Inc. Journal of Psychoeducational Assessment, 28(2), 167-174. https: / / doi.org / 10.1177 / 0734282909348217

[0053] Baldassano, C., Chen, J., Zadbood, A., Pillow, J. W., Hasson, U., & Norman, K. A. (2017a). Discovering Event Structure in Continuous Narrative Perception and Memory. Neuron, 95(3), 709-721. e5. https: / / doi.Org / 10.1016 / j.neuron.2017.06.041

[0054] Baldassano, C., Chen, J., Zadbood, A., Pillow, J. W., Hasson, U., & Norman, K. A. (2017b). Discovering Event Structure in Continuous Narrative Perception and Memory. https: / / doi.Org / 10.1016 / j.neuron.2017.06.041

[0055] Brouwer, H., Fitz, H., & Hoeks, J. (2012). Getting real about Semantic Illusions: Rethinking the functional role of the P600 in language comprehension. Brain Research, 1446, 127-143. https: / / doi.Org / 10.1016 / j.brainres.2012.01.055

[0056] Burkhardt, P. (2007). The P600 reflects cost of new information in discourse memory. NeuroReport, 18(17), 1851-1854. https: / / doi.org / 10.1097 / WNR.0b013e3282fla999

[0057] Calhoun, V D, Adali, T., Pearlson, G. D., & Kiehl, K. A. (2006). Neuronal chronometry of target detection: fusion of hemodynamic and event-related potential data. NeuroImage, 30(2), 544-553. https: / / doi.Org / 10.1016 / j.neuroimage.2005.08.060

[0058] Calhoun, Vince D., Adali, T., Kiehl, K. A., Astur, R., Pekar, J. J., & Pearlson, G. D. (2006). A method for multitask fMRI data fusion applied to schizophrenia. Human Brain Mapping, 27(7), 598-610. https: / / doi.org / 10.1002 / hbm.20204

[0059] Calhoun, Vince D, & Adali, T. (2009). Feature-based fusion of medical imaging data. IEEE Transactions on Information Technology in Biomedicine : A Publication of the IEEE Engineering in Medicine and Biology Society, 13(5), 711-720. https: / / doi.org / 10.1109 / TITB.2008.923773

[0060] Calhoun, Vince D, Liu, J., & Adali, T. (2009). A review of group ICA for fMRI data and ICA for joint inference of imaging, genetic, and ERP data. NeuroImage, 45(1 Suppl), S 163-72. https: / / doi.Org / 10.1016 / j.neuroimage.2008.10.057

[0061] Davey, J., Cornelissen, P. L., Thompson, H. E., Sonkusare, S., Hallam, G., Smallwood, J., & Jefferies, E. (2015). Automatic and Controlled Semantic Retrieval: TMSReveals Distinct Contributions of Posterior Middle Temporal Gyrus and Angular Gyrus. TheJournal of Neuroscience : The Official Journal of the Society for Neuroscience, 35(46), 15230- 15239. https: / / doi.org / 10.1523 / JNEUROSCI.4705-14.2015

[0062] Davey, J., Thompson, H. E., Hallam, G., Karapanagiotidis, T., Murphy, C., De Caso, I., ... Jefferies, E. (2016). Exploring the role of the posterior middle temporal gyrus in semantic cognition: Integration of anterior temporal lobe with executive processes.NeuroImage, 137, 165-177. https: / / doi.Org / 10.1016 / j.neuroimage.2016.05.051

[0063] DeLong, K. A., Quante, L., & Kutas, M. (2014). Predictability, plausibility, and two late ERP positivities during written sentence comprehension. Neuropsychologia, 61, 150-162. https: / / doi.Org / 10.1016 / j.neuropsychologia.2014.06.016

[0064] Dronkers, N. F., Ivanova, M. V., & Baldo, J. V. (2017, October 1). What do language disorders reveal about brain-language relationships? From classic models to network approaches. Journal of the International Neuropsychological Society, Vol. 23, pp. 741-754. https: / / doi.org / 10.1017 / S1355617717001126

[0065] Drozdick, L. W., Wahlstrom, D., Zhu, J., & Weiss, L. G. (2012). The Wechsler Adult Intelligence Scale — Fourth Edition and the Wechsler Memory Scale — Fourth Edition. In Contemporary intellectual assessment: Theories, tests, and issues, 3rd ed. (pp. 197- 223). New York, NY, US: The Guilford Press.

[0066] Dunn, L. M., & Dunn, D. (2007). Peabody Picture Vocabulary Test (5th ed.).Pearson.

[0067] Edwards, B. G., Calhoun, V. D., & Kiehl, K. A. (2012). Joint ICA of ERP and fMRI during error-monitoring. NeuroImage, 59(2), 1896-1903. https: / / doi.Org / 10.1016 / j.neuroimage.2011.08.088

[0068] Fengler, A., Meyer, L., & Friederici, A. D. (2016). How the brain attunes to sentence processing: Relating behavior, structure, and function. NeuroImage, 129, 268-278. https: / / doi.Org / 10.1016 / j.neuroimage.2016.01.012

[0069] Ferstl, E. C., Neumann, J., Bogler, C., & von Cramon, D. Y. (2008). The extended language network: a meta-analysis of neuroimaging studies on text comprehension. Human Brain Mapping, 29(5), 581-593. https: / / doi.org / 10.1002 / hbm.20422

[0070] Frankland, S. M., & Greene, J. D. (2015). An architecture for encoding sentence meaning in left mid-superior temporal cortex. Proceedings of the National Academy of Sciences of the United States of America, 112(37), 11732-11737. https: / / doi.org / 10.1073 / pnas.1421236112

[0071] Friederici, A. D. (2002). Towards a neural basis of auditory sentence processing. Trends in Cognitive Sciences, 6(2), 78-84. https: / / doi.org / 10.1016 / S1364- 6613(00)01839-8

[0072] Friederici, A. D. (2011). The brain basis of language processing: from structure to function. Physiological Reviews, 91(4), 1357-1392. https: / / doi.Org / 10.l 152 / physrev.00006.2011

[0073] Friederici, A. D. (2015). The Neuroanatomical Pathway Model of Language: Syntactic and Semantic Networks. In Neurobiology of Language. https: / / doi.org / 10.1016 / B978-0-12-407794-2.00029-8

[0074] Friederici, A. D., & Gierhan, S. M. E. (2013). The language network. Current Opinion in Neurobiology, 23(2), 250-254. https: / / doi.Org / 10.1016 / j.conb.2012.10.002

[0075] Friston, K. J., Holmes, A. P., Worsley, K. J., Poline, J. -P, Frith, C. D., & Frackowiak, R. S. J. (1994). Statistical parametric maps in functional imaging: A general linear approach. Human Brain Mapping, 2(4), 189-210. https: / / doi.org / 10.1002 / hbm.460020402

[0076] Graves, W. W., Binder, J. R., Desai, R. H., Conant, L. L., & Seidenberg, M. S. (2010). Neural correlates of implicit and explicit combinatorial semantic processing. NeuroImage, 53(2), 638-646. https: / / doi.Org / 10.1016 / j.neuroimage.2010.06.055

[0077] Hagoort, P. (2013). MUC (Memory, Unification, Control) and beyond. In Front Psychol (Vol. 4). https: / / doi.org / 10.3389 / fpsyg.2013.00416

[0078] Hagoort, Peter. (2005). On Broca, brain, and binding: a new framework. Trends in Cognitive Sciences, 9(9), 416-423. https: / / doi.Org / 10.1016 / j.tics.2005.07.004

[0079] Hagoort, Peter. (2013). MUC (Memory, Unification, Control) and beyond. Frontiers in Psychology, 4, 416. https: / / doi.org / 10.3389 / fpsyg.2013.00416

[0080] Himberg, J., Hyvarinen, A., & Esposito, F. (2004). Validating the independent components of neuroimaging time series via clustering and visualization. NeuroImage, 22(3), 1214-1222. https: / / doi.Org / 10.1016 / J.NEUROIMAGE.2004.03.027

[0081] Hoedemaker, R. S., & Gordon, P. C. (2017). The onset and time course of semantic priming during rapid recognition of visual words. Journal of Experimental Psychology: Human Perception and Performance, 43(5), 881-902. https: / / doi.org / 10.1037 / xhp0000377

[0082] James, A. G., Tripathi, S. P., & Kilts, C. D. (2014). Estimating brain network activity through back-projection of ICA components to GLM maps. Neuroscience Letters, 564, 21-26. http: / / dx.doi.Org / 10.1016 / j.neulet.2014.01.056

[0083] Jefferies, E. (2013). The neural basis of semantic cognition: converging evidence from neuropsychology, neuroimaging and TMS. Cortex; a Journal Devoted to the Study of the Nervous System and Behavior, 49(3), 611-625. https: / / doi.Org / 10.1016 / j.cortex.2012.10.008

[0084] Kuperberg, G. R. (2007). Neural mechanisms of language comprehension: Challenges to syntax. Brain Research, 1146, 23-49. https: / / doi.Org / 10.1016 / J.BRAINRES.2006.12.063

[0085] Kutas, M., & Federmeier, K. D. (2011). Thirty years and counting: finding meaning in the N400 component of the event-related brain potential (ERP). Annual Review of Psychology, 62, 621-647. https: / / doi.org / 10.1146 / annurev.psych.093008.131123

[0086] Luck, Steven. (2005). An Introduction to The Event-Related Potential Technique.

[0087] Mar, R. A. (2011). The Neural Bases of Social Cognition and Story Comprehension. Annual Review of Psychology, 62, 103-134. https: / / doi.Org / 10.l 146 / annurev-psych-120709-145406

[0088] McGrew, K. S., Schrank, F. A., & Woodcock, R. W. (2007). Woodcock- Johnson III Normative Update. Rolling Meadows, Illinois, United States: Riverside Publishing.

[0089] Mij ovic, B., Vanderperren, K., Novitskiy, N., Vanrumste, B., Stiers, P., Bergh, B. Van den, . . . Vos, M. De. (2012). The “why” and “how” of JointICA: results from a visual detection task. NeuroImage, 60(2), 1171-1185. https : / / doi . org / 10.1016 / j . neuroimage.2012.01.063

[0090] Milivojevic, B., Varadinov, M., Vicente Grabovetsky, A., Collin, S. H. P., & Doeller, C. F. (2016). Coding of Event Nodes and Narrative Context in the Hippocampus. The Journal of Neuroscience : The Official Journal of the Society for Neuroscience, 36(49), 12412- 12424. https: / / doi.org / 10.1523 / JNEUROSCI.2889-15.2016

[0091] Mollo, G., Jefferies, E., Cornelissen, P., & Gennari, S. P. (2018). Context- dependent lexical ambiguity resolution: MEG evidence for the time-course of activity in left inferior frontal gyrus and posterior middle temporal gyrus. Brain and Language, 177-178, 23- 36. https: / / doi.0rg / lO. lOl6 / j.bandl.2Ol8.Ol.OOl

[0092] Mueller, K. D., Hermann, B., Mecollari, J., & Turkstra, L. S. (2018, October 21). Connected speech and language in mild cognitive impairment and Alzheimer’s disease: A review of picture description tasks. Journal of Clinical and Experimental Neuropsychology, Vol. 40, pp. 917-939. https: / / doi.org / 10.1080 / 13803395.2018.1446513

[0093] Nelson, D. L., McEvoy, C. L., & Schreiber, T. A. (2004). The University of South Florida free association, rhyme, and word fragment norms. Behavior Research Methods, Instruments, and Computers, 36(3), 402-407. https: / / doi.org / 10.3758 / BF03195588

[0094] O’Sullivan, M., Brownsett, S., & Copland, D. (2019, October 1). Language and language disorders: Neuroscience to clinical practice. Practical Neurology, Vol. 19, pp. 380-388. https: / / doi.Org / 10.l 136 / practneurol-2018-001961

[0095] Olichney, J. M., Van Petten, C., Paller, K. A., Salmon, D. P., Iragui, V. J., & Kutas, M. (2000). Word repetition in amnesia: Electrophysiological measures of impaired and spared memory. Brain, 123(9), 1948-1963. https: / / doi.Org / 10.1093 / brain / 123.9.1948

[0096] Osterhout, L. (1997). On the Language Specificity of the Brain Response to Syntactic Anomalies: Is the Syntactic Positive Shift a Member of the P300 Family? Brain and Language, 59, 494-522. Retrieved from https: / / pdfs.semanticscholar.org / 4b9b / dfacb648516f8f472b49b9bdl6f 7534d9da.pdf

[0097] Osterhout, Lee, & Holcomb, P. J. (1992). Event-related brain potentials elicited by syntactic anomaly. Journal of Memory and Language, 31(6), 785-806. https: / / doi.org / 10.1016 / 0749-596X(92)90039-Z

[0098] Osterhout, Lee, Kim, A., & Kuperberg, G. (2012). The Neurobiology of Sentence Comprehension. Cambridge Handbook of Psycholinguistics, 365-389.

[0099] Ouyang, X , Chen, K., Yao, L., Hu, B., Wu, X , Ye, Q., & Guo, X. (2015). Simultaneous changes in gray matter volume and white matter fractional anisotropy in Alzheimer’s disease revealed by multimodal CCA and joint ICA. Neuroscience, 301, 553-562. https : / / doi . org / 10.1016 / j . neuroscience .2015.06.031

[0100] Polich J. Updating P300: an integrative theory of P3a and P3b. Clin Neurophysiol. 2007 Oct;l 18(10):2128-48. doi: 10.1016 / j.clinph.2007.04.019. Epub 2007 Jun 18. PMID: 17573239; PMCID: PMC2715154.

[0101] Pulvermuller, F. (2012). Meaning and the brain: The neurosemantics of referential, interactive, and combinatorial knowledge. Journal of Neurolinguistics, 25(5), 423- 459. https : / / doi . org / 10.1016 / j .j neuroling.2011.03.004

[0102] Riby, L. M., & Orme, E. (2013). A familiar pattern? Semantic memory contributes to the enhancement of visuo-spatial memories. Brain and Cognition, 81(2), 215— 222. https: / / doi.Org / 10.1016 / J.BANDC.2012.10.011

[0103] Rice, G. E., Caswell, H., Moore, P., Hoffman, P., & Lambon Ralph, M. A. (2018). The Roles of Left Versus Right Anterior Temporal Lobes in Semantic Memory: ANeuropsychological Comparison of Postsurgical Temporal Lobe Epilepsy Patients. Cerebral Cortex, 28(4), 1487-1501. https: / / doi.org / 10.1093 / cercor / bhx362

[0104] Richter, F. (2019). Prediction errors indexed by the P3 track the updating of complex long-term memory schemas. 1-40. https: / / doi.org / 10.1101 / 805887

[0105] Schrank, F. A., Mather, N., & McGrew, K. S. (2014). Woodcock- Johnson IV Test of Achievement. Rolling Meadows, IL: Riverside Publishing.

[0106] Selig, J. P., & Preacher, K. J. (2008, June). Monte Carlo method for assessing mediation: An

[0107] interactive tool for creating confidence intervals for indirect effects [Computer software]. Available from http: / / quantpsy.org / .

[0108] Shen, W., Fiori-Duharcourt, N., & Isel, F. (2016). Functional significance of the semantic P600. NeuroReport, 27(7), 548-558. https: / / doi.org / 10.1097 / WNR.0000000000000583

[0109] Simony, E., Honey, C. J., Chen, J., Lositsky, O., Yeshurun, Y., Wiesel, A., & Hasson, U. (2016). Dynamic reconfiguration of the default mode network during narrative comprehension. Nature Communications, 7(1), 12141. https: / / doi.org / 10.1038 / ncommsl2141

[0110] Skeide, M. A., Brauer, J., & Friederici, A. D. (2014). Syntax gradually segregates from semantics in the developing brain. NeuroImage, 100, 106-111. https: / / doi.Org / 10.1016 / j.neuroimage.2014.05.080

[0111] Smith, K. M., & Caplan, D. N. (2018, October 1). Communication impairment in Parkinson’s disease: Impact of motor and cognitive symptoms on speech and language. Brain and Language, Vol. 185, pp. 38-46. https: / / doi.Org / 10.1016 / j.bandl.2018.08.002

[0112] Thompson, C., Walenski, M., Europa, E., & Caplan, D. (2018). Neural networks for sentence comprehension and production: an ALE-based meta-analysis of neuroimaging studies. Frontiers in Human Neuroscience, 12. https: / / doi.org / 10.3389 / conf.fnhum.2018.228.00076

[0113] Turner, J. A., Calhoun, V. D., Michael, A., van Erp, T. G. M., Ehrlich, S., Segall, J. M., Gollub R. L., Csernansky, J., Potkin, S. G., Ho B., Bustillo, J., Schulz, S. C., & Wang L. (2012). Heritability of Multivariate Gray Matter Measures in Schizophrenia. Twin Research and Human Genetics, 15(3), 324-335. doi: 10.1017 / thg.2012.1.

[0114] Vabalas A, Gowen E, Poliakoff E, Casson AJ (2019) Machine learning algorithm validation with a limited sample size. PLoS ONE 14(11): e0224365. http s : / / doi . org / 10.1371 / j ournal .pone.0224365

[0115] Wager, T (2022). MediationToolbox(https: / / github.com / canlab / MediationToolbox), GitHub. Retrieved August 15, 2022.

[0116] Whitney, C., Huber, W., Klann, J., Weis, S., Krach, S., & Kircher, T. (2009). Neural correlates of narrative shifts during auditory story comprehension. NeuroImage, 47(1), 360-366. https: / / doi.Org / 10.1016 / j.neuroimage.2009.04.037

[0117] Whitney, C., Kirk, M., O’Sullivan, J., Lambon Ralph, M. A., & Jefferies, E. (2011). The neural organization of semantic control: TMS evidence for a distributed network in left inferior frontal and posterior middle temporal gyrus. Cerebral Cortex (New York, N.Y. : 1991), 21(5), 1066-1075. https: / / doi.org / 10.1093 / cercor / bhql80

[0118] Whitney, C., Kirk, M., O’Sullivan, J., Lambon Ralph, M. A., & Jefferies, E. (2012). Executive semantic processing is underpinned by a large-scale neural network: revealing the contribution of left prefrontal, posterior temporal, and parietal cortex to controlled retrieval and selection using TMS. Journal of Cognitive Neuroscience, 24(1), 133— 147. https : / / doi . org / 10.1162 / j ocn_a_00123

[0119] Zhang, S., & Li, C. R. (2012). Functional connectivity mapping of the human precuneus by resting state fMRI. NeuroImage, 59(4), 3548-3562. http s : / / doi . org / 10.1016 / j . neuroimage .2011.11.023CONCLUSION

[0120] While the methods and systems have been described in connection with preferred embodiments and specific examples, it is not intended that the scope be limited to the particular embodiments set forth, as the embodiments herein are intended in all respects to be illustrative rather than restrictive.

[0121] Unless otherwise expressly stated, it is in no way intended that any method set forth herein be construed as requiring that its steps be performed in a specific order. Accordingly, where a method claim does not actually recite an order to be followed by its steps or it is not otherwise specifically stated in the claims or descriptions that the steps are to be limited to a specific order, it is no way intended that an order be inferred, in any respect. This holds for any possible non-express basis for interpretation, including: matters of logic with respect to arrangement of steps or operational flow; plain meaning derived from grammatical organization or punctuation; the number or type of embodiments described in the specification.

[0122] Throughout this application, various publications may be referenced. The disclosures of these publications in their entireties are hereby incorporated by reference into this application in order to more fully describe the state of the art to which the methods and systems pertain.

[0123] It will be apparent to those skilled in the art that various modifications and variations can be made without departing from the scope or spirit. Other embodiments will be apparent to those skilled in the art from consideration of the specification and practice disclosed herein. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit being indicated by the following claims.

Claims

WHAT IS CLAIMED:

1. A method of individualized, non-invasive brain stimulation for enhanced learning comprising: obtaining brain data of a person; identifying one or more areas of a brain of the person for stimulation based on the brain data; stimulating the identified one or more areas of the brain, wherein the stimulation results in enhanced learning.

2. The method of claim 1, wherein the brain data are obtained while the person is performing a learning activity or other cognitive activity, not performing a task, or in a resting state.

3. The method of claim 2, wherein the other cognitive activities include emotional regulation, semantic and episodic memory, and other memory processes, executive functions, including but not limited to working memory, cognitive control, inferential processes, prediction error, reading and language comprehension / production, social cognition, vision and auditory, and motor processing.

4. The method of any one of claims 2 or 3, wherein the learning activity is a specific learning activity such as factual learning, or skill learning.

5. The method of any one of claims 1-4, wherein the brain data comprise high spatial- resolution brain data and / or high temporal-resolution brain data.

6. The method of claim 5, wherein the high spatial-resolution brain data and / or the high temporal-resolution brain data is predicted from low-spatial-resolution brain data and / or low temporal-resolution brain data using a trained artificial intelligence (Al) model.

7. The method of claim 6, wherein predicting the high spatial-resolution brain data and / or the high temporal-resolution brain data from low-spatial-resolution brain data and / or low temporal-resolution brain data using the trained Al model comprises: training the Al model by co-mapping existing low spatial-resolution brain data and / or low temporal-resolution brain data and existing high spatial-resolution brain data and / or existing high temporal-resolution brain data and using the trained Al model to identify patterns of co-occurrence between the existing low-spatial- resolution brain data and / or the existing low temporal-resolution brain data and theexisting high spatial-resolution brain data and / or the existing high temporal-resolution brain data; obtaining new low spatial-resolution brain data and / or new low temporalresolution brain data for the person; applying the trained Al model to the new low spatial-resolution brain data and / or the new low temporal-resolution brain data; and predicting new high spatial-resolution brain data and / or high temporalresolution brain data from the new low spatial-resolution brain data and / or the new low temporal-resolution brain data using the trained Al model.

8. The method of claim 7, wherein training the Al model by co-mapping the existing low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data and using the trained Al model to identify patterns of co-occurrence between the low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data comprises training the Al model by co-mapping existing functional magnetic resonance imaging (fMRI) data and / or structural MRI data, existing electroencephalogram (EEG) data, and functional near-infrared spectroscopy (fNIRS) data and using the trained Al model to identify the patterns of co-occurrence between the existing fMRI data and / or structural MRI data and the existing EEG and fNIRS data.

9. The method of claim 7 or claim 8, wherein obtaining the new low-resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining new EEG and fNIRS data for the person.

10. The method of claim 9, wherein applying the trained Al model to the new low-spatial- resolution brain data and / or the new low temporal-resolution brain data comprises applying the trained Al model to the new EEG and fNIRS data.

11. The method of claim 10, wherein predicting the new high spatial -resolution brain data and / or the new high temporal-resolution brain data from the new low spatial- resolution brain data and / or the new low temporal-resolution brain data using the trained Al model comprises predicting new fMRI data and / or structural MRI data of the person from the new EEG and fNIRS data using the trained Al model.

12. The method of claim 7, wherein training the Al model by co-mapping the existing low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data and using the trained Al model to identify patterns of co-occurrence between the low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data comprises training the Al model by co-mapping existing functional magnetic resonance imaging (fMRI) data and / or structural MRI data, and existing electroencephalogram (EEG) data and using the trained Al model to identify the patterns of co-occurrence between the existing fMRI data and / or structural MRI data and the existing EEG data.

13. The method of claim 12, wherein obtaining the new low-resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining new EEG data for the person.

14. The method of claim 13, wherein applying the trained Al model to the new low- spatial-resolution brain data and / or the new low temporal-resolution brain data comprises applying the trained Al model to the new EEG data.

15. The method of claim 14, wherein predicting the new high spatial -resolution brain data and / or the new high temporal-resolution brain data from the new low spatial- resolution brain data and / or the new low temporal-resolution brain data using the trained Al model comprises predicting new fMRI data and / or structural MRI data of the person from the new EEG data using the trained Al model.

16. The method of claim 7, wherein training the Al model by co-mapping the existing low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data and using the trained Al model to identify patterns of co-occurrence between the low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data comprises training the Al model by co-mapping existing functional magnetic resonance imaging (fMRI) data and / or structural MRI data, and existing functional near-infrared spectroscopy (fNIRS) data and using the trained Al model to identify the patterns of co-occurrencebetween the existing fMRI data and / or structural MRI data and the existing fNIRS data.

17. The method of claim 16, wherein obtaining the new low-resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining new fNIRS data for the person.

18. The method of claim 17, wherein applying the trained Al model to the new low- spatial-resolution brain data and / or the new low temporal-resolution brain data comprises applying the trained Al model to the new fNIRS data.

19. The method of claim 18, wherein predicting the new high spatial -resolution brain data and / or the new high temporal-resolution brain data from the new low spatial- resolution brain data and / or the new low temporal-resolution brain data using the trained Al model comprises predicting new fMRI data and / or structural MRI data of the person from the new fNIRS data using the trained Al model.

20. The method of claim 7, wherein training the Al model by co-mapping the existing low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data and using the trained Al model to identify patterns of co-occurrence between the low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data comprises training the Al model by co-mapping existing functional magnetic resonance imaging (fMRI) data and / or structural MRI data and existing electroencephalogram (EEG) and / or functional near-infrared spectroscopy (fNIRS) data and using the trained Al model to identify the patterns of co-occurrence between the existing fMRI data and / or structural MRI data and the existing EEG and / or fNIRS data.

21. The method of claim 20, wherein obtaining the new low-resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining new fMRI data and / or structural MRI data for the person.

22. The method of claim 21, wherein applying the trained Al model to the new low- spatial-resolution brain data and / or the new low temporal-resolution brain data comprises applying the trained Al model to the new fMRI data and / or structural MRI data.

23. The method of claim 22, wherein predicting the new high spatial -resolution brain data and / or the new high temporal-resolution brain data from the new low spatial- resolution brain data and / or the new low temporal-resolution brain data using the trained Al model comprises predicting new EEG and / or fNIRS data of the person from the new fMRI data using the trained Al model.

24. The method of claim 7, wherein training the Al model by co-mapping the existing low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data and using the trained Al model to identify patterns of co-occurrence between the low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data comprises training the Al model by co-mapping existing electroencephalogram (EEG) and functional nearinfrared spectroscopy (fNIRS) data and using the trained Al model to identify the patterns of co-occurrence between the existing EEG and the existing fNIRS data.

25. The method of claim 24, wherein obtaining the new low-resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining new EEG data for the person.

26. The method of claim 25, wherein applying the trained Al model to the new low- spatial-resolution brain data and / or the new low temporal-resolution brain data comprises applying the trained Al model to the new EEG data.

27. The method of claim 26, wherein predicting the new high spatial-resolution brain data and / or the new high temporal-resolution brain data from the new low spatial- resolution brain data and / or the new low temporal-resolution brain data using the trained Al model comprises predicting new fNIRS data of the person from the new EEG data using the trained Al model.

28. The method of any one of claims 7-27, wherein obtaining the existing low spatial- resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data comprises obtaining the existing low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data simultaneously or in parallel.

29. The method of any one of claims 7-28, wherein obtaining the new low spatial- resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining the new low spatial-resolution brain data and / or the new low temporal-resolution brain data for the person simultaneously or in parallel.

30. The method of any one of claims 7-29, wherein training the Al model by co-mapping existing low spatial-resolution brain data and / or low temporal-resolution brain data and existing high spatial-resolution brain data and / or existing high temporalresolution brain data and using the trained Al model to identify patterns of cooccurrence between the existing low-spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data comprises training the Al model by co-mapping existing low spatial-resolution brain data and / or low temporalresolution brain data and existing high spatial-resolution brain data and / or existing high temporal-resolution brain data and using transformers to identify patterns of cooccurrence between the existing low-spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data.

31. The method of any one of claims 7-29, wherein the Al model by co-mapping existing low spatial-resolution brain data and / or low temporal-resolution brain data and existing high spatial-resolution brain data and / or existing high temporal-resolution brain data and using the trained Al model to identify patterns of co-occurrence between the existing low-spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data comprises training the Al model by co-mapping existing low spatial-resolution brain data and / or low temporalresolution brain data and existing high spatial-resolution brain data and / or existing high temporal-resolution brain data and using neural architectures such as neural networks, recurrent neural networks (RNN), long short-term memory networks (LSTMs), convolutional neural networks (CNN), hybrid combinations of CNNs, RNNs and others, state space machines, autoencoders, generative adversarial networks (GANs), graph neural networks, and multi-modal fusion networks to identify patterns of co-occurrence between the existing low-spatial-resolution braindata and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data.

32. The method of any one of claims 31-32, wherein using the transformers to identify patterns of co-occurrence between the existing low-spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial- resolution brain data and / or the existing high temporal-resolution brain data comprises using transformer model architecture to perform cross-modal training in brain imaging data from different imaging modalities (e.g. EEG, fNIRS, and fMRI and / or structural MRI) to make predictions of how these imaging modalities correspond with one another.

33. The method of claim 32, wherein the cross-modal training comprises combining various pairings of imaging modality (e.g., fMRI / EEG / fNIRS and / or structural MRI) and single- and / or cross-modal training approaches within the transformer model architecture.

34. The method of claim 32, wherein the single- and / or cross-modal training approaches within the transformer model architecture include masked signal modelling, causal- signal modelling, cross-attention, multi-modal fusion, multi-modal pre-training, late fusion, and early fusion, among others.

35. The method of any one of claims 7-34, wherein obtaining the new low spatial- resolution brain data and / or the new low temporal-resolution brain data for the person comprises using continuous or discrete temporal and / or spatial signals from the new low spatial-resolution brain data and / or the new low temporal-resolution brain data.

36. The method of claim 35, wherein the new low spatial-resolution brain data and / or the new low temporal-resolution brain data is obtained using brain-imaging modalities and the continuous or discrete temporal and / or spatial signals are restricted based on the brain imaging modality used (for instance, only inputting spatial signals from fNIRS and temporal signals from EEG).

37. The method of any one of claims 6-36, wherein obtaining the new low spatial- resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining the new low spatial-resolution brain data and / or the new low temporal-resolution brain data for the person using whole-brain measurements (e.g. voxel), or region of interest measurements (e.g. atlas region in the brain).

38. The method of any one of claims 1-37, wherein stimulation for stimulating the identified one or more areas of the brain varies over time.

39. A computer program product comprising computer-executable instructions stored on a non-transient computer-readable medium for performing the method claimed in any one of claims 1-38.

40. A method of predicting high spatial-resolution brain data and / or the high temporalresolution brain data from low-spatial-resolution brain data and / or low temporalresolution brain data using a trained artificial intelligence (Al) model, comprising: training the Al model by co-mapping existing low spatial-resolution brain data and / or low temporal-resolution brain data and existing high spatial-resolution brain data and / or existing high temporal-resolution brain data and using the trained Al model to identify patterns of co-occurrence between the existing low-spatial- resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data; obtaining new low spatial-resolution brain data and / or new low temporalresolution brain data for a person; applying the trained Al model to the new low spatial-resolution brain data and / or the new low temporal-resolution brain data; and predicting new high spatial-resolution brain data and / or high temporalresolution brain data for the person from the new low spatial-resolution brain data and / or the new low temporal-resolution brain data using the trained Al model.

41. The method of claim 40, wherein training the Al model by co-mapping the existing low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data and using the trained Al model to identify patterns of co-occurrence between the low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data comprises training the Al model by co-mapping existing functional magnetic resonance imaging (fMRI) data and / or structural MRI data and existing electroencephalogram (EEG) data and functional near-infrared spectroscopy (fNTRS) data and using the trained Al Model toidentify the patterns of co-occurrence between the existing fMRI data and / or structural MRI data and the existing EEG and fNIRS data.

42. The method of claim 40 or claim 41, wherein obtaining the new low-resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining new EEG and fNIRS data for the person.

43. The method of claim 42, wherein applying the trained Al model to the new low- spatial-resolution brain data and / or the new low temporal-resolution brain data comprises applying the trained Al model to the new EEG and fNIRS data.

44. The method of claim 41, wherein predicting the new high spatial -resolution brain data and / or the new high temporal-resolution brain data from the new low spatial- resolution brain data and / or the new low temporal-resolution brain data using the trained Al model comprises predicting new fMRI data and / or structural MRI data of the person from the new EEG and fNIRS data using the trained Al model.

45. The method of claim 41, wherein training the Al model by co-mapping the existing low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data and using the trained Al model to identify patterns of co-occurrence between the low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data comprises training the Al model by co-mapping existing functional magnetic resonance imaging (fMRI) data and / or structural MRI data and existing electroencephalogram (EEG) data and using the trained Al model to identify the patterns of co-occurrence between the existing fMRI data and / or structural MRI data and the existing EEG data.

46. The method of claim 45, wherein obtaining the new low-resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining new EEG data for the person.

47. The method of claim 46, wherein applying the trained Al model to the new low- spatial-resolution brain data and / or the new low temporal-resolution brain data comprises applying the trained Al model to the new EEG data.

48. The method of claim 47, wherein predicting the new high spatial -resolution brain data and / or the new high temporal-resolution brain data from the new low spatial- resolution brain data and / or the new low temporal-resolution brain data using thetrained Al model comprises predicting new fMRI data and / or structural MRI data of the person from the new EEG data using the trained Al model.

49. The method of claim 40, wherein training the Al model by co-mapping the existing low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data and using the trained Al model to identify patterns of co-occurrence between the low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data comprises training the Al model by co-mapping existing functional magnetic resonance imaging (fMRI) data and / or structural MRI data and existing functional near-infrared spectroscopy (fNIRS) data and using the trained Al model to identify the patterns of co-occurrence between the existing fMRI data and / or structural MRI data and the existing fNIRS data.

50. The method of claim 49, wherein obtaining the new low-resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining new fNIRS data for the person.

51. The method of claim 50, wherein applying the trained Al model to the new low- spatial-resolution brain data and / or the new low temporal-resolution brain data comprises applying the trained Al model to the new fNIRS data.

52. The method of claim 51, wherein predicting the new high spatial -resolution brain data and / or the new high temporal-resolution brain data from the new low spatial- resolution brain data and / or the new low temporal-resolution brain data using the trained Al model comprises predicting new fMRI data and / or structural MRI data of the person from the new fNIRS data using the trained Al model.

53. The method of claim 40, wherein training the Al model by co-mapping the existing low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data and using the trained Al model to identify patterns of co-occurrence between the low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data comprises training the Al model by co-mapping existing functional magnetic resonance imaging (fMRI) data and / or structural MRI data and existing electroencephalogram (EEG) and / orfunctional near-infrared spectroscopy (fNIRS) data and using the trained Al model to identify the patterns of co-occurrence between the existing fMRI data and / or structural MRI data and the existing EEG and / or fNIRS data.

54. The method of claim 53, wherein obtaining the new low-resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining new fMRI data and / or structural MRI data for the person.

55. The method of claim 54, wherein applying the trained Al model to the new low- spatial-resolution brain data and / or the new low temporal-resolution brain data comprises applying the trained Al model to the new fMRI data and / or structural MRI data.

56. The method of claim 55, wherein predicting the new high spatial-resolution brain data and / or the new high temporal-resolution brain data from the new low spatial- resolution brain data and / or the new low temporal-resolution brain data using the trained Al model comprises predicting new EEG and / or fNIRS data of the person from the new fMRI data and / or structural MRI data using the trained Al model.

57. The method of claim 40, wherein training the Al model by co-mapping the existing low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data and using the trained Al model to identify patterns of co-occurrence between the low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data comprises training the Al model by co-mapping existing electroencephalogram (EEG) and functional nearinfrared spectroscopy (fNIRS) data and using the trained Al model to identify the patterns of co-occurrence between the existing EEG and the existing fNIRS data.

58. The method of claim 57, wherein obtaining the new low-resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining new EEG data for the person.

59. The method of claim 58, wherein applying the trained Al model to the new low- spatial-resolution brain data and / or the new low temporal-resolution brain data comprises applying the trained Al model to the new EEG data.

60. The method of claim 59, wherein predicting the new high spatial-resolution brain data and / or the new high temporal-resolution brain data from the new low spatial-resolution brain data and / or the new low temporal-resolution brain data using the trained Al model comprises predicting new fNIRS data of the person from the new EEG data using the trained Al model.

61. The method of any one of claims 40-60, wherein obtaining the existing low spatial- resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data comprises obtaining the existing low spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data simultaneously or in parallel.

62. The method of any one of claims 40-61, wherein obtaining the new low spatial- resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining the new low spatial-resolution brain data and / or the new low temporal-resolution brain data for the person simultaneously or in parallel.

63. The method of any one of claims 40-62, wherein training the Al model by comapping existing low spatial-resolution brain data and / or low temporal-resolution brain data and existing high spatial-resolution brain data and / or existing high temporal-resolution brain data and using the trained Al model to identify patterns of co-occurrence between the existing low-spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data comprises training the Al model by co-mapping existing low spatial-resolution brain data and / or low temporal-resolution brain data and existing high spatial-resolution brain data and / or existing high temporal-resolution brain data and using transformers to identify patterns of co-occurrence between the existing low-spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial- resolution brain data and / or the existing high temporal-resolution brain data.

64. The method of any one of claims 40-62, wherein training the Al model by comapping existing low spatial-resolution brain data and / or low temporal-resolution brain data and existing high spatial-resolution brain data and / or existing high temporal-resolution brain data and using the trained Al model to identify patterns of co-occurrence between the existing low-spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolutionbrain data and / or the existing high temporal-resolution brain data comprises training the Al model by co-mapping existing low spatial-resolution brain data and / or low temporal-resolution brain data and existing high spatial-resolution brain data and / or existing high temporal-resolution brain data and using neural architectures such as neural networks, recurrent neural networks (RNN), long short-term memory networks (LSTMs), convolutional neural networks (CNN), hybrid combinations of CNNs, RNNs and others, state space machines, autoencoders, generative adversarial networks (GANs), graph neural networks, and multi-modal fusion networks to identify patterns of co-occurrence between the existing low-spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial-resolution brain data and / or the existing high temporal-resolution brain data.

65. The method of any one of claims 63-64, wherein using the transformers to identify patterns of co-occurrence between the existing low-spatial-resolution brain data and / or the existing low temporal-resolution brain data and the existing high spatial- resolution brain data and / or the existing high temporal-resolution brain data comprises using transformer model architecture to perform cross-modal training in brain imaging data from different imaging modalities (e.g. EEG, fNIRS, fMRI, and / or structural MRI) to make predictions of how these imaging modalities correspond with one another.

66. The method of claim 65, wherein the cross-modal training comprises combining various pairings of imaging modality (e.g., fMRI / EEG / fNIRS and / or structural MRI) and single- and / or cross-modal training approaches within the transformer model architecture.

67. The method of claim 66, wherein the single- and / or cross-modal training approaches within the transformer model architecture include masked signal modelling, causal- signal modelling, cross-attention, multi-modal fusion, multi-modal pre-training, late fusion, and early fusion, among others.

68. The method of any one of claims 40-67, wherein obtaining the new low spatial- resolution brain data and / or the new low temporal-resolution brain data for the person comprises using continuous or discrete temporal and / or spatial signals from the new low spatial-resolution brain data and / or the new low temporal-resolution brain data.

69. The method of claim 68, wherein the new low spatial-resolution brain data and / or the new low temporal-resolution brain data is obtained using brain-imaging modalitiesand the continuous or discrete temporal and / or spatial signals are restricted based on the brain imaging modality used (for instance, only inputting spatial signals from fNIRS and temporal signals from EEG).

70. The method of any one of claims 40-69, wherein obtaining the new low spatial- resolution brain data and / or the new low temporal-resolution brain data for the person comprises obtaining the new low spatial-resolution brain data and / or the new low temporal-resolution brain data for the person using whole-brain measurements (e.g. voxel), or a region of interest measurement (e.g. atlas region in the brain).

71. A computer program product comprising computer-executable instructions stored on a non-transient computer-readable medium for performing the method claimed in any one of claims 40-70.

Images