Screening, monitoring and therapeutic mechanisms for focused ultrasound

Abstract

Systems and methods for providing focused ultrasound therapy are disclosed. A patient is screened to determine whether focused ultrasound therapy is suitable for the patient to treat a disease. The patient is monitored to measure a plurality of health-related parameters of the patient, and if it is determined that the focused ultrasound therapy is appropriate, an onset of a symptom associated with the disease is detected or predicted based on the plurality of health-related parameters. A personalized location of the focused ultrasound therapy is determined for the patient based on the at least one health-related parameter. At least one parameter associated with focused ultrasound therapy is selected based on the at least one health-related parameter. Focused ultrasound therapy is provided to the patient at the selected location using the selected at least one parameter. A p-health-related parameter is measured after the focused ultrasound therapy is provided to determine the effectiveness of the therapy.

Description

Cross-references to related fields

[0001] This application claims priority to U.S. Application No. 63 / 337,146 filed on May 1, 2022, U.S. Application No. 63 / 348,713 filed on June 3, 2022, U.S. Application No. 63 / 392,572 filed on July 27, 2022, U.S. Application No. 63 / 400,960 filed on August 25, 2022, and U.S. Application No. 63 / 435,456 filed on December 27, 2022, the subject matter of which is incorporated herein by reference in its entirety. Technical Field

[0002] The present invention relates generally to decision-making assistance systems and, more particularly, to mechanisms for focused ultrasound therapy. Background Art

[0003] Neuromodulation also refers to an emerging class of medical therapies that target the nervous system for restoring function, relieving pain, or controlling symptoms. These therapies consist primarily of targeted stimulation through various forms of energy. Electrical stimulation devices include deep brain stimulation systems, spinal cord stimulators, vagus nerve stimulators, and transcutaneous electrical nerve stimulation devices. Other methods for neuromodulation can use magnetic stimulation in a focused ultrasound system, such as transcutaneous magnetic stimulation, and sound. Neuromodulation appears to be promising in treating a variety of diseases, but the appropriate time, dose, and location of neuromodulation may be difficult to determine for a specific disease. Summary of the invention

[0004] According to one example, a method for providing focused ultrasound therapy is disclosed. A patient is screened to determine whether focused ultrasound therapy is appropriate for the patient to treat a disease. The patient is monitored to measure a plurality of health-related parameters of the patient, and if it has been determined that focused ultrasound therapy is appropriate for the patient, the onset of symptoms associated with the disease is detected or predicted based on the plurality of health-related parameters. Personalized positioning of the focused ultrasound therapy for the patient is determined based on at least one parameter of the plurality of health-related parameters. At least one parameter associated with the focused ultrasound therapy is selected based on at least one parameter of the plurality of health-related parameters. Focused ultrasound therapy is provided to the patient at the selected location using the selected at least one parameter. The plurality of health-related parameters are measured after providing the focused ultrasound therapy to determine the effectiveness of the focused ultrasound therapy.

[0005] According to another example, a system for generating clinical parameters for a user is provided. A physiological sensing device monitors a first plurality of health-related parameters representing the user over a defined time period. A portable computing device obtains a second plurality of health-related parameters representing the user through the portable computing device. A network interface retrieves a third plurality of health-related parameters representing the user from an electronic health record (EHR) system, the first plurality of health-related parameters, the second plurality of health-related parameters, and the third plurality of health-related parameters together forming a health-related parameter set. A feature aggregator generates an aggregated parameter set from the health-related parameter set, wherein each of the aggregated parameter sets includes a unique suitable subset of the health-related parameter set. A predictive model assigns clinical parameters to the user based on the subset of the aggregated parameter set.

[0006] According to another example, a method for generating a value representing one of risk and progression of a disease is provided. A first image representing a brain of a patient is acquired from a first imaging system, and a second image representing one of a retina, an optic nerve, and a vasculature associated with one of an optic nerve and a retina of the patient is acquired from a second imaging system. A representation of each of the first image and the second image is provided to a machine learning model. A value is generated at the machine learning model from the representation of the first image and the representation of the second image, and the patient is assigned to one of a plurality of intervention categories based on the generated value.

[0007] According to yet another example, a method for determining risk of disease by imaging a patient's brain is provided. A first image representing a structure of the brain is acquired from a first imaging system, and a second image representing connectivity of the brain is acquired from one of the first imaging system and a second imaging system. The first image is segmented into a plurality of sub-regions of the brain to generate a segmented first image, such that each of at least a subset of a plurality of voxels comprising the first image is associated with one of the plurality of sub-regions. A representation of the segmented first image and the second image is provided to a machine learning model trained on imaging data of a plurality of patients with known outcomes. A clinical parameter representing the patient's risk of disease is generated from the representation of the segmented first image and the second image. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 A method for providing therapy via focused ultrasound therapy is shown; Figure 2 Denotes a method used to screen patients for a specific disease; Figure 3 A method for using cues to generate features for screening patients for disease is shown; Figure 4 A method for extracting connectivity features to evaluate a patient for focused ultrasound therapy is shown; Figure 5A method for generating features from imaging to extract connectivity features for evaluating a patient for focused ultrasound therapy is shown; Figure 6 A method for monitoring a patient to predict or detect the onset of symptoms associated with a disease of the patient is shown; Figure 7 Another example of a method for assessing the onset of symptoms in a patient requiring treatment with focused ultrasound therapy is shown; Figure 8 A method for targeted focused ultrasound therapy in the brain of a patient to treat a disease is shown; Fig. 9 Another method of using acute feedback to target focused ultrasound therapy in a patient's brain to treat a disease is shown; Fig.10 A method for selecting a dose for focused ultrasound therapy is shown; Fig.11 A system for providing treatment to a patient via focused ultrasound therapy is shown; Fig.12 A system for targeting focused ultrasound therapy to treat or diagnose disease in a specific area of ​​a patient's brain is shown; Fig.13 A system for determining the presence or risk of a disease based on imaging of a patient's brain is shown; Fig.14 An example of a system for diagnosing and monitoring disease is shown; and Fig.15 is a schematic block diagram of an exemplary system illustrating hardware components. DETAILED DESCRIPTION

[0009] As used herein, the term "health" is intended to refer to the mental, physical, cognitive, behavioral, social, and emotional health of a user, and should be interpreted to encompass each of the user's health, function, balance, resilience, homeostasis, disease, and condition. In various examples herein, the user's health may relate to the user's readiness to perform work-related, athletic, or daily functions, to enter a state of flow or zone, the user's susceptibility to infectious diseases, the user's ability to resist the effects of addiction, the user's exacerbation of pain, the increase or decrease of stress, anxiety, the user's quality of life, and similar qualities of the user.

[0010] The "health-related parameter" is a parameter related to the user's health.

[0011] "Biorhythm" is any chronobiological phenomenon that affects a person, including, but not limited to, circadian rhythms, ultradian rhythms, hypodian rhythms, diurnal cycles, sleep / wake cycles, and life patterns.

[0012] "Focused ultrasound therapy" as described herein is any treatment that applies ultrasound to the brain for the purpose of achieving neuromodulation, as well as modulating neural activity or disrupting the blood-brain barrier to allow targeted access to the brain by therapeutic agents. Thus, focused ultrasound therapy may involve the use of ultrasound to modulate neural activity, the use of ultrasound in combination with microbubbles to disrupt the blood-brain barrier and allow suitable therapeutic agents to selectively enter the brain, or combining blood-brain barrier disruption with neuromodulation of suitable targets. It should be understood that the use of focused ultrasound to disrupt the blood-brain barrier will involve a certain degree of neuromodulation due to the effects of ultrasound on surrounding tissue, but focused ultrasound therapy involving a combination of neuromodulation and disruption of the blood-brain barrier refers to at least two separate applications of focused ultrasound, wherein the first application of focused ultrasound is targeted to the location where the desired therapeutic agent penetrates into the brain, and the second application of focused ultrasound is targeted to the location of brain tissue where neuromodulation is desired.

[0013] As used herein, "portable monitoring device" refers to a device worn, carried, or implanted in a user, the device comprising one or both of an input device and a user interface for receiving input from the user, and a sensor for monitoring health-related parameters or parameters that can be used to calculate or estimate health-related parameters.

[0014] As used herein, "index" is intended to cover a composite statistic derived from a set of observations and used as an indicator or measure. An index can be an ordinal, continuous, or categorical value representing observations and correlations, and should be understood to cover statistics traditionally referred to as "scores" as well as the more technical meaning of index.

[0015] As used herein, a "clinical parameter" may be any continuous or categorical parameter representing a user's mental, physical, cognitive, behavioral, social, and emotional health, and may represent any or all of a user's health, function, zone or flow state, balance, resilience, homeostasis, disease, and condition.

[0016] As used herein, a "continuous parameter" generally refers to a parameter that can take any value within a predetermined range to a predetermined accuracy. Thus, a value is referred to herein as a continuous parameter even if, in practice, it is limited to a finite number of discrete values ​​within the range of the resolution at which the value is measured, calculated, or stored.

[0017] As used herein, a "physiological sensing device" is a device that measures one or more physiological parameters and / or biorhythms. Although in some cases an external device may be used to capture physiological parameters, a physiological sensing device is typically implanted, ingested, or wearable.

[0018] As used herein, a "portable computing device" is a computing device that can be carried by a user, such as a smartphone, smartwatch, tablet, notebook, and laptop, which can measure health-related parameters through sensors on the device or through interaction with the user. The portable computing device may include, for example, a user interface for receiving input from the user, a motion sensor for measuring the user's activity, and location services to track the user's location.

[0019] As used herein, a "predictive model" is a mathematical model or machine learning model that predicts the future state of a parameter or estimates the current state of a parameter that cannot be directly measured.

[0020] Figure 1 A method 100 for providing treatment via focused ultrasound therapy is shown. At 102, a potential patient is screened to determine whether focused ultrasound (FUS) therapy is appropriate for the patient. In one embodiment, the screening is performed by collecting a plurality of health-related parameters of the patient that have been determined to be associated with a particular disease and providing them to a predictive model to assign to the patient clinical parameters that are associated with the likelihood that the patient will benefit from focused ultrasound therapy.

[0021] Figure 2 A method 200 for screening a patient for a specific disease is represented. At 202, multiple health-related parameters of the patient are collected. Parameters can be retrieved from an electronic health record (EHR) interface and / or other available databases, including, for example, employment information (e.g., position, department, shift), age, gender, home zip code, genomic data, nutritional information, medication intake, family information (e.g., family type, number of residents and age), social and psychosocial data, consumer spending and profile, financial data, food safety information, presence of physical abuse, and relevant medical history. Parameters can also be measured by diagnostic systems, imaging systems, wearable sensors, cognitive tests, questionnaires, and other means. As described above, health-related parameters can include at least physiological, cognitive, motor / musculoskeletal, sensory, sleep, biomarkers, and behavioral parameters. Table I provides non-limiting examples of physiological parameters that can be measured and exemplary tests, devices, and methods for measuring physiological parameters.

[0022] Table I

[0023] Physiological parameters can be measured in a clinical setting using suitable devices, or can also be measured in a non-clinical setting by wearable, implantable or portable devices. Some information can also be determined based on self-reporting by the user through an application in a mobile device or through interaction with an application on a mobile device. For example, a smart watch, ring or patch can be used to measure the user's heart rate, heart rate variability, body temperature, blood oxygen saturation, movement and sleep. In a non-clinical setting, these values ​​can also be subjected to day and night analysis to estimate variability. For example, eye tracking can be performed using a camera on a mobile device and dedicated software.

[0024] Table II provides non-limiting examples of cognitive parameters that are gamified and can be measured, as well as exemplary methods and tests / tasks for measuring such cognitive parameters. Cognitive parameters can be assessed through a series of cognitive tests, such as measuring executive function, decision making, working memory, attention, and fatigue.

[0025] Table II

[0026] These cognitive tests can be performed in a clinical / laboratory environment or in a natural non-clinical environment, such as when the user is at home, at work, or in other non-clinical environments. Intelligent devices (such as smart phones, tablet computers, or smart watches) can facilitate the measurement of these cognitive parameters in a natural non-clinical environment. For example, Erickson interference tasks, N-back tasks, and psychomotor vigilance tasks can be obtained through applications on smart phones, tablet computers, or smart watches. In one example, patients can be allowed to explore a virtual reality environment and collect items in the environment. Then the patient is asked to recalculate the position where each item is found in the virtual environment and the relationship between the position and the starting point, testing the ability of the patient to call the spatial relationship between virtual positions.

[0027] Table III provides non-limiting examples of parameters associated with a user's motion and activity, and exemplary tests, devices, and methods, referred to herein for ease of reference as "motor parameters," that may be measured. The use of portable monitoring, physiological sensing, and portable computing devices allows for the measurement of motor parameters. Using embedded accelerometers, GPS, and cameras, the user's motion may be captured and quantified to understand the health impact on them and the relationship to health-related parameters. Range of motion and gait analysis may be assessed in a clinical setting using appropriate motion capture and camera devices.

[0028] Table III

[0029] Table IV provides non-limiting examples of parameters associated with a user's sensory acuity, as well as exemplary tests, apparatuses, and methods, referred to herein for ease of reference as "sensory parameters," that may be measured.

[0030] Table IV

[0031] Table V provides non-limiting examples of parameters associated with a user's sleep amount, stage, and quality, as well as exemplary tests, devices, and methods, referred to herein for ease of reference as "sleep parameters," that may be measured.

[0032] Table V

[0033] Table VI provides non-limiting examples and exemplary tests, apparatuses, and methods of parameters extracted by locating biomarkers associated with a user, which are referred to herein as "biomarker parameters" for ease of reference, and which may be measured. Biomarkers may also include imaging and physiological biomarkers associated with chronic health states and improvements or deteriorations in chronic health states.

[0034] Table VI

[0035] Table VII provides non-limiting examples of psychosocial and behavioral parameters that may be measured, as well as exemplary tests, devices, and methods, which for ease of reference are referred to herein as "psychosocial parameters."

[0036] Table VII

[0037] Behavioral and psychosocial parameters can measure the user's functioning as well as subjective / self-report questionnaires. Subjective / self-report questionnaires can be collected in a clinical / laboratory setting or in a natural environment, in a field non-clinical setting (such as when the user is at home, work, or other non-clinical environment). Smart devices such as smartphones, tablets, or personal computers can be used to administer the subjective / self-report questionnaires. Using embedded accelerometers and cameras, these smart devices can also be used to capture facial expression analysis to analyze the user's facial expressions, which may indicate emotions, anxiety, depression, excitement, and fatigue. This emotion detection can be performed using a suitable predictive model trained on the face of a user who mimics or experiences a specified emotion or mental state.

[0038] Other elements of monitoring can include monitoring the user's compliance with the use of smartphones, televisions, portable devices, portable devices. For example, the system can send messages to users asking about their health level, overall mood, or the status of any other health-related parameters on the portable computing device. A measure of compliance can be determined based on the percentage of these messages (to which the user responds through a user interface on the portable computing device) and is used to indicate how likely the patient is to comply with the treatment plan and monitoring. In addition, in cases where the parameters are not easily extracted from a wearable or portable device, the parameters can be retrieved from an electronic health record (EHR) database. In particular, biomarkers and motion parameters can be retrieved from the EHR, as well as other parameters including medical history, prescribed medications, demographic parameters, age, height, weight, and other medically relevant parameters.

[0039] At 204, a feature set is determined from the extracted health-related parameters as categorical and continuous parameters representing the health-related parameters. In one example, the parameters may include descriptive statistics, such as measures of central tendency (e.g., median, mode, arithmetic mean, or geometric mean) and measures of deviation (e.g., range, interquartile range, variance, standard deviation, etc.) of the time series of the monitored parameters, as well as the time series itself. Specifically, for at least one parameter, the feature set provided to the prediction model may include: two values ​​representing the parameter values ​​at different times, or a single value, such as a measure of central tendency or a measure of deviation representing the values ​​of the parameter at multiple time points.

[0040] In one embodiment, features may be generated by wavelet transforming a time series of one or more parameter values ​​to provide a set of wavelet coefficients. It should be understood that the wavelet transform used herein is two-dimensional, and thus the coefficients may be conceived as a two-dimensional array across time and across frequency or scale.

[0041] For a specified parameter time series x i , the wavelet coefficients generated in the wavelet decomposition W a (n) Can be limited to: (Formula 1) in, Ψ is the wavelet function, M is the length of the time series, a and n Limit the coefficient calculation location.

[0042] A facial expression classifier (not shown) can evaluate recorded data from a camera and / or recorded images or videos of a user's face from a smartphone or other mobile device to assign an emotional state to the user at different times of the day. The extracted features can be categorical, representing the most likely emotional state of the user, or continuous, for example, as a time series of probability values ​​for various emotional states (e.g., anxiety, discomfort, anger, etc.) determined by the facial expression classifier. One or more image classifiers reduce the provided medical images to categorical or continuous features for use by the predictive model. For example, the image classifier can generate connectivity parameters representing the interconnectivity of different locations within the patient's brain, parameters representing the functional structure of the brain, parameters representing the vasculature of the brain or eye, and parameters representing eye movement and pupil dilation. It should be understood that each of the facial expression classifier and one or more image classifiers can be implemented using one or more models for predictive models discussed below.

[0043] In one embodiment, the extracted features from the health-related parameter set can be collected into an aggregate parameter set. It should be understood that each aggregate parameter can be a health-related parameter set, a function of a parameter from a health-related parameter set, or a weighted combination of features extracted from health-related parameters. Accordingly, a specified aggregate parameter can represent multiple health-related parameters, and typically, the multiple health-related parameters represented by each aggregate parameter will be related, so that the aggregate parameter represents a specific health area of ​​the user. Typically, each aggregate parameter can use parameters from different sources. In some embodiments, the aggregate parameters can be provided to multiple predictive models (not shown), each predictive model receiving a unique appropriate subset of the aggregate parameters. Each predictive model can provide different clinical parameters representing different aspects of the user's health, so that the aggregate parameters can be used for multiple purposes of assessing the user's health.

[0044] It should be understood that the specific parameters and features used for screening may vary with implementation. For example, screening for pain may focus on measured autonomic parameters, sleep parameters, data in patient medical records, functional brain imaging, and patient self-reports. In another example, screening for addiction, PTSD, phobias, panic disorders, anxiety, depression, and schizophrenia may focus on autonomic, behavioral, cognitive, and psychosocial parameters, data in patient medical records, and patient self-reports in various questionnaires. Screening for dementia may focus on brain and eye imaging, autonomic, psychosocial, sensory, biomarker, and cognitive parameters, and data in patient medical records. Patients' responses to cues appropriate for their disease (self-reported and in the form of measured physiological values) may also be recorded and used to generate features, particularly for addiction, depression, PTSD, anxiety, and phobias. Screening for movement disorders such as Parkinson's disease (PD) and Huntington's disease, or sequelae of stroke, can focus on motor, autonomic, psychosocial, and cognitive parameters. It should be understood that the methods herein focus on using our methods to monitor and improve the behavioral and cognitive and autonomic nervous system aspects of these diseases. As a specific example, our methods help improve visual deficits in Parkinson's disease, such as visual perception, blurred vision, judging distance, depth perception, and other visual and spatial deficits that affect so many PD patients. Screening for obsessive-compulsive disorder and neurodevelopmental disorders, such as autism and autism-related disorders, can focus on autonomic, psychosocial, and cognitive parameters.

[0045] Figure 3A method 300 for using cues to generate features for screening a patient's disease is shown. It should be understood that the method 300 describes eliciting a response to a cue in the context of screening a patient, and the response elicited by the cue can also be used to detect or predict the onset of symptoms in the patient, to determine, refine and optimize the appropriate target location or appropriate parameters for focused ultrasound treatment, or to evaluate the effectiveness of focused ultrasound treatment. The presentation of cues can also be used as part of the treatment of various diseases, which will be discussed in more detail below. At 302, a set of parameters representing the disease is collected from the patient as a first set of values. The specific parameters will vary with the disease for which the patient is screened, but in general, the parameters can include the patient's responses to various standard questionnaires and physiological parameters measured from the patient. At 304, cues associated with the disease are presented to the patient. It should be understood that cues specific to the disease will be selected and can include not only visual cues, but also auditory, taste, tactile and olfactory cues. For example, if the patient suffers from an addiction to alcohol, a cue can be selected to represent the sound of a beer bottle opening or the taste or smell of the patient's preferred alcoholic beverage. It should be understood that a virtual reality or augmented reality system can be used to simulate the cues. An example of presentation of cues can be found in U.S. Patent Publication US2021 / 0162217, filed on December 2, 2020, entitled "Methods and systems of entering and monitoring addiction using cue reactive", the entire contents of which are incorporated herein by reference in their entirety. At 306, after the cues are presented as a second set of values, a set of parameters representing the disease is collected from the patient. At 308, a set of screening features is determined from the first set of values ​​and the second set of values. In one example, the screening feature is determined as a function of a first value and a second value of a specified parameter, such as a difference or ratio between the values.

[0046] Other features can be extracted from imaging taken of patients. The feature set could represent the connectivity of different regions of the brain, as differences in connectivity could indicate different risk levels for various diseases. Figure 4A method 400 for extracting connectivity features to evaluate a patient for focused ultrasound therapy is shown. It should be understood that the method 400 describes extracting connectivity features in the context of screening a patient, which connectivity features can also be used to detect or predict the onset of symptoms in a patient, to determine the appropriate location or appropriate parameters for focused ultrasound therapy, or to evaluate the effectiveness of focused ultrasound therapy. At 402, a first image representing the structure of the brain is acquired from a first imaging system. At 404, a second image representing the connectivity of the brain is acquired from one of the first imaging system and the second imaging system. In one embodiment, the second image is acquired by diffusion tensor imaging. At 406, the first image is segmented into a plurality of sub-regions of the brain to generate a segmented first image, so that each of at least a subset of a plurality of voxels comprising the first image is associated with one of the plurality of sub-regions. It should be understood that for a given disease, not all portions of the first image may be of interest, and therefore only those voxels representing the plurality of sub-regions may be included in the segmented portion. At 408, features are extracted from the segmented first and second images. In one example, the image is provided directly to a deep learning system, such as a convolutional neural network, such that individual pixel intensities or chromaticity values ​​in the image are the extracted features. Alternatively, other connectivity parameters may be specific to the disease, representing the density of connections between a first region of interest associated with a possible focused ultrasound treatment target and other regions of interest within the brain.

[0047] In another embodiment, Figure 5 A method 500 for generating features from imaging of an eye to evaluate a patient for focused ultrasound therapy is shown. It should be understood that the method 500 describes extracting parameters from imaging of the eye in the context of screening a patient, and features representing the retina, optic nerve, and associated vasculature of these structures can also be used to detect or predict the onset of symptoms in a patient, for determining an appropriate location or appropriate parameters for focused ultrasound therapy, or for evaluating the effectiveness of focused ultrasound therapy. At 502, a first image representing a patient's brain is acquired from a first imaging system. In one example, the first imaging system is a magnetic resonance imaging (MRI) imager. Additionally or alternatively, the first image can be acquired by diffusion tensor imaging. In one example, the representation of the first image represents a cortical contour of the brain. In another example, the representation of the first image represents the vasculature of the brain. In another example, the representation of the first image represents protein plaques / tangles within the brain, such as beta-amyloid or tau protein. In yet another example, the representation of the first image represents the connectivity of the brain. At 504, a second image representing one of the patient's retina, optic nerve, and associated vasculature is acquired from a second imaging system. In one example, the second image is an optical coherence tomography angiography image.

[0048] At 506, features are extracted from each of the first and second images for use in a predictive model. In one example, it will be appreciated that the representation of each image may include chromaticity values ​​associated with each pixel or voxel in the image, thereby providing all or most of the data associated with each image to the machine learning system. Alternatively, a set of numerical features representing the content of each image, such as the total area or volume of a structure or tissue of interest, may be extracted. In one example, the patient's pupil may be imaged to provide additional parameters representing at least one of eye tracking data, eye movement, pupil size, and pupil size change. In one example, various biometric parameters extracted from the brain image include parameters related to connectivity, brain convolutions, volumes, vascular load, networks, white matter load, gray / white matter ratio, cortical thickness, sulcal depth, and other measurements. Images of the brain may also be segmented to determine these parameters for specific regions of the brain. Similarly, numerical parameters can be extracted from images of the eye, including the volume, thickness or texture of the retina or individual retinal layers at one or more locations, or the average thickness of the retina or one or more layers across locations, values ​​representing the vascular pattern and density of the retina or individual retinal layers, the size of the foveal avascular zone, the height, width or volume of the optic chiasm, the height, width or volume of the optic nerves within the orbit, the height, width or volume of the optic nerves within the cranium, and the total area of ​​the vasculature in the image.

[0049] return Figure 2 At 206, a clinical parameter is assigned to the patient from the extracted features by the prediction model. The prediction model may utilize one or more pattern recognition algorithms, each of which analyzes the extracted features or a subset of the extracted features to assign a continuous or categorical clinical parameter to the user, the continuous or categorical clinical parameter representing the likelihood that the patient will benefit from focused ultrasound therapy. In one example, the clinical parameter may be a continuous parameter representing the likelihood that the patient has a disease that can be treated by focused ultrasound therapy, the likelihood that the patient has a specific disease (such as Alzheimer's) that can be treated by focused ultrasound therapy, the likelihood that the patient will benefit from focused ultrasound therapy given a known diagnosis, or the likelihood that the patient will benefit from focused ultrasound therapy at a specific location or region of the brain. In another example, the clinical parameter can be a categorical parameter indicating whether the patient has a disease that can be treated by focused ultrasound therapy, a category indicating a change in symptoms associated with the condition or disease (e.g., "improved," "stable," "worsened"), a category indicating the predicted response to focused ultrasound therapy in general or at a specific location or region of the brain, the severity of whether the patient has a specific disease that can be treated by focused ultrasound therapy (such as Alzheimer's disease), whether the patient would benefit from focused ultrasound therapy given the diagnosis, or a category indicating a range of likelihood that the patient falls into one of these categories.

[0050] Where multiple classification or regression models are used, an arbitration element may be utilized to provide coherent results from the multiple models. The training process for a given classifier will vary with how it is implemented, but training typically involves statistically aggregating the training data into one or more parameters associated with the output class. The training process may be completed on a remote system and / or on a local device or wearable application. The training process may be implemented in a federated or non-federated manner. For rule-based models such as decision trees, domain knowledge, such as provided by one or more human experts or extracted from existing research data, may be used in place of or in addition to the training data to select rules for classifying users using extracted features. Any of a variety of techniques may be used for the classification algorithm, including support vector machines, regression models, self-organizing maps, fuzzy logic systems, data fusion processes, enhancement and packaging methods, rule-based systems, or artificial neural networks.

[0051] Federated learning (also called collaborative learning) is a prediction technique that trains algorithms on multiple decentralized edge devices or servers that save local data samples without exchanging their data samples. This approach contrasts with traditional centralized prediction techniques where all data samples are uploaded to one server, and more classical decentralized approaches that assume that local data samples are distributed identically. Federated learning enables multiple participants to build a common, stable prediction model without sharing data, thus solving key issues such as data privacy, data security, data access rights, and access to heterogeneous data. Its applications are distributed across many industries, including defense, telecommunications, IoT, or pharmaceuticals.

[0052] For example, an SVM classifier can conceptually divide boundaries in an N-dimensional feature space using multiple functions called hyperplanes, where each of the N dimensions represents an associated feature of a feature vector. The boundaries define the range of feature values ​​associated with each category. Therefore, the output category and associated confidence value of a given input feature vector can be determined based on its position relative to the boundary in the feature space. In one embodiment, SVM can be implemented using a kernel method using a linear or nonlinear kernel.

[0053] The ANN classifier includes multiple nodes with multiple interconnections. The values ​​from the feature vector are provided to multiple input nodes. Each input node provides these input values ​​to a layer of one or more intermediate nodes. The intermediate node is specified to receive one or more output values ​​from the previous node. The received values ​​are weighted according to a series of weights established during the training of the classifier. The intermediate node converts the value it receives into a single output according to the transfer function at the node. For example, the intermediate node can sum the received values ​​and perform a binary step function on the sum. The final node layer provides confidence values ​​for the output categories of the ANN, wherein each node has a related value representing the confidence of one of the associated output categories of the classifier. Another example is to use an autoencoder to detect abnormal values ​​in health-related parameters as an anomaly detector to identify when various parameters exceed the normal range of an individual.

[0054] Many ANN classifiers are fully connected and feed-forward. However, convolutional neural networks include convolutional layers in which nodes from the previous layer are connected to only a subset of nodes in the convolutional layer. Recurrent neural networks are a class of neural networks in which the connections between nodes form a directed graph along a time series. Unlike feed-forward networks, recurrent neural networks can incorporate feedback from states caused by earlier inputs, so that the output of a recurrent neural network for a given input can be a function not only of that input, but also of one or more previous inputs. For example, a long short-term memory (LSTM) network is an improved version of a recurrent neural network that makes it easier to keep past data in memory.

[0055] A rule-based classifier applies a set of logical rules to the extracted features to select an output class. Typically, the rules are applied in sequence, with the logical outcome of each step influencing the analysis of the following steps. The specific rules and their sequence can be determined based on any or all of the training data, simulated reasoning from previous situations, or existing domain knowledge. An example of a rule-based classifier is the decision tree algorithm, in which the values ​​of the features in the feature set are compared to corresponding thresholds in a hierarchical tree structure to select the class of the feature vector. The random forest classifier is a modification of the decision tree algorithm using a guided aggregation or "bagging" approach. In this approach, multiple decision trees are trained on a random sample of the training set, and the average (e.g., mean, median, or mode) results across the multiple decision trees are returned. For classification tasks, the results of each tree are classified, so the modal results can be used.

[0056] In some embodiments, the prediction model can be retrained to adjust various parameters of the model based on the accuracy of the predictions made by the model. Parameters associated with the model, such as internal weights and thresholds for generating categorical inputs or outputs from continuous values, can be adjusted based on the difference between the actual results and the predicted results. In one example, the actual value of the clinical parameter for a given patient can be determined as a categorical or continuous result, for example, based on the patient's degree of improvement, diagnosis, or other suitable results, and combined with the patient's feature set to provide additional training samples for the model.

[0057] Return to Figure 1If it is determined that focused ultrasound treatment is not suitable for the patient (N), the method proceeds to 104, where the patient is assigned to an alternative treatment. Alternative treatments may include medications, transcranial magnetic stimulation, deep brain stimulation, biologics, surgical interventions, changes in settings of existing spinal cord stimulators, behavioral and social interventions, digital interventions through portable devices, mindfulness methods, social media methods, home visits by care providers, directing individuals to clinics, emergency rooms or hospitals, or directing users to undergo additional testing. If it is determined that focused ultrasound treatment is suitable (Y), the method proceeds to 105, where a specific focused ultrasound treatment regimen is selected. It should be understood that the specific focused ultrasound treatment regimen may vary with the severity of the disease. If a patient is determined to be at high risk for the disease or is determined to have a mild or moderate disease, neuromodulation may be performed on them via focused ultrasound. Patients with more severe disease manifestations may receive more intensive treatments, such as drugs or antibodies with targeted destruction of the patient's blood-brain barrier by focusing ultrasound combined with a microbubble dose provided to the patient's bloodstream, but it should be understood that in addition to neuromodulation by focused ultrasound, patients can also receive focused ultrasound treatment to destroy the blood-brain barrier. In one example, for the treatment of neurodegeneration, the treatment can utilize a therapeutic agent to remove protein plaques or other pathological substances (i.e., tau tangles) in the brain, such as anti-β-amyloid antibodies (e.g., Aducanumab, Lecanemab, etc.). In the case where the therapeutic agent is an anti-β-amyloid antibody (e.g., targeting β-amyloid and its components), such an antibody (e.g., Aducanumab) can be intravenously injected, for example, at a dose of at least 0.1 mg / kg, 1 mg / kg, 3 mg / kg, or 6 mg / kg. For other antibodies, other doses may also be injected. In some aspects, the therapeutic agent is Lecanemab, and its dose is, for example, 10 mg / kg every two weeks. Lecanemab targets the soluble "protofil" form of amyloid-β, and also binds (albeit more weakly) to extracellular amyloid deposits known as plaques, to which Aduhelm primarily binds. Other therapeutic agents may be used in combination with targeted disruption of the blood-brain barrier for other diseases.

[0058] The method then proceeds to 106 where the patient is monitored to predict or detect the onset of symptoms. It should be appreciated that the timing of focused ultrasound therapy may be individualized and dependent upon the onset of symptoms and progression of the disease as well as various treatment regimens associated with the focused ultrasound modality and any therapeutic agents introduced during therapy.

[0059] Figure 6A method 600 for monitoring a patient to predict or detect the onset of symptoms associated with a disease of the patient is shown. At 602, parameters of the patient are monitored. It should be understood that the monitored parameters can include any of the parameters discussed in Tables I-VII, and these parameters can be monitored, for example, using a wearable device, a portable device (such as a cell phone, tablet computer, and personal computer), or by periodically querying an electronic health record system. At 604, suitable features are generated based on the monitored parameters. At least a subset of the monitored parameters can be represented by features representing changes in the parameters over time, and can be represented as a time series or a descriptive statistic, such as a measure of central tendency or a measure representing the deviation of the values ​​of the parameter at multiple time points.

[0060] At 606, the extracted parameters may be provided to a predictive model to determine whether the patient is experiencing or is expected to begin experiencing symptoms associated with a specified disease that may be addressed by focused ultrasound therapy. For example, the predictive model may predict or detect the onset of craving in a patient receiving addiction treatment, the onset of pain in a patient receiving chronic pain treatment, the worsening of damage caused by a neurodegenerative disease, hallucinations or delusions associated with schizophrenia, or acute episodes associated with depression, PTSD, drug craving, behavioral addiction, or anxiety. In one example, the predictive model may assign a continuous parameter corresponding to the likelihood that the user has or will soon experience symptoms associated with the disease, an increase in patient stress, an increase in patient anxiety, the likelihood that the user will experience an increase in symptoms associated with the disease, the likelihood that the user will use an addictive substance or engage in addictive behavior during rehabilitation or treatment, an increase in reliance on an addictive substance or addictive behavior, a current or expected user pain level, a user's expected performance level associated with a current or future time in a particular activity or occupation, or a change in symptoms associated with a condition or disease.

[0061] In another example, the predictive model may assign classification parameters corresponding to the range of possibilities described above, the presence or expected presence of a particular condition or disease, a set of categories representing a patient's readiness for a particular activity or occupation, categories representing changes in symptoms associated with a condition or disease (e.g., "improved," "stable," "worsened"), or categories representing user status (e.g., "normal," "stressed," "sick"). In one embodiment, the predictive model may include a component model that predicts future values ​​of an aggregate parameter, such as a convolutional neural network having one or more two-dimensional arrays of wavelet transform coefficients as input. The wavelet coefficients detect not only changes in time, but also changes in temporal patterns, and therefore, may reflect changes in a user's normal biorhythm. It should be understood that the specified component model may use data other than the aggregate parameter, such as other extracted features, to provide these predictions. Additionally or alternatively, the predictive model may use a component model that predicts current or future values ​​of the aggregate parameter, and then use these metrics as features for generating the predictive model output.

[0062] In one example, the prediction model is an anomaly detection model that detects deviations from expected values ​​within feature space and determines when these deviations are significant. The anomaly detection model can be trained on data from the user (which establishes a baseline of the user's expected values) and / or on data collected from other users. In one example, the prediction model is initially trained on data collected from other users while values ​​for a subset of the set of health-related parameters are collected from the user over a period of time. Once sufficient data is available for the user, the prediction model is retrained on the collected values ​​for the subset of the set of health-related parameters.

[0063] In one example, when the disease is an addiction, phobia, anxiety disorder, or post-traumatic stress disorder, the predictive model can utilize parameters such as self-reported anxiety and parameters associated with anxiety (such as physiological parameters and sleep parameters). In another example, where the disease is a neurodegenerative disease, the predictive model can utilize cognitive tests and other parameters presented to the patient in a clinical or non-clinical setting (e.g., via a smartphone, tablet, or home computer). In addition, sleep parameters and imaging biomarkers can be used to track the progression of the disease and the need for additional treatment. In another example, when the disease is depression, chronic pain, or schizophrenia, patients self-report through questionnaires and physiological data collected from wearable and portable devices. Sleep data can also be used to monitor depression and chronic pain.

[0064] Figure 7Another example of a method 700 for assessing the onset of symptoms in a patient requiring treatment by focused ultrasound therapy is shown. The result of the method is a clinical parameter indicating whether the patient is experiencing or is expected to experience symptoms that will require focused ultrasound therapy. At 702, a first plurality of health-related parameters representing a user are monitored at a physiological sensing device over a defined period of time. In one example, the first plurality of health-related parameters may include parameters representing the user's autonomic function and parameters representing the user's sleep and circadian rhythm. At 704, a second plurality of health-related parameters representing the user are obtained by a portable computing device. In one example, the second plurality of health-related parameters may include parameters representing the user's cognitive and / or social behavioral health. At 706, a third plurality of health-related parameters representing the user are retrieved from an electronic health record (EHR) system. The third plurality of health-related parameters may include, for example, parameters representing the user's musculoskeletal health, genome, and various biomarkers. The first plurality of health-related parameters, the second plurality of health-related parameters, and the third plurality of health-related parameters together form a health-related parameter set.

[0065] At 708, aggregated parameter sets are generated from the health-related parameter sets, each of which includes a unique suitable subset of the health-related parameter set. In one example, the aggregated parameter set includes at least a first aggregated parameter representing the user's autonomic function, a second aggregated parameter representing the user's cognitive function, and a third aggregated parameter representing the user's movement and musculoskeletal health. In another example, the aggregated parameter set includes at least a first aggregated parameter representing the user's sleep and circadian rhythm, a second aggregated parameter representing the user's social behavioral function, and a third aggregated parameter representing the user's biomarkers and genome.

[0066] At 710, clinical parameters are assigned to users by a predictive model based on a subset of the aggregated parameter set. In one example, the clinical parameter is a value representing the overall health status of the user, and the subset of the aggregated parameter set includes the entire aggregated parameter set. In another example, the subset of the aggregated parameter set is a suitable subset. It should be understood that the aggregated parameters can be provided to multiple predictive models, each predictive model receiving a unique subset of the aggregated parameter set. In one example, the predictive model is an anomaly detection model that detects deviations from expected values ​​within a feature space and determines when these deviations are significant. The anomaly detection model can be trained on the user's data (which establishes a baseline for the user's expected values) or on data collected from other users. In one example, the predictive model is initially trained on data collected from other users while values ​​of a subset of health-related parameter sets are collected from the user over a period of time. Once there is a sufficient amount of data available for the user, the predictive model can be retrained on the collected values ​​of the subset of health-related parameter sets.

[0067] In one example, wavelet decomposition is performed on the time series for at least one aggregation parameter to provide a set of wavelet coefficients, and the set of wavelet coefficients or one or more values ​​derived from the set of wavelet coefficients can be provided to the prediction model. Additionally or alternatively, a predicted value representing a future value of a specified aggregation parameter can be assigned to a user based on the value of the subset of aggregation parameters, and the value assigned to the user can be assigned based on the predicted value.

[0068] Additionally or alternatively, feedback in the form of self-reported symptom levels, records of subsequent clinical visits, or measured future values ​​of parameters can be used to refine the predictive model. For example, the self-reported or measured values ​​can be compared to the values ​​assigned to the user by the predictive model, and the parameters associated with the predictive model can be changed based on the comparison. In one example, this can be achieved by generating rewards that reinforce the learning process based on the similarity of the measured results to the values ​​assigned to the user, and changing the parameters through the reinforcement learning process.

[0069] It should be understood that each health value and the value assigned to the user can be provided to one or more of the user, the user's health care provider, the user's care team, the research team, the user's workplace, the user's sports team, the insurance company, or other entities of interest, for example, through a user interface or a network interface. This allows the value to be used to determine the user's care and activities. The feedback provided to the user can be used to improve the user's awareness, perception, and interpretation of being in an overall positive state and a negative state, thereby allowing the user to learn strategies for avoiding negative states and inducing positive states. The health data provided can also be used for improvement or optimization of cognitive, motor, sensory, and behavioral functions, and generally attempts to improve the user's quality of life by suggesting actions for the user in response to changes in clinical parameters. For example, when a clinical parameter exceeds a predetermined range of values, a message suggesting a course of action for the user can be sent to the user's portable computing device.

[0070] return Figure 1If it is determined that the patient is not experiencing symptoms that should be addressed by focused ultrasound therapy (N), the method remains at 106. If it is determined that focused ultrasound therapy is desirable, the method proceeds to 108, where appropriate parameters for focused ultrasound therapy are determined. It should be understood that appropriate parameters for focused ultrasound therapy may include a target location for focused ultrasound therapy and parameters related to the dose of focused ultrasound therapy, such as the total duration of modulation, the duration of each application, the number of applications, and the intensity or power of each application. Example parameters for focused ultrasound therapy may include sonication dose, power (<150 W), ultrasound duration (<60 min), frequency direction, repetition time (e.g., five seconds on and ten seconds off), pulse duration (e.g., 100 ms on, 900 ms off, continuous or burst, energy per minute (e.g., 0 J / min-290 J / min), frequency (e.g., 1 MHz-3 MHz), number of elements (e.g., 1-1024), and waveform form.

[0071] Figure 8 A method 800 for targeting focused ultrasound therapy in a patient's brain to treat a disease is shown. At 802, a first image representing a structure of the brain is acquired from a first imaging system. At 804, a second image representing connectivity of the brain is acquired from one of the first imaging system and a second imaging system. In one embodiment, the second image is one of a set of graphs representing connectivity of the brain, each image being acquired by applying focused ultrasound therapy at a location within a region of interest associated with the image and measuring an activity level in at least one of a plurality of subregions of the brain in response to the applied focused ultrasound therapy. Additionally or alternatively, the second image may be acquired by diffusion tensor imaging.

[0072] At 806, the first image is segmented into a plurality of sub-regions of the brain to generate a segmented first image, such that each of at least a subset of the plurality of voxels comprising the first image is associated with one of the plurality of sub-regions. It should be understood that for a given disease, not all portions of the first image may be of interest, and thus only those voxels representing the plurality of sub-regions may be included in the segmented portion. At 808, a location within the region of interest of the brain is selected as a target for focused ultrasound therapy based on the segmented first and second images. It should be understood that the target is selected not only to be close to and connected to the brain region where the modulation is expected to be beneficial, but also to avoid connecting to the brain region that may cause undesirable effects. For example, accidental stimulation of the amygdala may cause anxiety or fear, which may be counterproductive in the treatment of many diseases. Small changes in the modulation location can greatly change the amount of accidental modulation applied to undesirable regions and the amount of modulation applied to desired regions of the brain.

[0073] In order to treat neurodegenerative diseases by opening the blood-brain barrier using focused ultrasound, specific targets can be determined by imaging, particularly by locating areas with high concentrations of protein plaques or other pathological substances. Treatment of neurocognitive diseases can generally focus on the hippocampus, olfactory cortex, frontal lobe, parietal lobe, spinal belt, any cortical area with protein plaques or other pathological substances in the brain, nuclei, entorhinal cortex, thalamic nuclei, nucleus accumbens, fornix, mammillary bodies, central internal capsule and mammillary thalamic tract. For the treatment of addiction or anxiety, the target can be within one of the following: nucleus accumbens, insula, cingulate cortex, dorsal prefrontal cortex, interlaminar nucleus of the thalamus and anterior subthalamic nucleus. In order to treat cognitive and behavioral aspects of Parkinson's disease, particularly the visual and spatial deficits commonly associated with the disease, the selected targets can be within the subthalamic nucleus, globus pallidus, lenticular nucleus, nucleus accumbens, ventral striatum, globus pallidus, ventral internal capsule and thalamic transmission nucleus. For treatment of cognitive and behavioral aspects of Huntington's disease, targets may be within the caudate nucleus, lentiform nucleus, nucleus accumbens, ventral striatum, nucleus pallidus, ventral capsule, and thalamic nuclei.

[0074] For the treatment of PTSD, treatment targets can be selected from any of the following: amygdala, prefrontal cortex, hippocampus, basolateral complex of amygdala, lateral nucleus, basal nucleus, sensory nucleus of thalamus, ventromedial prefrontal cortex, basolateral complex, basal infralimbic subregion of prefrontal cortex, and nucleus accumbens. For the treatment of chronic pain, treatment targets can be selected in the sensory lateral thalamus, inner laminar nucleus of thalamus, sensory nucleus, internal capsule, pericardial / pericardial gray matter, anterior cingulate cortex, spinal cord, vagus nerve, insula, and cingulate cortex. For the treatment of schizophrenia, targets can be selected in the nucleus accumbens, inner laminar nucleus, thalamic nucleus, olfactory, sensory, visual, gustatory cortex and corresponding thalamic transmitter nuclei, medial and lateral geniculate bodies, transmitter nuclei of all pathways, angular cortex, and Herschel's gyrus. For obsessive-compulsive disorder, suitable targets can be found in the nucleus accumbens, ventral striatum, and ventral internal capsule. Suitable targets for treating cognitive and behavioral aspects of neurodevelopmental disorders such as autism and autism-related disorders may include the nucleus accumbens, ventral striatum, ventral capsule, and thalamic nuclei. To aid in stroke recovery, the specific target will depend on the area damaged by the stroke.

[0075] Fig. 9Another method 900 for targeting focused ultrasound therapy in a patient's brain to treat a disease using acute feedback is shown. At 902, a first image representing a structure of the brain is acquired from a first imaging system. At 904, a second image representing connectivity of the brain is acquired from one of the first imaging system and a second imaging system. In one embodiment, the second image is one of a set of graphs representing connectivity of the brain, each image being acquired by applying focused ultrasound therapy at a location within a region of interest associated with the image, and measuring an activity level in at least one of a plurality of subregions of the brain in response to the applied focused ultrasound therapy. Additionally or alternatively, the second image may be acquired by diffusion tensor imaging.

[0076] At 906, the first image is segmented into a plurality of sub-regions of the brain to generate a segmented first image such that each of at least a subset of a plurality of voxels comprising the first image is associated with one of the plurality of sub-regions. It should be appreciated that for a given disease, not all portions of the first image may be relevant to the given disease, and thus only those voxels representing the plurality of sub-regions may be included in the segmented portion. At 908, a location within a region of interest of the brain is selected as an initial target for focused ultrasound therapy based on the segmented first and second images. It should be appreciated that the given therapy may be performed at a plurality of separate locations, and the analysis may be performed for each of those locations.

[0077] At 910, focused ultrasound therapy is applied to the selected location, and patient feedback on the applied focused ultrasound therapy is measured at 912. Feedback can include clinician observations of the patient's appearance and behavior, patient self-reports of disease symptoms, electrical activity measured in the brain, and biometric parameters, as described in Tables I-VII above. In particular, imaging can be used to determine effectiveness and potential side effects at a designated location. For example, when focused ultrasound therapy includes introducing antibodies into the brain, a fluid-attenuated inversion recovery (FLAIR) MRI sequence can indicate the extent of antibody delivery and blood-brain barrier opening. If the FLAIR signal at the location is too high, further application of energy to the location in the designated treatment can be avoided. Contrast MRI can be used in a similar manner. For the treatment of neurocognitive diseases, amyloid positron emission tomography (PET) can be used to assess protein reduction at a designated location. Another imaging quantitative feedback is Scanned GRE, which can show A dark spot in an MRI scotomas that can indicate damage and therefore a need to avoid further energy application to that location. Acoustic feedback can also be used in focused ultrasound applications to assess the safety and effectiveness of treatment. These imaging assessments provide a quantitative personalized assessment of treatment safety, dosage, and benefit and are akin to digital fingerprinting.

[0078] At 914, the effectiveness of the focused ultrasound therapy is determined based on the measured feedback. This can be achieved, for example, by a rule-based approach or by providing measured feedback parameters to a predictive model. If the modulation is determined to be effective (Y), the position of the focused ultrasound therapy is maintained at 916 for subsequent focused ultrasound therapy, and the method terminates. If invalid (N), the selected position is determined to be invalid, and the method returns to 908 to select a new position within the region of interest as a target for the focused ultrasound therapy. The position of the focused ultrasound therapy target is personalized and can be modified in real time from acute feedback during treatment, or for subsequent treatment. The dose of the focused ultrasound therapy can also be adjusted.

[0079] Fig.10 A method 1000 for selecting a dose for a focused ultrasound treatment is shown. It should be understood that the dose for the focused ultrasound treatment is defined by parameters such as the frequency of application of the focused ultrasound treatment, the energy distribution of the focused ultrasound treatment, the shape of the energy field, the direction of the energy field, the pulse rate of the focused ultrasound, the duration of discrete applications of energy, the number of discrete applications of energy, the total duration of the treatment, the dose of microbubbles provided to the patient for selective opening of the blood-brain barrier, the type or dose of the therapeutic agent provided in conjunction with the treatment, and the intensity or power of the treatment. At 1002, an initial dose for modulation can be selected. It should be understood that the initial dose can be standard across patients, but in the illustrated embodiment, the initial dose is patient-specific and can be determined based on factors such as the severity of the disease, the size of the patient's head, the type of target tissue, the patient's skull density ratio, and the location of the focused ultrasound treatment relative to the skull. For example, patients with lower skull density ratios may require higher doses (e.g., higher intensity or power), and the impact of location can vary from treatment to treatment, where locations close to the skull require different settings in the focused ultrasound because fewer transducers can effectively target the location.

[0080] At 1004, focused ultrasound therapy is applied to the selected location, and at 1006, patient feedback on the applied focused ultrasound therapy is measured. This feedback may include clinician observations of the patient's appearance and behavior, patient self-reports of disease symptoms, electrical activity measured in the brain, and biometric parameters, as described in Tables I-VII above. In particular, imaging can be used to determine the effectiveness and potential side effects of a given treatment. For example, when focused ultrasound therapy includes introducing antibodies into the brain, a fluid attenuated inversion recovery (FLAIR) MRI sequence can indicate the extent of antibody delivery and blood-brain barrier opening. If the FLAIR signal at the location is too high, the subsequent portion of the treatment or the dose of the subsequent treatment (e.g., power, duration, microbubble dose, etc.) can be reduced. Contrast MRI can be used in a similar manner. To treat neurocognitive diseases, amyloid positron emission tomography (PET) can be used to assess the reduction of protein. Dark spots in MRI scotomas can indicate damage and therefore the need for a lower dose or a different approach. Acoustic feedback can also be used in focused ultrasound applications to assess the safety and effectiveness of treatment. In particular, the size and expansion of microbubbles provided to patients to allow the blood-brain barrier to be opened by applied ultrasound can be monitored to ensure the safety and effectiveness of the treatment.

[0081] At 1008, the effectiveness of the focused ultrasound therapy is determined based on the measured feedback. For example, this can be achieved through a rule-based approach or by providing the measured feedback parameters to a predictive model. If the modulation is determined to be effective (Y), then at 1010, the selected dose parameters for the focused ultrasound therapy are retained for subsequent energy applications, and the method terminates. If not effective (N), then the selected dose parameters are determined to be invalid, and the method returns to 1002 to select new dose parameters for the therapy.

[0082] return Figure 1, once treatment has been applied, feedback is collected from the patient to determine whether the treatment has been effective. It should be understood that this can be performed during or immediately after treatment ("acute feedback"), in a short time after treatment (e.g., five hours to five days) ("subacute feedback"), or in a longer time after treatment (e.g., more than five days after treatment) ("chronic feedback"). For example, feedback can be obtained as any parameter listed in Tables I-VII. In order to collect acute feedback, tasks can be presented immediately during or after treatment to determine whether the patient's performance will increase, decrease or otherwise change due to treatment. The feedback collected can also include self-reports from patients, clinician's observations, measured biometric parameters (such as heart rate variability and blood pressure) and other related parameters. Typically, parameters can be adapted according to clinician's review, rule-based systems (the system generates suggested changes based on initial parameters and measured feedback) or by predictive models trained on feedback data, treatment parameters and previous patient's clinical results. In the context of addiction, an example of the use of feedback in treating the disease can be found in U.S. patent application publication No. 2021 / 0162216, entitled “Neuromodulation Methods for Improving Addition using Multi-dimensionalFeedback,” the entire contents of which are incorporated herein by reference in their entirety.

[0083] For example, acute and subacute feedback of the presentation application of the treatment can be obtained by measuring physiological parameters. In some examples, the physiological parameter can be the response of the patient's autonomic nervous system to the focused ultrasound treatment, and multiple physiological parameters can be measured during any given assessment session. The physiological parameters can be measured by a wearable device (such as a ring, watch, or belt) or by a smartphone or tablet (for example, in a natural non-clinical environment, such as when the patient is at home, work, or other non-clinical environment). Exemplary physiological parameters include heart rate, heart rate variability, sweat, saliva, blood pressure, pupil size, changes in pupil size, eye movement, brain activity, skin electrodermal activity, body temperature, and blood oxygen saturation level. Table I above provides non-limiting examples of physiological parameters that can be measured and exemplary tests for measuring physiological parameters. Sleep parameters can also be measured using wearable devices.

[0084] In particular, when treating neurocognitive diseases, cognitive parameters can be assessed by a series of cognitive tests that measure, for example, executive function, decision-making, working memory, attention, and fatigue. Table II above provides non-limiting examples of cognitive parameters that are gamified and can be measured and exemplary methods and tests / tasks for measuring such cognitive parameters. These cognitive tests can be performed in a clinical / laboratory environment or in a natural nonclinical environment, such as when the user is at home, at work, or in other nonclinical environments. Intelligent devices (such as smartphones, tablet computers, or smart watches) can facilitate measuring these cognitive parameters in a natural, nonclinical environment. For example, Erickson interference tasks, N-back tasks, and psychomotor alertness tasks can be performed by an application on a smartphone, tablet computer, or smart watch.

[0085] Behavioral and psychosocial parameters (such as those described in Table III above) can measure the user's function (such as the user's movement) through wearable devices and subjective / self-report questionnaires. Subjective / self-report questionnaires can be collected in clinical / laboratory environments or in natural environments, in field non-clinical environments (such as when the user is at home, work or other non-clinical environments). Smart devices such as smartphones, tablets or personal computers can be used to manage subjective / self-report questionnaires. Using embedded accelerometers and cameras, these smart devices can also be used to capture the user's movements and facial expression analysis to analyze the user's facial expressions, which can indicate emotions, anxiety, depression, excitement and fatigue. Wearable or portable devices can also be used to measure parameters representing sleep length, sleep depth, length of sleep stages and heart rate variability.

[0086] In the illustrated embodiment, chronic feedback can be obtained by imaging the brain and eye (particularly the retina, optic nerve, and associated vasculature) to observe changes relative to the presence or progression of cognitive disease. To this end, one or more PET scans (including, for example, PET scans using fluorodeoxyglucose (FDG) PET scans, or PET scans with radioactive fluorine-labeled ligand-linked markers), gradient-recalled echo GRE ) scan, magnetization T1-prepared sequence scan, T2-weighted fluid-attenuated inversion recovery (FLAIR) scan, susceptibility-weighted imaging (SWI) scan, T1 contrast scan (including, for example, T1 gadolinium contrast scan), arterial spin labeling (ASL) scan, and / or dynamic contrast-enhanced (DCE) MR perfusion scan to determine the effect of the treatment on the patient.

[0087] In some examples, where the image is a PET scan, the scan can be analyzed to determine changes in the concentration of beta-amyloid, tau protein, and / or other biomarkers of neurodegenerative diseases based on, for example, a previous PET scan or a predetermined threshold / baseline value of these biomarkers. Treatment can be provided or adjusted based on changes in concentration. For example, treatment can be adjusted or provided when a decrease in the concentration or density of beta-amyloid, tau protein, and / or another biomarker of a neurodegenerative disease is detected. Alternatively, treatment can be adjusted or provided when no decrease in the concentration / density of beta-amyloid, tau protein, and / or another biomarker of a neurodegenerative disease is detected; or when a decrease in the concentration / density of beta-amyloid, tau protein, and / or another biomarker of a neurodegenerative disease that is insufficient to improve a human patient's neurodegenerative disease is detected. This determination can be made acutely, subacutely, or chronically after providing focused ultrasound therapy. Therefore, these signals can provide immediate / acute feedback during treatment, subacute feedback after a few minutes / hours / days of providing treatment, and long-term feedback after weeks / months of providing treatment.

[0088] At 110, focused ultrasound therapy is provided to the patient. It should be understood that focused ultrasound therapy may include neuromodulation by focused ultrasound, destruction of the blood-brain barrier by microbubbles provided to the patient's bloodstream in combination with administration of a therapeutic agent to the bloodstream, or neuromodulation and administration of a therapeutic agent combined with targeted destruction of the blood-brain barrier. In one embodiment, the administration of neuromodulation may be prior to or accompanied by initiation, wherein stimulation associated with the patient's disease is presented to increase neural activity in a brain region associated with the disease. For example, a patient with a neurodegenerative disease may be asked to perform a cognitive task immediately before or during neuromodulation. Patients with addiction may be presented with cues associated with their addiction that may induce or enhance dependence and / or anxiety, such as smells, tastes, images, videos, sounds, or tactile stimuli. Patients with PTSD, obsessive-compulsive disorder, phobias, or anxiety disorders may be presented with cues that may trigger or enhance symptoms of their disease, such as cues associated with trauma. Patients with depression may be presented with cues intended to induce or enhance negative effects. Patients with chronic pain may be stimulated by pain. Patients with Huntington's disease, Parkinson's disease, stroke, or autism may be presented with stimuli that are associated with cognitive or behavioral deficits that the patient exhibits. For example, depth perception is often a problem for Parkinson's disease patients, and asking patients to perform tasks that rely on depth perception could provide a priming effect for treating that deficit.

[0089] At 112, the treatment provided to the patient is adjusted based on the measured feedback. Typically, the treatment can be adapted based on a review of the feedback by the clinician, a rule-based system that generates suggested changes based on the initial parameters and the measured feedback, or by a predictive model trained on the feedback data, the treatment parameters, and the clinical results of previous patients. Feedback after the focused ultrasound treatment can affect whether to provide further focused ultrasound treatment and how the provided focused ultrasound treatment should be adjusted. In terms of adjusting the treatment in the context of focused ultrasound treatment, the method may involve adjusting the parameters or dosage of the focused ultrasound treatment, such as the duration, frequency, or intensity of the focused ultrasound treatment. When the collected data indicates that the patient's condition has not improved, the method may involve adjusting the focused ultrasound treatment so that the focused ultrasound treatment is more effective. For example, if the patient previously had five minutes of focused ultrasound (FUS) during a treatment course, the patient may then have twenty minutes of FUS during each treatment course, or if the patient has FUS every thirty days, the patient may then have FUS every two weeks. In contrast, if the parameter measurement indicates improvement, the focused ultrasound treatment parameters may not need to be adjusted, and subsequent focused ultrasound treatment sessions, for example, may be used primarily as maintenance sessions, or the intensity, frequency, or duration of the focused ultrasound treatment may be reduced. The above scenarios are exemplary only and are provided to illustrate that the presence and types of changes in a patient's physiological parameter measurements during and after treatment can affect whether treatment should be adjusted or terminated.

[0090] In addition, the extent of a patient's physiological, cognitive, psychosocial, or behavioral parameter measurements during or after treatment can affect the parameters of subsequent focused ultrasound treatment. For example, if a particular patient seeking treatment has a physiological, cognitive, psychosocial, or behavioral parameter measurement during or after treatment that is higher than the average parameter measurement for the same patient population, subsequent treatment may be more aggressive. Conversely, if a particular patient's parameter measurement during or after treatment is lower than the average parameter measurement for the same patient population, subsequent treatment may be less aggressive. In other words, the severity or extent (as well as baseline values ​​and levels) of a patient's physiological, cognitive, psychosocial, or behavioral parameter measurements during or after focused ultrasound treatment can be correlated with the extent or aggressiveness of future focused ultrasound treatment. The above scenarios are exemplary only and are provided to illustrate that the extent of changes in a patient's physiological parameter measurements during and after focused ultrasound treatment can affect the parameters of subsequent treatment.

[0091] In certain aspects, acute, subacute, and chronic feedback are each determined based on one or more combinations of physiological, cognitive, psychosocial, and behavioral parameters of the patient after treatment. For example, a measurement of a baseline value of one or more combinations of the patient's physiological, cognitive, psychosocial, and behavioral parameters can be obtained. The patient can then be exposed to an initial focused ultrasound signal to the patient's neural target site. Subsequent measurements of the result values ​​of one or more combinations of the patient's physiological, cognitive, psychosocial, and behavioral parameters can be obtained during or after the application of the initial focused ultrasound signal. The result values ​​can be compared with the baseline values ​​to determine whether the patient's cognitive and / or behavioral function has improved. When it is determined that the patient's cognitive and / or behavioral function has not improved, the focused ultrasound treatment can be adjusted. For example, if it is determined that the focused ultrasound treatment is unsuccessful, the focused ultrasound treatment can be provided to a different target location. Once the new treatment parameters are adjusted, the method returns to 106 to continue monitoring the patient.

[0092] Fig.11 A system 1100 for providing treatment to a patient by focused ultrasound therapy is shown. The system 1100 includes multiple data sources that provide data to a central server 1110. One or more imaging systems 1102 are configured to provide images of the patient's brain or eyes. In practice, the imaging system 1102 may include a computed tomography image, a magnetic resonance imaging (MRI) system, a positron emission tomography (MRI) system, an optical coherence tomography system, an angiography imager, and visible and infrared cameras. The images provided from the one or more imaging systems may be provided directly to the central server 1110 or to an electronic health record (EHR) database 1104 that may be accessed by the central server 1110. Portable monitoring devices 1106, 1108 may include wearable devices that can be used to measure physiological parameters as well as sleep parameters.

[0093] The patient may also be provided with portable monitoring devices 1106, 1108, which include sensors for monitoring the system that tracks the user's health-related parameters. It should be understood that a given portable monitoring device (e.g., 1106) may communicate directly with a central server 1110 to provide health-related parameters to the server, or communicate with another portable monitoring device (e.g., 1108) that relays the health-related parameters to the server. In one example, multiple portable monitoring devices may include physiological sensing devices and portable computing devices. By using portable monitoring devices 1106, 1108, measurements can be taken continuously from any one of the user's home, classroom, work, or sports field (indeed, in any case). Additional parameters may be retrieved from the EHR database 1104 and / or other available databases via a network interface 1112 associated with the server 1110. These parameters can include, for example, employment information (e.g., title, department, shift), age, gender, home zip code, genomic data, nutritional information, medication intake, household information (e.g., household type, number of residents, and ages), social and psychosocial data, consumer spending and profile, financial data, food safety information, presence of physical abuse, and relevant medical history.

[0094] Between one or more imaging systems 1102, electronic health database 1104, and portable monitoring systems 1106, 1108, each parameter listed in Tables I-VII above can be measured or accessed. For example, many behavioral and psychosocial parameters can be determined from the user's functions and activities and subjective / self-report questionnaires. Subjective / self-report questionnaires can be collected in clinical / laboratory environments or in natural environments, in field non-clinical environments (such as when the user is at home, work, or other non-clinical environments). One of the portable monitoring systems 1106, 1108 can be a smart device, such as a smartphone, tablet computer, or personal computer, which can be used to manage subjective / self-report questionnaires. By using embedded accelerometers and cameras, these smart devices can also be used to capture facial expression analysis to analyze the user's facial expressions, which can indicate emotions, anxiety, depression, excitement and fatigue, and measurements of patient activity. Similarly, cognitive parameters can be determined based on tests administered on smart devices. Physiological parameters and sleep parameters can be measured at the portable monitoring system. Movement and biomarker parameters can be measured in a clinical setting and recorded in the EHR database 1104 for later retrieval, as can any of the physiological, sleep, behavioral, cognitive, and psychosocial parameters that cannot be measured using portable monitoring devices.

[0095] The central server 1110 analyzes data collected by the portable monitoring devices 1106, 1108 and any clinical data received from the EHR system 1104 and the imaging system 1102 at the network interface 1112. The central server 1110 can be implemented as a dedicated physical server or part of a cloud server arrangement. In addition to the remote server, the data can be analyzed in whole or in part on the local device itself and / or in a federated learning mechanism, in which case any data from the EHR 1104 can be provided to the local device through a suitable network interface. The information received from the portable monitoring devices 1106, 1108 and the network interface 1112 is provided to a feature extractor 1114 that extracts multiple features. In one embodiment, these features are aggregated at a feature aggregator 1116 to provide an aggregated parameter set. In this example, the aggregated parameters are then used directly as features of the prediction model 1120, or used to derive features of the prediction model.

[0096] The feature extractor 1114 determines categorical and continuous parameters representing health-related parameters. In one example, the parameters may include descriptive statistics, such as measures of central tendency (e.g., median, mode, arithmetic mean, or geometric mean) and measures of deviation (e.g., range, interquartile range, variance, standard deviation, etc.) of the time series of the monitored parameters, as well as the time series itself. Specifically, for at least one parameter, the feature set provided to the prediction model may include two values ​​or a single value representing the value of the parameter at different times, such as a measure of central tendency or a measure of deviation representing multiple parameter values. In addition, the feature set may use features associated with other patients associated with the patient (such as a spouse, child, colleague, or friend) as additional features.

[0097] In other examples, a feature may represent a deviation of a user from an established pattern of features. For example, the value of a specified parameter may be tracked over time, and a measure of central tendency may be established as a whole or for a specific time period. The collected features may represent a deviation of a specified parameter from a measure of central tendency. For example, a change in the user's activity level measured by either or both of a motion sensor and a global positioning system (GPS) tracking may be used as a health-related parameter. Other elements monitored may include monitoring the user's compliance with the use of a smartphone, TV, portable device, or portable device. For example, the system may send a message to the user asking about their health level, overall mood, or the state of any other health-related parameter on a portable computing device. A measure of compliance may be determined based on the percentage of these messages to which the user responds through a user interface on the portable computing device, and the measure of compliance may be used as a feature of the prediction model 1120.

[0098] In one embodiment, the feature extractor 1114 may perform a wavelet transform on the time series of values ​​of one or more parameters to provide a set of wavelet coefficients. It should be understood that the wavelet transform used herein is two-dimensional, so that the coefficients can be conceived as a two-dimensional array across time and across frequency or scale.

[0099] The feature extractor 1114 may also include a facial expression classifier (not shown) that evaluates recorded data from the camera and / or recorded images or videos of the user's face from one of the portable monitoring devices 1106, 1108 (such as a smartphone or other mobile device) to assign an emotional state to the user at different times of the day. The extracted features may be categorical, representing the most likely emotional state of the user, or continuous, for example, as a time series of probability values ​​of various emotional states (e.g., anxiety, discomfort, anger, etc.) determined by the facial expression classifier. The feature extractor 1114 may also include one or more image classifiers that reduce the medical image provided from the imager system 1102 to categorical or continuous features for the predictive model. It should be understood that each of the facial expression classifier and the one or more image classifiers may be implemented using one or more models discussed above for the predictive model. In one example, the feature extractor 1114 may provide the medical image to a convolutional neural network (CNN) and extract one or more potential values ​​from the CNN as features.

[0100] In an example using aggregated features, the feature aggregator 1116 generates an aggregated parameter set from a set of health-related parameters collected by the portable monitoring devices 1106, 1108 and any clinical data received from the EHR system 1104 at the network interface 1112. It should be understood that each aggregated parameter can be a weighted combination of a function of a set of health-related parameters or parameters from the set of health-related parameter sets. Accordingly, a specified aggregated parameter can represent multiple health-related parameters, and typically, the multiple health-related parameters represented by each aggregated parameter will be related so that the aggregated parameter represents a specific health domain of the user. Typically, each aggregated parameter can use parameters from a variety of sources, so that the specified aggregated parameter can be a combination of features from two or more of the portable monitoring devices 1106, 1108 and the network interface 1112. In some embodiments, the system 1100 may include multiple predictive models (not shown), each predictive model receiving a unique appropriate subset of aggregated parameters. Each predictive model can provide different clinical parameters representing different aspects of the user's health, so that the aggregated parameters can be used for multiple purposes of assessing the user's health.

[0101] The prediction model 1120 may utilize one or more pattern recognition algorithms, each of which analyzes the extracted features or a subset of the extracted features to assign a continuous or categorical clinical parameter to the user. It should be understood that the prediction model 1120 may use any of the various pattern recognition algorithms described above for the prediction model. In one example, the prediction model 1120 may assign a continuous parameter corresponding to the likelihood that the user has or is at risk for a disease, the likelihood that the user is experiencing the effects of aging, the likelihood that the user is experiencing the onset of dementia, the likelihood that the user has or will have a neurodegenerative disease, the likelihood that the user will experience an exacerbation or "attack" of symptoms of a chronic disease (such as chronic pain or anxiety), the likelihood that the user will use addictive substances during rehabilitation or treatment, the current or predicted user pain level, the user's expected performance level associated with the current or future time of a particular activity or occupation, the change in symptoms associated with a condition or disease, the current or predicted response to treatment, the likelihood that the user has experienced increased stress, or the user's overall health level. It should be understood that the specific nature of the clinical parameter assigned by the prediction model will vary depending on the problem to which the model is applied, such as screening, symptom detection, or predicting or determining the effectiveness of an applied focused ultrasound treatment.

[0102] In another example, the predictive model 1120 may assign classification parameters corresponding to a range of possibilities as described above, the presence or risk of a particular disease, a set of categories representing the user's readiness for a particular activity or occupation, categories representing changes in symptoms associated with a condition or disease (e.g., "improved," "stable," "worsened"), categories representing current or predicted responses to treatment, categories representing user status (e.g., "normal," "stressed," "sick"), or categories indicating that a particular action should be suggested to the user. The generated parameters may be stored in a non-transitory computer-readable medium, for example, as part of a record in an electronic health record database, or used to suggest a course of action to the user.

[0103] In one embodiment, the prediction model 1120 may include a component model that predicts future values ​​of health-related parameters or aggregated parameters, such as a convolutional neural network with one or more two-dimensional arrays of wavelet transform coefficients as input. The wavelet coefficients detect not only changes in time, but also changes in time patterns, and therefore can reflect changes in the user's normal biorhythms. Additionally or alternatively, the prediction model may use a component model that predicts current or future values ​​of health-related or aggregated parameters, and these measurements are then used as features for generating the output of the prediction model. These data can also be used to group users with users who react similarly to these parameters, and data fed back from users within a specified group is used to better customize the model for the user.

[0104] The predictive model 1120 may also be used to facilitate a feedback strategy to one or more designated recipients, which may include a user, health care provider, family, friends, social system, care team, supervisor, coach, caregiver, and other entities, where, for example, a course of action is suggested in response to a change in the individual's health. Examples of diseases for which detection, or disease, or risk of disease exacerbation may trigger intervention include, but are not limited to, anxiety, depression, suicidal thoughts, stress, mood, agitation, obsessive-compulsive disorder, OCD, Parkinson's disease, tremor, and chronic pain. Feedback or intervention may include feedback from the system to increase awareness of a detected problem or education about the problem, initiate or modify current treatment, or direct the user to additional testing, where current treatment includes, but is not limited to, focused ultrasound therapy, medications, biologics, surgical interventions, behavioral interventions and social interventions, digital interventions via portable devices, home visits by care providers, directing individuals to clinics, support groups, emergency rooms, or hospitals that provide directions or locations of clinics, support groups, emergency rooms, or hospitals.

[0105] It should be understood that the recommended course of action can be any course of action designed to enhance the health of the user or others, and can include, for example, performing prescribed meditation, performing prescribed exercises, stopping current activities, and contacting medical professionals. In one example, when it is determined that the user is experiencing significant stress, the feedback can include a "digital reset" in which the user is instructed to perform deep breathing or other stress reduction techniques. If this is ineffective, or the problem recurs, the feedback can be promoted to a digital intervention prescribed by a medical professional, in which the user regularly participates in guided stress reduction techniques. If this is also ineffective, the user can be instructed to seek help from a counselor, coach, or therapist. It should be understood that the model is predictive, so interventions for stress, pain, and similar problems can be recommended before the user is aware of the problem. In another example, guidance on sleep hygiene can be received in response to an indication of insufficient or restless sleep, or a sleep study can be guided, or a doctor can be referred for analysis. It should be understood that interventions associated with sleep can also be assigned for other detected problems, such as decreased cognitive function, decreased exercise, work, or other performance, or increased risk of stroke and heart disease.

[0106] In some embodiments, the prediction model 1120 may include a feedback component 1124 that can adjust various parameters of the prediction model 1120 based on the accuracy of the predictions made by the model. In one example, the feedback component 1124 can be shared by multiple prediction models 1120, where the results of the user associated with each prediction model are compared to the results predicted by the output of the model. Parameters associated with the model can be adjusted based on the difference between the actual results and the predicted results, such as, for example, thresholds for converting continuous values ​​into categorical inputs or outputs, or parameters associated with the pattern recognition algorithm that includes the prediction model.

[0107] Alternatively, the prediction model 1120 can obtain feedback at the individual model level. For example, in a prediction model 1120 that uses a composition model to predict future values ​​of health-related parameters, once the health-related parameters are measured, the model receives consistent feedback about the accuracy of these predictions. This feedback can be used to adjust the parameters of the model, including the personalized threshold of the user, to generate a classification input or output from a continuous value or a baseline value of a biorhythm associated with the user. Alternatively, feedback can be provided from the final output of the model, and the feedback can be compared with other data (such as the state reported by the user) to provide feedback to the model. In one embodiment, a reinforcement learning method can be used to adjust the model parameters based on the accuracy of the predicted future values ​​of the health-related parameters or the output of the prediction model at the intermediate stage of the prediction model 1120. For example, the decision threshold for generating a classification output from a continuous index generated by the prediction model 1120 can be set to an initial value based on feedback from multiple models of previous users, and the decision threshold is adjusted by the reinforcement model to generate a user-specific decision threshold.

[0108] Fig.12 A system 1200 for targeting focused ultrasound therapy to treat or diagnose a disease in a specific area of ​​a patient's brain is shown. The system 1200 includes a processor 1202 and a non-transitory computer-readable medium 1210 that stores executable instructions for targeting focused ultrasound therapy to treat or diagnose a disease. The executable instructions include 1212, which receives a first image representing the structure of the patient's brain and a second image representing the connectivity of the brain from one or more associated imaging systems (not shown). In one embodiment, the first image is a T1 magnetic resonance imaging (MRI) image and the second image is a diffusion tensor imaging (DTI) image generated using an MRI imager. The imaging interface 1212 may include suitable software components for communicating with an imaging system (not shown) or a repository (not shown) storing images over a network via a network interface (not shown) or via a bus connection.

[0109] The first image is provided to a registration component 1214, which segments the first image into a plurality of sub-regions of the brain. The identified sub-regions may include, for example, the frontal pole, the temporal pole, the superior frontal region, the medial orbitofrontal region, the caudal anterior cingulate region, the rostral anterior cingulate region, the olfactory cortex region, the parahippocampal gyrus region, the pericune gyrus region, the lingual gyrus region, the cuneus region, the isthmus region, the anterior cuneus region, the paracentral lobule region, and the fusiform gyrus region. In one example, the registration component 1214 registers the first image to a standard atlas to provide a segmented portion. In another embodiment, a convolutional neural network trained on a plurality of annotated image samples may be used to provide a segmented image. An example of such a system can be found in the work by Huo et al., "3D Whole Brain Segmentation Using Spatially Localized Atlas Network Blocks" (available at https: / / doi.org / 10.48550 / arxiv.1903.12152), which is incorporated herein by reference in its entirety. The registration component 1214 can also register the second image with the first image so that the locations of nodes in the connectome of the brain are known.

[0110] Each of the segmented first image and the second image can be provided to a targeting component 1216, which selects a location and intensity distribution for focused ultrasound treatment. It should be understood that before the segmented first image is provided to the targeting component, it can be registered with the second image. The targeting component 1216 generates a connection group of the brain representing neural connections within the brain from the second image. A region of interest can be defined within the first image based on known sub-regions, and the location and intensity distribution of focused ultrasound treatment within the region of interest can be selected based on the generated connection group. It should be understood that the connection group can be determined as a passive connection group or an active connection group, the passive connection group represents the physical connectivity between parts of the brain, and the active connection group represents the activity induced in parts of the brain in response to energy provided at a specific location.

[0111] In one example, the targeting component 1216 can operate in conjunction with the focused ultrasound therapy system 1217 to generate a map of the brain's active connectome. Specifically, energy can be applied at various locations within the region of interest, and activity within the brain can be determined by a suitable functional imaging modality. In one embodiment, the activity measured in response to each location within the brain is recorded as a result of the focused ultrasound therapy for that location. It should be understood that when energy is provided, multiple locations can be affected, and the detected activity can be attributed to each location, for example, expressed as a voxel, based on the percentage of energy received at each voxel. Alternatively, the results of multiple measurements can be compared, for example, by solving one or more n-dimensional linear systems, where n is the number of voxels within the region of interest. Therefore, the active connectivity associated with each location within the region of interest can be determined.

[0112] In one embodiment, the region of interest can be divided into a set of voxels, and each voxel can be assigned a cost based on its connection to other regions of the brain in the connectome to form a cost map. For example, a positive cost can represent a location within the region of interest that is connected to a part of the brain that does not generally need stimulation, and a negative cost can represent a location within the region of interest that is connected to a part of the brain that generally needs stimulation. It should be understood that this can in turn create a "utility map" where positive values ​​represent locations that need stimulation.

[0113] Each location has an associated intensity distribution around a reference point (such as a center point) that represents the amount of energy provided to the region for the specified location of the reference point. In one example, each voxel can be assigned a value normalized by the maximum intensity, and this value can be used to weight the contribution of the voxel to the total cost associated with the location and intensity distribution. An optimization process such as gradient descent can be used to search for an optimal or near-optimal location and intensity distribution for the region of interest, and the resulting location and intensity distribution can be provided to the treatment planning system 1220 for use in generating a treatment plan for the patient.

[0114] Fig.13 A system 1300 for determining the presence or risk of a disease based on imaging of a patient's brain is shown. The system 1300 includes a processor 1302 and a non-transitory computer-readable medium 1310 storing executable instructions for determining the risk or presence of a disease based on imaging of a patient's brain. The executable instructions include an imaging interface 1312 that receives a first image representing the structure of the brain and a second image representing the connectivity of the brain for the patient. In one embodiment, the first image is a T1 magnetic resonance imaging (MRI) image and the second image is a diffusion tensor imaging (DTI) image generated using an MRI imager. The imaging interface 1312 may also receive functional images representing activity within the brain, for example from the same or a different MRI imager. The imaging interface 1312 may include suitable software components for communicating with an imaging system (not shown) or a repository (not shown) storing images over a network via a network interface (not shown) or via a bus connection.

[0115] The first image is provided to a registration component 1314, which segments the first image into multiple sub-regions of the brain. The identified sub-regions may include, for example, the frontal pole, the temporal pole, the superior frontal region, the medial orbitofrontal region, the caudal anterior cingulate region, the rostral anterior cingulate region, the olfactory cortex region, the parahippocampal gyrus region, the pericune gyrus region, the lingual gyrus region, the cuneus region, the isthmus region, the anterior cuneus region, the paracentral lobule region, and the fusiform gyrus region. In one example, the registration component 1314 registers the first image to a standard atlas to provide a segmented portion. In another embodiment, a convolutional neural network trained on multiple annotated image samples may be used to provide a segmented image. The registration component 1314 registers the second image with the first image to know the location of the nodes within the connection group within the brain.

[0116] The registered second image is provided to a prediction model 1316. The prediction model 1316 analyzes the provided connectome image to assign a clinical parameter to the user, the clinical parameter representing one of the following: the likelihood that the patient has or will have a problem with a general disease category within a specified time period, the likelihood that the patient has or will have a problem with a disease within a specified time period, the likelihood that the patient will generally respond to treatment for the disease, or the likelihood that the patient will respond to a specific treatment for the disease (such as focused ultrasound treatment). It should be understood that the clinical parameters can be categorical or continuous. The prediction model 1316 can utilize one or more pattern recognition algorithms, such as those described above, implemented as classification and regression models. Regardless of the specific model employed, the clinical parameters generated by the prediction model 1316 can be provided to the user at the display 1320 through a user interface, or stored on a non-transitory computer-readable medium 1310, such as in an electronic medical record associated with the patient.

[0117] Fig.14An example of a system 1400 for diagnosing and monitoring disease is shown. The system 1400 includes a processor 1402 and a non-transitory computer-readable medium 1410 storing executable instructions for receiving data representing a patient from a plurality of sources and determining a clinical parameter representing the likelihood that the patient has or will have a cognitive disease. The executable instructions include an imaging interface 1412 that receives a first image representing the structure or connectivity of the patient's brain and a second image representing one of the patient's retina, optic nerve, and associated vasculature from a corresponding imaging system (not shown). In one embodiment, the first image is a T1 magnetic resonance imaging (MRI) image representing the structure of the brain. In other embodiments, the first image may be a diffusion tensor imaging (DTI) image generated using an MRI imager, a position emission tomography (PET) image acquired using a tracer labeled with radioactive fluorine, glucose or beta-amyloid or tau protein. The second image may be any image from which the retina, optic nerve, or associated vasculature is extracted, such as an optical coherence tomography (OCT) image, an OCT angiography image, and an image acquired by fundus photography. In some examples, imager interface 1412 may also receive an image or video of a patient's pupil to provide parameters representative of at least one of eye tracking data, eye movement, pupil size, and pupil size change.

[0118] The imaging interface 1412 may include suitable software components for communicating with an imaging system (not shown) or a repository (not shown) storing images through a network interface (not shown) or through a bus connection through a network. In some embodiments, the imaging interface 1412 may segment the first image into multiple sub-regions of the brain, so that each of at least a portion of the pixels or voxels comprising the image is associated with one of the multiple sub-regions. The identified sub-regions may include, for example, the frontal pole, the temporal pole, the superior frontal region, the medial orbitofrontal region, the caudal anterior cingulate region, the rostral anterior cingulate region, the olfactory cortex region, the parahippocampal gyrus region, the pericune gyrus region, the lingual gyrus region, the cuneus region, the isthmus region, the anterior cuneus region, the paracentral lobule region, and the fusiform gyrus region. In one example, the imaging interface 1412 registers the first image to a standard atlas to provide a segmented portion. In another embodiment, a convolutional neural network trained on multiple annotated image samples may be used to provide a segmented image. In the case of providing multiple images representing a brain taken with different imaging modalities, the imager interface 1412 may register the images to each other. For example, an image representing structure can be registered with an image representing connectivity of the brain, thereby knowing the locations of nodes within connectomes within the brain.

[0119] The prediction model 1416 receives a representation of the first image and a representation of the second image, and generates a value representing a clinical characteristic associated with one or more diseases of the patient. For example, the generated value is a categorical or continuous value representing the likelihood that the patient currently has a disease. In another example, the generated value is a categorical or continuous value representing the likelihood that the patient will develop a disease. In another example, the generated value is a categorical or continuous value representing the likelihood that the patient will respond to a specific treatment for the disease. In another example, the generated value is a categorical value representing the expected optimal treatment, optimal target location, or optimal parameter set for the treatment of the patient's disease. In another example, the generated value is a continuous or categorical value that represents the patient's disease progression in general or specifically in response to an applied treatment (such as focused ultrasound therapy).

[0120] The representation of each of the first image and the second image provided to the prediction model 1416 may include any of the images themselves, represented as chromatic values ​​from pixels or voxels comprising the images, an image or mask derived from the images, or a set of digital features extracted from the images. For example, the representation of the first image may be any of a representation of a cortical contour of the brain, a representation of the vasculature of the brain, a representation of a brain beta amyloid contour, and a representation of the connectivity of the brain from the first image. It should be understood that the prediction model 1416 may also receive clinical parameters representing the patient, for example, measured by one or more sensors and / or retrieved from a medical health record database, such that values ​​representing clinical characteristics associated with one or more cognitive diseases of the patient are calculated from the representation of the first image, the representation of the second image, and the clinical parameters.

[0121] The prediction model 1416 can utilize one or more pattern recognition algorithms, such as classification and regression models, each of which analyzes the provided data to assign values ​​to the user. It should be understood that the prediction model 1416 can use any of the pattern recognition algorithms described above for the prediction model. Regardless of the specific model employed, the categorical or continuous values ​​generated by the prediction model 1416 can be provided to the user at the display 1420 via a user interface, or stored on a non-transitory computer-readable medium 1410, such as in an electronic medical record associated with the patient.

[0122] Fig.15 1 is a schematic block diagram of an exemplary system 1500 that illustrates hardware components that can implement examples of the systems and methods disclosed herein. The system 1500 may include various systems and subsystems. The system 1500 may be a personal computer, a laptop, a workstation, a computer system, an appliance, an application specific integrated circuit (ASIC), a server, a server blade center (Server BladeCenter), a server cluster, etc.

[0123] The system 1500 may include a system bus 1502, a processing unit 1504, a system memory 1506, memory devices 1508, 1510, a communication interface 1512 (e.g., a network interface), a communication link 1514, a display 1516 (e.g., a video screen), and an input device 1518 (e.g., a keyboard, a touch screen, and / or a mouse). The system bus 1502 may communicate with the processing unit 1504 and the system memory 1506. Other memory devices 1508, 1510, such as a hard drive, a server, a separate database, or other non-volatile memory, may also communicate with the system bus 1502. The system bus 1502 interconnects the processing unit 1504, the memory devices 1506-1510, the communication interface 1512, the display 1516, and the input device 1518. In some examples, the system bus 1502 also interconnects additional ports (not shown), such as a universal serial bus (USB) port.

[0124] The processing unit 1504 may be a computing device and may include an application specific integrated circuit (ASIC). The processing unit 1504 executes a set of instructions to implement the operations of the examples disclosed herein. The processing unit may include a processing core.

[0125] Additional memory devices 1506, 1508, 1510 can store data, programs, instructions, text or database queries in compiled form, and any other information that may be needed to operate the computer. Memory 1506, 1508, 1510 can be implemented as a computer-readable medium (integrated or removable), such as a memory card, a disk drive, a compact disk (CD), or a server accessible via a network. In some examples, memory 1506, 1508, 1510 may include text, images, video, and / or audio, portions of which may be available in a format that humans can understand.

[0126] Additionally or alternatively, the system 1500 may access external data sources or query sources through a communication interface 1512 , which may be in communication with the system bus 1502 and a communication link 1514 .

[0127] In operation, system 1500 can be used to implement one or more parts of a system for providing focused ultrasound therapy to a patient according to the present invention. According to some examples, computer executable logic for implementing the system resides in system memory 1506 and one or more of memory devices 1508, 1510. Processing unit 1504 executes one or more computer executable instructions originating from system memory 1506 and memory devices 1508, 1510. The term "computer-readable medium" used herein refers to a medium that participates in providing instructions to processing unit 1504 for execution. The medium may be distributed over multiple discrete components that are all operably connected to a common processor or a group of related processors. Specific details are given in the above description to provide a thorough understanding of the embodiments. However, it should be understood that the embodiments can be practiced without these specific details. For example, physical components can be shown in block diagrams to avoid obscuring the embodiments with unnecessary details. In other cases, well-known circuits, processes, algorithms, structures, and techniques can be shown without unnecessary details to avoid obscuring the embodiments.

[0128] The implementation of the above-mentioned techniques, blocks, steps and devices can be accomplished in various ways. For example, these techniques, blocks, steps and devices can be implemented in hardware, software or a combination thereof. For hardware implementation, the processing unit can be implemented in one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, other electronic units designed to perform the above-mentioned functions and / or combinations thereof.

[0129] In addition, it should be noted that the embodiments can be described as processes shown in flow charts, flow diagrams, data flow diagrams, structure diagrams, or block diagrams. Although a flow chart can describe the operations as a sequential process, many operations can be performed in parallel or concurrently. In addition, the order of the operations can be rearranged. When the operations of the process are completed but may have additional steps not included in the figure, the process terminates. The process can correspond to a method, function, procedure, subroutine, subprogram, etc. When the process corresponds to a function, its termination corresponds to the function returning to the calling function or main function.

[0130] In addition, the embodiment can be implemented by hardware, software, scripting language, firmware, middleware, microcode, hardware description language and / or any combination thereof. When implemented with software, firmware, middleware, scripting language and / or microcode, the program code or code segment performing the necessary tasks can be stored in a machine-readable medium such as a storage medium. Code segment or machine executable instruction can represent any combination of a procedure, function, subroutine, program, routine, subroutine, module, software package, script, class, or instruction, data structure and / or program statement. Code segment can be coupled to another code segment or hardware circuit by transmitting and / or receiving information, data, independent variable, parameter and / or memory content. Information, independent variable, parameter, data, etc. can be transmitted, forwarded or transmitted by any suitable means, including memory sharing, message transmission, ticket transmission, network transmission, etc.

[0131] For firmware and / or software implementations, the methods may be implemented with modules (e.g., procedures, functions, etc.) that perform the functions described herein. Any machine-readable medium that tangibly embodies instructions may be used to implement the methods described herein. For example, software code may be stored in a memory. The memory may be implemented within a processor or external to a processor. As used herein, the term "memory" refers to any type of long-term, short-term, volatile, non-volatile, or other storage medium, and is not limited to any particular type of memory or amount of memory, nor to the type of medium in which the memory is stored.

[0132] In addition, as disclosed herein, the term "storage medium" may refer to one or more memories for storing data, including read-only memory (ROM), random access memory (RAM), magnetic RAM, core memory, magnetic disk storage media, optical storage media, flash memory devices, and / or other machine-readable media for storing information. The term "machine-readable medium" includes, but is not limited to, portable or fixed storage devices, optical storage devices, wireless channels, and / or various other storage media capable of storing or carrying instructions and / or data.

[0133] What has been described above is an example. Of course, it is impossible to describe every conceivable combination of parts or methods, but those of ordinary skill in the art will recognize that many other combinations and arrangements are possible. Therefore, the present disclosure is intended to cover all such changes, modifications and variations that fall within the scope of the present application including the appended claims. As used herein, the term "includes" means including but not limited to the term "includes" means including but not limited to. The term "based on" means at least partially based on. In addition, where the present disclosure or claims mention "one", "a", "first" or "another" element or its equivalent, it should be interpreted as including one or more such elements, neither requiring nor excluding two or more such elements.

Claims

1. A method for providing focused ultrasound therapy, comprising: Screening a patient to determine whether focused ultrasound therapy is appropriate for treating a condition in the patient; monitoring the patient to measure a plurality of health-related parameters of the patient and, if it has been determined that focused ultrasound treatment is appropriate for the patient, detecting or predicting the onset of symptoms associated with the disease based on the plurality of health-related parameters; When the onset of symptoms is detected or predicted, selecting a personalized location of a brain target for the patient for focused ultrasound treatment based on at least one of the plurality of health-related parameters; selecting at least one parameter associated with the focused ultrasound therapy based on at least one of the plurality of health-related parameters; providing focused ultrasound therapy to the patient at the selected location using the selected at least one parameter; and After providing the focused ultrasound therapy, the plurality of health-related parameters are measured to determine the safety and effectiveness of the focused ultrasound therapy.

2. The method according to claim 1, wherein: The disease is addiction to a substance or behavior, and measuring the plurality of health-related parameters includes measuring a subset of the plurality of health-related parameters following a cue associated with the substance or behavior.

3. The method according to claim 2, wherein: The cues are provided to the patient via one of a physical object, a computer monitor, a virtual reality, and an augmented reality system.

4. The method according to claim 2, wherein: The cues include two or more of visual cues, auditory cues, taste cues, tactile cues, interoceptive cues, and olfactory cues.

5. The method according to claim 1, wherein: The disease is a neurodegenerative disease, and screening a patient to determine whether focused ultrasound treatment is suitable for the patient includes: capturing an image of the patient's eye; and A set of parameters is determined that represents one of a retina of the patient, an optic nerve of the patient, and vasculature associated with the retina or the optic nerve.

6. The method according to claim 5, wherein: A parameter set representing one of the patient's retina, the patient's optic nerve, and the vascular system associated with the retina or the optic nerve includes: the volume of the retina in the image, the thickness of the retina, the texture of the retina, the thickness of the retina layer, the volume of the retinal layer, the texture of the retinal layer, a value representing the vascular pattern, a value representing the vascular density, the size of the foveal avascular zone, the width of the optic chiasm, the height of the optic nerve in the eye, the width of the optic nerve in the cranium, and one of the total area of ​​the vascular system.

7. The method according to claim 1, wherein: The condition is chronic pain, and monitoring the patient to detect or predict the onset of the symptoms includes monitoring the patient to predict the onset of a pain episode before the patient is aware of the symptoms.

8. The method according to claim 1, wherein: The disorder is one of post-traumatic stress disorder (PTSD), panic attacks, phobias, depression, anxiety disorders, and schizophrenia, and screening the patient to determine whether focused ultrasound treatment is suitable for the patient includes: capturing images representing the brain's connectivity; and At least one parameter is determined from the image.

9. The method according to claim 8, wherein: The image is a first image, and screening the patient to determine whether focused ultrasound therapy is suitable for the patient further comprises: acquiring a second image representing the structure of the brain; segmenting the second image into a plurality of sub-regions of the brain to generate a segmented second image such that each of at least a subset of a plurality of voxels comprising the second image is associated with one of the plurality of sub-regions; providing the first image and the segmented representation of the second image to a machine learning model trained on imaging data of a plurality of patients whose screening results for the disease are known; and A clinical parameter indicative of the patient's risk of suffering from the disease is generated based on the segmented representations of the first and second images.

10. The method according to claim 9, further comprising: The second image is registered with the first image to provide a registered connection group, wherein the registered connection group represents the positions of nodes within the connection group relative to the multiple sub-regions, and providing the segmented second image and the first image representation to the machine learning model includes: providing the registered connection group to the machine learning model.

11. The method according to claim 1 also includes generating an aggregate parameter set from multiple health-related parameters, each aggregate parameter including a unique suitable subset of the multiple health-related parameters, and providing the aggregate parameter set to a prediction model, which assigns clinical parameters representing the possibility of symptom onset associated with the disease based on the subset of the aggregate parameter set.

12. The system according to claim 10, wherein: The prediction model is a first prediction model representing a first disease, the clinical parameter is a first clinical parameter, and the subset of the aggregated parameter set is the first subset of the aggregated parameter set. The system also includes a second prediction model, which assigns a second clinical parameter representing a second disease to the user through the second prediction model based on the second subset of the aggregated parameter set, and the second subset of the aggregated parameter set is different from the first subset of the aggregated parameter set.

13. The method according to claim 1, further comprising: providing digital intervention via a portable device if focused ultrasound therapy is determined not to be appropriate for said patient; The digital intervention includes: a support tool that assists in one of education, mindfulness, improved sleep, and pain prevention, and a message to a care provider to contact the patient and the location of a clinic, emergency room, support group, or hospital.

14. The method according to claim 1, wherein: After providing focused ultrasound therapy, measuring the multiple health-related parameters to determine the effectiveness of the focused ultrasound therapy includes: measuring a first subset of the multiple health-related parameters as acute feedback, measuring a second subset of the multiple health-related parameters as subacute feedback, and measuring a third subset of the multiple health-related parameters as chronic feedback.

15. The method according to claim 1, wherein: After providing the focused ultrasound therapy, measuring the plurality of health-related parameters to determine the effectiveness of the focused ultrasound therapy includes: generating one of a magnetic resonance imaging (MRI) image and a positron emission tomography (PET) image of the patient's brain; and At least one of the plurality of health-related parameters is extracted from one of the magnetic resonance imaging (MRI) image and the positron emission tomography (PET) image.

16. The method according to claim 15, wherein: Generating one of the magnetic resonance imaging (MRI) image and the positron emission tomography (PET) image of the brain includes generating one of a functional magnetic resonance imaging (MRI) image and a metabolic magnetic resonance imaging (MRI) image.

17. The method according to claim 15, wherein: Generating one of the magnetic resonance imaging (MRI) image and the positron emission tomography (PET) image of the brain includes generating the magnetic resonance imaging (MRI) image while the patient performs a cognitive task.

18. The method according to claim 1, wherein: Monitoring the patient to measure the plurality of health-related parameters includes measuring one of pupil size, change in pupil size, and eye movement.

19. The method according to claim 1, wherein: Providing focused ultrasound therapy to the patient at the selected location includes providing stimulation associated with the disease to the patient before or during the focused ultrasound therapy.

20. The method according to claim 19, wherein: The disease is one of a neurodegenerative disease, autism, an autism spectrum disorder, stroke, Parkinson's disease, and Huntington's disease, and providing the stimulation includes presenting one of a cognitive task, a motor task, and a behavioral task to the patient.

21. The method according to claim 19, wherein: The disease is obsessive-compulsive disorder, the disease is one of post-traumatic stress disorder (PTSD), phobia, anxiety, depression and addiction, and providing the stimulus includes: presenting one of an image, video, taste, sound, smell or tactile stimulus selected to induce or exacerbate symptoms of the disease.

22. The method according to claim 1, wherein: Providing focused ultrasound therapy to the patient at the selected location comprises: introducing microbubbles into the patient's blood; introducing a therapeutic agent into the blood of the patient; and Focused ultrasound is provided to a location within the brain to open the blood-brain barrier where it is needed to allow the therapeutic agent to penetrate into the brain.

23. The method of claim 22, further comprising providing focused ultrasound to a location within the brain where neuromodulation is desired.

24. The method according to claim 1, wherein: The selected location is within one of the nucleus accumbens, ventral striatum, and ventral capsule of the patient.

25. The method of claim 1, wherein: The disease is one of autism, autism spectrum disorder, Parkinson's disease and Huntington's disease, wherein, based on at least one of the multiple health-related parameters, selecting personalized positioning for focused ultrasound treatment for the patient includes: selecting personalized goals to address cognitive and behavioral deficits associated with the disease.

26. A system for generating clinical parameters for a user, the system comprising: a physiological sensing device that monitors a first plurality of health-related parameters representative of the user over a defined period of time; a portable computing device that obtains a second plurality of health-related parameters representative of the user; a network interface that retrieves a third plurality of health-related parameters representing the user from an electronic health record (EHR) system, the first plurality of health-related parameters, the second plurality of health-related parameters, and the third plurality of health-related parameters collectively forming a health-related parameter set; a feature aggregator that generates an aggregated parameter set from the health-related parameter set, each parameter in the aggregated parameter set comprising a unique suitable subset of the health-related parameter set; as well as A prediction model is provided to assign the clinical parameter to the user based on a subset of the aggregated parameter set.

27. The system of claim 26, wherein: The prediction model is a first prediction model, the clinical parameter is a first clinical parameter, and the subset of the aggregated parameter set is the first subset of the aggregated parameter set. The system also includes a second prediction model, and the second prediction model assigns a second clinical parameter to the user through the second prediction model based on the second subset of the aggregated parameter set, and the second subset of the aggregated parameter set is different from the first subset of the aggregated parameter set.

28. The system of claim 26, wherein: The aggregated parameter set includes at least: a first aggregated parameter representing the sleep and circadian rhythm of the user, a second aggregated parameter representing the social behavior function of the user, and a third aggregated parameter representing the biomarker and genome of the user.

29. The system of claim 26, the portable computer device providing feedback intervention to the user based on the value of the clinical parameter.

30. The system of claim 26, wherein: The clinical parameter is a value representative of the overall health of the user, and the subset of the aggregated parameter set includes the entire aggregated parameter set.

31. A method for generating a value representing one of risk and progression of a disease, comprising: acquiring a first image representing a brain of a patient from a first imaging system; acquiring a second image from a second imaging system, the second image representing one of a retina, an optic nerve, and vasculature associated with one of the optic nerve and the retina of the patient; providing a representation of each of the first image and the second image to a machine learning model; generating a value in the machine learning model based on the representation of the first image and the representation of the second image; as well as The patient is assigned to one of a plurality of intervention categories based on the generated value.

32. The method of claim 31 , further comprising providing clinical parameters extracted from an electronic health record (EHR) database to the machine learning model, wherein The clinical parameter represents one of the patient's medical history, a treatment prescribed for the patient, and a measured biometric parameter of the patient, and generating the value in the machine learning model includes: generating the value based on the clinical parameter, the representation of the first image, and the representation of the second image.

33. The method according to claim 31, wherein: The second image is an optical coherence tomography (OCT) image, an optical coherence tomography (OCT) angiography image, and an image generated by fundus photography.

34. The method of claim 31, wherein: The representation of the second image includes a parameter representing one of the volume of the retina in the image, the thickness of the retina, the texture of the retina, the thickness of a retinal layer, the volume of a retinal layer, the texture of a retinal layer, a value representing a vascular pattern, a value representing a vascular density, the size of the foveal avascular zone, the width of the optic chiasm, the height of the optic nerve within the eye, the width of the optic nerve within the cranium, or the total area of ​​the vasculature.

35. The method of claim 31 , further comprising imaging the patient's pupil to provide parameters representative of at least one of eye tracking data, eye movement, pupil size, and pupil size change, wherein: Generating the value in the machine learning model includes: generating the value based on the representation of the first image, the representation of the second image and the parameter.

36. A method for determining the risk of a disease by imaging the brain of a patient, comprising: acquiring a first image representing the structure of the brain from a first imaging system; acquiring a second image representing connectivity of the brain from one of the first imaging system and the second imaging system; segmenting the first image into a plurality of sub-regions of the brain to generate a segmented first image such that each of at least a subset of a plurality of voxels comprising the first image is associated with one of the plurality of sub-regions; providing the segmented representations of the first and second images to a machine learning model trained on imaging data of a plurality of patients with known outcomes; and A clinical parameter representing the risk of the patient suffering from the disease is generated based on the segmented representations of the first and second images.

37. The method of claim 36, further comprising registering the second image with the first image to provide a registered connectome representing positions of nodes within the connectome relative to the plurality of sub-regions, and providing the representation of the segmented first and second images to a machine learning model comprises: The registered connectome is provided to the machine learning model.

38. The method of claim 36, wherein: Providing the representation of the segmented first and second images to a machine learning model includes providing the segmented first and second images to the machine learning model.

39. The method of claim 36, wherein: Acquiring the second image includes acquiring the second image by diffusion tensor imaging.

40. The method of claim 36, wherein: The clinical parameters indicate the likelihood that a patient will respond to treatment for impaired neuropsychiatric function.

Images