WO2017189725A1 - Therapeutic recovery analytics system and method of evaluating recovery - Google Patents

Abstract

A therapeutic recovery analytics system and a method for evaluating recovery of an operator includes providing a controller and at least one device in communication with the controller and collecting data representative of one or more attributes of an operator with the at least one device and analyzing the one or more attributes with the controller to determine the state of the operator. The controller and the at least one device identifies a shutter condition in the operator and gathers additional attributes of the operator. One or more biomarker values are collected from the operator and are analyzed with the operator attributes to develop an operator profile. The controller generates one or more outputs for use by the operator in response to the operator profile to affect a recovery state in the operator.

Description

THERAPEUTIC RECOVERY ANALYTICS SYSTEM AND METHOD OF

EVALUATING RECOVERY

CROSS REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Patent Application No.

62/328,306, entitled "PSYCHOLOGICAL PERFORMANCE P2 BEFORE THE IMPACT PROACTIVE MENTAL HEALTHCARE FOR MILITARY SPECIAL OPERATIONS WARRIORS AND FIRST RESPONDERS" and filed on April 27, 2016, which is incorporated by reference in its entirety in this disclosure.

INTRODUCTION

[0002] This disclosure relates to a therapeutic recovery analytics system and a method of evaluating therapeutic performance of a subject.

[0003] Patient assessment and treatment methods usually stem from traditional disease- based medical care practices and procedures. Put another way, traditional healthcare and treatment models are reactionary in nature, focusing on mitigation of symptoms and are not designed for full, highly-functional recovery. These healthcare treatment models focus on managing a lifetime of symptoms by prescribing powerful medications, which, in the case of mental health patients may cause deeper instability in a patient. This treatment model demands further hypervigilance, caution and concern on the part of the medical staff.

[0004] For example, if a patient has endured long periods of failed, disturbed or impaired cognitive function, the patient may develop disorders, breakdowns, attacks ranging from depression, anxiety, substance abuse, and possibly physical wounds or injuries. Further, mental healthcare practice typically treats each case on an individual basis, and treatments focus on common protocols and standards.

[0005] In January 2017, the Veterans Affairs (VA) system estimates the incidence of Post-Traumatic Stress Disorder (PTSD) remained unchanged since 2008, with 10-20% of post 9/11 combat veterans experiencing PTSD. Early identification and treatment are the primary indicators of improved outcomes following mental health injury from trauma.

However, barriers within the traditional mental health system result in 60%-80% of soldiers who need mental healthcare failing to seek and receive treatment for over a year after injury. This failure within the mental health system to recognize and deliver timely care leads to debilitating life-long disease, billions of dollars in annual costs of health-care and lost productivity.

[0006] The traditional mental health approaches fail to proactively assess and treat subjects that may become susceptible to mental conditions to prevent potential health issues. For example, the average mental health span of modern combat soldiers is around 10 years, with burnout commonly occurring at age 31. Individuals that are subj ected to enhanced human performance conditions, such as extreme athletes, or soldiers in high threat combat situations, such as special warfare troops may experience shorter spans of mental health before burnout. Current techniques fail to evaluate mental and emotional performance of special operations warfighters, intelligence warriors, and medical professionals in high threat environments before the impact of trauma.

[0007] Additionally, utilizing human intuition to make deployment and mental health decisions does not support the fast-paced life of modem special operations. Leadership groups cannot afford to wait for different personalities between professional mental healthcare providers and experts to resolve conflicts among themselves to address the high levels of threat and exposure.

[0008] While some high-risk, high-threat situations common to military and first responder scenarios are obvious, such as explosions in the battlefield or downed helicopters, many others are difficult to identify, such as life-or-death decisions made away from anything conspicuous, complex interaction of events with an individual's personal experiences, and the cumulative effect of "minor events." Similarly, routine mental health checkup systems overburden mental health staff and resources through unnecessary focus on unaffected individuals, resulting in delayed treatment for affected individuals and overconsumption of resources. The optimal solution to providing proper mental healthcare relies on the ability to passively monitor individuals for signs of mental injury, and provide immediate warning to mental health professionals to allow early intervention.

SUMMARY

[0009] A therapeutic recovery analytics system and a method for evaluating recovery of an operator is provided. In one embodiment, a method of evaluating performance and recovery in an operator includes providing a controller and at least one device in

communication with the controller and collecting data representative of a first set of one or more attributes of an operator with the at least one device. The first set of operator data is analyzed with the controller to determine a first state of the operator.

[0010] Data representative of a second set of one or more attributes of an operator is collected with the at least one device and the second set of operator data is analyzed with the controller to determine a second state of the operator. The controller evaluates the second set of operator data with the first set of operator data to develop a first operator profile and identifies, with the controller and the at least one device, a shutter condition in the operator to gather a third set of one or more attributes of the operator.

[0011] One or more biomarker values are collected from the operator with the at least one device and are analyzed with the third set of one or more attributes of the operator against the first operator profile with the controller to develop a second operator profile. The controller generates one or more outputs for use by the operator in response to the second operator profile to affect a third state in the operator.

[0012] The one or more attributes of the operator collected by the at least one device and analyzed by the controller may include one or one or more physical, physiological and psychological values of the operator that are collected as data with the at least one device. The method may also include the step of providing a database containing one or more data resources used by the controller to generate the first and second operator profiles. The controller may implement one or more algorithms to analyze the second and third sets of one or more attributes of the operator with the one or more data resources in the database to generate the first and second operator profiles and may further implement machine learning instructions in response to outputs from the one or more algorithms to develop evaluation protocols for use in evaluating and generating operator profiles.

[0013] The method may also include the step of monitoring the operator with the controller and at least one device in response to the generation of the second operator profile to affect the third state in the operator. The method may also include the step of transmitting at least one behavioral modification command to the output in response to evaluation of the one or more attributes of the operator following detection of the shutter condition.

[0014] In another embodiment of the disclosure, a method of evaluating performance and recovery in an operator includes providing a controller and at least one device in

communication with the controller and collecting data representative of a first set of one or more attributes of an operator with the at least one device. The first set of operator data is analyzed with the controller to determine a first state of the operator. [0015] Data representative of a second set of one or more attributes of an operator is collected with the at least one device and the second set of operator data is analyzed with the controller to determine a second state of the operator. The controller evaluates the second set of operator data with the first set of operator data to develop a first operator profile and identifies, with the controller and the at least one device, a shutter condition in the operator to gather a third set of one or more attributes of the operator.

[0016] One or more biomarker values are collected from the operator with the at least one device and are analyzed with the third set of one or more attributes of the operator against the first operator profile with the controller to develop a second operator profile. The controller generates one or more outputs for use by the operator in response to the second operator profile to affect a third state in the operator and transmits at least one behavioral modification command to the output in response to evaluation of the one or more attributes of the operator following detection of the shutter condition.

[0017] The one or more attributes of the operator collected by the at least one device and analyzed by the controller may include one or one or more physical, physiological and psychological values of the operator that are collected as data with the at least one device. The method may also include the step of providing a database containing one or more data resources used by the controller to generate the first and second operator profiles.

[0018] The controller may implement one or more algorithms to analyze the second and third sets of one or more attributes of the operator with the one or more data resources in the database to generate the first and second operator profiles and may further implement machine learning instructions in response to outputs from the one or more algorithms to develop evaluation protocols for use in evaluating and generating operator profiles. The method may also include the step of monitoring the operator with the controller and at least one device in response to the generation of the second operator profile to affect the third state in the operator.

[0019] In another embodiment of the disclosure, a system for evaluating a state of health in an operator includes a data gathering component including at least one device configured to obtain one or more attributes about the health of the operator. A controller in

communication with the data gathering component receives and analyzes the one or more attributes of the operator. A discovery and analytics module is in communication with the controller and is configured to generate at least one output for use by the operator in response to data analysis received from the controller. The controller generates at least one operator profile in response to the analysis of the one or more attributes of the operator that is transmitted by the discovery and analytics module to the operator affect a change in the health of the operator.

[0020] The at least one device of the data gathering component collects one or one or more physical, physiological and psychological values of the operator indicative of the health state of the operator. The controller generates at least one operator profile in response to collection of data representative of a first set of one or more attributes of an operator with the at least one device and analysis of the first set of operator data with the controller to determine a first state of the operator.

[0021] The controller collects data representative of a second set of one or more attributes of an operator with the at least one device and analyzes the second set of operator data with the controller to determine a second state of the operator. The controller evaluates the second set of operator data with the first set of operator data to develop a first operator profile and identifies a shutter condition in the operator to gather a third set of one or more attributes of the operator. The controller analyzes one or more biomarker values and the third set of one or more attributes of the operator against the first operator profile with the controller to develop a second operator profile and generates one or more outputs for use by the operator in response to the second operator profile to affect a third state in the operator.

[0022] The controller includes a database containing one or more data resources used by the controller to generate the at least one operator profile. The controller implements one or more algorithms to analyze the one or more attributes of the operator with the one or more data resources in the database to generate the at least one operator profile. The controller implements machine learning instructions in response to outputs from the one or more algorithms to develop evaluation protocols for use in evaluating and generating the at least one operator profile.

[0023] The above features and advantages, and other features and advantages of the disclosure, will be readily apparent from the following detailed description of the

embodiment(s) and best mode(s) for carrying out the disclosure when taken in connection with the accompanying drawings and appended claims.

BRIEF DESCRIPTION OF THE DRAWINGS [0024] FIGURE 1 is a flowchart detailing a method of evaluating the therapeutic recovery of a patient using the therapeutic recovery analytics system in accordance with the disclosure;

[0025] FIGURES 2A-2D are schematic illustrations representative of results of monitored accumulated operator stress over a specified time in accordance with the system and method of the present disclosure;

[0026] FIGURE 3 is a schematic illustration of the therapeutic recovery analytics system in accordance with the disclosure;

[0027] FIGURE 4 is a schematic illustration of another embodiment of the therapeutic recovery analytics system;

[0028] FIGURE 5 is a schematic illustration of the system learning and prediction modeling using a reference baseline signal in accordance with the disclosure;

[0029] FIGURE 6 is a schematic illustration of a psychological impairment prediction model of the system using an estimate of the number of potentially -traumatic events that an operator has encountered during a window of time;

[0030] FIGURE 7 is a schematic illustration the learning and predictive phases of classification in the system in accordance with the disclosure;

[0031] FIGURE 8 is a schematic illustration of the learning phase of a classifier for predicting changes in baseline from physiology from the number of potentially -traumatic events the operator has experienced; and

[0032] FIGURE 9 is a schematic illustration of a model to predict impairment from various longitudinal features in the biomarker in accordance with the disclosure

DETAILED DESCRIPTION

[0033] Reference will now be made in detail to several embodiments of the disclosure that are illustrated in accompanying drawings. Whenever possible, the same or similar reference numerals are used in the drawings and the description to refer to the same or like parts or steps. The drawings are in simplified form and are not to precise scale. For purposes of convenience and clarity, directional terms such as top, bottom, left, right, up, over, above, below, beneath, rear, and front, may be used with respect to the drawings. These and similar directional terms are not to be construed to limit the scope of the disclosure.

[0034] The components of the disclosed embodiments, as generally described and illustrated in the Figures herein, could be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the disclosure is not intended to limit the scope of the disclosure, as claimed, but is merely representative of possible embodiments of the disclosure.

[0035] In addition, while numerous specific details are set forth in the following description to provide a thorough understanding of the embodiments disclosed herein, some embodiments may be practiced without some or all of these details. Moreover, for clarity, certain technical material in the related art has not been described in detail to avoid unnecessarily obscuring the disclosure.

[0036] The present disclosure relates generally to a therapeutic recovery analytics system and method for evaluating the recovery of a subject. The system may include a virtual intelligence system and be termed "Therapeutic Recovery Analysis - Variable Intelligence System" or "TRA-ViS" for short. The system and method uses or applies data and machine learning to enhance healthcare treatment, both mental healthcare and overall patient healthcare. The system and method operate in real time to establish a database of individual and peer group baselines, parameterize deviations from these baselines and use these deviations to predict previously correlated psychological impairment by using interpretable machine learning and analytic models to develop and produce actionable data for medical

professionals and evaluate intuitive patterns in the data.

[0037] In one non-limiting embodiment, the system and method provide real-time and historical insight about subjects or operators, such as special warfare soldiers, to medical professionals before the impact of trauma of warfare on the operators to eliminate or reduce mental health injuries associated with high-threat operational environments. This insight enables command leadership to plan and take actions to extend the operational shelf-life of operators, and prevent burnout to extend the operational mental health life span of the subjects into at least the ages of 45-50.

[0038] The system and method of the present disclosure assists medical professionals and automates certain aspects of healthcare treatment, generates valuable insights, and reduces human uncertainty in patient evaluation. This may mitigate incidents of subjective or questionable mental healthcare that can be used to proactively anticipate injuries and prevent wounds, based on ongoing environmental and biosensor data in real-time, and equip healthcare professionals and leadership individuals to intervene immediately or within hours of an event, versus months and years after trauma where the wounds have grown and evolved to the point of becoming effectively untreatable. Collectively, this system and method provides the opportunity to establish objective metrics of operational readiness, adaptation potential and resilience for military operators. These data can be further evaluated to optimize physical training, nutritional interventions and psychological treatment needs during operator training, deployment and combat.

[0039] The system and method of the present disclosure will operate in real time for the early prediction of trauma-induced impairment and establish a database of individual and peer group baselines, parametrize deviations from the individual and peer group baselines and use these deviations to predict previously correlated psychological impairment. The system and method will strike a balance between interpretable machine learning models, so that when such predictions are made, an actionable, succinct summary can be issued to a mental health professional as well as the operator's direct leadership and "black-box" models capable of identifying complicated if not intuitive patterns in the data. The system and method may be designed as a modular platform with plug-and-play capabilities, allowing it to incorporate and adjust to new multivariate signals enabled by advances in medical technologies and psychological evaluation, and experiment with different machine learning algorithms.

[0040] Referring to the drawings, wherein like reference numbers correspond to like or similar components throughout the several Figures, Figure 1 illustrates one non-limiting embodiment of a method for evaluating recovery of a subject. The method may include evaluating therapeutic adjustments to the treatment, care and training of an operator in accordance with the present disclosure. The method, generally represented by reference numeral 100, evaluates a variety of psychological or mental, physical or biological and physiological variables associated with healthcare and high-threat environments by monitoring, collecting and displaying health, physiological and biological information in data form for analysis and treatment.

[0041] As will be explained in greater detail below, the method may monitor physical actions and how close the subject is to physical and/or mental exhaustion to evaluate whether the operator is performing at a high level of cognitive function to carry out a full range of requested or required actions. For example, soldiers may experience a wide range of stress that accompanies initial enlistment, the completion of basic training, the continued training continuum, and potential deployment and sustained operations. Therefore, providing additional monitoring techniques and biomarkers that can assist in pre-identifying candidate qualifications and/or resilience may act to streamline training effectiveness and potential treatment. This may also translate to increased economy of training and when coupled with other assessment tools and strengthen candidate selection for advanced assignments.

[0042] The system and method may be used to prevent traumatic physical and/or mental health injuries associated with trauma in high threat environments, such as combat, warfare, or other form of exertion before the impact of trauma or mental health injuries including anxiety disorders, stress, depression, Post-Traumatic Stress disorders, and self-medication or substance abuse. The method will help the operators endure long-term exposure to extreme levels of life-threatening stress and cope with situations which demand full-spectrum awareness, hyper vigilance, focus and concentration, when the operator is positioned in high- stress environments. This proactive method may assist a medical practitioner or leader in learning and understanding a subj ect to develop a persistent, proactive, and personalized mental healthcare strategy before, during or after each mission or battle in real time. It is understood that the system and method may also be used in everyday or normal/non-high threat environments to accomplish the objectives of this disclosure.

[0043] A flowchart illustrating the method 100 of the present disclosure as Figure 1 is described in greater detail. The method 100 need not be applied in the specific order recited herein and it is further understood that one or more steps may be eliminated. The method may begin at box or step 101. At step or box 102, data representing one or more attributes or values about a subj ect's physical, physiological and/or psychological state are collected from the operator or subject. In one non-limiting example, a subject or operator may include a military service operator, extreme athlete or the like, though it is understood that that the subjects may be humans from any demographic or position. For purposes of simplicity, the subject will be described as a military service operator or a soldier.

[0044] From an operator's recruitment until his or her retirement, the operator's physical, psychological and physiological values or data will be collected, analyzed and monitored. At the recruit stage, until the operator is officially selected, basic data and information is collected. As the operator elevates in responsibility, deeper and more persistent data collection will take place. It is contemplated that a more experienced operator becomes a higher priority asset, and scarcer in availability. As such, the operator demands closer scrutiny and greater allocation of assets to his or her resources, equipment and support.

[0045] The one or more values collected at step 102 may be initial assessments of the operator, including, but not limited to, demographic information, height, weight, body composition or hydro densitometry, psychological assessments and collection of physical specimens. A typical battery of psychological assessments serves as a measure of the cognitive aspects of the participants' psychological health. The National Institutes of Health (NIH) Toolbox Cognition Domain battery assesses six regularly recognized facets of cognition; executive function, attention, episodic memory, language, processing speed and working memory.

[0046] A physical specimen such as a small hair sample of about 3 centimeters will be obtained near the scalp from the posterior vertex region of the head of the subject. The inclusion of a 3-centimeter sample represents approximately a 3-month period for extended Cortisol exposure. The feasibility of physiological monitoring will be coupled with measures of psychological resilience, acute and long-term Cortisol. This approach will serve as initial pilot data to better determine how extended physiological monitoring may be useful in pre- identifying stress signatures otherwise exclusively aligned with more invasive biomarker collections. Cortisol is a steroid hormone released by the adrenal glands.

[0047] Moreover, the use of hair sampling and the measures of long term Cortisol exposure will be aligned with physical and psychological metrics aligned with fitness capacity and resilience. While previous work has suggested that Cortisol levels (urine, blood, saliva) demonstrate rapid and transient changes to varied modalities of stress (obesity, physical exercise, cognitive/mental challenge), it is difficult to obtain the necessary number of samples to adequately quantify extended exposure to all-cause Cortisol exposure. It is hypothesized that higher initial extended Cortisol values will be positively correlated with measures of fitness and heat tolerance because of exercise and training. It is also

hypothesized that real-time monitoring will identify unique stressor signatures that may serve to correlate with unique psychological outcomes.

[0048] After the completion of the initial hair sampling and physical attribute testing, operators may be monitored during a 10 to 15-day period using one or more activity monitoring devices as will be described in greater detail below. For example, measurement of a heat tolerance protocol may be determined from continuous recordings of rectal and skin temperature, heart rate, ratings of perceived exertion and physiological strain index (PSI). This protocol can accurately determine heat tolerance of the operator, which is tightly linked to initial aerobic fitness capacity.

[0049] It is hypothesized that higher initial extended Cortisol values will be positively correlated with measures of fitness and heat tolerance because of exercise/training history over the 3-month period prior to study enrollment. It is also hypothesized that real-time monitoring will identify unique stressor signatures that may serve to correlate with unique psychological outcomes as will be described below.

[0050] At step or box 104, the one or more values collected at step 102 of the operator data is analyzed with a controller. The controller, as will be described in greater detail below, may comprise one or more independent or integrated modules or components that cooperate to function as the controller. The controller measures and analyzes the physical, physiological and/or psychological value of the operator to determine a first state of the operator.

[0051] At step or box 106, at least one device is positioned on or adjacent to the operator to collect data representative of a second set of one or more attributes of the operator. The at least one device may include a variety of technologies, including but not limited to algorithms, bio-sensors, wearable technologies, micro-computing devices, communication devices, encryption, software, and the like that are used to monitor and obtain real-time performance values or attributes of the operator. It is contemplated that, in addition to the technologies listed above, that the method step 106 may also include reports or data gathered using one or more methodologies for evaluating physical and/or psychological performance of the operator. For example, the methodologies may include evaluation of military mission planning meetings, mission rehearsals or practice events, evaluating operational footage from drone or individual video imaging devices, satellite communications, and updates after mission reports and the like.

[0052] At step or box 108, the controller analyzes the one or more values obtained from the operator with the controller. The analysis may include evaluation of the values based upon previous reports or readings obtained from the subject as described in step 102 or may include evaluation of peer data. Through use of the combination of the methodologies and values obtained by the technology from the operator, the controller may conduct real-time analysis of the subject to detect, forecast, predict, and anticipate emerging mental health concerns and injuries common with special operations warfare, combat, law enforcement, etc., before, during, and after each mission.

[0053] The box referenced as numeral 109 generally represents the above-described operator data collection steps (steps 102 and 106) and data analysis steps (steps 104 and 108). It is contemplated that the data collection and data analysis steps are continuously run for an operator during the operator's deployment or service time. Box 109 graphically represents this operation deployment time provides an initial framework for the advancement of 24-7 physiological monitoring to develop unique stress signatures indicative of positive or negative adaptation to the training and/or operational environment. When coupled with other biomarker collections and minimally invasive psychological evaluations, an operator profile may be developed that may assist the operator and/or one or more medical professionals to identify an at risk profile of the operator to minimize post operations/deployment clinical concerns for the operator.

[0054] This data will be analyzed for trends that link initial Cortisol levels and physical fitness as predictors to training completion as an outcome. These results may indicate the potential for hair Cortisol as a metric to gauge a potential operator's likelihood to successfully complete training, enabling refined admission or exclusion criteria. It is also hypothesized that the continuous physiological monitoring will reveal stressor signatures indicative of an operator's ability to complete training. These stressor signatures will be further analyzed for indications of changes in an operator's psychological profile, either through changes in the operator's psychological assessments, or through development of profiles that have been previously shown to indicate abnormal psychological stresses.

[0055] Stress is welcomed and natural, however, too much unregulated stress left unchecked or managed over too long a period of long exposure to high-risk, high-threat environments is proven to reduce human performance even causing Post Traumatic Stress Injuries (PTSi), anxiety disorders, depression and worse, substance abuse or self-medication compensating for loss of concentration, creating unnecessary distractions that compromises the performance of the subject or operator.

[0056] Integral to an algorithmic approach to predicting injury or impairment is the concept of a model, or a schematic representing the inter-relationship of different elements of the system, in particular the relationships connecting data to be collected and context to be detected or predicted. As an example of a simple model connecting exposure to stressful events and psychological injury / impairment, an operator experiences stressful events. These events will occur with different frequencies depending on the setting, for example depending on whether the operator is on the battlefield or making high-risk decisions removed from the lines of fire.

[0057] As a proxy for the accumulated stress experienced by an operator at any given time, consider the total of the number of stressful events experienced by the operator within the most recent time interval. The size of the interval then serves as a stand-in for the inverse of an operator's resilience, in that the longer the time window, the longer a stressful event impacts the operator, and thus the more likely the stressful event will be compounded by subsequent events in the same time window. As shown in Figures 2A-2D, plots of accumulated stress represented by the x-axis and reference numeral 54 versus time on the y- axis and reference numeral 56 or four different hypothetical operators. Starting with Figure 2A, a time window of 5 is adopted, while Figure 2B uses a time window of 15, with Figure 2C using a time window of 25 and finally a time window of 35 in Figure 2D. As the time window increases, the accumulated stress increases. If a value representing the onset of psychological injury occurs at an accumulated stress of about 3, indicated by the line 58 on the plots, a given sequence of stressful events resulting in psychological injury depends on the individual operator's resilience, modeled by the size of the time window.

[0058] At step or box 110, the controller, in cooperation with the at least one device, identifies a shutter condition and captures data about the operator or subject at a specific moment in time representing a third set of one or more attributes of the operator. The moment in time may be referenced as "shutter" and may be in response to an event affecting the operator, such as a moment of trauma, whether it be a firefight, battle, death, killing, high-risk decision making, or the like. In a "shutter" condition, all data gathering tools will freeze- capture a digital snapshot of the operator and his or her physical, physiological and psychological conditions at the moment of the event and help determine how the trauma occurred. It is contemplated that the overall value of shutter may be in anticipating mental and emotional injuries or conditions, and the alert or the alarm notification of the subject and/or other related parties upon detection of injuries or conditions for injuries. If managing staff can project what happened and see what is happening, the staff can anticipate or predict what will happen before the impact of trauma.

[0059] At step or box 1 12, one or more biomarker values may be captured by at least one data gathering device. A biomarker is a biological molecule found in blood, other body fluids or tissues that is a sign of a normal or abnormal process, or of a condition or disease. A biomarker may be used to see how well the body responds to a treatment for a disease or condition. Exemplary biomarker values may be captured by technology such as biosensors or wearable technologies, or may be captured using medical devices to gather bodily fluids such as blood, urine, saliva and the like. This data will be collected in real-time, and communicated over micro-computing and wearable devices via all forms of communication to the controller. The data gathering usually occurs immediately following recovery of the subject at a safe location by nurses and pararescuemen, sometimes termed PJs, or team medics. [0060] Correlations between measurable physiological variables and mental health injury indicate the potential to track the onset of mental illness through monitoring of these indicators. When compared with their peers, individuals diagnosed with anxiety disorders, depression, and mood disorders display significantly different gross physiological signatures in response to stresses, and/or while at rest. Physiological responses do not manifest identically among different kinds of psychological stresses. For example, individuals diagnosed with major depression display a depressed heart rate variability compared to healthy peers, like that observed in individuals diagnosed with PTSD, however they do not display significant differences in the rate at which heart rate modulates in response to a stressor or during recovery.

[0061] Thus, to draw accurate conclusions about a subject's psychological state, we must perform concurrent analysis of numerous physiological parameters. For example, an operator suffering from an anxiety disorder may suffer an increased resting heart-rate, increased number of events with deviation from baseline, an increased heart-rate recovery following stressor, decreased heart rate variability and the like. Certain small molecule biomarkers provide additional insight into an individual's psychological performance capability. For example, analysis of an individual's resting Cortisol level in conjunction with how Cortisol fluctuates in response to stimulus can predict the person's psychological readiness.

[0062] The correlation with psychological injury and impairment may be with changes in the baseline for given physiological signals. Psychological injury may be correlated with changes to the baseline of physiological signals. Thus, it may be expected that at the point where the threshold for psychological injury is reached, there would be a corresponding change in the baseline for the physiological signal. The immediate responses to stress are different. Thus, different individuals will have different instantaneous responses, and the characteristics of an operator's stress response will be part of his or baseline. That is, the baseline for an operator will include not only the characteristic shape of their unperturbed biomarker signal, but also the characteristic responses to different stimuli. For example, the response bumps in this figure can be parametrized by the height and width of the bump. Further, these stress responses themselves can change once a threshold is crossed in accumulated stress.

[0063] At step or box 114, data collected from the operator and communicated to the controller from the one or more data gathering devices along with the third set of one or more attributes of the operator collected at the shutter condition will be analyzed by the controller and prepared for analysis by a medical or consulting team. For example, doctors can assess in real-time the severity of the operator's mental, emotional and physical responses and project the level of mental health injuries or wounds. Similarly, leadership of the operator can assess the subj ect's ability to recover, and determine whether the subject can confidently continue with their active mission. If necessary, the leadership can have the operator pull back to recover.

[0064] As part of step or box 1 14, the controller may implement one or more algorithms or set of instructions that may be coded to monitor, collect, and analyze the data from one or more of the steps above and, thereby, implement machine learning protocols to adaptively learn from the data sets relating to the operator. The data sets may pull information for evaluation from a plurality of resources. For example, resources may include, but are not limited to, databases, communications such as from biosensors, wearable technologies, and the like. In response, the algorithms executed by the controller may be answering questions, assessing, displaying data, tracking, monitoring, making predictions, and recommending courses of action by processing Big Data, sweeping large datasets, and assessing key variables associated with the human mind and emotion.

[0065] Based on historical and real-time data about the subj ect, the controller may generate forecasts of future subject health concerns and recommend courses of action through the interface within hours of the initial diagnosis request, thereby reducing or eliminating mental health wounds and injuries, not months as is the current medical convention. The controller then takes actions to improve upon them by suggesting courses of action, and thus mitigates risks contributing to mental health injuries. Response time in the present disclosure is usually within minutes and hours of traumatic experiences, not months or years after trauma as is the convention.

[0066] At step or box 1 16, the controller generates one or more outputs for use by the operator based upon the second operator profile developed by the controller in response to the shutter condition to affect a third state in the operator. The one or more outputs may include an actionable output including an operator assessment and action report that includes one or more instructions to treat the operator or subject. The actionable output may include a continued monitoring of the operator as represented by line 1 17. Alternatively, as is represented by box or step 118, a user interface in communication with the controller will display and allow access to the findings generated by the controller. The user interface may be identified by the name ORANGE and include a custom engineered user experience interface. Alternatively, the output may be in the form of a report or other tangible result.

[0067] The user interface may include one or more electronic dashboards or cockpits to display data from a point of impact or trauma of a subject to a medical professional or leadership team member. The user interface may allow the medical professional to further analyze an operator's prior data entries including historical data, from the point of impact/trauma backwards by minute, hour, days, weeks and months, potentially even years forward and backwards. The comprehensive collection and analysis of data also allows the ORANGE system to evaluate and integrate predictive analytics to thereby intervene immediately before the impact of trauma.

[0068] In one non-limiting example, medical professionals evaluating the operator, such as experts, doctors and rehabilitative therapists, will be involved from the personal and professional levels and follow and monitor the operators from mission planning, through to practice, and throughout operation. Dashboard or cockpit displays may show a grade system for each operator, according to a plurality of factors, including but not limited to: how long they have actively operated or performed special operations missions, length of deployment, years of experience, levels of physical, emotional and mental maturity, characteristics, technical skillsets, medical, family and financial history, and the like.

[0069] As a result, the controller factors in hundreds of variables deemed important to assessments to allow for immediate intervention with the operator, advising and helping the operators improve their physical, mental and emotional situation and prevent further damage. Proactive, persistent reliable medical intervention not only keeps the operators alive, but preserves their strengths, mitigating risks and injuries. Such intervention can immediately aid in the healing and recovery process, and improve upon the conditions of operators' weaknesses, injuries or wounds.

[0070] At step or box 120, the controller may develop and transmit at least one behavioral modification command to the operator. Cranial Alert Technologies (CAT) are a behavioral modification technology, such an example of a first line or frontline bio-sensor. Example of CAT devices may include audio communications, video communications or the like. Head mounted communication devices may transmit through the ear of the subject to alert the subject of dangerous responses to stress, trauma, their actions and behaviors, and possibly choices and decisions in battle. Examples of warning factors include but are not limited to: elevated heart and pulse rate, blood circulation throughout missions, etc. [0071] If such variables are elevated too long, at too high of levels, an audio alarm will sound off in the operator's ear, slowly entering their conversations as static until it is gradually louder, motivating operators to make immediate changes in behavior. If the teams of medical professionals or leaders do not see real-time improvement, or, from the operator's perspective or if their body mounted sensors or wearable technologies do not convey improvement in performance, the operator will immediately experience interruption or cutoff in radio communications after the operator's heart rate elevates or spikes above critical levels for too long of a time span.

[0072] Elevated or spiked rates in heart rate beats-per-minute for sustained periods of 10 time increases tension, stress and anxieties. Because audio communications are vital to operator's performance, if they lose radio communications and thus unable to transmit voice and data or monitoring transmissions, they will be motivated to change behavior, such as slow down, pause, take a break until radio communications continue. It is understood that the method may also be implemented without the use of step 120.

[0073] Referring now to Figures 3-4, schematic representations of the therapeutic recovery analytics system are provided. Reference numeral 10 generally refers to the therapeutic recovery analytics system of the present disclosure. The system 10 may be developed with a flexible service based architecture, based upon micro service design principles, to provide the technical baseline to collect, analyze, and act on physiological data and potential PTSD indicators and will support short term research and development while enabling long term extensibility. The micro-service architecture of system 10 enables the use of a service bus, generally reference by numeral 12, for integration with wearable heart rate monitors and other Internet of Things (IoT) devices referenced by numeral 14.

[0074] The system service bus 12 may provide the message routing, orchestration, and linkage of loosely coupled processing and service components shown in Figure 2. Through open service interface design, the bus 12 will allow cooperation between services through message transfer rather than dependent linkages between consumers, providers, and functional endpoints. All system 10 architectural modules will interface in this space, allowing integration of the IoT device platform 14, the deep learning or Before Impact Therapeutic Engine (BITE), generally referenced by numeral 16, algorithm based analy sis processes, generally referenced by numeral 1 8 persistence architecture, generally referenced by numeral 20. Discovery and analytics, generally referenced by numeral 22, will be enabled through these interfaces, as will integration with a communications network, including, but not limited to, a cellular network, satellite communications network, Wi-Fi network, wired network and/or the Tactical Assault Kit (TAK) network, TAK servers, and ATAK devices that may be part of the Cranial Alert Technologies (CAT) referenced by numeral 24.

[0075] Referring now to Figure 4, the system 10 is illustrated with a universal front-end data gathering and formatting component, generally referenced by numeral 14, a controller 16 having a data analytics component or bus generally referenced by numeral 12 and in communication with a deep learning system and analytics discovery and presentation component generally referenced by numeral 22. The IoT device platform 14 allows the use of wearable technology devices to provide real time monitoring and analysis of an operator's vital signs and biofeedback. A variety of devices may be contemplated for use with this platform, including, but not limited to, heart rate monitors, blood pressure monitors, voice recognition technology, facial recognition technology, biometric identification technology, retina scanning devices and the like. The operator is generally referenced by block 60.

[0076] The IoT platform 14 may be a modular, plug-and-play system that will adapt to the system 10 as the system collects, parses, and analyzes data. This system 10 relies on a front-end data formatting process in platform 14 that recognizes data files from all devices approved for use with system 10, such as devices that have proven capable of obtaining scientifically accurate data, and converts the data stream into a standard form used by the system data analysis algorithms. The system analytics retains this plug-and-play functionality by adjusting analytical methods according to the types of data available. For example, if heart-rate and heart-rate variability are reliably available throughout an event, each individual data-set will be analyzed, as well as analyses involved correlations between the two data streams. However, if only heat rate data is reliable throughout an event, analyses and conclusions will be limited to those of that single data stream.

[0077] The controller 16 shown in Figures 3 and 4 cooperates with bus 12 of the system 10 and may utilize highly adaptable models that can be operated on several platforms. In operation, the deep learning platform will operate as part of the controller described above to collect data from operators and develop training data sets with the appropriate format to be used with the deep learning framework. The controller 16 is designed to aid objectivity by reducing the subjective nature of human analysis when intuition, instinct or human based assessments are applied. Deeper learning improves the efficiency of the system 10, resembling the actions of a human brain. The system 10 engages in learning, changing, adapting, growing and evolving even, transforming in real-time. This eventually elevates the intelligence of the platform 16.

[0078] The controller 16 processes and transmits biosensor and biomarker data collected from every possible resource, such as databases, wearable technologies, biosensors, microcomputing devices and the like. The controller 16 is an intelligent research module capable of efficiently and expediently analyzing more solution data from more sources for more simultaneous queries than the human mind and sources optimal solutions to top-priority resource requirements to maintain the operator's capabilities.

[0079] The controller 16 uses an Active Intelligence Engine (AIE) and works by applying cognitive and linguistic algorithms to semantically structured data. It utilizes a common database vocabulary, and has sets of rules for handling the information it collects from news organizations, government, academia, the public web, business/professional journals, press releases, and the like. The controller 16 enables computers to do more useful work through intelligent machine-led gathering, interpretation, and organizing of structured, semi- structured and unstructured information from government, the Internet, industry, and academic sources.

[0080] The controller 16 handles multiple types of data. It can handle cataloged data, which are known innovations and represent the basis for the engine's data stores and can also handle real-time data constantly brought in from the one or more devices for analysis. The platform's reasoning ability dynamically grows and improves over time while challenge queries in the algorithm identify and reported problems that are collected and monitored to identify potential patterns. As field challenges and resource requirements are populated into the controller 16, the controller 16 runs the input query against the latest real-time information crawled from the web, as well as data stores already catalogued and structured.

[0081] A challenge query may be returned in a familiar search interface format and can be organized by solution type, area or domain, TRL Level, subject matter expert, and confidence level. Challenge queries remain live and actively monitored until terminated or graduated to a solution set file by the system administrator. The challenge query can be graduated to a file containing supporting documentation organized in a solution set dashboard where the possible solution can be vetted, analyzed and tracked all within the system.

[0082] It is conceivable that solution set files are automatically created based on the system's 10 confidence level about a solution. In addition, it is also possible that challenge queries can be mechanically constructed from the analysis of after action reports, and/or other means of resource reporting. Finally, the open solution registry of the system allows government and academic research labs, corporate research and development teams, and entrepreneurs to share their solution's advancement and progress with governmental entities. The registry is cross-referenced against existing data stores and incoming real time data, to produce an index accounting for relevance, viability and solvency.

[0083] Another part of the system 10 will be analyzing individual and peer-group baselines. Here, a baseline could refer to a reference biomarker signal over some undisturbed window of time, or a characteristic bump in the biomarker signal in response to an external event. The system 10 will identify any responses that are "abnormal" with respect to the person's historical norms, which will be further analyzed for how the data is trending, such as moving closer or farther away from the medical norm, and information on what that type of adjustment means, and/or what it means if that trend continues, and then provide that information to mental health teams as a guideline instead of a warning.

[0084] Incoming biomarkers from wearable devices on the operator 60 through the IoT platform 14 such as accelerometer and sleeping heart rate data will be recorded by wearable devices, or sampled periodically by a healthcare professional. Data from wearable technology may be uploaded onto the IoT platform 14, from which it will go into a database. From the database, the data will be passed to the discovery and analytics pipeline or component 22 where basic signal processing and statistics will be performed to extract features from the data.

[0085] Some of these features are short-term properties of the data, such as a measureable response to an external event, while other features will contain information about the data over a much longer period, for example, how many measureable responses were there in the past twenty -four hours. The analytics pipeline will also determine a reference baseline signal for each operator as well as peer group reference baselines. Statistics on the baseline are maintained in a database, so that the model is continuously updated. Both instantaneous and longitudinal features will be used to classify the state of an operator as "impaired" or "not impaired". The classifier will have been initially developed through coordinated experiments pairing physiological data with psychological evaluations, but trained in such a way that it will continue to learn and update its models over time.

[0086] As part of the development and use of the platform, algorithms 18 may be utilized on data sets gathered from the IoT devices 14 to generate an initial model for the deep learning platform 16. The initial model will be used by the platform 16 and will be adjusted based upon its accuracy and error ratios over as many time periods as necessary until diminishing returns are noticed in the model. In one non-limiting example, the algorithms 18 may be derived from combining the five most effective algorithm learning techniques, including analogizers, evolutionaries, Bayesians, connectionists, and symbolists. The Master Algorithm is a mix or fusion of the five most common and effective algorithms: inverse deduction from philosophy, backpropagation from neuroscience and biology, genetic programming from genetics and evolution biology, probabilistic inference and Bayesian inference, and vector machines from psychology and math.

[0087] The present disclosure in its preferred embodiment will utilize a rule of three. No one algorithm is the best. However, conditions dictate which one or whether all are more appropriate. Ideally, in the disclosure's preferred embodiment, the design will use no less than three approaches averaging out three results or answers. Higher risks may be accepted with only one result. Using awareness, experience, analogies, Bayesian logic, etc., the system 10 can learn, assess, and problem-solve the most complex problems. This includes monitoring the mental health of highly functional and active tactical athletes, who are required to perform at the highest levels of performance.

[0088] In one non-limiting example of the present disclosure, the model may be refined and implemented as an application on a mobile device to perform inferences upon input information received from attached sensors that may be intemal or extemal to the device to provide real time wamings and indicators to psychological staff at a remote location. Once a psychologist at the remote site receives notification of an issue with the operator, the psychologist may download the data from the operator' s device to confirm the validity of the model's assessment.

[0089] It is contemplated that the psychologist's actual diagnosis, the deep l earning model's diagnosis and the model data from the device are then prepared into a result archive that is transmitted back to the controller of the system 14 to be used to further enhance the model. The system's 10 iterative cycle of creating a model, deploying it onto mobile devices, using it within a field environment to allow it to generate issues, and then testing its validity through expert opinion and knowledge will provide a method of early detection of mental and physical illness within military operators.

[0090] Algorithms ultimately act on data, and as a result, algorithms 18 should be trained on data, in particular, data that looks like the data the eventual algorithm will be acting on. This is especially true when the algorithms 18 will be performing some sort of prediction based on the data. The system 10 may use predictions based on data, of the likelihood that an operator will acquire PTSD. As such, for developing this platform, it is best to do so in conjunction with joint physiological and psychological evaluation.

[0091] Thus, "machine learning" of the controller 16, for the purposes of this section, is the collection and interworking of a set of modules, that each perform some essential role in collectively mapping an input such as a data stream from a wearable device, or an uploaded Cortisol sample, to an output such as the likelihood that an operator will develop PTSD without an intervention during the next specified interval, for example in the next day or next week. The system 10 may establish a machine learning pipeline for detecting time-marked events from incoming multivariate biomarker statistics.

[0092] Features may be extracted from the code, for example the establishment of a "baseline" signal, and parametrizing deviations from this baseline, in response to a stressor presented as part of an experiment or evaluation. We distinguish between instantaneous features that appear and disappear within a short interval of time, say thirty seconds or so, and longitudinal features that capture a statistic or categorical variable describing the signal over a longer interval of time.

[0093] The persistence architecture 20 of the system 10 includes a specialized data store to capture and tag Personal Health Information (PHI) that will be used to enhance PTSD analysis and prediction capabilities. The data within the store shall be protected by the obfuscation algorithms. A searchable index for both simple word queries, along with highly faceted search criteria is created to allow the persistence architecture 20 to evolve during the performance of this effort, to become highly tuned for highly transaction input and output for IoT devices 14.

[0094] The discovery and analytics module 22 may enable medical professionals to uncover new relationships between data elements gathered by IoT devices 14 that have been recorded and related data within the system 10 data store. Analytical algorithms will combine related data indicators to transform the data into new quantitative representations that rely upon the deep learning framework 16 to draw inferences and suggest predictions or indicators for medical professionals to act on intervention in certain situations where combined factors could result in degradation of an operator's mental health. The discovery and analytics module provides a feature rich representation of medical factors to achieve actionable data for medical decision makers. [0095] Referring to Figures 5-9, the discovery and analytics modules evaluate various data sets, including biomarkers, variable of interest, standard psychological metrics and the like. Figure 5 graphically represents a series of biomarkers 26 analyzed by the system 10. The biomarkers measure a plurality of variables to provide full situational awareness on the condition of operators. Alternative or future embodiments of the present disclosure may include other variables not explicitly listed here.

[0096] Measuring variables include the following: Cortisol as a blood serum marker, catecholamines such as dopamine, norephinephrine and epinephrine, urine, baseline hypertension proxy mark, stress hormones such as testosterone, etc., vital signs via devices and the like. Other variables may include operator history, including mental, physical, personal, operational subcategories of history, schedules including follow up, personal, work history, time, body weight, pharmaceuticals/prescriptions/medication, Global Positioning System (GPS), location time stamp, exertion levels, vital signs such as pulse, heart beat per minute, blood sugar, etc.

[0097] Variables of interest may be regulated, standard protocols for collecting data and established medical guidelines for standard values. Variables may include but are not limited to: pulse rate, blood pressure, blood oxygen levels, respiration rate, basic metabolic panel, drug screening, etc. Notably, this is not to be used for disciplinary consideration.

Additionally, knowledge of an operator's need to use substances indicates the existence of a psychological issue to be addressed.

[0098] In the disclosure's preferred embodiment, the disclosure will utilize standard protocols on a weekly basis to collect data. This will be conducted through visits with a trained nursing staff. This may involve approximately 5 minutes of the patient's time and a blood draw procedure. The future embodiments of the disclosure will implement new technologies to monitor these statistics on a real-time basis, like current wearable devices.

[0099] Monitoring of the standard physiological variables will give a view into the general functioning of the operator at any one time, and be able to compare their current state with their resting state, or with previously measured baselines. There is a wealth of information about how each physiological parameter reacts to stressful situations, and can affect overall physical and emotional health. The more important point of information will be how an operator's "resting" vital statistics change between deployments. Significant changes in the "resting" values of these metrics will be indicative of something that has changed within the operator over time. This will signal the need for assessment and potential intervention.

[00100] Figure 5 illustrates the effect of the startle or "SHUTTER" function on an operator's vitals, generally represented by numeral 28. A set of standard psychological metrics includes variables having standardized protocols and reporting methods, either through clinical guidelines put forth through the Diagnostic and Statistical Manual of Mental Disorders (DSM) and the International Statistical Classification of Diseases and Related Health Problems (ICD), or through standard research practice within the psychological community. These metrics may include but are not limited to: sleep - Pittsburgh Sleep Quality Index (PSQI), amount per day, time of day, quality, REM cycles, traumatic stress - PTSD Checklist 5 (PCL-5), anxiety - Beck Anxiety Inventory (BAI), depression - Beck Depression Inventory (BAI), and general neurological function - Neurobehavioral Symptom Inventory (NSI).

[00101] In a preferred embodiment, the system 10 will use standard protocols to collect data every 2 months. Minimalist self-reports will also be made every week, with the operator stating whether each parameter, sleep, anxiety, depression, etc., has changed or remain unchanged. A significant change to any parameter will signal the need for assessment and potential intervention. When possible, continual monitoring of these parameters will be done using wearable devices, e.g. monitoring of sleep patterns. Alternative or future embodiments of the disclosure may utilize newer, industry -standard wearables as they become available.

[00102] The process may also check chemical biomarkers. Variables within this section show promise within the scientific literature for their role in regulating/signaling psychological function. Each variable will be measured by taking a weekly blood draw from the operator, and submitting it for testing through standard processes. Variables may include, but are not limited to: Cortisol for stress, as elevated levels of Cortisol signal increased stress levels. This can also be evaluated through concentration in tears.

[00103] Norepinephrine is another possible indicator of stress. Elevated levels of norepinephrine indicate increased stress and/or hyper-awareness. Sustained elevated levels while at rest indicate nervous system injury or psychological dysfunction. Sustained elevated levels may also lead to physical harm. Glial Fibrillary Acidic Protein can be a potential indicator of traumatic brain injury. Preliminary studies suggest elevated levels of GFAP in the blood can indicate TBI 1 -7 days post-injury. Glucocorticoid Receptor values can indicate a potential link to PTSD and depression. Elevated serum levels correlate with increased PTSD and depression symptoms.

[00104] Other variables include endorphins as a measure of pleasure. Level monitoring can indicate psychological state/changes of operator. Sustained low-levels of endorphins can indicate depression or other mood-disorders. Serotonin can be a measure of happiness.

Decreased serum levels of serotonin correlate with onset of depression and other mood- disorders. Testosterone can be an indicator of aggression. Increased levels of testosterone lead to difficulty controlling aggression. Decreased levels of testosterone lead to malaise. Sustained abnormal levels indicate existence and/or risk of psychological dysfunction.

[00105] In addition to extracting useful features from the biomarker signals, the system and method identifies time, location and other components corresponding to an event. An event might be an acoustic shock presented to an operator. The biomarker data at and just after the time of the shock, will likely a show an abrupt increase. The event labels the biomarker data at the corresponding time or startle. Alternatively, the event of interest may be the "impaired" condition and will apply in cases recommended by medical literature and system analysis, as it pertains to psychological impairment or job functionality.

[00106] The system 10 may predict the "impaired" even from patterns in the asynchronous multivariate data from wearable devices and lab samples or in classifying psychological impairment. Alternatively, parametrized deviations from baseline where baseline could mean a reference signal over some undisturbed window of time, or baseline could refer to the characteristic bump response to some stressor.

[00107] A biomarker feature is some quantity that can be computed from the collected data that can be shown to be highly correlated with another outcome, for example the time- marked labels that could mean "stressor applied or not applied" or "psychological impairment or no psychological impairment." Features are a core input to the prediction component of the system 10 and care must be taken in selecting features for a predictive model. For example, as is shown in Figure 5, a reference baseline signal 26 may be undisturbed by an extemal event, or may appear as a bump 28 in the signal due to a startle, and the extraction of the bump from the signal using some signal processing as generally referenced by 30.

[00108] For example, it could parametrize a bump by height and width, then identify bumps within the signal by applying a band-pass filter to the biomarker signal. The filter response will give the times at which a bump occurred. Therefore, a feature signal is created where at each time the value of the feature is, for example, the height and width of the filter. From clinical perspective, it is important to establish a baseline for both the individual and the individual's peer-group. Peer-group statistics on the multivariate signal determine a rough mapping between levels of biomarkers and normative standards in the medical literature. Such statistics are often themselves functions of what are called features.

[00109] A database will be used to store parametrizations of responses to different stimuli. These parameterizations will serve as the baselines for an individual's physiological responses to different sources of trauma, startle, and distraction. Other features that we will be interested in are measures of deviation from an established baseline. Either a reference for the signal over an unperturbed interval of time, or a reference for the characteristic response to stimuli. The bump is parametrized by its height and width, and the height, width pairs for each of the operators may be stored in the database.

[00110] A classifier 36 classifies or detects a startle event from extracted bump features. This will be useful not only for the exercise in using data to predict or classify events but also for being able to detect potentially -traumatic events from the biomarker signal, which will be an important subroutine for more advanced downstream processing. For instance, as is shown in Figure 6, an estimate of the number of potentially -traumatic events that an operator has encountered during the past window of time, could be used as a feature for predicting psychological impairment.

[00111] Extracting parametrized bumps 34 from the biomarker signal 32, to be used as features in a classifier for predicting psychological impairment involves a classifier 36. For example, a characteristic shape in the biomarker data occurs at a time when an external shock was presented. The presence of this bump 34 in the signal to indicate that the operator experienced a startling event. Thus, the deep learning controller and related algorithms 18 of the system may consider for each time point in the data signal, consider the rectangle determined by the largest window, beginning at that time point, such that the biomarker signal all time points within that window exceed some minimum threshold. That is, for each point, a height/width pair will be extracted. In addition to the height/width features computed for the biomarker, the event labels indicate whether a "startle" or other stressing scenario was presented to the participant.

[00112] Figure 7 illustrates both the learning and predictive phases of classification. In the learning phase of building a model, both the sequence of height/width pairs from the data 32, 34, and the corresponding sequence of "startle'V'no startle" labels, are passed to a learning module that fits the parameters of a classifier 36. In most cases, a good place to start for building a classifier is clustering and logistic regression. These are both straightforward conceptually and there are many built-in routines for performing building these classifiers. In the case of regression, the learned model is in the form of weights that get applied to height and width of the rectangle which scale their relative significance in predicting a startle.

[00113] In the predictive phase, the classifier 36 is being used to perform

classification. The weights that were learned previously are used to predict whether a startle has taken place using the extracted height/width pairs from the biomarker signal. There are different metrics for evaluating the predictive performance of the classifier 36. A common one is the area under the ROC curve, where the ROC curve plots true positive rate, the percentage of startle events that were correctly identified, versus the false positive rate, the number non-startle events that were incorrectly identified as being startle events.

[00114] Once a model has been learned 18, the parameters or weights of the model tell us how to use extracted height/width features in detecting startles or potentially -traumatic events. From the instantaneous features of local height and width, and the detection of startling events 34, the system 10 can compute longitudinal or global features for example the number of startles 38 an operator has experienced in the past hour, or day. It is reasonable to presume that the density of potentially -traumatic events an operator has experienced recently correlates with changes in baseline physiology, which is assumed to be correlated with psychological impairment.

[00115] Figure 8 illustrates the learning phase of a classifier for predicting changes in baseline from physiology from the number of potentially -traumatic events the operator has experienced. In developing a classifier 36 for predicting changes in baseline from longitudinal features it is important to have good "baseline has changed" labels. One component of establishing reference and response baselines is computing statistics, for instance the mean and standard deviation, the minimum and maximum, of the signal over some window of time. Then, some form of anomaly detection can be used to indicate that the statistics have or have not changed. However, we are interested in baseline physiology because that changes in baseline are correlated with the event "impairment."

[00116] Thus an important component of identifying what longitudinal features should be baselined in the first place are those that give good prediction of psychological impairment, generally referenced by 40. Thus, for feature extraction, whether for instantaneous features such as the height and width of the local biomarker signal, and longitudinal features such as baseline statistics over a longer window of time generally referenced by 42, establishing the predictive utility as generally referenced by 44 of the features is a primary focus. If change in baseline to a given biomarker feature does not predictively useful, then our system should not care about the feature.

[00117] Figure 9 illustrates learning a model to predict impairment from various longitudinal features in the biomarker, such as detected changes in baseline features and patterns in the experienced startles. The "impairment" / "no impairment" labels will come from administered psychological evaluations referenced by 46 and the features will be extracted from the collected biomarker data 48. A natural place to start as far as features to baseline are those in the medical literature that have been shown to correlate with impairment. However, to identify novel patterns in biomarker data that are predictive of impairment, joint physiological data and psychological labels will be needed to train and test a classifier to effect an agnostic actuarial approach in which meaningful patterns are identified as those that have predictive value.

[00118] One approach is using visual inspection referenced by 50. However, this approach will break down when it comes to inline processing for a couple of reasons. One, keeping a human in the loop to constantly evaluate incoming data streams will be inefficient and subject to error, regardless of the difficulty in keeping up with the volume. Another shortcoming is that there is no adaptability to changing correlations between visually identified patterns and testing positive on a psychological test. The long-term impairment may be predicted by patterns in the multivariate statistics, which may include a combination of longitudinal and instantaneous features. However, the system may predict impairment as shown by 52 can be associated with distinct subpopulations that may be identified early on from the distinct patterns that are detected.

[00119] In a non-limiting example of the disclosure, the system 10 may conduct a primary analysis based on the operator's heart rate, heart rate variability and Cortisol levels. The data obtained from the operator through the at least one device may be parsed by the controller into events that are defined by the time period between departure from then return to the baseline value for the operator. Among the factors that may be evaluated by the controller are the determination of the baseline, duration of the event, the rate of increase of one or more of the values, the rate of decrease of one or more of the values and plateau levels of the values and the like.

[00120] The event parameters may be analyzed by the controller in light of how the values compare to the operator's normal levels as well as medically recognized normal levels based upon a review of information stored in the system database. The controller may evaluate the data and generate an operator profile based on a number of factors, including, but not limited to, is the baseline value atypically high, is the maximal rate observed for this biomarker abnormal, is the rate of increase or decrease abnormal for the operator and the like.

[00121] The controller may conduct additional analysis of the operator's measured values or data. For example, the controller may utilize the deep learning modules of the system to analyze for unpredicted trends in the data. Alternatively, the controller may access and use a global analysis of events to look for trends when something is flagged as abnormal to determine if there are unseen predictors or psychological outcomes that occur as a result of an occurred event.

[00122] Referring back to Figure 3, the TAK services module 24 may represent one potential output of the system 10. As described above, the system 10 transmits at least one behavioral modification command to the operator. Cranial Alert Technologies (CAT) are a behavioral modification technology, such an example of a first line or frontline bio-sensor. The system 10 may be incorporated into the ATAK module.

[00123] The ATAK module was developed for use as a moving map capability that has since transformed into a large scale multi-purpose application with real-time video applications, targeting plugins, and landing zone preparation tools. The system 10 will create a plugin to the ATAK module through the use the TAK approved network for transmitting geolocation data of military members. The system will use the TAK network to transmit diagnosis data back to medical professionals.

[00124] Additionally, the ATAK module will gain a plugin that works within their current program that allows commanders and other military members on the ground to know the current health status of the operators. Due to the risk of false positives that can only be verified by a psychologist, no mental illness diagnoses will be transmitted to anyone except psychological staff. The only information that will be shared to all members within the TAK network is general physiological information, for example heart rate or oxygen level, even with this minor limitation in place this plugin would allow ATAK to not only know where personnel is located but also their current health. In a high pace operational environment, messaging is often not feasible and counterproductive to mission accomplishment, this plugin will alleviate some of the callback requirements of forward deployed operators.

[00125] Alternatively, wearable technologies connected to mobile communications devices such as via cellular, Wi-Fi, satellite communications and the like may be implemented to monitor warrior bodily functions, vitals, and biomarkers and transmit data to the operators over encrypted links while the system and medical professionals will monitor, collect, analyze and transmit data, responses, advice, or listen to the operators. That data will be prominently displayed on dashboards or a virtual diagnostic cockpit over smart contact lenses or "Google Glasses." This exchange of data coming off the operator, where doctors monitor their actions, decisions and behavior combined with biomarkers provide real time insights allowing operators and doctors to interact with each other making appropriate changes as required to improve human performance.

[00126] The detailed description and the drawings or figures are supportive and descriptive of the disclosure, but the scope of the disclosure is defined solely by the claims. While some of the best modes and other embodiments for carrying out the claimed disclosure have been described in detail, various alternative designs and embodiments exist for practicing the disclosure defined in the appended claims. Furthermore, the embodiments shown in the drawings or the characteristics of various embodiments mentioned in the present description are not necessarily to be understood as embodiments independent of each other. Rather, it is possible that each of the characteristics described in one of the examples of an embodiment may be combined with one or a plurality of other desired characteristics from other embodiments, resulting in other embodiments not described in words or by reference to the drawings. Accordingly, such other embodiments fall within the framework of the scope of the appended claims.

Claims

1. A method of evaluating performance and recovery in an operator comprising: providing a controller and at least one device in communication with the controller;

collecting data representative of a first set of one or more attributes of an operator with the at least one device;

analyzing the first set of operator data with the controller to determine a first state of the operator;

collecting data representative of a second set of one or more attributes of an operator with the at least one device;

analyzing the second set of operator data with the controller to determine a second state of the operator;

evaluating, with the controller, the second set of operator data with the first set of operator data to develop a first operator profile;

identifying, with the controller and the at least one device, a shutter condition in the operator to gather a third set of one or more attributes of the operator;

collecting one or more biomarker values from the operator with the at least one device;

analyzing the one or more biomarker values and the third set of one or more attributes of the operator against the first operator profile with the controller to develop a second operator profile; and

generating, with the controller, one or more outputs for use by the operator in response to the second operator profile.

2. The method of claim 1 wherein the one or more attributes of the operator further comprise one or one or more physical, physiological and psychological values of the operator that are collected as data with the at least one device.

3. The method of claim 1 further comprising the step of providing a database containing one or more data resources used by the controller to generate the first and second operator profiles.

4. The method of claim 3 wherein the controller implements one or more algorithms to analyze the second and third sets of one or more attributes of the operator with the one or more data resources in the database to generate the first and second operator profiles.

5. The method of claim 4 wherein the controller implements machine learning instructions in response to outputs from the one or more algorithms to develop evaluation protocols for use in evaluating and generating operator profiles.

6. The method of claim 1 further comprising the step of monitoring the operator with the controller and at least one device in response to the generation of the second operator profile to affect the third state in the operator.

7. The method of claim 1 further comprising the step of transmitting at least one behavioral modification command to the output in response to evaluation of the one or more attributes of the operator following detection of the shutter condition.

8. A method of evaluating performance and recovery in an operator comprising: providing a controller and at least one device in communication with the controller;

collecting data representative of a first set of one or more attributes of an operator with the at least one device;

analyzing the first set of operator data with the controller to determine a first state of the operator;

collecting data representative of a second set of one or more attributes of an operator with the at least one device;

analyzing the second set of operator data with the controller to determine a second state of the operator;

evaluating, with the controller, the second set of operator data with the first set of operator data to develop a first operator profile;

identifying, with the controller and the at least one device, a shutter condition in the operator to gather a third set of one or more attributes of the operator; collecting one or more biomarker values from the operator with the at least device;

analyzing the one or more biomarker values and the third set of one or more attributes of the operator against the first operator profile with the controller to develop a second operator profile;

generating, with the controller, one or more outputs for use by the operator in response to the second operator profile to affect a third state in the operator; and

transmitting at least one behavioral modification command to the output in response to evaluation of the one or more attributes of the operator following detection of the shutter condition.

9. The method of claim 8 wherein the one or more attributes of the operator further comprise one or one or more physical, physiological and psychological values of the operator that are collected as data with the at least one device.

10. The method of claim 8 further comprising the step of providing a database containing one or more data resources used by the controller to generate the first and second operator profiles.

1 1. The method of claim 10 wherein the controller implements one or more algorithms to analyze the second and third sets of one or more attributes of the operator with the one or more data resources in the database to generate the first and second operator profiles.

12. The method of claim 1 1 wherein the controller implements machine learning instructions in response to outputs from the one or more algorithms to develop evaluation protocols for use in evaluating and generating operator profiles.

13. The method of claim 8 further comprising the step of monitoring the operator with the controller and at least one device in response to the generation of the second operator profile to affect the third state in the operator.

14. The method of claim 8 further comprising the step of transmitting at least one behavioral modification command to the output in response to evaluation of the one or more attributes of the operator following detection of the shutter condition.

15. A system for evaluating a state of health in an operator comprising: a data gathering component including at least one device configured to obtain one or more attributes about the health of the operator;

a controller in communication with the data gathering component to receive and analyze the one or more attributes of the operator; and

a discovery and analytics module in communication with the controller configured to generate at least one output for use by the operator in response to data analysis received from the controller,

wherein the controller generates at least one operator profile in response to the analysis of the one or more attributes of the operator that is transmitted by the discovery and analytics module to affect a change in the health of the operator.

16. The system of claim 15 wherein the at least one device of the data gathering component collects one or one or more physical, physiological and psychological values of the operator indicative of the health state of the operator.

17. The system of claim 15 wherein the controller generates at least one operator profile in response to:

collecting data representative of a first set of one or more attributes of an operator with the at least one device,

analyzing the first set of operator data with the controller to determine a first state of the operator,

collecting data representative of a second set of one or more attributes of an operator with the at least one device,

analyzing the second set of operator data with the controller to determine a second state of the operator,

evaluating, with the controller, the second set of operator data with the first set of operator data to develop a first operator profile, identifying, with the controller and the at least one device, a shutter condition in the operator to gather a third set of one or more attributes of the operator,

collecting one or more biomarker values from the operator with the at least one device,

analyzing the one or more biomarker values and the third set of one or more attributes of the operator against the first operator profile with the controller to develop a second operator profile, and

generating, with the controller, one or more outputs for use by the operator in response to the second operator profile to affect a third state in the operator.

18. The system of claim 15 wherein the controller further comprises a database containing one or more data resources used by the controller to generate the at least one operator profile.

19. The system of claim 18 wherein the controller implements one or more algorithms to analyze the one or more attributes of the operator with the one or more data resources in the database to generate the at least one operator profile.

20. The system of claim 19 wherein the controller implements machine learning instructions in response to outputs from the one or more algorithms to develop evaluation protocols for use in evaluating and generating the at least one operator profile.