A NEUROLOGICAL ASSESSMENT OF A SUBJECT

CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of priority of U.S. Provisional Application

No., 62/687,206 filed June 19, 2018 and U.S. Provisional Application No. 62/752,833, filed October 30, 2018 the entire disclosures of which are hereby incorporated herein by reference.

TECHNICAL FIELD

[0002] The disclosure relates generally to providing a neurological assessment of a subject. The subject can be a person diagnosed with or suspected of suffering from a neurological condition, or their caregiver(s). In one aspect, a computerized system and method for processing signals and data obtained from individuals, generating mathematical models of the data and signals to compute risk assessment scores to predict, propose and/or evaluate recommendations and resource allocation for a specific subject and/or its caregiver are provided.

BACKGROUND

[0003] Individuals with neurological and neurocognitive diseases have long experienced impediments to the diagnosis, treatment recommendations, costs of treatment and resource allocation for them and their caregivers. In many cases, these issues are caused by the lack of necessary tools that accurately capture, classify, and track the underlying symptomology of the disease. Typically, in neurology, dementia is used as a higher-level term to describe the neurological condition of someone with a difficulty in maintaining their activities of daily living (ADL). However, there are many subtypes of the disease. For example, the most common subtype of dementia is Alzheimer’s disease. Other subtypes include dementia with Lewy body, vascular, and at least fifteen other variations. Interestingly, most brain autopsies have shown that 75% of dementia cases are mixed pathologies. Moreover, care recommendations, mortality, and estimated cost of care differ among the various subtypes of dementia, which makes it critical that the proper diagnoses are provided at the outset.

[0004] Further, it is well established that specialists in dementia neurology are scarce.

For example, there are only an estimated 600 dementia neurologists across the United States that provide an expert diagnosis and care plan, and approximately 1000 dementia neurologists globally. According to research estimates there are roughly 10 million cases in the United States with dementia, thereby suggesting that neurologists would have to see 17,000 patients individually to make a difference. Further, neurological diseases, such as dementia, are considered one of the highest trending causes of disability and mortality worldwide. This trend is expected to continue in the future given the expanding aging populations. Indeed, conditions such as dementia and stroke will be the most impactful on societies worldwide, further supporting the need for geriatric neurology education.

[0005] Nonetheless, the supply of specialized neurologists continues to decrease. This is further showcased, by a research titled,“Dementia Neurology Deserts,” which compiled an index score that represents the supply-demand mismatch between the number of cases of dementia projected in 2025 by U.S. state and the number of projected neurologists by location. For example, Wyoming had the largest index score, representing an area where patients and families are most in need of specialist care. As discussed above, these types of shortages have a major impact on the level and quality of care that patients are able to access and receive, leading to extended wait times and misdiagnoses. [0006] The effects of misdiagnosis of dementia-related neurological diseases, combined with the lack of specialized expertise, impose a great burden on the patients and their caregivers. For example, clinical data from the Alzheimer’s Association estimates that 50% of all dementia cases are undiagnosed. Assisting dementia patients in ADLs is cited as the number one priority for eager family caregivers who are looking for ways to help their loved one. Without assistance from the health care system due to neurologist shortages, many family caregivers turn to non expert sources of information, which may have a detrimental result to the patient. Moreover, it has been established that caregiver mortality rates can increase when caring for an individual that suffers, among other things, a neurological disease. Specifically, in the case of dementia, caregiver mortality reportedly increases by 60%. As a result, for long term care insurance (LTCI) carriers, assisting family caregivers represents an opportunity to engage, educate, and bend the cost curve.

[0007] In addition, the 50% of undiagnosed dementia patients can also represent a misclassified risk to care-related businesses such as hospitals, pharmaceutical companies, and insurance companies that must allocate or reserve resources to cater to such patients. For example, a claim for arthritis from a hip fracture may be undiagnosed dementia thereby representing higher claims paid as the duration of claim payment will be longer for dementia claims. Moreover, research shows that almost half of dementia patients have other chronic medical conditions. In those cases, when patients have other chronic conditions, such as diabetes or chronic kidney disease, dementia exacerbates the costs of those conditions. In addition, non disclosure of diagnosis by physicians due to attitudes around dementia, time constraints, and operational challenges of medical practice can also translate into higher claims processing costs. SUMMARY

[0008] This disclosure relates generally to providing a neurological assessment of a subject using, for example, a computerized system and method for processing signals and data obtained from individuals, generating mathematical models of the data and signals to compute risk assessment scores to predict, propose and/or evaluate recommendations and resource allocation for the subject and/or their caregiver. In one aspect, the system and method generate models that provide granular identification and overall assessment for neurological conditions, or any other condition that may impact the daily living activities, of an individual. Dynamically generated risk assessment scores can be adapted to address specific needs associated with the care of the individual. In some implementations, the computed risk assessment scores can be used to assess an individual’s current and future states, predict the individual’s progress, and propose and/or evaluate suitable courses of treatment and resource allocation for the individual and/or its caregiver.

[0009] Moreover, in some implementations, the systems and methods provide a dynamically generated care-related recommendation dashboard that can be transmitted to different entities associated with the care of the individual and/or its caregiver in order to inform their decision-making.

[0010] In some implementations, the disclosure provides a data inclusive approach into the individual’s health assessment, risk exposure, and recommended treatment by generating unique and adaptive assessments for both the individual and its caregiving network to the different entities involved in the overall process. In one aspect, the system and method use the individual’s and caregiver’s observed responses in conjunction with the individual-specific characteristics, and generated mathematical models to account for the gaps in the current process. Accordingly, in some implementations, the system and method use observed responses and clinical signals of the specific individual and caregiver to refine generated models, compute risk scores and related forecasts, and to effectively personalize the models. In some implementations, the personalized models can be used to forecast expected health improvements and/or declines, caregiver recommendations and resource allocation by insurers. By using the observed response data to personalize the models, the models are modified to account for individual variability and mixed underlying pathologies that is not accounted for in conventional models.

[0011] Indeed, in some implementations, the system and method provide the health care professional with the ability to monitor and enhance a personalized and adaptable treatment plan for the individual. Moreover, as discussed above, in many neurological diseases the access to specialized health professionals is limited given the lack of training and available specialists. In some implementations, the system and method address this gap by providing adaptive mathematical models that offer assessments and identify underlying pathologies while minimizing the false positive and false negative rates. Further, in some implementations, the disclosed system and method can generate multiple risk scores that would provide a quick, accurate, and efficient overview of the individual’s current and projected state. For example, according to some implementations, the disclosed system and method dynamically generate a mortality score, a morbidity score and/or any other suitable risk score for the individual. In addition, in some implementations, the disclosure provides a health assessment and accompanying risk scores for the individual’s caregiver. Indeed, as discussed above, the caregiver’s health is highly affected by the individual’s underlying pathology and progress of treatment. Accordingly, in some implementations, the disclosed systems and methods generate a collateral health assessment of the caregiver.

[0012] Moreover, as indicated previously, in one aspect, the system and method generate mathematical models to assist in the provision of an individual’s health assessment. Specifically, in some implementations, the system and method use existing clinical criteria that are dynamically adapted based on clinical signals and additional information provided by the individual in order to provide an accurate assessment of underlying pathologies. The models can be trained off-line in a supervised manner, on-line by taking into account the dynamic changes of the individual, or in any other suitable way.

[0013] In some implementations, the system and method generate models that can offer an assessment at multiple clinical resolutions. For example, in the case of dementia, a mathematical model can be created using existing clinical criteria and further refined based on the individual’s clinical and other data obtained through testing. Using specified and dynamically generated rules, such model can provide an efficient and more accurate identification of the subtype of dementia and its associated severity. The model can be supplemented with additional clinical criteria to provide a more granular identification of, for example, specific proteins and genes that may be activated and can be part of the originating factors of the underlying pathology. Indeed, such a key insight can assist in offering highly targeted treatment plans or even assist in changing the understanding of the processes underlying specific pathologies.

[0014] Generally, the system and method involves the creation and dynamic adaptation of mathematical models developed from clinical criteria along with data and clinical signals gathered from patients, processing the models to create a composite health assessment of the individual, including risk scores and guided predictive analyses, and determining specific recommendations at different levels of granularity based on the recipient of the health assessments and risk scores.

[0015] Specifically, in some implementations, the system and method generates a recommendation dashboard that provides customized information depending on the recipient. For example, the patient and its caregiver may receive information on practical advice for increasing the quality of their daily life activities, a recommendation for contacting a specialist health professional, a recommendation for nutritional changes and/or any other suitable recommendation. In addition, in some implementations of the disclosure, the primary health care professional overseeing the treatment of the individual can also receive recommendations from the system. Specifically, in some implementations, the health care professional can receive recommendations such as potential changes in prescribed medications, onset of any new underlying pathologies, progress indicators associated with current treatment plans followed by the individual or any other suitable recommendation. Further, if the patient has authorized the release of their health assessment to health care related third parties the disclosed system and method can provide, for example, predicted costs of care to an insurance company, potential adjustments in claim processing for efficient allocation of resources and/or provide statistics to a pharmaceutical company on potential changes in protein and gene activations based on the adaptive health assessments.

[0016] The disclosed system and method provide an assessment of a neurological condition of a subject. More specifically, in one aspect, the system and method provide for one or more processors operatively coupled to a non-transient computer-readable storage medium storing a reference database having clinical data relating to the neurological condition. The one or more processors can adaptively extract, from a capture module in communication with the one or more processors, a plurality of signals associated with one or more neural parameters of the subject. In one implementations, the one or more neural parameters are selected from a plurality of neural parameters responsive to a prior clinical assessment of the neurological condition of the subject and based on the clinical data in the reference database, and the plurality of signals are stored to the computer-readable storage medium.

[0017] In one implementation, the one or more processors dynamically generate, using a neural assessment module operatively coupled to the reference database, one or more neurological risk scores for the subject based on the plurality of signals by computing, at least in part, a set of weights for combining the plurality of signals, wherein the weights are computed based at least in part on the clinical data in the reference database, and wherein the one or more neurological risk scores are indicative of a probability that the subject suffers or is likely to suffer from the neurological condition. As also indicated previously, the processor subsequently stores the one or more neurological risk scores to the computer-readable storage medium.

[0018] Moreover, in one implementation, the system and method automatically adjust, through the processor, the selection of the one or more neural parameters based on the one or more neurological risk scores and the clinical data in the reference database, and provide a set of care-related recommendations based on at least one of the one or more neurological scores.

[0019] According to some implementations, the system and method provide for the one or more neurological risk scores to indicate at least a classification score for a neurodegenerative disease. Indeed, as discussed above, in many cases a patient suffers from a mixed pathology of diseases that can include multiple types of neurological diseases and/or other chronic conditions. [0020] To account for this uniqueness of any particular patient together with the high instances of false positive diagnosis of the underlying neurological diseases, in some implementations, the system and method use the one or more processors to compare, using the neural assessment module, the prior clinical assessment of the neurological condition of the individual with the one or more neurological risk scores, and determine whether the prior clinical assessment of the neurological condition is a false positive assessment and/or a false negative assessment.

[0021] Conceptually, according to some implementations of the disclosure, the caregiver is provided with the ability to input information about the patient, such that the one or more signals include answers to a set of recurring questions posed to a caregiver, a subject and/or both whereby one or more of the recurring questions is adjusted based on said answers. Further, in some implementations, the system and method assess the quality of the answers provided by the caregiver so as to minimize any false classifications of the underlying neurological disease. Indeed, this is accomplished by having the one or more processors configured to generate, using the capture module, a confidence index associated with the answers provided by the caregiver, wherein the confidence index is indicative of the quality of answers provided by the caregiver.

[0022] As discussed above, in some implementations, the system and method also provide recommendations to assist the patient and/or caregiver. Specifically, the one or more processors are further configured to select, based on the one or more neurological risk scores, the set of recommendations from a pool of recommendations, that, when applied to the subject, modify at least some of the plurality of signals, such that the one or more neurological risk scores are reduced or improved. Indeed, the system can also generate a predicted reduction and/or predicted improvement in the one or more neurological risk scores based on the set of recommendations.

[0023] In some implementations, the system and method involve providing a set of recommendations to the patient, and/or caregiver, and/or health care related third party based on the neurological assessment and existing protocols associated with the neurological assessment. Specifically, according to some implementations of the disclosure, the one or more processors compare the one or more neurological risk scores with a set of pre-determined thresholds to determine the set of recommendations and subsequently transmit the recommendations to a health provider.

[0024] In addition, in some implementations, the system and method provide customized adjustment of the recommendations displayed to the patient/caregiver by causing the one or more processors to transmit the set of recommendations to the neural assessment module, identify, using the neural assessment module, one or more pre-determined neurological risk scores associated with the set of recommendations, compare the one or more pre-determined neurological risk scores with the generated one or more neurological risk scores, generate a progress indicator for at least one of the one or more neurological risk scores based on the comparison, and adjust the set of recommendations based on the progress indicator for the at least one or more neurological risk scores.

[0025] In some implementations, the system and method provide a dynamic analysis window by causing the neural assessment module to compute a score trend for at least one of the one or more neurological risk scores based on the one or more signals and the clinical data in the reference database, compare a set of parameters obtained from the score trend to a set of pre- determined thresholds, and determine at least one recommendation for at least one of the one or more neurological risk scores based on the comparison.

[0026] As also discussed above, according to some implementations, the system and method provide efficient reserving and allocation of health related resources. Specifically, the system and method can provide reserving and allocation of such care related resources including, but not limited to the type of human resources that a subject may need (e.g., registered nurse, physical therapist etc.), the duration that a resource may be needed (e.g., number of sessions of physical therapy, reserving of specialized equipment etc.). For example, the one or more processors can reserve one or more care-related resources based on the one or more neurological risk scores. Such scores can also be used by the one or more processors to generate a brain tissue pathology, genetic and biomarker indicator, and/or mortality and/or morbidity indicator for the subject and/or a caregiver of the subject.

[0027] In some implementations, the system and method generate one or more neurological risk scores for a caregiver based on the one or more neurological risk scores of the subject. Indeed, as discussed above, the caregiver’s health assessment is highly affected and dependent on the patient’s condition and state. In order to make such determinations, the caregiver has the ability to input information to the system that is combined with the one or more signals using a capture module communicatively coupled to one or more sensors. In some implementations, these sensors are selected from the group consisting of: biochemical sensors, GPS location sensors, respiratory rate, electrocardiography, electroencephalography, gyroscope, heart rate measurement, accelerometer, electrooculography, electromyography, augmented/virtual reality sensors, and blood pressure measurement. Moreover, the captured signals can be selected from the group consisting of audio, video, text, and binary signals. [0028] Furthermore, in some implementations, the system and method provide for a health assessment wherein the one or more neurological risk scores include at least one severity score for a neurodegenerative disease and that such information, including at least the one or more neurological risk scores, can be transmitted in the form of one or more alerts to a patient/caregiver or a user of the system.

[0029] In one implementation, the system and method apply advanced machine learning techniques in order to identify sets of weights that contribute to the computations of the neurological risk scores. Indeed, the set of weights is computed by the neural assessment module using a model selected from the group consisting of: a deep network, fuzzy logic, gradient descent optimization, Bayesian inference.

[0030] In one implementation, a method for providing an assessment of a neurological condition of a subject is provided. The method comprises the steps of providing one or more processors operatively coupled to a non-transient computer-readable storage medium storing a reference database having clinical data relating to the neurological condition. Further the method proceeds by adaptively extracting, from a capture module in communication with the one or more processors, a plurality of signals associated with one or more neural parameters of the subject. Further, the one or more neural parameters are selected from a plurality of neural parameters responsive to a prior clinical assessment of the neurological condition of the subject and based on the clinical data in the reference database. The method further provides for storing the plurality of signals to the computer-readable storage medium.

[0031] In one implementation, the method includes the step of dynamically generating, using a neural assessment module operatively coupled to the reference database, one or more neurological risk scores for the subject based on the plurality of signals by computing, at least in part, a set of weights for combining the plurality of signals. The weights are computed based at least in part on the clinical data in the reference database, and the one or more neurological risk scores are indicative of a probability that the subject suffers or is likely to suffer from the neurological condition.

[0032] Further, in some implementations, the method includes the steps of storing the one or more neurological risk scores to the computer-readable storage medium, automatically adjusting the selection of the one or more neural parameters based on the one or more neurological risk scores and the clinical data in the reference database, and providing a set of care-related recommendations based on at least one of the one or more neurological scores.

[0033] Some advantages of the foregoing include providing an automated and correct identification of the dementia subtype, which greatly increases the type and effect of care provided to individuals. In addition, impacting the efficient allocation of health-related resources. For example, current methods to evaluate cognitive dysfunction in a benefits eligibility assessment offer a rudimentary diagnostic capability. Specifically, most long-term care insurance (LTCI) carriers use a mental status exam called the MMSE, MOCA, clock drawing, or verbal recall. Although these tests are helpful in understanding if the person has normal aging or dementia, they do not classify the subtype of dementia or provide any risk assessment, and/or care recommendations. Accordingly, the foregoing provides the ability to predict, propose and/or evaluate suitable courses of treatment and resource allocation for a specific individual and/or its caregiver, as well as an opportunity for insurers to improve their assessment of dementia costs using a model that can adjust for the complexities around diagnosis and risk. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] An understanding of the following description will be facilitated by reference to the attached drawings, in which:

[0035] Figure l is a block diagram illustrating an exemplary system for the assessment of a neurological condition of a subject, in accordance with some implementations of the disclosed subject matter.

[0036] Figure 2 is a block diagram illustrating an exemplary capture module that receives the different input signals in the system, in accordance with some implementations of the disclosed subject matter.

[0037] Figure 3 is a schematic diagram of an exemplary health assessment system during operation, in accordance with some implementations of the disclosed subject matter.

[0038] Figure 4 is a flow chart illustrating a process for assessing the quality of the inputs to the capture module, in accordance with some implementations of the disclosed subject matter.

[0039] Figures 5A-5B are flow charts illustrating the process of the diagnosis engine of the health assessment system, in accordance with some implementations of the disclosed subject matter.

[0040] Figures 6A-6B are flow charts illustrating an exemplary process of the diagnosis engine related to dementia, in accordance with some implementations of the disclosed subject matter.

[0041] Figures 7A-7B are flow charts illustrating an exemplary process of the predictive engine of the health assessment system, in accordance with some implementations of the disclosed subject matter. [0042] Figures 8A-8B are flow charts illustrating an exemplary process of the operation of the recommendation module of the health assessment system, in accordance with some implementations of the disclosed subject matter.

[0043] Figure 9 illustrates an exemplary user interface for the capture module of the health assessment system, in accordance with some implementations of the disclosed subject matter.

[0044] Figure 10 illustrates an exemplary user interface for the recommendation dashboard of the health assessment system, in accordance with some implementations of the disclosed subject matter.

[0045] Figure 11 illustrates an exemplary user interface for the recommendation dashboard of the health assessment system, in accordance with some implementations of the disclosed subject matter.

[0046] Figure 12 illustrates an exemplary user interface for the display of a subject’s health assessment, in accordance with some implementations of the disclosed subject matter.

DETATEED DESCRIPTION

[0047] The system and method take into account variability between individuals that is typically unaccounted for by conventional mathematical models. Further, the present system and method adapt the mathematical models to a specific individual providing unique and dynamically generated forecasting and analysis, to identify, predict, propose and/or evaluate different aspects of the individual’s health assessment. Notably, the system and method can be used prospectively to assess and recommend a proposed treatment plan even before administering it to the individual, or to identify a plan for the individual that will achieve a desired pre-determined outcome. [0048] An exemplary implementation of the present disclosure is discussed below with reference to Figure 1. Specifically, Figure 1 shows a block diagram 100 illustrating an exemplary system for providing an assessment of a neurological, or any other condition, of an individual including a composite profile of potential underlying conditions. By way of example, the system 100 may include conventional hardware and software typical of a general purpose desktop, laptop/notebook or tablet computer. However, in accordance with some implementations of the disclosure, the system 100 is further specially-configured with hardware and/or software comprising microprocessor-executable instructions for specially-configuring the system 100 to carry out the method described below.

[0049] Referring again to Figure 1, system 100 can include one or more processor 108 capable of executing machine instructions. In some implementations, processor 108 can be any of a microprocessor, microcontroller, ASIC or any suitable combination thereof. Processor 108 is in communication with memory 110 that stores executable code for the operation of system 100. In some implementations, memory 110 is a non-volatile memory such as a ROM, EEPROM, magnetic hard drive, flash memory, and/or solid state memory. Further, processor 108 is coupled to network interface 114 enabling access to the Internet 116. In some implementations, network interface 114 can be hardware and/or software such as a transceiver, WAN, Bluetooth, Near Field Communication (NFC), wireless adapter or any other suitable combination thereof. In some implementations, system 100 includes software modules or libraries that are modular in nature, and that are manipulated by a software program comprising microprocessor-executable instructions for specially-configuring the system 100 to carry out the disclosed method. Specifically, in some implementations, system 100 includes a capture module 102 that is capable of receiving input signals and data from an individual and/or caregiver and subsequently processing the signals and information to identify one or more characteristics. In addition, capture module 102 is associated with user interface 118 presented in a display 120 that provides the individual and/or caregiver with the ability to easily input the requested signals and data. For example, user interface 118 can display a set of customized questions that the individual can answer, and/or prompts for transferring images, video or any other suitable signal. In some implementations, capture module 102 receives periodic updates of the clinical signals and data from a user of system 100 (e.g., individual and/or caregiver).

[0050] Further, capture module 102 is in communication with database 112 and stores the received clinical signals and data provided by the individual and/or the caregiver or any other suitable user. In some implementations, database 112 also includes clinical criteria defining, for example, one or more pathological conditions such as dementia or any other neurological disease. In some implementations, database 112 also includes clinical criteria used to identify the potential candidates of underlying conditions and the severity of each of the stored conditions. In some implementations, database 112 also includes clinical criteria at different resolutions and granularity. For example, database 112 can include a list of areas of the brain, proteins and/or genes that are activated during the presence of symptoms associated with one or more underlying pathologies that may be identified in the individual. For example, in some implementations, the above-described list can be provided as shown in Table 1 below whereby a model can be generated to provide not only the different subtypes of a neurological disease and associated health assessment risk scores, but also the areas of the brain and activated proteins that are associated with said assessments.

Table 1 : Vulnerability in neurodegenerative dementia

[0051] Further, database 112 can include a set of recommendations to be chosen for display to the individual and/or caregiver based on decisions made by the recommendation module 106 discussed in detail below. In some implementations, database 112 can be a relational database, non-relational database, document database or any other suitable combination thereof.

[0052] Referring again to system 100, capture module 102 is in communication with neural assessment module 104 which receives the clinical signals, data and clinical criteria stored in database 112. Neural assessment module 104 is capable of generating and applying one or more mathematical models associated with the identified clinical criteria stored in database 112. Specifically, neural assessment module 104 can generate one or more mathematical models associated with one or more neurological diseases such as dementia, Alzheimer’s or any other disease. In some implementations, the one or more mathematical models may be generated using supervised, non-supervised or any other suitable method associated with machine learning. For example, neural assessment module 104 may generate a mathematical model for identifying different subtypes of dementia using a multi-class neural network, support vector machine, convolutional neural networks or any other deep architecture using supervised learning. Further, neural assessment module 104 can use unsupervised methodologies such as clustering, auto encoders, principal component analysis (PCA) or any other suitable combination thereof. In addition, in some implementations, neural assessment module 104 can use fuzzy logic by transforming the provided inputs (e.g., clinical signals, data and clinical criteria) from capture module 102 into fuzzified inputs using membership functions and generate rules that identify and match the inputs to one or more neurological diseases. In some implementations, the generated rules can be dynamically adapted based on the provided clinical signals and data of the individual, thus creating a customized model for that individual.

[0053] For example, in some implementations, neural assessment module 104 can generate rules for neurological conditions whereby matches with the provided signals and data from the individual can be made based on disease rules for core, supportive, and exclusion criteria. As discussed above, the rules can be stored in database 112 accessible by the neural assessment module 104 and can be generated for any assigned condition. For example, common stroke syndromes (ischemic and hemorrhagic) and neuromuscular syndromes (based on at nerve plexus, roots, division, cord, branch segments) are matched based on rules using reported symptoms and abilities with neuroanatomical principles in the central and peripheral nervous system. Epilepsy syndromes are matched based rules with reported seizure semiology patterns known in the literature. Further, rules may also be based on neuroanatomical principles that have been encoded and stored in database 112. In some implementations the clinical criteria for generating the adaptable rules are obtained from the National Institutes of Health and Aging, American Academy of Neurology, consortium guidelines from respective neurological societies globally (example: Parkinson’s Disease Society United Kingdom) or any other suitable source.

[0054] In some implementations, neural assessment module 104 applies rules and models associated with multiple conditions to obtain a composite profile of potential underlying diseases affecting the individual. In some implementations, the composite profile includes likelihoods of each of the identified underlying conditions and can be obtained using probabilistic techniques such as Bayesian inference, Bayesian averaging, Expectation-Maximization (EM) or any other suitable probabilistic framework. In some implementations, the composite profile consists of a set of weights that identify the membership of each of the underlying diseases obtained by applying the generated models. In some implementations, neural assessment module 104 can optimize the set of weights by applying any suitable optimization technique such as gradient descent, and/or sparse optimization.

[0055] Further, as discussed above, in some implementations the neural assessment module’s 104 rules and associated weights can be stored in database 112 and may be automatically and dynamically adjusted by system 100. For example, in neurological diseases including dementia, rules can be altered by the system’s ability to correctly diagnose previously clinically diagnosed conditions. For example, if system 100 identifies that a number of individuals have pre-determined Alzheimer’s disease dementia based on the clinical signals and data (such as biomarkers, imaging, genomics, and/or clinical examinations) provided to the capture module 102, the system’s rules can be adjusted so that core, supportive, and exclusion criteria rules are dynamic to capture previously diagnosed patients.

[0056] As discussed above, neural assessment module 104 can use any suitable mathematical model. Generally, system 100 generates a mathematical model that involves developing a mathematical function that defines a curve that best“fits” or describes the observed clinical data, as will be appreciated by those skilled in the art.

[0057] Nonetheless, system 100 also provides for the manual adjustment of the rules and associated weights by a system’s administrator in order to accommodate for continually, modifiable, up-to-date clinically accepted evidence-based criteria, and/or best clinical practice. For example, a system administrator can change or adapt the rules generated by neural assessment module 104 based on previously diagnosed conditions. [0058] Moreover, neural assessment module 104 can generate one or more risk assessment scores that provide an overall assessment on the health of the individual and/or the caregiver. For example, in some implementations the one or more risk assessment scores can be a mortality score, a morbidity score, a long term care (LTC) score, a fiduciary score (e.g., associated with the cost of care) or any other suitable score. Further, in some implementations the risk assessment score is generated using a weighted average associated with the likelihoods obtained from the individual’s composite profile. In addition, as will be discussed below in reference to Figure 3, neural assessment module 104 can also provide predictive health assessments of the individual by generating and processing trends. In some implementations, one or more risk scores for the caregiver can be obtained based on dependencies in the health expectancy between the individual and its caregiver. For example, in some implementations, capture module 102 obtains information and data on the caregiver (e.g., medical history, age, etc.) and generates an initial caregiver morbidity score. Subsequently, the cargiver morbidity score can be modified based on the morbidity score of the individual who is receiving the care.

[0059] In some implementations, these models may be pre-stored in database 112, can be added to the system on a periodic basis, e.g., as part of distributed updates periodically stored on the system, or may be downloaded from the Internet, electronic media, and/or any another network on-demand. In addition, in some implementations, the one or more mathematical models can be hard-coded into, or otherwise be an integral part of, a unitary software program.

[0060] In some implementations, system 100 can generate, using recommendation module 106, a library of recommendation rules or personalized care suggestions associated with the generated health assessment and composite profile of the individual and/or caregiver. For example, in some implementations, care suggestions may be organized by behavioral change, informational, and sensor recommendations. Using behavioral change theory principles, those recommendation rules that involve a user’s behavior change may be made differently than those that involve informational care rules. Further, the individual and/or caregiver can use user interface 112 to provide real-time feedback in the form of like (positive) or dislike (negative) comments to the care suggestions, through a slider or any other suitable interface in order to further tune the specificity of the engine’s approaches and suggestions. In some implementations, users of system 100 can also add personal care suggestions that can be mined using artificial intelligence or machine learning principles to find patterns. As discussed above, system 100 can provide recommendations of different granularity to different users of the system. For example, recommendation module 106 can generate one or more alerts based on the obtained risk assessment scores to a health care professional. For example, such alerts can include recommendations for additional clinical testing, change in prescription medications or any other suitable recommendation.

[0061] Continuing with Figure 2, as discussed above, system 100 includes a capture module 102 that receives the clinical signals and data from the individual and/or caregiver. Specifically, capture module 102 is coupled to user interface 118 and display 120 and presents requests for signals and data from the user. Specifically, in some implementations, capture module 102 can receive sensor signals 202, textual and/or medical records 214, image/video records 226 or any other suitable data such as binary data. Further, the types of signals and data to be provided to capture module 102 can be suggested by the recommendation module 106 using prior user input (symptoms, abilities, staging, disease matches, past medical history, social history, mood history, or genomic history). [0062] More specifically, sensor signals 202 can include data from active and passive sensing devices with diagnostic and monitoring capabilities. For example, active sensors can provide signals that include, but are not limited to GPS location signals 204, biochemical signals 206, biomedical signals 208, such as respiratory rate, electrocardiography, electroencephalography, heart rate measurements, electrooculography, electromyography, or blood pressure measurement, gyroscope 210, and accelerometer 212. Further, capture module 102 can also obtain signals from passive tracking sensors that include ambient sensor technology, e-textile systems, remote monitoring, virtual/augmented reality sensors, or identifiable embedded computing devices within household objects or objects of everyday use.

[0063] In some implementations active and passive sensors can be used to obtain the following exemplary signals and/or data:

• Biochemical sensors can be used to detect electrolyte levels in one’s body and alert the patient to hydrate more with water.

• Accelerometers can be used to measure arm swing amplitude and track Parkinsonian symptoms over time.

• Noise-level detectors can be used to alert users to high noise environments

that may agitate patients with neurodegeneration.

• GPS tracking can be used to help patients navigate one’s home or track a patient with a visual neurodegenerative disease/disorder.

[0064] Moreover, capture module 102 can receive textual data and medical records 214.

For example, as discussed above, capture module 102 can receive information from answers provided to a survey of questions displayed to the individual and/or caregiver. Specifically, in some implementations the questions can be contained in database 112 and can be based on neuroanatomy, neurobehavioral, and current clinically accepted or evidence based criteria within each disease category. In addition, capture module 102 can generate a library of questions transformed into common non-medical language whereby phrasing of questions can be optimized through real-time user testing. Further, natural language processing and other pattern recognition software executed via a processor may be used to re-phrase questions based on user feedback.

[0065] Moreover, in some implementations, the survey of questions 216 presented to the caregiver can include questions associated with the two types of activities of daily living, instrumental and basic. Specifically, answers to these questions can assist in identifying potential neurological diseases. For example, if patients have impairment in instrumental activities of daily living, they may fall into the category of normal aging, mild cognitive impairment (MCI) or dementia, assuming system 100 finds a match based on the clinical criteria. As will be explained below, in reference to Figures 6A-6B, system 100 can label individuals with dementia when there is significant impairment in instrumental activities of daily living with inability to maintain basic activities of daily living.

[0066] Further, in certain disease/disorders, instrumental daily activities maybe normal, although social functioning maybe impaired. For example, in neurodegeneration, certain dementias such as semantic dementia variant of primary progressive aphasia, progressive non fluent variant of primary progressive aphasia, logopenic variant of primary progressive aphasia, behavioral variant of frontotemporal dementia, or posterior cortical atrophy, do not have memory as a predominant symptom, and hence individuals may score as independent in instrumental activities of daily living, although significant impairment in social, language, or visuospatial skills affects day to day function resulting in a diagnosis of dementia. [0067] In some implementations, the survey questions 216 that are answered by a caregiver and/or individual can capture the real-time brain function status within cognitive domains such as language, memory, visuospatial, behavioral, executive, and motor. In addition, sleep screening, mood screening, activities of daily living questions and relevant medical history 218 including, but not limited to relevant past medical history, prescription records 220, family, genomic testing history 224, medical equipment history 222 and social history questions 218 can be provided to capture module 102. In some implementations, questions can be nested in database 112 and activated or asked based on the individual’s and/or caregiver’s answers to prior questions. In some implementations, choosing which questions to nest is based on incidence data of diseases/disorders to evaluate for less common disease/disorder subtypes. For example, less common criteria for rare diseases/disorders are nested to identify what type of disease or disorder subtype. In some implementations, users responses’ to questions may include, but are not limited to“yes”,“no”, and“don’t know”. The responses from users’“yes” answers may be categorized based on the cognitive domain that they originated from (e.g., language, motor, visuospatial, etc.).

[0068] As discussed above, capture module 102 can also receive image and/or video records such as voxel morphometry 228, results from functional MRIs 230, Positron Emission Tomography (PET) 232 and any other suitable image/video record to assist in determine the health assessment of an individual and/or its caregiver including one or more risk scores.

[0069] Continuing, Figure 3 shows a block diagram of an exemplary implementation of the disclosed system and the interactions of its different components. For example, in some implementations, a patient and/or caregiver 302 utilizing the system is periodically presented through display 120 with a set of survey questions 216 that can include requests for one or more sensor signals 202. Indeed, in some implementations the set of questions presented to the user of the system can be selected by capture module 102 (e.g., survey selection 306) based on different criteria. Subsequently, the signals and data provided by the individual and/or caregiver are processed by capture module 102. For example, in the case of survey questions 216, capture module 102 generates a survey quality index 308 to determine whether the data provided can be utilized for assessing the health of the individual and computing the one or more risk scores. In addition, the received signals and data can be normalized and transformed into a suitable format by processor 108. Further, capture module 102 identifies signals correlated to specific clinical criteria 304, thus providing an initial estimate of potential underlying diseases and also determines if the individual has been pre-diagnosed with one or more diseases.

[0070] In some implementations, the individual and/or caregiver are requested to provide clinical signals and data to capture module 102 periodically based on a pre-determined interval. In some implementations, the individual and/or caregiver are prompted to provide such updated signals and data through a generated alarm. In some implementations, the periodic time interval for providing updated clinical signals and data can be adjusted based on criteria associated with, for example, previous data submissions and/or current analysis of data and/or any other suitable criterion.

[0071] As discussed above, capture module 102 transmits the signals and data to neural assessment module 104 and specifically to diagnosis engine 310. Diagnosis engine 310 generates one or more models and provides an identification and/or a composite profile of potential underlying conditions. In some implementations, diagnosis engine 310 can apply a set of rules for obtaining likelihoods of diseases. In some implementations, diagnosis engine 310 dynamically generates a set of weights corresponding to the membership of each identified disease in the composite profile. Specifically, the set of weights are dynamically generated by refining the one or more mathematical models using the individual’s provided signals and data. In addition, diagnosis engine 310 can also generate a set of weights that provide for different severity levels for each of the identified diseases.

[0072] Following, neural assessment module 104 generates one or more risk assessment scores using risk scoring 312. For example, as discussed above, risk scoring 312 can involve the generation of a long-term care (LTC) score, a mortality score, a morbidity score, a fiduciary score or any other suitable score. For example, an LTC risk score can be used to project fiduciary risk (z.e., cost of care) for enterprise companies, families that are caregiving, policymakers, and/or state/federally run reserves. In addition, in some implementations, LTC risk score can be used by actuaries and insurers for distribution of claims and health care resources such as e.g., human resources, specialized equipment, duration of specialized care etc. In addition, risk scoring 312 can use the derived LTC score and suspected underlying neurological condition (identified either through diagnosis engine 310 or as a pre-established diagnosis), a neuro-mortality risk is calculated. In some implementations, such score can be calculated by assigning to neurological conditions a longevity index score using data and published reports from neurology death banks, published longitudinal studies, and other public health databases.

[0073] In some implementations, risk scoring 312 derives the one or more risk scores as a composite value from multiple sub-scores using machine learning and computer-assisted predictive technologies. Further, in some implementations, the one or more risk scores generated by risk scoring 312 in neural assessment module 104 can be stored in database 112 along with additional risk scores from multiple enterprise companies such as insurance or corporations of the insured population.

[0074] Further, in some implementations, risk scoring 312 can generate a predictive health assessment and health composite profile for a subject and/or caregiver. Specifically, in some implementations, risk scoring 312 can determine based on the data provided from capture module 102 whether the subject is at risk in developing one or more neurological diseases. For example, in some implementations, if neural assessment module 104 has determined, through data provided by the subject to capture module 102, that they suffer from two out of the three elements required for a determination of dementia, then risk scoring 312 can generate a predictive health assessment indicating that the subject is likely to suffer dementia with a likelihood score of 66% and can further indicate a future likely symptom associated with the third unmet criterion.

[0075] In some implementations, risk scoring 312 can generate additional risk scores for the caregiver based on the individuals generates risk scores. For example, in some implementations, such dependent caregiver risk scores can be used to underwrite caregiver life insurance by calculating caregiver morbidity and mortality based on the care recipient’s disease burden and overall care need.

[0076] Specifically, risk scoring 312 can calculate the mortality of the caregiver and individual for use in determining premiums of an insurance policy. For example, capture module 102 receives data for both the individual and caregiver. In some implementations, the data is used to access actuarial tables stored in database 112 and generate a base mortality score for the caregiver. Subsequently, the mortality score of the individual (e.g., care recipient) can be generated and both scores can be stored in memory 112. In some implementations, the individual’s morbidity score then modifies the caregiver’s score to result in a modified caregiver mortality score. The resulting caregiver mortality score can be then used to generate an insurance amount and insurance premium (either monthly or yearly) outputted to a display 120 or stored in memory 110.

[0077] In some implementations, the insurance amount is based upon the cost of care for the individual’s (e.g., care recipient’s) life expectancy if the caregiver dies. Further, the insurance can be used to pay the same amount to the caregiver if the care recipient passes away to help offset part of the cost through re-investing the premium paid into an annuity or other financial investment products of care given over the lifetime of the care recipient.

[0078] In addition, further insights can be derived from the analysis of the stored health risk scores providing for a temporal view that pertains to trends of neuro-mortality, LTC risk, cost of care risk, and morbidity. To that end, neural assessment module 104 includes a predictive engine 314 that generated trends of the computed risk scores for data mining and analysis. Specifically, in some implementations users of system 100 are periodically presented with requests for providing clinical signals and data into capture module 102. The periodic updating of the signals and data provides for the refinement of the one or more mathematical models by diagnosis engine 310, the updated calculation of risk scores by risk scoring 312, and the subsequent generation of trends using predictive engine 314. In some implementations, the generated trends can be used for computing one or more progress indicators and/or localize areas of concern for the individual, assess the quality of recommendations and treatment plans. In some implementations, such information can be used to model reserves, adjust underwriting in real time, or offer new insurance products. [0079] Moreover, once neural assessment module 104 has generated a composite profile for the individual using the one or more generated mathematical models, a set of risk scores and their associated trends during the individual’s monitoring time, the information is further transmitted to recommendation module 106. In some implementations, recommendation module generates a recommendation/treatment dashboard 318 that can be transmitted to one or more users of the system. Specifically, recommendation module 106 can generate customizable dashboards adapting the presented recommendations. For example, recommendation dashboard 318 can be transmitted to a health care provider and can include, for example, prescription medication information. Further, the recommendation dashboard 318 can be transmitted to the individual and/or caregiver and include different suggestions associated with the severity of the one or more identified diseases. For example, level 1 suggestions are based on early, moderate, or late stages for each disease/disorder or dementia subtype. Level 2 suggestions involve crossing each possible symptom with each internal diagnosis/disease match, leading to additional, more granular care suggestions. Lastly more refined suggestions take into account ways to maximize abilities for each level 2 suggestions. In some implementations, to generate a personalized recommendation dashboard for display on a computer display or printed out via a printer in a report format, the system searches a recommendation library database for care suggestions that match the individual caring profile. These recommendations can be accessed by professional and lay care navigators of the system. In addition, recommendation dashboard 318 can also be provided to a third party in the health care industry such as a pharmaceutical company, insurance company or any other suitable third party that can benefit from dynamic reserving and allocation of care related resources, such as parties involved in financial markets, annuities, and underwriting (see e.g., 320). For example, in some implementations, based on the granularity of the analysis conducted by neural assessment module 104, the recommendation dashboard can provide a set of proteins and/or genes activated due to an underlying condition. Further, recommendation module 106 can also provide estimation of resource allocation 316 based on the individual’s composite profile and predictive engine’s 314 results.

[0080] Figure 4, illustrates a flow chart describing process 400 for generating a survey quality index 308 for assessing the usefulness and quality of the answers provided to capture module 102 by the individual and/or caregiver. At step 402 capture module 102 selects a set of questions from database 112 to be presented to the individual and/or caregiver on display 120 (see e.g., step 404). As discussed above, in some implementations, the selection of survey questions is adapted based on previously answered questions, previously computed composite disease profiles and risk scores, and/or previously provided recommendations by recommendation module 106. At step 406, capture module 102 receives the answers to survey questions 216.

[0081] At step 408, capture module 102 transforms the answers to the survey questions in a suitable format such as a real value, binary value and/or any other suitable value. In some implementations, the transformation of the answers to the survey questions is based on clinical criteria 304. For example, certain clinical criteria 304 may be scored higher than others in their significance. In some implementations, the survey questions 216 capture in non -medical terms clinical criteria 304, thus the answers provided by the individual and/or caregiver can be correlated back to the clinical criteria 304. As a result, if one clinical criterion is determined to be of high significance it will be scored with a higher real value if the individual’s answer to the related survey question indicated its existence. [0082] Upon transforming the answers into real values at step 408, capture module 102 proceeds at step 410 to determine the number of answers with a zero value. For example, in some implementations, users responses’ to questions can include but are not limited to answers including“yes”,“no”, and“I don’t know”. In some implementations, the answer“I don’t know” may be interpreted as a zero value at step 408. At step 410 and once users have completed a set of questions or portions of the electronic questionnaire, capture module 102 determines the number of answers provided with a zero value and computes a confidence index. In some implementations, capture module 102 calculates a confidence index such that:

[Total answers -‘Don’t know’ answers] ÷ [Total answers] = Confidence index

If the confidence index is less than a pre-determined confidence threshold (e.g., NO at step 412) then capture module 102 returns to step 402 and selects additional questions to be answered in order to ensure that there is enough information provided to generate an accurate composite profile and associated risk scores. In some implementations, a low confidence index can trigger a prompt through user interface 118 for more observations about the patient or person. For example, if a user has more than an allowed amount of“don’t know” answers, the system will alert the user to go back and observe or ask others for help. If, however, the confidence index is larger than a pre-defmed confidence threshold (e.g., YES at step 412) then capture module 102 proceeds to mine the provided data, compute statistics and perform an initial analysis based on clinical criteria 304. In some implementations, a pre-determined confidence threshold may be manually provided by the system’s administration. In some implementations, a pre-determined confidence threshold may be automatically generated based on, for example, the one or more risk scores and their trends as obtained from neural assessment module 104. [0083] Referring back to Figure 3, the capture module 102 subsequently provides the clinical signals and data to neural assessment module 104 whereby diagnosis engine 310 generates the composite profile of the individual. In some implementations diagnosis engine 310 uses the process described in Figures 5A-5B. Specifically, at step 502 neural assessment module 104 receives the formatted clinical signals and data (e.g., survey answers) from capture module 102. At step 504 diagnosis engine 310 identifies the set of clinical criteria associated with the provided signals and information. For example, a set of provided data can be identified within the clinical scope of neurological diseases. At step 506, neural assessment module further identifies a class of conditions that may be candidates for the individual based on the provided data. Upon identifying a set of candidate conditions, diagnosis engine 310 applies rules and/or one or more mathematical models for each of the candidate conditions at step 508 and generates likelihoods for each of them at step 510, thus creating a composite health assessment profile for the individual. As discussed above, in some implementations, the rules and mathematical models can be refined based on previously provided data. In some implementations, the generated likelihoods can be represented as a set of weights obtained by the one or more models through the use of machine learning algorithms, fuzzy logic and/or any other suitable algorithm. At step 512, diagnosis engine 310 determines whether a pre-existing diagnosis has been provided by capture module 102.

[0084] If diagnosis engine 310 has not received a pre-existing diagnosis (see e.g., NO at step 512) then it stores the composite health profile and likelihoods into database 112. However, if a pre-existing diagnosis has been provided to diagnosis engine 310 (see e.g., YES at step 512) then at step 516 diagnosis engine 310 compares the pre-existing diagnosis with the set of conditions in the composite health profile of the individual. Subsequently, if diagnosis engine determines that there is no match ( see e.g., NO at step 518) then the composite profile is stored in database 112 along with the pre-existing diagnosis. In some implementations, an alert can be send to a health provider to indicate the lack of match and recommend additional testing.

[0085] If, however, diagnosis engine 310 determines a match between the pre-existing diagnosis and a diagnosis provided in the composite profile of the individual (see e.g., YES at step 518) then at step 522 diagnosis engine 310 identifies the data associated with the matched clinical condition. For example, in some implementations, diagnosis engine 310 identifies the set of survey questions that are associated with the matched condition. At step 524, diagnosis engine 310 adjusts the mathematical model and/or rules associated with the clinical condition based on the identified answers so as to create a model customized to the individual. For example, in some implementations observations from the identified answers may lead to additional rules and/or the elimination of other rules for a specific condition for the individual.

[0086] For example, in the case of a neurological condition, Figures 6A-6B illustrate the above described process in more detail. Specifically, as discussed above, survey questions 216 include questions associated with instrumental and basic types of activities of daily living. Specifically, if an individual is impaired in instrumental activities of daily living, then they may fall into the category of normal aging, mild cognitive impairment (MCI) or dementia, assuming the system finds a match. Further, the system labels those with dementia when there is significant impairment in instrumental activities of daily living with the inability to maintain basic activities of daily living as dementia. In some implementations, the basic activity of daily living questions are nested and only activated when there is a score that exceeds a pre- determined threshold in the instrumental activity of daily living questions. [0087] More specifically, at step 602, diagnosis engine 310 receives the answers associated with the instrumental activities of daily living (IADL) and computes an associated score at step 604. Subsequently, if the computed score exceeds a pre-determined threshold (see e.g., YES at step 606) then diagnosis engine 310 determines that the underlying condition falls under the class of dementia and proceeds to analyze the answers provided to the questions associated with the basic activities of daily living (BADL) at step 608. At step 610, diagnosis engine 310 compares the resulting score with the set of candidate conditions. If at step 612, diagnosis engine determines that there is a match and that the set of candidate conditions includes dementia (e.g., YES at step 612) then diagnosis engine 310 identifies a dementia subtype at step 614. If, however, diagnosis engine 310 determines that there is no match (see e.g., NO at step 612) then diagnosis engine 310 identifies the condition as dementia not- otherwise specified (NOS).

[0088] Returning back to step 606, if diagnosis engine 310 determines that the score obtained from the IADL answers does not exceed a pre-determined threshold (see e.g., NO at step 606) then at step 618 diagnosis engine determines if the IADL score is greater than zero. If it is determined that the IADL score is not greater than zero (see e.g., NO at step 618) then at step 620 the diagnosis engine 310 determines that the individual is undergoing normal aging. If, however, it is determined at step 618 that the IADL score is greater than zero (see e.g., YES at step 618) then diagnosis engine 310 identifies an underlying class of dementia and checks at step 622 the set of candidate diagnoses. If at step 624 there is a match (see e.g., YES at step 624) then diagnosis engine 310 determines at step 628 the subtype of dementia. If, however, there is no match (see e.g., NO at step 624) then at step 626 diagnosis engine 310 determines that the individual is undergoing normal aging (NA) and/or mild cognitive impairment (MCI). [0089] For example in some implementations, the following process may be used for determining the neurological condition of dementia and its subtypes:

• If total IADL score is 9 or more, proceed with BADL questions

• If total IADL score is 8 or less, do not proceed with BADL questions Classification:

• If a total IADL score is 1-8, check candidate diagnoses. If a dementia match occurs, override“normal aging/MCI” diagnosis and add dementia subtype. If no match occurs, system will output normal aging or MCI.

• If the total score is 9 or more, the system will ask BADL questions and further classify dementia using the following:

o Mild: score of 9-16

o Moderate: score of 17-30

o Severe: score of 31 -42

In some implementations, the above scores and thresholds may be modified based on previously diagnosed patients’ staging or user feedback.

[0090] Moreover, the following table provides exemplary IADL questions and their associated scores that can be dynamically adjusted by system 100.

[0091] In addition, the following table provides exemplary BADL questions and their associated scores that can be dynamically adjusted by system 100.

[0092] As discussed above, neural assessment module 104 includes predictive engine

314 for providing trends and analysis on the computed health risk scores and composite profile of the individual and/or caregiver. Specifically, Figures 7A-7B illustrate an exemplary process employed by health scoring 312 and predictive engine 314. As discussed above, system 100 prompts the user to provide periodic updates with respect to the provided clinical signals and data that is received by neural assessment module 104 at step 702. At step 704, diagnosis engine 310 calculates health risk scores for each of the periodic updates and generates trend graphs at step 706. At step 708, a set of points can be selected on one or more trend graphs. In some implementations, the set of points can represent a temporal window that a health provider may want to inspect. In some implementations, the set of points can be automatically selected by predictive engine 314 based on statistics of the one or more trend graphs. For example, the set of points can represent an area of the graph that indicates extreme fluctuation (e.g., exhibits increased curvature). At step 710, predictive engine 314 generates one or more progress indicators associated with the one or more trend graphs. For example, a progress indicator can be computed using the slope of the trend graph, the derivative of the function, and/or the difference of values between the set of points. In some implementations, a progress indicator can be a cumulative indicator and/or a weighted average obtained from the different trend graphs.

[0093] At step 712 predictive engine 314 determines whether the one or more progress indicators exceed a pre-determined threshold. In some implementations, the pre-determined threshold may be set by the administrator of the system. In some implementations, the pre- determined threshold may represent a desired progress and can be set by a health provider. In some implementations, the pre-determined threshold can be automatically set by predictive engine 314 based on the individual’s previous progress.

[0094] If at step 712, the progress indicator does not exceed the pre-determined threshold

(see e.g., NO at step 712) then predictive engine 314 generates an intermediate projected risk score by increasing the current risk score of the individual at step 714 and transmits an alert to recommendation module 106 that in turn forwards the alert to the individual, the caregiver and/or the health care provider.

[0095] If at step 712 the progress indicator exceeds the pre-determined threshold (see e.g., YES at step 712) then predictive engine 314 proceeds to step 718 and determines whether the intermediate projected risk score is smaller than a target risk score. In some implementations, target risk score can be set by a health provider, and/or the caregiver of the individual. In some implementations, target risk score may be automatically set based on a step function and knowledge of the individual’s underlying condition composite profile.

[0096] If at step 718 the intermediate projected risk score is smaller than a target risk score (see e.g., YES at step 718) then, at step 720, predictive engine 314 decreases the current risk score of the individual. In some implementations, predictive engine 314 decreases the current risk score proportionally to the decrease of the projected risk score. In some implementations, the decrease of the current risk score is based on a pre-determined scales associated with the conditions identified in the individual’s composite profile.

[0097] If, however, at step 718 the intermediate projected risk score is greater than a target risk score (see e.g., NO at step 718) then, at step 724, predictive engine 314 determines from the composite health assessment profile of the individual the condition with the least progress. In some implementations, predictive engine 314 identifies the condition with the least progress by generating trend graphs and computing multiple progress indicators for each of the identified conditions present in the composite health assessment profile and subsequently sorting the progress indicators.

[0098] At step 728, predictive engine 314 adjusts the current risk score based on the identified condition with the least progress. For example, in some implementations, the current risk score is modulated based on the rate of progress of the identified condition. Further, at step 730, neural assessment module 104 and predictive engine 314 transmit the adjusted current risk score and the identified condition with the least progress to recommendation module 106. Subsequently, recommendation module 106, generates an alert specific to the condition that was identified as having the least progress and transmits the alert to the individual, the caregiver and/or a health care provider.

[0099] Referring back to Figure 3, as discussed above, the disclosed system includes a recommendation module 106 that provides different levels of recommendations based on the individual’s composite health assessment profile and generated health risk scores. In some implementations, recommendation module 106 determines the type of recommendations based on the recipient. For example, in some implementations recommendation module can send alerts and/or recommendations to the individual, the caregiver, a health provider, a third party related with the health care of the individual or any other suitable recipient.

[00100] Specifically, Figures 8A-8B illustrate process 800 performed by recommendation module 106. At step 802, recommendation module receives the one or more health risk scores and composite health assessment profile of the individual. Subsequently, at step 804, recommendation module 106 determines whether the risk score is greater than a pre-determined threshold. In some implementations, the pre-determined threshold can be obtained by existing actuarial tables. In some implementations, the pre-determined threshold can be manually provided by a health care professional and/or the system administrator.

[00101] If at step 804 recommendation module 106 determines that the risk score is not greater than a pre-determined threshold (see e.g., NO at 804), then, at step 808, recommendation module 106 identifies and selects a set of recommendations based on the composite health assessment profile of the individual, generates a recommendation dashboard at step 808 and transmits the dashboard the individual and/or the caregiver at step 810.

[00102] If, however, at step 804 recommendation module 106 determines that the risk score is greater than a pre-determined threshold (see e.g., YES at 804), then, at step 812, recommendation module 106 determines the severity of each of the conditions identified within the composite health assessment profile, and at step 814 selects a set of recommendations based on the identified severity. Subsequently, recommendation module 106, generates a customized recommendation dashboard for the user and/or caregiver at step 816 and a separate customized recommendation dashboard for health care related third party including a request for feedback, at step 818. At step 820, recommendation module 106 transmits the customized recommendation dashboard to the patient and/or caregiver. Further, at step 822 recommendation module 106 receives the feedback from the health care related third party and at step 824, recommendation module 106 adjusts the recommendation dashboard previously provided to the patient and/or caregiver and transmits said adjustments back to the patient and/or caregiver.

[00103] In some implementations, recommendation module 106 can include a social media component that creates groups of individuals based on concordance rate of“yes” answers between composite health assessment profiles, medical/social/family history, and biographical information such as relationship to patient, age, occupation, gender, and location. Specifically, in some implementations, affinities can be represented by an affinity index.

[00104] In some implementations, the affinity index can be used to pair individuals, either one on one, or in groups. In some implementations, recommendation module 106 can generate online group discussions organized by identified conditions, and/or symptoms, abilities, pharmaceutical, non-pharmaceutical use history, and procedure history topics. Anonymous data from discussion threads can be mined for patterns using machine learning, artificial intelligence and/or data analytics, and the information can be aggregated to develop predictive analytics.

[00105] Figures 9-12 illustrate an exemplary user interface for capture module 102 and the recommendation dashboard generated by recommendation module 104. As shown in Figure 9, in some implementations, capture module 102 can request input from a subject and/or caregiver by displaying in a suitable user interface an avatar in the form of a chat. In some implementations, the avatar can be a“virtual nurse”, a live health provider and/or it can have any other suitable form. In some implementations, capture module 102 can request input in the form of a displayed survey, in audible format, sensory format (e.g., digital braille) or any other suitable form. In some implementations, capture module 102 can request information directly from one or more sensors and receive sensor signals 202. In some implementations, capture model 102 can request input form a caregiver and/or user through the manipulation of a slider, thus providing a set of granular data. For example, in some implementations such granularity can be used by the capture module 102 to adapt in real-time the selection of questions to be displayed to the user and/or caregiver. In some implementations, the use of a slider can provide one or more initial real valued weights provided to the neural assessment module 104 in order to assist risk scoring 312 and predictive engine 314. [00106] Figure 10, shows an exemplary output from recommendation module 106 presented on display 120 using a suitable user interface. In some implementations, recommendation module 106 can provide customized recommendations to a caregiver and/or the subject by providing nested results associated with the different activities of daily living that may be impacted. For example, Figure 10 shows a set of recommendations directed to a caregiver and thematically grouped based on the health assessment results of the individual. In some implementations, customized recommendations can include visual aids such as diagrams, animations, and/or videos. In some implementations, the user interface for the displayed recommendations can include a search engine box for a user to identify additional recommendations that may be similar. Further, in some implementations the user interface can provide the recommendations in the form of a“virtual nurse” (e.g., hot), text-to-speech format or any other suitable format.

[00107] In addition, in some implementations, displayed recommendations can include a displayable solicitation for feedback, as shown in Figure 11. For example, in some implementations such feedback can be in the form of starts, value ratings, a slider and/or any other suitable indicator displayable in the user interface. In some embodiments, solicitation of feedback can be provided using audible tones, via speech and/or any other suitable way to accommodate sensory impairment of the subject. In some implementations, the collected feedback is further utilized to refine current and/or future recommendations. In some implementations, the collected feedback can be provided to capture module 102 in order to refine the selection of survey questions provided to the subject and/or caregiver.

[00108] Figure 12, shows an exemplary user interface displaying a health assessment for a subject and accompanying health composite profile. In some implementations, the user interface can provide customizable displays of the health assessment associated with the recipient of the results. For example, a system administrator and/or health provider can receive a confidence index with respect to the survey responsiveness and quality of input data. Moreover, in some implementations, the user interface can provide the composite profile including a risk score for one or more neurological diseases along with a brief explanation. In some embodiments, the health assessment may be generated and transmitted in the form of an alert, email and/or in any other suitable form. In some implementations, the displayed health assessment can include interactive elements that allow for the retrieval of any additional pertinent information such as underlying data, answers and/or any other suitable information.

[00109] In an alternative implementation, the methods and systems described herein can be delivered as web services via a network computing model. In such an implementation, the system may be implemented via a physician/user-operable client device and a centralized server carrying out much of the functionality described above. For example, in some implementations, the network computing environment can include a system operatively connected to a plurality of client devices via a communications network, such as the Internet or a proprietary wireless mobile telephone network. In some implementations, the disclosed system can include hardware and software conventional for web/application servers, but can be further configured in accordance with some implementations of the disclosure to provide the processing functionality described above with reference to system 100, and for interacting with the client devices. By way of example, client devices may be a personal computer, a mobile telephone/smartphone, or a tablet PC, which may have substantially conventional hardware and software for communicating with the server system via the communications network. For example, such devices may be configured for accessing a website or web interface maintained by the server system, such that the physician/user may operate the client device to provide input and/or receive output described above, and to communicate with the server system which performs the associated processing described herein. In these implementations, the client devices may not require any special- purpose software; rather, all special-purpose software is incorporated into the server system, and the client devices are used merely to communicate with inventive server system.

[00110] Alternatively, the client device may be a smartphone, tablet PC or other computing device configured with a specially-configured native software application running on the client device, and communicating with the server system. In such an implementation, some or all of the structure and/or processing described above with reference to system 100 may be provided at client computing device, which may be operate by the user/physician, and which may communicate with the server system to provide the functionality described herein.

[00111] It will be understood and appreciated that one or more of the processes, sub- processes, process steps or approaches described in connection with system 100 may be performed by hardware, software, or a combination of hardware and software on one or more electronic or digitally-controlled devices. The software may reside in an application memory in a suitable electronic processing component or system such as, for example, one or more of the functional systems, devices, components, modules, or sub-modules. The application memory may include an ordered listing of executable instructions for implementing logical functions. The instructions may be executed within a processing module, which includes, for example, one or more microprocessors, general purpose processors, combinations of processors, digital signal processors (DSPs), field programmable gate arrays (FPGAs), or application-specific integrated circuits (ASICs). Further, the schematic diagrams describe a logical division of functions having physical (hardware and/or software) implementations that are not limited by architecture or the physical layout of the functions. The example systems described in this application may be implemented in a variety of configurations and operate as hardware/software components in a single hardware/software unit, or in separate hardware/software units.

[00112] It is also understood that the term database is used to include traditional databases and relational database, flat files, data structures. Examples of some databases include SQL, MySQL, Microsoft Access to give but a few examples.

[00113] The executable instructions may be implemented as a computer program product having instructions stored there in which, when executed by a processing module of an electronic system, direct the electronic system to carry out the instructions. The computer program product may be selectively embodied in any non-transitory computer-readable storage medium for use by or in connection with an instruction execution system, apparatus, or device, such as an electronic computer-based system, processor-containing system, or other system that may selectively fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. In the context of this document, computer-readable storage medium is any non- transitory means that may store the program for use by or in connection with the instruction execution system, apparatus, or device. The non-transitory computer-readable storage medium may selectively be, for example, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device. A non-exhaustive list of more specific examples of non-transitory computer readable media include: an electrical connection having one or more wires (electronic); a portable computer diskette (magnetic); a random access, i.e., volatile, memory (electronic); a read-only memory (electronic); an erasable programmable read-only memory such as, for example, Flash memory (electronic); a compact disc memory such as, for example, CD-ROM, CD-R, CD-RW (optical); and digital versatile disc memory, i.e., DVD (optical). Note that the non-transitory computer-readable storage medium may even be paper or another suitable medium upon which the program is printed, as the program may be electronically captured via, for instance, optical scanning of the paper or other medium, then compiled, interpreted, or otherwise processed in a suitable manner if necessary, and then stored in a computer memory or machine memory.

[00114] Other devices, apparatus, systems, methods, features and advantages of the invention will be or will become apparent to one with skill in the art upon examination of the following figures and detailed description. It is intended that all such additional systems, methods, features and advantages be included within this description, be within the scope of the invention, and be protected by the accompanying claims.

[00115] Having thus described a few particular implementations of the invention, various alterations, modifications, and improvements will readily occur to those skilled in the art. Such alterations, modifications and improvements as are made obvious by this disclosure are intended to be part of this description though not expressly stated herein, and are intended to be within the spirit and scope of the invention. Accordingly, the foregoing description is by way of example only, and not limiting. The invention is limited only as defined in the following claims and equivalents thereto.