Methods and systems for evaluating mental conditions

Abstract

Provided are systems, methods, computer-readable media, and techniques for assessing a mental condition, including: (A) obtaining, from a first source, first input data corresponding to a patient; (B) obtaining, from a second source, second input data corresponding to the patient; (C) analyzing the first input data and the second input data using one or more machine learning models to generate a plurality of patient scores for at least two mental conditions for the patient; (D) generating at least two patient scorecards (e.g., clinical rating scales) for the patient based at least in part on the plurality of patient scores, wherein each patient scorecard of the at least two patient scorecards corresponds to at least one of the at least two mental conditions, and wherein each patient scorecard of the at least two patient scorecards is at least partially complete; and (E) outputting or storing the at least two patient scorecards.

Description

METHODS AND SYSTEMS FOR EVALUATING MENTAL CONDITIONSCROSS-REFERENCE

[0001] This application claims the benefit of U.S. Provisional Patent Application No. 63 / 599,836, filed on November 16, 2023, which is entirely incorporated herein by reference.BACKGROUND

[0002] Psychiatric evaluations are comprehensive evaluations that may be conducted by clinician (e.g., mental health professionals) to assess a patient’s mental, emotional, psychological, and behavioral well-being. In a psychiatric evaluation, the clinician may gather useful information about the patient’s mental health status, diagnose psychiatric disorders, formulate appropriate treatment plans, modify existing treatment plans, assess physical conditions that may cause symptoms, determine utility of a psychiatric treatment, identify long-term challenges, determine patient competency regarding their care, assess the patient’s risk of harm to self or others, gather information for expert or witness testimony for a legal proceeding, etc.

[0003] During a psychiatric evaluation, the clinician may gather information through various techniques, such as interviews, questionnaires, scorecards (e.g., clinical rating scales), and standardized assessments. The information gathering techniques may explore the patient’s personal history, family history, current symptoms, and relevant life events. Evaluations such as scorecards (e.g., clinical rating scales) or questionnaires may be completed by the clinician, the patient, or, in some cases, in collaboration with other healthcare professionals, family members, or caregivers to gain a broader perspective on the patient’s mental health.SUMMARY

[0004] One aspect of the present disclosure provides a method for assessing mental condition of a patient, comprising: (a) obtaining, from a first source, first input data corresponding to the patient; (b) obtaining, from a second source, second input data corresponding to the patient; (c) analyzing the first input data and the second input data using one or more machine learning models to generate a plurality of patient scores for at least two mental conditions for the patient; (d) generating at least two patient scorecards for the patient based at least in part on the plurality of patient scores, wherein each patient scorecard of the at least two patient scorecards corresponds to at least one of the at least two mental conditions, and wherein each patient scorecard of the at least two patient scorecards is at least partially complete; and (e) outputting or storing the at least two patient scorecards.

[0005] In some embodiments, the first input data comprises vocal data of the patient and the second input data comprises wearable biometric data of the patient. In some embodiments, thefirst input data voice tone data of the patient and the second input data comprises vocal language data of the patient. In some embodiments, the first source is the same device as the second source. In some embodiments, the first source is a different device than the second source. In some embodiments, the first input data comprises vocal data of the patient. In some embodiments, the method further comprises: preprocessing the vocal data. In some embodiments, preprocessing the vocal data comprises: filtering third-party voices from the vocal data, thereby generating filtered vocal data; and generating a transcription based at least in part on the filtered vocal data. In some embodiments, preprocessing the vocal data comprises: timestamping the vocal data. In some embodiments, preprocessing the vocal data comprises: decrypting the vocal data. In some embodiments, the first source comprises a user device, and the method further comprises: causing the user device to prompt the user to record the vocal data via an audio sensor of the user device, wherein either an audio file or a video file comprises the vocal data. In some embodiments, analyzing the first input data comprises one or both of: analyzing voice tone of the vocal data for sentiment markers, wherein the sentiment markers correspond to at least one patient score of the plurality of patient scores, or analyzing language of the vocal data for words and meaning, wherein the words and meaning correspond to at least one patient score of the plurality of patient scores. In some embodiments, the second input data comprises wearable biometric data of the patient. In some embodiments, the wearable biometric data comprises one or more of: sleep data, activity data, routine data, temperature data, or photographic data. In some embodiments, the wearable biometric data comprises the photographic data and the method further comprises: prompting the patient to capture a facial image, wherein the photographic data comprises the facial image. In some embodiments, the second source comprises one or more of: a smartwatch, a smartphone, smart glasses, a fitness tracker, or a sleep tracker. In some embodiments, the method further comprises: preprocessing the wearable biometric data. In some embodiments, preprocessing the wearable biometric data comprises: timestamping the wearable biometric data. In some embodiments, analyzing the second input data comprises: analyzing the wearable biometric data for patient routine, wherein the patient routine corresponds to at least one patient score of the plurality of patient scores. In some embodiments, the one or more machine learning models comprises one or more neural networks. In some embodiments, the one or more neural network comprise one or more large language model. In some embodiments, the one or more large language models use a transformer architecture. In some embodiments, the one or more machine learning models are trained at least in part by: obtaining a set of training data for a plurality of training patients, wherein the training data comprises (i) one or both of training vocal data or training wearable biometric data for the plurality of training patients and (ii) a plurality of training patient scores for the plurality of training patients; classifying the set oftraining data into a plurality of classified subsets that each correspond to a different training patient score of the plurality of training patient scores or a different range of training patient scores of the plurality of training patient scores; and generating the one or more machine learning models using the plurality of classified subsets. In some embodiments, the one or more machine learning models comprises at least two machine learning models, and wherein each of the at least two machine learning models corresponds to each of the at least two mental conditions. In some embodiments, the one or more machine learning models consists of one machine learning model, wherein the one machine learning model corresponds to each of the at least two mental conditions. In some embodiments, the plurality of patient scores comprise at least three different patient scores. In some embodiments, the plurality of patient scores comprise at least four different patient scores. In some embodiments, the plurality of patient scores comprise at least five different patient scores. In some embodiments, the plurality of patient scores comprise a scale of discrete ratings. In some embodiments, the at least two mental conditions comprise bipolar disorder. In some embodiments, the at least two mental conditions comprise schizophrenia. In some embodiments, the at least two mental conditions comprise two or more of: mania, depression, post-traumatic stress disorder, premenstrual dysphoric disorder, suicidal ideation, psychosis, dementia, attention-deficit / hyperactivity disorder, anti-social behavior, bullying, anxiety, stress, self-analysis, personality, communication styles, or leadership. In some embodiments, the at least two patient scorecards comprise two or more of: a Young Mania Rating Scale scorecard, a Bipolar Depression Rating Scale scorecard, a Positive and Negative Syndrome Scale scorecard, a Patient Health Questionnaire-9 scorecard, a Brief Psychiatric Rating Scale scorecard, or a Myers-Briggs scorecard (or other personality trait scorecards). In some embodiments, each patient scorecard of the at least two patient scorecards is fully complete. In some embodiments, outputting or storing the at least two patient scorecards comprises: causing the at least two patient scorecards to be presented to one or more of: the patient, a caregiver of the patient, or a family member of the patient. In some embodiments, the patient is undiagnosed with both of the at least two mental conditions. In some embodiments, the second input data comprises remote sensing data. In some embodiments, the method further comprises: after a period of time elapsed since performing one or more of operations (a)-(e): (f) obtaining, from the first source, updated first input data corresponding to the patient; (g) obtaining, from the second source, updated second input data corresponding to the patient; (h) analyzing the updated first input data and the updated second input data using the one or more machine learning models to generate a plurality of updated patient scores for the at least two mental conditions for the patient; (i) updating the at least two patient scorecards for the patient based at least in part on the plurality of updated patient scores, thereby generating at least twoupdated patient scorecards for the patient; and (j) outputting or storing the at least two updated patient scorecards. In some embodiments, the period of time is at least about 24 hours. In some embodiments, the period of time is at least about 72 hours. In some embodiments, the period of time is at least about one week. In some embodiments, the period of time is at least about one month.

[0006] Another aspect of the present disclosure provides a non-transitory computer readable medium comprising machine executable code that, upon execution by one or more computer processors, implements any of the methods or techniques above or elsewhere herein.

[0007] Another aspect of the present disclosure provides a system comprising one or more computer processors and computer memory coupled thereto. The computer memory comprises machine executable code that, upon execution by the one or more computer processors, implements any of the methods or techniques above or elsewhere herein.

[0008] Additional aspects and advantages of the present disclosure will become readily apparent to those skilled in this art from the following detailed description, wherein only illustrative embodiments of the present disclosure are shown and described. As will be realized, the present disclosure is capable of other and different embodiments, and its several details are capable of modifications in various obvious respects, all without departing from the disclosure.Accordingly, the drawings and description are to be regarded as illustrative in nature, and not as restrictive.INCORPORATION BY REFERENCE

[0009] All publications, patents, and patent applications mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent, or patent application was specifically and individually indicated to be incorporated by reference. To the extent publications and patents or patent applications incorporated by reference contradict the disclosure contained in the specification, the specification is intended to supersede or take precedence over any such contradictory material.BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The novel features of the invention are set forth with particularity in the appended claims. A better understanding of the features and advantages of the present invention will be obtained by reference to the following detailed description that sets forth illustrative embodiments, in which the principles of the invention are utilized, and the accompanying drawings (also “Figure” and “FIG.” herein), of which:

[0011] FIG. 1 shows an example of a process for assessing a mental condition of a patient;

[0012] FIG. 2 shows an example of a workflow for assessing a mental condition of a patient;

[0013] FIG. 3 shows an example of another workflow for assessing a mental condition of a patient;

[0014] FIG. 4 shows an example of another workflow for assessing a mental condition of a patient;

[0015] FIG. 5 A shows an example of a user device and server for assessing a mental condition of a patient;

[0016] FIG. 5B shows an example of a user device for assessing a mental condition of a patient;

[0017] FIG. 6A shows an example of a user interface for presenting lifestyle score and mood measures of a patient;

[0018] FIG. 6B shows an example of a user interface for presenting a summary of mania ratings of a patient;

[0019] FIG. 6C shows an example of a user interface for presenting a Positive and Negative Syndrome Scale data of a patient;

[0020] FIG. 6D shows an example of a user interface for presenting instructions on recording a vocal sample of a patient;

[0021] FIG. 6E shows an example of a user interface for capturing a vocal sample of a patient;

[0022] FIG. 6F shows an example of a user interface for presenting a dashboard summary for a patient;

[0023] FIG. 6G shows an example of a user interface of a more detailed presentation of mood and speech of a patient;

[0024] FIG. 6H shows an example of a user interface of a more detailed presentation of sleep of a patient;

[0025] FIG. 61 shows an example of a user interface of a more detailed presentation of activity of a patient;

[0026] FIG. 6 J shows an example of a user interface of a more detailed presentation of routine of a patient;

[0027] FIG. 6K shows an example of a user interface for managing notifications for a patient;

[0028] FIG. 6L shows an example of a user interface for notifying a patient of an alert;

[0029] FIG. 6M shows an example of a user interface for presenting weekly insights;

[0030] FIG. 6N shows an example of a user interface for presenting a weekly summary of a patient;

[0031] FIG. 7 shows an example of a method for assessing a mental condition of a patient;

[0032] FIG. 8 shows an example of a computer system that is programmed or otherwise configured to the methods disclosed herein;

[0033] FIG. 9 A shows an example of an unfilled patient scorecard for the Young Mania Rating Scale;

[0034] FIG. 9B shows an example of an unfilled patient scorecard for the Brief Psychiatric Rating Scale;

[0035] FIG. 9C shows an example of an unfilled patient scorecard for the Patient Health Questionnaire-9;

[0036] FIGs. 10A-10B show an example of a generated scorecard for the Young Mania Rating Scale;

[0037] FIGs. 10C-10F show an example of a generated scorecard for Positive and Negative Syndrome Scale; and

[0038] FIG. 11 shows an example of performance data of generating scorecards.DETAILED DESCRIPTION

[0039] While various embodiments of the invention have been shown and described herein, it will be obvious to those skilled in the art that such embodiments are provided by way of example only. Numerous variations, changes, and substitutions may occur to those skilled in the art without departing from the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed.

[0040] The systems, the methods, the computer-readable media, and the techniques disclosed herein provide a platform for patients and caregivers (e.g., patient families and healthcare providers) to better monitor, treat, and diagnose patients for a variety of mental conditions such as schizophrenia and bipolar disorder. This platform may monitor patients using voice and biometric recording devices that may provide valuable insight into patient speech, behavior, routine, and physiology. These insights can be used as by machine learning models, such as large language models, to determine severity scores relating to various mental conditions. These severity scores may then be used to automatically populate clinical scorecards (e.g., clinical rating scales) that healthcare providers have historically used to diagnose or otherwise assess patients. The systems, the methods, the computer-readable media, and the techniques disclosed herein may be further used to monitor the efficacy of treatments and protect patients by enabling more frequent monitoring of the patients’ mental conditions. This monitoring may be a valuable prophylaxis in preventing the onset or worsening of mental conditions for which a patient may be predisposed or diagnosed with.

[0041] The systems, the methods, the computer-readable media, and the techniques disclosed herein may be used to assess a mental condition of a patient. In some cases, the mental condition may include mental illnesses. In some cases, mental condition may include mental traits, such as general personalities or characteristics. In some cases, the mental conditions may include one ormore of: absence seizure, abulia, abuse, acute stress disorder:, addiction and dependency, adjustment disorders:, adverse effects of medication, age-related cognitive decline, akiltism, alcohol addiction, Alzheimer’s disease, amnesia, amphetamine addiction, anger, anorexia nervosa, anterograde amnesia, anti-social behavior, anxiety, Asperger’s syndrome, attention deficit disorder, attention deficit hyperactivity disorder, autism spectrum disorder, autophagia, avoidant personality disorder, barbiturate related disorders, benzodiazepine-related disorders, bereavement, binge eating disorder, bipolar disorder, body dysmorphic disorder, borderline personality disorder, breathing-related sleep disorder, brief psychotic disorder, bruxism, bulimia nervosa, bullying effects on mental health, caffeine addiction, cannabis addiction, catatonic disorder, childhood disintegrative disorder, childhood onset fluency disorder, claustrophobia, communication styles, conduct disorder, confidence issues, conversion disorder, coronavirus effects on mental health, cotard delusion, crisis, cyclothymia, delerium, delusional disorder, dementia, dependent personality disorder, depersonalization disorder, depression, depressive personality disorder, derealization disorder, dermotillomania, desynchronosis, developmental coordination disorder, diogenes syndrome, discrimination effects on mental health, dissocial personality disorder, dissociative amnesia, dissociative fugue, dissociative identity disorder, down syndrome, drug dependency, dyslexia, dyspareunia, dysthymia, eating disorder, Ekbom’s syndrome, emotionally unstable personality disorder, encopresis, enuresis, erotomania, exam stress, factitious disorder, fitness effects on mental health, folie a deux, fregoli delusion, frotteuristic disorder, fugue state, gambling addiction, Ganser syndrome, gender dysphoria, general adaptation syndrome, grandiose delusions, hallucinogen addiction, haltlose personality disorder, hearing voices, histrionic personality disorder, hoarding, Huntington’s disease, hyperkinetic syndrome, hypersomnia, hyperactive sexual desire disorder, hypoactive sexual desire disorder, hypochondriasis, hypomania, hysteria, impulse control disorder, inhalant addiction, insomnia, intellectual development disorder, intermittent explosive disorder, joubert syndrome, kleptomania, Korsakoff’s syndrome, lacunar amnesia, language disorder, leadership, learning disability support, learning disorders, LGBTQIA+ effects on mental health, loneliness, major depressive disorder, malingering, mania, medication introduction, medication stopping, medication-related disorder, melancholia, mindfulness, misophonia, money effects on mental health, morbid jealousy, multiple personality disorder, munchausen syndrome, narcissistic personality disorder, narcolepsy, neurocognitive disorder, neuroleptic-related disorder, nightmare disorder, non rapid eye movement, non-epileptic attack disorder , obsessive-compulsive disorder (OCD), oneirophrenia, onychophagia, opioid addiction, oppositional defiant disorder, orthorexia, panic attacks, paranoia, paranoid personality disorder, parenting effects on mental health, Parkinson’s disease, partner relational problem, passive-aggressive personality disorder,pedophilic disorder, perfectionism, perinatal anxiety, perinatal OCD, persecutory delusion, personality disorders, pervasive developmental disorder, phencyclidine related disorder, phobias, phobic disorder, phonological disorder, pica, polysubstance related disorder, postnatal depression and perinatal mental health, postpartum depression, postpartum psychosis, post-traumatic embitterment disorder (PTED), post-traumatic stress disorder (PTSD), premenstrual dysphoric disorder, premenstrual dysphoric disorder (PMDD), primary hypersomnia, psychogenic amnesia, psychoneurotic personality disorder, psychosis, pyromania, reactive attachment disorder, recurrent brief depression, relational disorder, rem sleep behavior disorder, restless leg syndrome, retrograde amnesia, Retts disorder, rumination syndrome, sadistic personality disorder, schizoaffective disorder, schizoaffective disorder, schizoid personality disorder, schizophrenia, schizotypal personality disorder, seasonal affective disorder (SAD), selective mutism, self- defeating personality disorder, self-esteem, self-harm, separation anxiety disorder, sexual addiction, sexual disorders, sleep paralysis, sleep problems, sleep terror disorder, social anxiety disorder, social media effects on mental health, somatization disorder, Stendhal syndrome, stereotypic movement disorder, stimulant addiction, stress, student life effects on mental health, stuttering, suicidal feelings, suicidal ideation, suicide attempt, tardive dyskinesia (TD), tobacco addiction, Tourette’s syndrome, transient global amnesia, transient tic disorder, transitions effects on mental health, transvestic disorder, trauma, trichotillomania, undifferentiated somatoform disorder, vagus nerve stimulation (VNS), workplace effects on mental health, etc.Technical Advantages of the Platform

[0042] Advantageously, the systems, the methods, the computer-readable media, and the techniques disclosed herein provide benefits to patients, patient families, and healthcare providers by enabling monitoring, early detection, and treatment of various mental conditions that may be challenging to treat due to issues with lack of patient self-awareness to their state of illness, patient adherence to treatment, depth of healthcare provider expertise, tracking of patient progress, resource limitations such as health care provider time and availability, and reliance on subjective patient memories during healthcare provider evaluations and patient interviews. Benefits such as monitoring and early detection provide a prophylaxis benefit to patients and enable healthcare providers to provide treatment to prevent the onset or worsening of various mental conditions.

[0043] Also advantageously, the systems, methods, computer-readable media, and techniques disclosed herein provide private and secure evaluation of patients while permitting more continuous monitoring and analysis of their conditions. For example, a healthcare provider (e.g., physician) will often not be at home with the patient and will only be able to evaluate the patient when in the often contrived environment of the healthcare provider’s office. The systems, themethods, the computer-readable media, and the techniques disclosed herein enable evaluating a patient while the patient moves throughout the world, such as, in their home, at work, at the store, while sleeping, interacting with others, etc.

[0044] Also advantageously, in generating clinical scorecards (e.g., clinical rating scales) that comport with the clinical scorecards healthcare providers have been using for years, the systems, the methods, the computer-readable media, and the techniques disclosed herein may streamline integration of machine learning techniques into the mental health workflow. For example, this streamlined integration may improve efficiency and speed at which healthcare providers can identify and treat patients. Also advantageously, this streamlined integration may help reduce network traffic for communications involved in the identification and treatment of mental conditions in a patient.

[0045] Also advantageously, the systems, the methods, the computer-readable media, and the techniques disclosed herein generate two or more scorecards independently. Simultaneous generation of two or more scorecards is impossible to be completed by a physician. Further, independent generation of two or more scorecards, each corresponding to a different mental condition is impossible for a physician to complete. Even the best physicians cannot practically retain complete independence between filling out two or more scorecards. For example, if a physician believes a patient likely has a first mental condition corresponding to a first scorecard, then when the physician fills out the second scorecard corresponding to the second mental condition, the physician’s bias (e.g., confirmation bias) will influence their completing of the second scorecard (e.g., reducing explanatory power). This problem of bias becomes increasingly strong as a third, fourth and fifth scorecard are added to the analysis process.

[0046] This bias problem cannot be solved by humans alone and requires the use of the systems, the methods, the computer-readable media, and the techniques disclosed herein. Practically, cost, convenience, scheduling, etc. limit a patient’s evaluation by multiple physicians to remove the issue of confirmation bias. However, even if a patient were to schedule two different appointments with two different physicians to perform two different evaluations for two different conditions (in an attempt to remove confirmation bias), the two different physicians would see the patient at different times or different days, experiencing different manifestations of symptoms. This prevents the two physicians from seeing the same perspective or presentation for episodic conditions. Each physician seeing the patient at different times or days may be especially problematic for conditions with high variability in symptoms across different times. For example, an undiagnosed bipolar patient visiting a first physician at a first time may be prescribed antidepressants if the undiagnosed bipolar patient is experiencing a depressive episode at the first day, A bipolar patient prescribed antidepressants may induce a manic episode.Similarly, if the undiagnosed bipolar patient is experiencing a less severe or more mixed symptoms on a second day when visiting a second physician, the patient may again be misdiagnosed and misprescribed medication. Accordingly, there is a need to understand multiple characteristics of the patient, which can lead physicians to delay diagnosis and treatment until they have seen the same patient multiple times. Advantageously, the systems, the methods, the computer-readable media, and the techniques disclosed herein generate two or more scorecards independently whilst in an at-home setting and without physician interaction, enabling multiple observations to be collected and providing improved diagnostic information enabling higher confidence in an earlier diagnosis.

[0047] Further, this problem of confirmation bias also cannot be solved by having the two physicians in the same room, each performing a different evaluation for a different condition. Firstly, this is practically difficult to coordinate and does not scale for evaluating more and more conditions. Secondly, this does not enable a retrospective look back at the patient to consider whether the evaluations administered were appropriate. For example, a patient could be evaluated by a first physician for a first condition and by a second physician for a second condition, where the first physician and second physician are in the same room. However, if in the future, the physicians believe the patient should have instead (or in addition) been evaluated for a third condition and a fourth condition, there is no ability to look back; a new series of evaluations, inherently performed at a later time, would have to be performed. Conversely, using the systems, the methods, the computer-readable media, and the techniques disclosed herein, a large number of evaluations may be performed, reducing the probability of omitting an important evaluation. However, even if an important evaluation is omitted, the systems, the methods, the computer- readable media, and the techniques disclosed herein may look back at the previously-collected data from the first round of evaluations and apply the omitted evaluation to the previously- collected data, without needing to recollect data.Examples of Processes for Assessing a Mental Condition

[0048] FIG. 1 illustrates an example of a process 100 for assessing a mental condition of a patient for the evaluation of the patient for one or more disorders. The process depicted by FIG. 1 uses data which may inform one or more machine learning models on whether a patient presents with one or more disorders, the one or more disorders may be indicated by entries on a mental condition scorecard (e.g., clinical rating scale) which may be informed by the one or more machine learning models. Input into the one or more machine learning models may comprise patient voice data 110 and patient biometric data 115. The voice data 110 and patient biometric data 115 may be analyzed by a model module 120 which may comprise the one or more machine learning models and may be used to inform a scorecard completion module 130. The scorecardcompletion module 130 may be used to provide an ultimate assessment or aid in ultimate assessment of a patient and whether they present with a mental condition.

[0049] Patient voice data 110 may comprise a data representation of the speech of the patient and may comprise vocabulary, syntax, vocal tone, sentiment, emotion, volume, speech rate, sentence structure, coherence, abstraction, concreteness, repetition and vocal language. The patient voice data 110 may be obtained via a recording device with may include both audio and video recording media. Recording devices may include laptops, smart phones, desktop computers, wearable electronics, tablets, cameras, audio sensors, or other means of capturing voice recordings. Patient voice data 110 may be captured actively (e.g., via guided interview or remote call), ambiently (e.g., via cameras in a location, smart device such as an Amazon Echo, Ring devices, or Google Nest, or via wearable recording device), remotely, locally, or any combination thereof. The active capturing of patient voice data 110 may comprise prompting the patient with the recording devices. Patient voice data 110 may be obtained via pre-recorded audio or video files uploaded to a website, phone application, cloud platform, or other means of volatile or nonvolatile storage for later access. In some cases, patient voice data 110 may be obtained from the audio component of a video conference call (e.g., Zoom, Skype, FaceTime, etc.).

[0050] Advantageously, ambient data collection devices may be used to increase the frequency, quality, and objectivity of data collected. Ambient data collection can be used to obtain highly temporally resolved data that can provide detailed insight into patient behavior that may not be obtainable via other common means of data collection (e.g., interviewjournal, form, log).

[0051] Patient voice data 110 may be preprocessed or otherwise adjusted, fixed, or augmented to enhance the quality of the patient voice data 110. The preprocessing may include transcription of the patient voice data 110 to a text representation of the patient voice data 110. The preprocessing may further comprise, timestamping, encrypting, decrypting, or filtering the patient voice data 110. The filtering may comprise removing third-party voices from the patient voice data 110. The removal of third-party voice data may be used to aid the transcription of the patient voice data 110.

[0052] Patient biometric data 115 may comprise a data representation of the physiological state or behaviors of the patient. A device used for gathering patient biometric data 115 may comprise one or more devices and may comprise some, all, or none (e.g., be entirely separate) of the recording devices used for recording patient voice data 110. A device used for gathering patient biometric data 115 may comprise a smartwatch, a smartphone, smart glasses, pedometer, blood pressure monitor, a fitness tracker, a sleep tracker, or any combination thereof. The patient biometric data 115 may comprise one or more of sleep data, activity data, routine data, temperature data, or photographic data. The patient biometric data 115 may be obtained viamotion sensing technology which may track the patient’s movement over time. The tracking of patient’s movement may provide insight into patterns of behavior which may be of use in assessing the mental condition of the patient. Patient biometric data 115 may be preprocessed or otherwise adjusted, fixed, or augmented to enhance the quality of the patient biometric data 115. The preprocessing may comprise timestamping the patient biometric data 115. The photographic data (e.g., facial images) may be gathered via prompts to the patient by a device which may or may not be the recording device used to obtain patient voice data 110. The photographic data may be requested at the time of login for an app associated with the process 100. The biometric data 115 may be gathered actively, ambiently, remotely, locally, or any combination thereof.

[0053] Advantageously, the patient voice data 110 may be deleted after analysis to protect patient privacy.

[0054] The model module 120 take as input the patient voice data 110 and the patient biometric data 115. The model module 120 may comprise generic service models and proprietary service models for the analysis of patient data. The model module 120 may be responsible for generating scores that are then used in the scorecard completion module 130.

[0055] The generic service models of the model module 120 may comprise a voice sentiment method and a voice to text method. The voice sentiment methods, which may comprise machine learning models capable of correlating patient voice data 110 to sentiment (e.g., happy, sad, anxious, manic, etc.), may be used to analyze patient voice data 110 and produce sentiment markers which may correspond to at least one patient score relevant to the scorecard completion module 130. The sentiment methods, which may also be referred to as sentiment analysis, may comprise both voice to sentiment and text to sentiment methods. The generic service models may be comprised of commercially available or open-source APIs.

[0056] The proprietary models of the model module 120 may comprise text analysis and text benchmarking algorithms. The algorithms may comprise machine learning models, where the machine learning models may comprise large language models. Text analysis algorithms can be used to analyze text data, which may comprise transcriptions of patient voice data 110. The text analysis algorithms may be used to analyze the patient voice data 110 via the transcriptions to learn the meaning of the patient voice data 110. The meaning of words, sentences, concepts, topics, ideas and communication style used by the patient may correspond to at least one patient score relevant to the scorecard completion module 130.

[0057] The machine learning models used in the model module 120 may comprise neural networks, deep neural networks, convolutional neural networks, residual neural networks, long short-term memory neural networks, perceptrons, multi-layer perceptrons, or other non-neuralnetwork-based machine learning model. In some cases, the machine learning models may be substituted by statistical (non-machine learning) models.

[0058] The machine learning models used in the model module 120 may further comprise neural network based large language models (LLMs), which may comprise an architecture called a transformer architecture (also referred to as transformers). LLMs may be trained to identify and assign values to specific speech and cognitive behaviors which may correspond to the mental condition of the patient (e.g., schizophrenia, bipolar disorder). The values assigned to the behaviors may be associated with a presence of the one or more disorders and a severity may be assigned to the values and may be used to assess the mental condition of the patient. LLMs may be trained on psychiatric, psychological, and other clinical data to establish a benchmark model that may be used to improve consistency among healthcare practitioners when assigning a diagnosis or mental condition to the patient. Individual LLMs may be trained to better recognize a specific disorder and may be trained to assign severity scores related to the specific disorder. LLMs may further be trained to establish benchmark or baseline models based on the patient individually, these LLMs may be used to measure improvement / deterioration in the mental condition of the patient. LLMs may be used to generate scores to be used by the scorecard completion module 130 to fill out patient scorecards (e.g., clinical rating scales) which may indicate the mental condition of the patient and the presence or absence of the one or more disorders.

[0059] The LLMs may comprise both open-source and service-based algorithms comprise large foundational models (LFMs). LFMs may include Azure OpenAI (complete and chat APIs for GPT-3, GPT-3.5, and GPT-4, used in ChatGPT), OpenAI (complete and chat APIs for GPT-3, GPT-3.5 and GPT-4), Google Vertex Al (e.g., PaLM, PaLM-2), Meta's LLaMa family of models, as well as BLOOM, Ernie 3.0 Titan, Anthropic's Claude 2, FLAN-T5, OpenAssistant, RoBERT a, MiniLM, and MPNet.

[0060] The machine learning models used in the model module 120 may be trained on a training dataset, the training dataset describing a plurality of patients from actual or synthetic patient data. In some cases, the synthetic patient data may include patient data generated that exhibits signs of particular mental conditions. The synthetic patient data may be used, for example, with adversarial techniques (e.g., generative adversarial networks) to test model scoring or model training. In some cases, the synthetic patient data may enable scaling data and protecting privacy of real patients.

[0061] The training dataset comprises voice data, biometric data, and score data which may be used to train the machine learning models to make associations between the voice data and biometric data of the training dataset and the score data of the training dataset. The trainingdataset may comprise a series of subsets for the plurality of patients and may be used to train one or more machine learning models to accomplish one or more tasks. The one or more tasks may comprise the assignment of one or more scores to one or mental conditions. The one or more mental conditions may comprise bipolar disorder, schizophrenia, schizophrenia spectrum disorders, mania, bipolar depression, depression, post-traumatic stress disorder, premenstrual dysphoric disorder, suicidal ideation, psychosis, dementia, frontotemporal dementia, attention deficit / hyperactivity disorder, anti-social behavior, bullying, anxiety, stress, self-analysis, Alzheimer’s, personality, communication styles, or leadership.

[0062] In some cases, the scorecard completion module 130 may obtain the scores provided by the model module 120 and organize the scores into a plurality of scorecards (e.g., clinical rating scales). The plurality of scorecards may be used by patients and clinicians to obtain insight into the mental condition of the patient. The plurality of scorecards can be benchmarked against existing clinical data and may be rapidly understood by practicing clinicians. The plurality of scorecards may comprise a Young Mania Rating Scale scorecard, a Bipolar Depression Rating Scale scorecard, a Positive and Negative Syndrome Scale scorecard, Hamilton Rating Scale for Depression, Premenstrual Dysphoric Disorder Scale scorecard, DAPHNE Frontotemporal Dementia scorecard, Frontal Behavioral Inventory scorecard, Clinical Dementia Rating scorecard, Suicide Intent Scale scorecard, a Patient Health Questionnaire-9 scorecard, a Brief Psychiatric Rating Scale scorecard, a Myers-Briggs scorecard (or other personality trait scorecards), etc. Each scorecard of the plurality of scorecards may be fully or partially completed by the scores received from the model module 120. The plurality of scorecards may be presented to one or more of the patient, a caregiver of the patient, or a family member of the patient. In some cases, the patient may have prior diagnoses, yet in other cases the patient may receive an indication of a presence of the one or more disorders based on the plurality of scorecards.

[0063] Advantageously, the plurality of scorecards (e.g., clinical rating scales) may obtain a higher reliability, accuracy, and consistency over time due to the learning of the algorithms used in the model module 120. The higher reliability, accuracy, and consistency can be attributed to, for example, collecting patient voice data 110 and patient biometric data 115 as the data is generated, without relying on the patient’s subjective recall / memory. Additionally, the higher reliability, accuracy, and consistency can be attributed to, for example, collecting patient voice data 110 and patient biometric data 115 more frequently because the scorecards are generated automatically and to not rely on the time of a medical professional for manual completion.

[0064] Also advantageously, the plurality of scorecards (e.g., clinical rating scales) may assess a greater number of disorders at once. The model module 120 may learn to assess the mental condition of the patient for multiple disorders. The plurality of scorecards may be more reliablethan a human practitioner who may not have expertise in each disorder of the set of disorders assessable by the process 100.

[0065] The plurality of scorecards (e.g., clinical rating scales) generated by the scorecard completion module 130 may comprise clinically validated scorecards that are structured, interpretable, and in use by psychiatrists. Each scorecard of the plurality of scorecards may be completed by the scorecard completion module 130 and may be automatically distributed to the patient, family members of the patient, or healthcare providers for assessment, monitoring, or diagnosis. The plurality of scorecards may include symptom severity scores and other psychological assessments as derived from the scores of the model module 120. The plurality of scorecards may be used to generate alerts when one or more score of the scores breaches a threshold of a set of thresholds that indicate a presence of the one or more disorders. The plurality of scorecards may be generated on an iterative basis for the patient.

[0066] Advantageously, a process 100 of gathering patient voice data 110, patient biometric data 115, generating a plurality of scores based on an analysis of the patient voice data 110 and biometric data 115 via the model module 120, and automatically organizing the plurality of scores into the plurality of scorecards using the scorecard completion module 130 to present to patients, caregiver, or family member, may be repeated iteratively. An iterative version of the process involves gathering updated patient voice data 110 and patient biometric data 115 to obtain updated pluralities of scores for tracking changes to the patient’s mental condition over time. The process 100 may be repeated on an iterative basis in intervals comprising 24 hours, 72 hours, one week, and one month.

[0067] Advantageously, the collection of patient voice data 110 and patient biometric data 115 may be collected in a secure manner to protect patient privacy and confidentiality between patient and patient’s family and healthcare professional. The security of the data may be ensured via encryption technology used to protect data from those not including the patient, family members of the patient, and healthcare professionals monitoring the health of the patient. Similar protocols may be used for protection of the patient’s data when it is transferred between devices used for collection of patient voice data 110 and patient biometric data 115 and the model module 120. The protocols may further be used to protect data as it is transferred from the model module 120 to the scorecard completion module 130. The protocols may be further used to protect data that is presented by the scorecard completion module 130 to the patient, the patient’s family, or the patient’s healthcare provider.

[0068] Advantageously, the automatic / iterative nature of the scorecard completion module 130 may enable healthcare providers to exploit their experience without the need to use time and resources manually filling out the plurality of scorecards (e.g., clinical rating scales) mentionedherein. Reduction of the use of time and resources may enable earlier intervention, allow patient self-monitoring, improve treatment adherence, reduce high-risk situations for the patient, and improve patient safety.Examples of Workflows for Assessing a Mental Condition

[0069] FIG. 2 shows an example of a workflow 200 for assessing a mental condition of a patient 205. The workflow 200 may be used to implement the process 100 of FIG. 1. At a high level, the workflow may include an application user interface 220, a questionnaire runner 230, a questionnaire transformer 240, and workhorses 250. One or more of the application user interface 220, the questionnaire runner 230, the questionnaire transformer 240, or the workhorses 250 may be embodied on computer-readable media executable on hardware. In some cases, the hardware may be a single device or plurality of co-located devices, such as a server. In other cases, the hardware may be dispersed over a plurality of non-co-located devices, such as having the application user interface 210 on a user device (e.g., a phone, a laptop, a desktop, etc.) and one or more of the questionnaire runner 230, the questionnaire transformer 240, or the workhorses 250 on a server.

[0070] The application user interface 220 may be configured to interact with the patient 205 and a caregiver 210 (e.g., a clinician, a doctor, a psychiatrist, a psychologist, a therapist, a social worker, nurse practitioner, a nurse, a family member, etc.). The application user interface 220 may be embodied on a user device accessible to the patient 205 or the caregiver 210. Accordingly, the application user interface 220 may comprise inputs and outputs (e.g., graphical, textual, video, audio, vocal, etc.) to enable obtaining data from the patient 205 or the caregiver 210 or presenting data to the patient 205 or the caregiver 210. The application user interface 220 may collect various types of data corresponding to the patient 205, such as vocal data, image data, video data, and biometric data.

[0071] In some cases, the application user interface 220 may be configured to collect data from one or more wearable devices configured to provide physical (e.g., medical, activity, nutrition, sleep, etc.) data about the patient 205. For example, the wearable devices can be a smart watch, a sleep tracker, a fitness tracker, a pedometer, a smartphone, smart glasses, a blood pressure monitor, etc. For example, the wearable devices can be a smartwatch or smartphone with an installed physical statistic monitoring application to use one or more sensors of the smartwatch, smart glasses or smartphone to collect biometric data such as sleep data, activity data, routine data, temperature data, photographic data, etc. In some cases, in addition or in alternative to wearable devices, one or more ambient devices may collect biometric data of the user. For example, ambient devices may include smart-home technology that collects biometric data, such as activity data or routine data of the patient 205 as they move about their home.

[0072] In some cases, the patient 205 or the caregiver 210 may configure settings and preferences for data collection through the application user interface 220 (e.g., turn one or more data collection instruments 110( 1 )- 110(N) on / off, change a sampling rate, etc.). In some cases, the patient 205 or the caregiver 210 may be able to provide feedback relating to information presented at the application user interface 220 to the application user interface 220. For example, in some cases, the patient 205 or the caregiver 210 may report at the application user interface 220 that the patient 205 has received a treatment, an intervention, a health condition, etc.

[0073] In some cases, the application user interface 220 may be configured to obtain data from the patient 205 or the caregiver 210 directly. For example, the application user interface 220 may obtain data from the patient 205 or the caregiver 210 through a prompting the patient 205 or the caregiver 210 with a questionnaire. In another example, the application user interface 220 may be configured to obtain data from the patient 205 through prompting the patient 205 to provide a vocal recording. This vocal recording may be in response to a speaking prompt. Alternatively, this vocal recording may be without a speaking prompt.

[0074] In some cases, the application user interface 220 may be configured to present information, such as alerts (e.g., auditory, haptic, visual, etc.) to the patient 205 or the caregiver 210. For example, the application user interface 220 may present scores or scorecards (e.g., clinical rating scales) corresponding to the patient 205. Alerts may be generated to indicate to the patient 205 or the caregiver 210 that information is ready for presentation. Alerts may be generated to indicate to the patient 205 or the caregiver 210 that the application user interface 220 is requesting input information.

[0075] The application user interface may communicate with the questionnaire runner 230. The questionnaire runner 230 may be embodied on a computing device. In some cases, the computing device may be accessible to the patient 205 or the caregiver 210. In other cases, the computing device may be inaccessible to the patient 205 or the caregiver 210.

[0076] In some cases, the data (e.g., biometric data, questionnaire data, vocal data, etc.) collected by the application user interface 220 may transmitted to the questionnaire runner 230, such as via manually uploading by the patient 205 or the caregiver 210, or automatically. For example, the patient 205 or the caregiver 210 may authorize the application user interface 220 to transmit data to the questionnaire runner 230. In some cases, the patient 205 or the caregiver 210 can interact with the questionnaire runner 230 via the application user interface 220. For example, a user may be able to update information about themself (e.g., physical statistics, health conditions, behavior, etc.) or biometric information.

[0077] In some cases, the questionnaire runner 230 may be configured to process data obtained from the application user interface 220. For example, the questionnaire runner 230 may request,e.g., from the workhorses 250, a particular combination of voice, text, or image analysis that corresponds to the format of the scorecard (e.g., clinical rating scale) that the questionnaire runner 230 has been configured to complete for that patient at that time. In some cases, the questionnaire runner 230 may batch data collected by the application user interface 220. In some cases, the questionnaire runner 230 may analyze an output. For example, analyzing the output may include aggregating scores produced from the workhorses 250 into complete scorecards and comparing aggregate scores to clinically relevant and patient historic benchmarks. The benchmarks may be benchmarks for a population (e.g., a general population, a condition population, a study population, etc.), or benchmarks for the patient 205. For example, the output may be compared against an “average normal” or against “my normal” for the patient 205. Outputs from the questionnaire runner may be transmitted to the application user interface 220 (e.g., for the application user interface 220 to present the output).

[0078] In some cases, the questionnaire runner 230 may operate in conjunction with the questionnaire transformer 240. The questionnaire transformer 240 may be remote or co-located with the questionnaire runner 230. For example, the questionnaire transformer 240 may be embodied on a computing device (e.g., server, mobile phone, computer, etc.) that maybe the same computing device as the questionnaire runner 230. In another example, the questionnaire transformer 240 may be embodied on a computing device (e.g., server, mobile phone, computer, etc.) that maybe different than the computing device embodying the questionnaire runner 230. The questionnaire transformer 240 may transform questionnaires into data analysis queries. The data analysis queries may be sent to the questionnaire runner 230. In some cases, depending on the computing capacity of the questionnaire runner 230 and the workhorses 250, the urgency of the request, which may vary through time, the questionnaire transformer 240 may present analysis to the questionnaire runner 230 in smaller or larger queries using smaller or larger computing capacity.

[0079] In some cases, the questionnaire runner 230 may operate in conjunction with the workhorses 250. The workhorses 250 may be remote or co-located with the questionnaire runner 230. For example, the questionnaire workhorses 250 may be embodied on a computing device (e.g., server, mobile phone, computer, etc.) that maybe the same computing device as the questionnaire runner 230. In another example, the workhorses 250 may be embodied on a computing device (e.g., server, mobile phone, computer, etc.) that maybe different than the computing device embodying the questionnaire runner 230. The workhorses 250 may perform one or more data analysis or prediction operations. For example, the workhorses 250 may perform one or more machine learning techniques disclosed herein (e.g., training a machinelearning model, using a machine learning model, tuning a machine learning model, validating a machine learning model, etc.).

[0080] As illustrated, the workhorses 250 comprise a plurality of modules, including a privacy engine, a file splitter, a text engine, a tone engine, an image engine, and a video engine. The privacy engine may be used for various privacy enhancing operations, such as identifying, within vocal data, the voice of the patient 205 and isolating the voice of the patient 205 from other sounds (e.g., other voices) in the vocal data. In some cases, the privacy engine may be used for encrypting and decrypting data (e.g., at rest). The file splitter may be used in pre-processing, such as by splitting and batching data (e.g., text data, vocal data, image data, video data, biometric data, etc.). The text engine may be used for processing and analyzing text data (e.g., text input from a questionnaire, voice to text data, etc.), such as by generating text analysis. The tone engine may be used for processing and analyzing tone data (e.g., tonality, cadence, inflection, etc.) from vocal data to generate a sentiment analysis. The image engine may be used for processing and analyzing image data (e.g., images of the patient 205) via image classification or proximity scoring. The video engine may be used for processing and analyzing video data (e.g., videos of the patient 205), such as by isolating vocal data or image data from the video data.

[0081] FIG. 3 shows an example of another workflow 300 for assessing a mental condition of a patient. FIG. 3 illustrates four levels of the workflow 300: user hardware, user application, cloud / server, and sentiment analysis. The four levels of the workflow 300 may be embodied as computer-readable media executing, for example, on one or more computing devices (e.g., servers, personal computers, smartphones, etc.). In some cases, the one or more computing devices are a single computing device. In some cases, the one or more computing devices are a plurality of co-located computing devices. In some cases, the one or more computing devices are a plurality of remote computing devices.

[0082] The user hardware may be one or more of a smartphone, a smartwatch, an activity tracker, a sleep tracker, a routine tracker, a video recorder, a camera, an audio recorder, etc. The user hardware may be the same as or similar to various examples of user hardware (e.g., wearable devices, ambient devices, personal computing devices, etc.) disclosed herein.

[0083] In some cases, the user application (e.g., the application user interface 220) may run on the user hardware. In some cases, the user application may run on one or more computing devices other than the user hardware. The user application may perform various functions. For example, the user application may perform voice identification and voice recording of the patient’s voice. The user application may obtain (e.g., generate, receive, etc.) timestamped voice data. In some cases, the user application may upload the voice data to the cloud / server. In some cases, the user application may generate user processing requests (e.g., with a subscription status)to transmit to the cloud / server. In some cases, the user application may obtain a questionnaire score from the cloud / server. For example, in response to obtaining the questionnaire score, the user application may generate a prompt with further questions (e.g., depending on the questionnaire score obtained). In some cases, the user application may present (e.g., to a patient or caregiver) a dashboard or scorecard (e.g., clinical rating scale). In some cases, the user application may generate one or more alerts (e.g., for prompting a patient or caregiver to provide data).

[0084] In some cases, the cloud / server may store data. For example, the cloud / server may store vocal data, video data, biometric data, etc. The data may be stored, in some cases, temporarily and discarded after use. In some cases, the cloud / server may consolidate or sample data (e.g., vocal data, video data, biometric data, etc.) for questionnaire completion. In some cases, the cloud / server, based at least in part on the data, may generate (e.g., via the questionnaire runner 230) a questionnaire score. Generating the questionnaire score based at least in part on the data may include analyzing sentiment data, text data, biometric data, etc. The questionnaire score may be sent to the user application. In some cases, as additional data is obtained, such as at a repeating interval (e.g., 12 hours, 24 hours, 48 hours, 72 hours, 96 hours, weekly, bi-weekly, monthly, bi-monthly, twice yearly, yearly, bi-yearly, etc.), the cloud / server may process the additional data. Processing the additional data may include updating the scorecard (e.g., clinical rating scale) or dashboard. The cloud / server may transmit the updated scorecard or updated dashboard to the user application for presentation (e.g., to a patient or a caregiver).

[0085] In some cases, the sentiment analysis may include analyzing audio data. In some cases, the audio data may be included in video data. In some cases, the audio data may be isolated from other data types (e.g., video data). In some cases, the audio data may include vocal data. In some cases, the vocal data may be of the patient. In some cases, the vocal data of the patient may be isolated from other audio data (e.g., vocal data of others, ambient noises, etc.). In some cases, the vocal data of the patient may not be isolated from other audio data. In some cases, the audio data may include text data (e.g., a transcription of vocal data). Based at least in part on the audio data, the sentiment analysis may generate sentiment scores. The sentiment scores may correspond to vocal intonation, tonality, cadence, inflection, etc. The sentiment scores may, in addition or in alternative, correspond to analysis of text data corresponding to vocal data. For example, the sentiment scores may correspond to meaning (e.g., explicit or implicit) of the text data.

[0086] FIG. 4 shows an example of another workflow 400 for assessing mental condition of a patient. The assessed mental condition may include, for example, one or more of: bipolar disorder, bipolar mania, bipolar depression, depression, premenstrual syndrome (PMS) / premenstrual dysphoric disorder (PMDD), schizophrenia, psychosis, suicidal ideation,dementia, ADHD, anti-social behavior, bullying, anxiety, stress, self analysis, personality, communication styles, leadership (or any other mental condition disclosed herein). The workflow 400 may be embodied on computer-readable media executable on hardware. In some cases, the hardware may be a single device or plurality of co-located devices, such as a server. In other cases, the hardware may be dispersed over a plurality of non-co-located devices, such as a user device (e.g., a phone, a laptop, a desktop, etc.) and on a server.

[0087] At a high level, the workflow 400 includes generating filled out (or at least partially filled out) symptom severity scorecards (e.g., clinical rating scales). The symptom severity scorecards may be the same as or similar to those that have been developed by the psychiatry and psychology profession to assess many mental disorders listed in the Diagnostic and Statistical Manual of the American Psychiatric Association. Such scorecards may help to standardize symptom assessment between doctors, between patients and over time. Some of these scorecards have large bodies of published clinical evidence built up over many years through independent studies. Such scorecards may be well understood decision support tools for the use in managing patients with mental disorders. The workflow 400 may enable automatic generation of such scorecards from, for example, voice and sensor data collected from patients.

[0088] In some cases, the workflow 400 for automating scorecard (e.g., clinical rating scale) generation may be the same or similar for different types of scorecards. However, in other cases, for some scorecards, specific sensor input may be of higher value. To improve high quality output from automated generation of scorecards, each scorecard may have a specific element of corresponding executable computer-readable media. Corresponding executable computer- readable media may be tested (e.g., validated) for efficacy by comparing the output of automatically generated scorecards with the output generated manually by trained professionals. As a result of this software development and testing process, the menu of scorecards may be expanded over time, from the same or similar foundational techniques.

[0089] In some cases, scorecards (e.g., clinical rating scales) may be developed to address bipolar disorder. Notably, patients with bipolar disorder can exhibit both high mood, low mood, and psychotic symptoms. As a result, substantial elements of scorecards for other mental disorder scorecards can be completed by re-using techniques used for bipolar disorder. Additionally, scorecards for PMS, PMDD and Schizoprenia may be implemented by the workflow 400. In some cases, the same or similar scorecard infrastructure, customizable to multiple specific mental conditions, may then be used on standard phycological, communication, learning styles, personality tests, etc. In some cases, the same or similar scorecard infrastructure, customizable to multiple specific mental conditions, may then be used on personal development and leadership tests. In some cases, with each additional scorecard, an LLM may be implemented. For example,the LLM may be specific to the scorecard. In another example, with each additional scorecard, an LLM that may be the same as or similar to LLMs for other scorecards may be used. In some cases, the workflow 400 may enable wellness and personal coaching for a scorecard population through automated chat interactions and targeting and tracking of scorecard outputs.

[0090] As illustrated the workflow 400 may obtain data from various monitoring tools (e.g., wearable devices, ambient devices, personal devices, etc.) to determine one or more of text data, voice data, image data, activity data, etc. Based on the obtained data, the workflow 400 may generate scores. These scores may be used for generating a scorecard (e.g., clinical rating scales), such as, as disclosed herein.

[0091] In some cases, the generated scorecard (e.g., clinical rating scale) may be configured to be readable by a patient or a caregiver. In some cases, the scorecard may be used in generating a treatment. The treatment may be automatically provided on a user device. In some cases, the treatment may be personalized to the patient based at least in part on their scorecard. The personalized treatment may include social treatment, medicinal treatment, coaching treatment, therapeutic treatment, digital treatment, etc. The treatment may include, for example, one or more ohf personalized coaching, motivational training, gamified training, chatbots, or other forms of personalized treatment. The treatments may rely on coaching tools or therapy tools. In some cases, the treatment may include a cognitive behavioral therapy. For example, if the systems, the methods, the computer-readable media, and the techniques disclosed herein identify (e.g., via a scorecard) that a patient is at increased risk of mania, the patient may be automatically presented with a meditation activity. In another example, if the systems, the methods, the computer- readable media, and the techniques disclosed herein identify (e.g., via a scorecard) that a patient is at increased risk of schizophrenia, the patient may be automatically presented with a question and answer activity.Examples of Systems for Assessing a Mental Condition

[0092] FIG. 5 A shows an example of a user device 510A and server 520A for assessing a mental condition of a patient. The user device 510A and the server 520 A are included in the system 500A. The system 500A illustrates a cloud computing / server example of the systems, the methods, the computer-readable media, and the techniques disclosed herein. Specifically, the system 500A may illustrate the example of having the user device 510A (e.g., smartphone, smartwatch, laptop, wearable device, etc.) that collects data (e.g., wearable data, sleep data, activity data, biometric data, routine data, etc.) and provides the data to the server 520A for processing and analysis. Once the server 520A performs the processing and analysis, the output may be transmitted back to the user device 510A (e.g., for presentation or storage).

[0093] One or more components of the user device 510A or the server 520A may be the same as or similar to the components of the computer system of FIG. 8. Further, one or more components of the user device 510A or the server 520A may be configured to perform one or more of the process 100 of FIG. 1, the workflow 200 of FIG. 2, the workflow 300 of FIG. 3, the workflow 400 of FIG. 4, or the method 700. Further, one or more components of the user device 510A may be configured to present the user interfaces 600A-600K of FIGs. 6A-6K.

[0094] In some cases, the user device 510A may include an input / output module 512A that may include one or more sensors for receiving data input, such as camera, microphone, keyboard, accelerometer, gyroscope, temperature sensor, location sensor, etc. The input / output module 512A may also be configured to present data to a user, such as on a graphical user interface or via speakers, haptic devices, etc. The user device 510A may further include a user application module 514A. The user application module 514A may operate in conjunction with the input / output module 512A to obtain data or output data (e.g., via display). The user application module 514A may be built into the user device 510A or may be downloaded to the user device 510A. In some cases, the user device 510A, while illustrated as a single device may multiple devices in practice. For example, in some cases, a first device may be used to obtain data for sending to the server 520A as input data while a second device may be used to present data sent from the server 520A as output data.

[0095] The user device 510A may be communicatively coupled to the server 520 A via a network 530A. In some cases, the network 530A may enable connection to the Internet (e.g., via a router). The network 530A may connect the user device 510A and the server 520A wirelessly (or, in some cases, via wired connection). In some cases, the network 530A may comprise an intranet. In some cases, the network 530A may operate via a more localized connection such as BlueTooth®, personal area network (PAN), local area network (LAN), wide area network (WAN), or other suitable network connections.

[0096] The server 520A may include a model module 522A and a data storage module 524A. The model module 522A may include one or more models disclosed herein. For example, the models may be machine learning models. These machine learning models may be neural network based models, such as large language models (LLMs). In some cases, the models may be statistical models (not machine learning based). The model module 522A may be configured to analyze data to generate a plurality of patient scores. In some cases, the model module 522A may analyze more than one type of data. For example, the model module 522 A may analyze both vocal data and wearable biometric data (e.g., routine data, activity data, sleep data, temperature data, photographic data, etc.). In some cases, based at least in part on the analysis of the data, the model module 522A may generate a prediction for one or more mental conditions of the patient.For example, the model module 522 A may generate a prediction for at least two mental conditions for said patient. In some cases, the model module 522A may be configured to generate (e.g., at least partially fill out) one or more patient scorecards (e.g., clinical rating scales) for the patient based at least in part on patient scores generated by the model module 522 A. For example, the model module 522 A may be generate at least two patient scorecards, e.g., each corresponding to a different mental condition for the patient. Advantageously, the model module 522A may be able to analyze a patient’s likelihood of having a plurality of mental conditions based on a plurality of different data types - improving over human analysis of a single type of data to assess a single mental health condition and fill out a single patient scorecard. In some cases, the data storage module 524A may store input data to the model module 522A. In some cases, the data storage module 524A may store output data from the model module 522A. For security and data privacy, the data storage module 524 A may erase stored data (e.g., after use of the stored data or after a certain period of elapsed time).

[0097] FIG. 5B shows an example of a user device 51 OB for assessing a mental condition of a patient. The user device 510B of system 500B may be the same as or similar to the user device 510A of the system 510A. However, the system 500B illustrates the example in which at least some of the functionality of the server 510A is embodied on the user device 51 OB. The system 500A may have certain advantages of processing speed and computational efficiency due to offloading more computationally heavy tasks to the server 510A for cloud computing. However, the system 500B have certain advantages of data privacy and data security by keeping all data local to the user device 51 OB.

[0098] Accordingly, the user device 51 OB may include an input / output module 512B and the user application module 514B that may be the same as or similar to the input / output module 512A and the user application module 514A, respectively. As illustrated, the user application module 514B may further include a model module 522B that may be the same as or similar to the model module 522A. The user device 51 OB may store input and output data of the user application module 514B and the model module 522B at a data storage module 524B that may be the same as or similar to the data storage module 524A.Example User Interfaces for Assessing a Mental Condition

[0099] FIGs. 6A-6K show examples of user interfaces 600A-600K for assessing a mental condition of a patient. The user interfaces 600A-600K may be displayed on a user device of the patient or a caregiver. In some cases, the user interfaces 600A-600K may be displayed using an application, such as the application user interface 220 of FIG. 2, the user application of FIG. 3, or the user application 514A / 514B of FIG. 5A / 5B. While FIGs. 6A-6K illustrate using the user interfaces 600A-600K to present information to a user (e.g., a physician, a patient, a familymember of a patient, a care provider of a patient, etc.), in some cases, other means of presenting information may be used in addition or in alternative. For example, in some cases, speakers or haptic output may be used to present information a user.

[0100] The user interface 600A illustrates an overview for the patient of multiple underlying condition relevant scores for the patient’s lifestyle, which may contribute positively or negatively to the risk of more severe symptoms, or the patients mood or state of mind, which may reflect actual symptom severity indications from scorecard (e.g., clinical rating scale) computations. For example, in bipolar disorder, the patient may be following a stable routine of regular sleep and exercise and may be sleeping an average of seven or more hours per night, producing a Lifestyle score >100%. Furthermore the patient may be exhibiting speech patterns that indicate higher than normal levels of excitement, and so, the patient may be at risk of entering a manic episode, as indicated by the 135% Mood Measures score. These summary measures may enable the patient or caregiver to take precautionary steps, such as more closely monitoring mood in the coming days.

[0101] The user interface 600B illustrates a daily, average weekly, average monthly and trended score for the Youngs Mania Rating scorecard (e.g., clinical rating scale) for bipolar disorder. This score gives the patient or caregiver a view of the patient’s symptom severity currently and over time. This time series view illustrates that the patient exhibited higher levels of mania around day 9 but that those levels stayed below a clinically significant threshold of 30. In some cases, if the patient had crossed the threshold, an alert may be generated. For example, this alert may alert the patient, the patient’s family, or a healthcare provider. In some cases, different thresholds may correspond to different types of alerts. For example, a lower threshold may alert the patient, a higher threshold may alert the patient’s family, and a yet higher threshold may alert the patient’s healthcare provider. This information may be helpful for the patient or caregiver to judge whether further intervention is required in order to manage the patient’s condition.

[0102] The user interface 600C illustrates a daily, average weekly, average monthly and trended score for the Positive and Negative Syndrome Scale scorecard (e.g., clinical rating scale). This scorecard is used to assess schizophrenia symptoms and the PANSS 14 scorecard is used for the positive and negative schizophrenia symptoms, whereas a PANSS 30 scorecard includes additional general psychiatric assessment topics. The user interface 600C illustrates that the patient did experience an increase in severity of symptoms around day 9 but those symptoms have now moderated. This information may help the patient or caregiver review whether a medication regimen is being adhered to or whether there are other factors that have led to this pattern of symptoms.

[0103] FIG. 6D shows an example of a user interface 600D for presenting instructions on recording a vocal sample of a patient. The vocal sample may be the same as or similar to the patient voice data 110 of FIG. 1. As explained in the instructions of user interface 600D, the patient may read the instructions themselves out loud into the recording of the vocal sample. Reading the prescribed instructions may aid in calibration, such as interpreting the words spoken, identifying and removing background sounds, adjusting gain, understanding different accents or dialects, identifying a baseline, etc.

[0104] In some cases, calibration from reading the instructions in the user interface 600D may further include learning to identify a patient’s voice. For example, identifying the patient’s voice may enable identifying the patient’s voice in a recording and removing other sounds (e.g., third party voices, ambient sounds, dog barking, etc.) from the recording. In cases of the third party voice being identified as a health care provider (e.g., a therapist, physician, nurse, etc.), the third party voice may not be removed from the recording, enabling, for example, analysis of not only the patient’s speech, but analysis of the patient’s speech in the context of the third party’s speech. In cases of the third party voice being identified as a non-health care provider (e.g., a friend, a family member, a coworker, a stranger, etc.), the third party voice may be removed from the recording, preserving the privacy of the third party.

[0105] In some cases, the user may be prompted to regarding-read the instructions one or more times to ensure full calibration. As shown in the user interface 600D, the instructions may clarify that the recording will identify the patient’s voice, but will filter out other noises, e.g., voices of other people. The instructions may further clarify privacy and security standards for the recording, such as how the recording will be permanently deleted after analysis.

[0106] FIG. 6E shows an example of a user interface 600E for capturing a vocal sample of a patient. In response to the patient activating a start button, the audio recording may begin to generate a vocal sample that may be the same as or similar to the patient voice data 110 of FIG.1. As illustrated in the user interface 600E, in some cases, the recording may be of a fixed length (e.g., 1 minute, 90 seconds, 2 minutes, 3 minutes, 4 minutes, 5 minutes, 6 minutes, 7 minutes, 8 minutes, 9 minutes, 10 minutes, etc.). Once the recording has reached the fixed length of time (or if a user ends the recording early) the recording may cease. In other cases, the recording may have a variable length (e.g., limited only by the storage of the user device).

[0107] In some cases, the user interface 600E may display audio amplitude indicators (e.g., real time or approximately real time) that cue a user as to the input volume levels. In some cases, the user interface 600E may display other indications for improving the recording, such as instructing the patient to move to a quieter area, instructing the patient to speak up, instructing the patient to speak more clearly or more slowly, instructing the patient to repeat something, etc.

[0108] In some cases, the user interface 600E may prompt the patient to discuss particular things. In some cases, prompts may be text-based. For example, prompts may ask a patient to describe their day or week, ask a patient about something they recently enjoyed, ask a patient about how they are sleeping, ask a patient about recent interactions with others, ask a patient about their job, ask a patient about current events, ask a patient about things they read, ask a patient about a memory, etc. In some cases, prompts may be image-based. For example, prompts may be images that are relatively emotive (e.g., images of a lion with cubs, a classical painting, a sunset, a flower, etc.). In some cases, the prompts may be one or more of text-based, image-based, audiobased, or video-based.

[0109] In some cases, the prompts of the user interface 600E may be randomized (e.g., pseudorandomized). Advantageously, in randomizing the prompts, confirmation bias may be reduced compared to a healthcare provider asking new prompts based on previous answers where the new prompts (consciously or subconsciously) confirm the healthcare provider’s hunches or suspicions. In some cases prompts may be generated using machine learning models that receive user data as input. For example, if patient biometric data (that is the same as or similar to the patient biometric data 115 of FIG. 1) suggests a patient has a reduced step count, a machine learning model may generate prompts to inquire about if the patient is physically ill (e.g., experiencing flu-like symptoms), if the patient is depressed, if the patient has been unusually busy at work, if the patient has an injury, etc. Accordingly, prompts may be adaptively selected based on the ability of the prompts to differentiate between a plurality of likely causes or diagnoses. In some cases, to help avoid confirmation bias with adaptive prompts, a plurality of parallel “pathways” of different causes or diagnoses may be provided to the patient (in series or in parallel).

[0110] FIG. 6F shows an example of a user interface 600F for presenting a dashboard summary for a patient. As illustrated, the dashboard summary may include a summary of a patient’s mood and speech, sleep, activity, and routine over a day. For mood and speech, a patient’s behavior may be indicated as low, balanced, or high, which be based on enthusiasm, irritability, calmness, vulnerability, confusion, etc. of the patient. In some cases, an ideal patient has a mood and speech indicated as medium. For sleep, a patient’s behavior may be indicated as short, balanced, or long, which may be based on a length of time of sleep, bedtime or wakeup routine, a quality of sleep, a number of times awoken during sleep, a length of time in light sleep, a length of time in deep sleep, a length of time in rapid eye movement (REM) sleep, etc. of the patient. In some cases, an ideal patient has a sleep indicated as medium. For activity, a patient’s behavior may be indicated as low, balanced, or high, which be based on a length of activity, calories burned, steps taken, heart rate, etc. of the patient. In some cases, an ideal patient has an activity indicated asbalanced. For routine, a patient’s behavior may be indicated as repetitive, balanced, or varied, which be based on a wake up time, a bed time, a time of activity, a meal time, etc. of the patient. In some cases, an ideal patient has a routine indicated as balanced In some cases, from the user interface 600F, a user can select mood and speech, sleep, activity, or routine, to view a more detailed presentation of behavior of the patient.[OHl] In some cases, the plurality of patient behavior, tracking, mood and speech, sleep, activity, and routine enable more accurate predictions. Any one of these four behaviors on their own may not be determinative of a condition, but the plurality of the four behaviors together may be informative. For example, if a user has a high value for activity for a particular day, it may be difficult to tell, based solely on activity level, whether the particular day was the patient’s day for working out or that the patient was having a manic episode. However, having a high activity for a day, coupled with a low sleep, a high mood and speech, and a low routine, may help clarify this day was an instance of a manic episode rather than a workout day.

[0112] FIGs. 6G-6J show example user interfaces 600G-600J of more detailed presentations of behavior of a patient. In some cases, a user may access any one of user interfaces 600G-600J by selecting the corresponding patient behavior in the dashboard summary of user interface 600F. The user may view more detailed information about each patient behavior of the user interfaces 600G-600J for a day (e.g., yesterday), a week (e.g., last week), a month (e.g., last 30 days), etc. Within each of the user interfaces 600G-600J, the user can view points (“MN points”) that represent the patient’s behavior over a period of time, where the Y-axis of the plot is a measure of the balance (or lack of balance) in the patient’s behavior.

[0113] FIG. 6K shows an example of a user interface 600K for managing notifications for a patient. As illustrated, the user (e.g., the patient) may be able to control one or more of inapplication notifications, email notifications, text or call notifications, etc. In-application notifications may include one or more of general notifications, reminders, indications reports are available, indications of emerging trends. Email notifications may provide insights including sending scheduled (e.g., daily, weekly, monthly, etc.) reports to the patient or to a trusted contact of the patient. Short message service (SMS) notifications may provide insights including sending, when emerging trends are important (e.g., concerning), notifications to the patient or to a trusted contact of the patient. The trusted contact may be someone selected by the patient or a healthcare provider of the patient. FIG. 6L shows an example of a user interface 600L for notifying a patient of an alert. As illustrated in the user interface 600L, a patient may be notified that their data (“MN Points” or “Mind Numbers”) are higher than usual, prompting the patient with the opportunity to generate a report to send to themselves, and, optionally, to the patient’s trusted contact.

[0114] FIG. 6M shows an example of a user interface 600M for presenting weekly insights. As illustrated, the weekly insights may provide data (e.g., high, medium, low, etc.) to indicate a patient’s behavior for mood and speech, sleep, activity, and routine for each day of the week. A goal for a patient may be to near the medium (e.g., balanced) level for each of the patient behaviors. Further, the user interface 600M may provide a measure of data completeness for the patient based on, for example, how much wearable data was provided or how much vocal data was provided. In some cases, the user interface 600M may provide a measure of medications taken as prescribed that receives input (e.g., from a user directly, from a smart medication dispenser, from a healthcare provider, etc.) an indication of the medications taken by the patient.

[0115] In some cases, the user interface 600M may provide an Achievements section to provide a written interpretation of the patient’s behavior data and highlight positive progress meant to motivate. In some cases, the user interface 600M may provide an Opportunities section to provide a written interpretation of the patient’s behavior data and highlight opportunities to improve the patient’s health or the data collection process. In some cases, the user interface 600M may provide an Insight for the Week section to provide a written interpretation of the patient’s behavior data and present tailored insights. These insights may be tailored to the patient in the sense that the insights are relevant to conditions, progress, behavior, etc. of the patient based on data gathered about the patient. In some cases, the insights may be presented as questions, chatbots, or links. For example, if a patient displays lack of sleep consistent with bipolar, a patient may be presented the question of “how important is sleep in bipolar disorder?” This question may be answered by a linked chatbot, or by a link to an information source (e.g., a video, an image, data, a webpage, etc.), such as, the International Bipolar Foundation.

[0116] FIG. 6N shows an example of a user interface 600N for presenting a weekly summary of a patient. The weekly summary of the user interface 600N may have certain similarities to the weekly insights of the user interface 600M of FIG. 6M. In some cases, the weekly summary may enable viewing trends for a week or 30 days. In some cases, the weekly summary of the user interface 600N (or the weekly insights of the user interface 600M) may be generated automatically on a weekly basis.Example Method for Assessing a Mental Condition

[0117] FIG. 7 shows an example of a method 700 for assessing a mental condition of a patient. At a high level, the method 700 includes: (A) obtaining, from a first source, first input data corresponding to a patient (block 705); (B) obtaining, from a second source, second input data corresponding to the patient (block 710); (C) analyzing the first input data and the second input data using one or more machine learning models to generate a plurality of patient scores for at least two mental conditions for the patient (block 715); (D) generating at least two patientscorecards (e.g., clinical rating scales) for the patient based at least in part on the plurality of patient scores, wherein each patient scorecard of the at least two patient scorecards corresponds to at least one of the at least two mental conditions, and wherein each patient scorecard of the at least two patient scorecards is a least partially complete (block 720); and (E) outputting or storing the at least two patient scorecards (block 725).

[0118] The one or more operations disclosed above with respect to the method 700 may be performed in any order. Further, at least one of the one or more operations disclosed above with respect to the method 700 may be repeated, e.g., iteratively.Examples of Machine Learning Techniques

[0119] As disclosed herein, in some cases, the systems, the methods, the computer-readable media, and the techniques disclosed herein may implement one or more machine learning techniques to assess a mental condition of a patient. In some cases, machine learning may generally involve identifying and recognizing patterns in existing data in order to facilitate making predictions for subsequent data. ML may include a ML model (which may include, for example, a ML algorithm). Machine learning, whether analytical or statistical in nature, may provide deductive or abductive inference based on real or simulated data. The ML model may be a trained model. ML techniques may comprise one or more supervised, semi-supervised, selfsupervised, or unsupervised ML techniques. For example, an ML model may be a trained model that is trained through supervised learning (e.g., various parameters are determined as weights or scaling factors). ML may comprise one or more of regression analysis, regularization, classification, dimensionality reduction, ensemble learning, meta learning, association rule learning, cluster analysis, anomaly detection, deep learning, or ultra-deep learning. ML may comprise: k-means, k-means clustering, k-nearest neighbors, learning vector quantization, linear regression, non-linear regression, least squares regression, partial least squares regression, logistic regression, stepwise regression, multivariate adaptive regression splines, ridge regression, principal component regression, least absolute shrinkage and selection operation (LASSO), least angle regression, canonical correlation analysis, factor analysis, independent component analysis, linear discriminant analysis, multidimensional scaling, non-negative matrix factorization, principal components analysis, principal coordinates analysis, projection pursuit, Sammon mapping, t-distributed stochastic neighbor embedding, AdaBoosting, boosting, gradient boosting, bootstrap aggregation, ensemble averaging, decision trees, conditional decision trees, boosted decision trees, gradient boosted decision trees, random forests, stacked generalization, Bayesian networks, Bayesian belief networks, naive Bayes, Gaussian naive Bayes, multinomial naive Bayes, hidden Markov models, hierarchical hidden Markov models, support vector machines, encoders, decoders, auto-encoders, stacked auto-encoders, perceptrons, multi-layerperceptrons, artificial neural networks, feedforward neural networks, convolutional neural networks, recurrent neural networks, long short-term memory, deep belief networks, deep Boltzmann machines, deep convolutional neural networks, deep recurrent neural networks, large language models, vision transformers, or generative adversarial networks.

[0120] Training the ML model may include, in some cases, selecting one or more untrained data models to train using a training data set. The selected untrained data models may include any type of untrained ML models for supervised, semi-supervised, self-supervised, or unsupervised machine learning. The selected untrained data models may be specified based upon input (e.g., user input) specifying relevant parameters to use as predicted variables or other variables to use as potential explanatory variables. For example, the selected untrained data models may be specified to generate an output (e.g., a prediction) based upon the input. Conditions for training the ML model from the selected untrained data models may likewise be selected, such as limits on the ML model complexity or limits on the ML model refinement past a certain point. The ML model may be trained (e.g., via a computer system such as a server) using the training data set. In some cases, a first subset of the training data set may be selected to train the ML model. The selected untrained data models may then be trained on the first subset of training data set using appropriate ML techniques, based upon the type of ML model selected and any conditions specified for training the ML model. In some cases, due to the processing power requirements of training the ML model, the selected untrained data models may be trained using additional computing resources (e.g., cloud computing resources). Such training may continue, in some cases, until at least one aspect of the ML model is validated and meets selection criteria to be used as a predictive model.

[0121] In some cases, one or more aspects of the ML model may be validated using a second subset of the training data set (e.g., distinct from the first subset of the training data set) to determine accuracy and robustness of the ML model. Such validation may include applying the ML model to the second subset of the training data set to make predictions derived from the second subset of the training data. The ML model may then be evaluated to determine whether performance is sufficient based upon the derived predictions. The sufficiency criteria applied to the ML model may vary depending upon the size of the training data set available for training, the performance of previous iterations of trained models, or user-specified performance requirements. If the ML model does not achieve sufficient performance, additional training may be performed. Additional training may include refinement of the ML model or retraining on a different first subset of the training dataset, after which the new ML model may again be validated and assessed. When the ML model has achieved sufficient performance, in some cases, the ML may be stored for present or future use. The ML model may be stored as sets ofparameter values or weights for analysis of further input (e.g., further relevant parameters to use as further predicted variables, further explanatory variables, further user interaction data, etc.), which may also include analysis logic or indications of model validity in some instances. In some cases, a plurality of ML models may be stored for generating predictions under different sets of input data conditions. In some embodiments, the ML model may be stored in a database (e.g., associated with a server).Examples of Neural Networks

[0122] The systems, the methods, the computer-readable media, and the techniques disclosed herein may implement one or more neural networks (NNs) in the evaluation of data corresponding to a subject, calculation of scores, or generation of scorecards (e.g., clinical rating scales). NNs are a subset of machine learning and are often at the core of many deep learning algorithms. Neural networks may comprise node layers, each which may comprise one or more of an input layer, one or more hidden layers, and an output layer. Each node of a neural network may connect to another node of the neural network. Each node of a neural network may have an associated weight and threshold. In some cases, if an output from any individual node of a neural network is above a specified threshold value, that node is activated, thereby sending data to the next layer of the neural network; otherwise, no data is passed along to the next layer of the neural network.

[0123] Convolutional neural networks are a type of neural network that may be, in some cases, implemented by the systems, the methods, the computer-readable media, and the techniques disclosed herein. CNNs are often used for classification and computer vision tasks. Prior to CNNs, manual, time-consuming feature extraction methods were used to identify objects in images. However, CNNs provide a more scalable approach to image classification and object recognition tasks, leveraging principles from linear algebra, specifically matrix multiplication, to identify patterns within an image. That said, CNNs can be computationally demanding, using graphical processing units (GPUs) to train models.

[0124] CNNs may be distinguished from other neural networks by their superior performance with image, speech, or audio signal inputs. CNNs may comprise three main types of layers: convolutional layers, pooling layers, and fully-connected (FC) layers. The convolutional layer may be the first layer of a CNN. While convolutional layers can be followed by additional convolutional layers or pooling layers, the fully-connected layer may be the final layer of the CNN.

[0125] When applied to computer vision tasks within images, with each layer, the CNN increases in its complexity, identifying greater portions of the image. Earlier layers of a CNN may focus on simple features of an image, such as colors and edges. As the image data progresses through thelayers of the CNN, the CNN starts to recognize larger elements or shapes of objects in the image until the CNN identifies the intended object.

[0126] The convolutional layer is a core building block of a CNN and may be where much of the computation of the CNN occurs. Convolution layers may use components including input data, a filter, and a feature map. Provided, for example, the input data comprises a color image (which e.g., includes a matrix of pixels in 3D), the input may have three dimensions — a height, width, and depth — which correspond to RGB in an image. CNNs may further comprise a feature detector (also known as a kernel or a filter), which moves across receptive fields of the image, checking if a feature is present. This process may be known as a convolution.

[0127] The feature detector may include a filter that is a two-dimensional array of weights, which represents part of an image. Filters of feature detectors may vary in size (e.g., 3x3 matrix), and the size may determine the size of the receptive field. The filter may be applied to an area of the image, and a dot product may be calculated between input pixels and the filter. The dot product may then be fed into an output array. Afterwards, the filter may shift by a stride, repeating the process until the filter has swept across the entire image. The final output from the series of dot products from the input and the filter may be known as a feature map, activation map, or a convolved feature. After each convolution operation, a CNN may apply a Rectified Linear Unit (ReLU) transformation to the feature map, introducing nonlinearity to the CNN.

[0128] In some cases, another convolution layer can follow the initial convolution layer of the CNN. For example, the structure of the CNN can become hierarchical as the later layers can see the pixels within the receptive fields of prior layers. As an example, assume a CNN used to determine if an image contains a face. Each individual part of the face (e.g., eyes, mouth, nose, etc.) makes up a lower-level pattern in the CNN, and the combination of the parts represents a higher-level pattern, creating a feature hierarchy within the CNN.

[0129] The pooling layers, also known as downsampling, are further layers of a CNN. Pooling layers may conduct dimensionality reduction, reducing the number of parameters in the input (e.g., image, video, audio, etc.). Similar to the convolutional layer, the pooling layer sweeps a filter across the entire input, but, unlike the convolution layers, the filters of the pooling layers do not have any weights. Instead, the filters of the pooling layers apply an aggregation function to values within the receptive field, populating the output array. There are two main types of pooling: max pooling and average pooling. Max pooling may comprise moving the filter across the input to select the pixel with the maximum value to send to the output array. Average pooling may comprise moving the filter across the input to calculate the average value within the receptive field to send to the output array. While a lot of information is lost in the pooling layer,the pooling layer also has a number of benefits to the CNN. For example, pooling layers may help to reduce complexity CNN, improve efficiency, and limit risk of overfitting of the CNN.

[0130] Fully-connected layers are the final layer of a CNN. As previously disclosed, pixel values of an input image are not directly connected to output layers in partially connected layers. However, in the fully-connected layer, each node in the output layer connects directly to a node in the previous layer. The FC layer performs the task of classification based on the features extracted through the previous layers and their different filters. While convolutional layers and pooling layers tend to use ReLu functions, FC layers may leverage a softmax activation function to classify inputs appropriately, producing a probability from 0 to 1.

[0131] In some cases, the systems, the methods, the computer-readable media, and the techniques disclosed herein may implement deep neural networks. DNNs are neural networks with numerous layers. For example, some neural networks may be classified as DNNs provided the neural network has 3 or more layers, 4 or more layers, 5 or more layers, 6 or more layers, 7 or more layers, 8 or more layers, 9 or more layers, 10 or more layers, etc. In some cases, these layers may include an input and output layer. Accordingly, in some cases, a CNN may be considered an instance of a DNN.

[0132] Hyperparameters may be used in CNN architectures to indicate information such as a number of kernels in a convolutional layer, a size of kernels in a convolutional layer, a size of stride, a size of kernels in a pooling layer, etc. Hyperparameters may further be used in a CNN to indicate information such as how the CNN is trained; for example, learning rate, weights, biases, momentum, decay, etc. In some cases, hyperparameters for a CNN may be set prior to training the CNN.Examples of Language Models

[0133] The systems, the methods, the computer-readable media, and the techniques disclosed herein may implement a language model in the evaluation of data corresponding to a subject, calculation of scores, or generation of scorecards (e.g., clinical rating scales). At a high level, a language model is a probabilistic model of a natural language that can generate probabilities of a series of words, based on text corpora in one or multiple languages it was trained on. Example applications of language models may include speech recognition, machine translation, natural language generation, optical character recognition, handwriting recognition, grammar induction, information retrieval, malware analysis, content creation, search structuring, machine assistance, customer support, cyber defense, code development, transcription generation, market research, keyword research, sales automation, sentiment analysis, risk analysis, copywriting, autocompletion, shell command generation, regex generation, SQL generation, code review,database optimization, DevOps automation, frontend or website generation, product documentation generation, personalized tutoring, decision making, the methods and the systems.

[0134] In some cases, language models may implement statistical or probabilistic techniques. Statistical language models may use statistical patterns in data to make predictions about the likelihood of specific sequences of words. For example, one technique for building a probabilistic language model is to calculate n-gram probabilities (e.g., a sequence of words, where n is a number greater than zero). In some cases, a statistical language model may calculate likelihood of different n-grams (word combinations) in a text. One example technique for calculating likelihood may include counting the number of times each word combination appears to obtain a first value. Then, the first value may be divided by the number of times the previous word appears. This technique is based on Markov assumption (e.g., probability of a word combination (the future) depends only on the previous word (the present) and not the words that came before it (the past)). Types of n-gram models may include unigrams (e.g., evaluate each word independently), bigrams (e.g., consider probability of a word given the previous word), trigrams (e.g., consider probability of a word given the previous two words), etc. While n-grams may be efficient, n-gram language models may be challenged in considering long-term context of words in a sequence. Other statistical language models may include, for example, exponential language models (e.g., maximum entropy language models), skip-gram language model the methods and the systems etc.

[0135] In some cases, language models may implement machine learning techniques. For example, language models may be neural language models. Advantageously, neural language models may be able to capture context better than statistical models. Also, neural language models may handle more complex language structures and longer dependencies between words than statistical language models.

[0136] In some cases, neural language models may implement deep neural networks (DNN). DNN language models use DNNs to predict the likelihood of a sequence of words. DNN language models may be trained on a large corpus of text data and are capable of learning the underlying structure of the language. DNN language models can handle large vocabularies and deal with rare or unknown words by using distributed representations. One simple example of a DNN language model may comprise an input layer, a hidden layer, another hidden layer, and an output layer.

[0137] In some cases, DNN language models may use recurrent neural networks (RNNs) or transformer networks. A RNN language model may “remember” previous outputs when receiving next inputs. This is in contrast to other neural networks in which inputs and outputs are independent of each other. RNN language models are particularly useful in predicting the nextword in a sentence, as RNN language models consider previous words in an input (e.g., sentence). One feature of RNN language models is a hidden state vector that remembers information about a sequence. This “memory” allows RNN language models to remember information that has been previously calculated and use this information to make future predictions. The hidden state is maintained by a hidden layer in the RNN language model.

[0138] In some cases, RNN language models can be computationally expensive for very long input sequences. This means that the RNN language model’s ability to make accurate predictions based on the information from the initial words of the sentence decreases. In some cases, this the “vanishing gradients” problem may be addressed using, for example, a Long Short-term Memory (LSTM) architecture that may be the same as or similar to the LSTM disclosed herein. In some cases, RNN language models generate continuous representations or embeddings of words. Such continuous space embeddings may help to alleviate the “curse of dimensionality,” which is the consequence of the number of possible sequences of words increasing exponentially with the size of the vocabulary, furtherly causing a data sparsity problem. In some cases, RNN language models reduce this challenge by representing words as non-linear combinations of weights in a neural net.

[0139] A large language model (LLM) is a language model characterized by its large size. LLMs may be based on machine learning techniques. In some cases, LLMs may be implemented as data structures, programs, algorithms, architectures, or the like, that may be designed to interpret natural language. LLMs may generate text responses in response to text based prompts. LLMs may be neural networks trained on large collections of natural language source documents. Accordingly, in some cases, LLMs may be trained to generate predictive responses based on provided prompts. LLM prompts may include context information, examples, or the like, that may enable LLMs to generate responses directed to specific queries.

[0140] LLMs may be very big and may run on specialist hardware. LLMs may be expensive to train and run and use massive amounts of data. LLMs may accomplish such a large size via artificial intelligence accelerators, which are able to process vast amounts of text data, mostly scraped from the Internet. LLMs may use neural networks with, for example, millions, billions, trillions the methods and the systems, etc. weights. In some cases, LLMs may be (pre-)trained using self-supervised learning and semi -supervised learning. In some cases, LLMs may be (predrained using unsupervised learning. In some cases, LLMs may use a transformer architecture for faster training. In some cases, LLMs may use alternative architectures such as mixture of experts (MoE).

[0141] In some cases, LLMs obtain an input text and repeatedly predict the next token or word. In some cases, fine tuning may be used to adapt LLMs to accomplish specific tasks. LargerLLMs sized models may be prompt-engineered to achieve similar results. LLMs may acquire embodied knowledge about syntax, semantics, and “ontology” inherent in human language corpora, but also inaccuracies and biases present in the corpora.

[0142] In some cases, a LLM may embed a plurality of large foundational models (LFMs), including both closed-source LFMs and open-source LFMs. For example, LFMs may include Azure OpenAI (complete and chat APIs for GPT-3, GPT-3.5, and GPT-4, used in ChatGPT), OpenAI (complete and chat APIs for GPT-3, GPT-3.5 and GPT-4), Google Vertex Al (e.g., PaLM, PaLM’2), Meta's LLaMa family of models, as well as BLOOM, Ernie 3.0 Titan, and ’Anthropic's Claude 2. For example, open-source LFMs may include LLaMa 2, FLAN-T5, OpenAssistant, RoBERT a, MiniLM, and MPNet.Examples of Computing Systems

[0143] Referring to FIG. 8, a block diagram is shown depicting an example machine that includes a computer system 800 (e.g., a processing or computing system) within which a set of instructions can execute for causing a device to perform or execute any one of the systems, the methods, the computer-readable media, and the techniques for static code scheduling of the present disclosure, such as in the evaluation of data corresponding to a subject, calculation of scores, or generation of scorecards (e.g., clinical rating scales). In some cases, one or more components of FIG. 2 may be included in or may be implemented by one or more components of the computer system 800. The components in FIG. 8 are examples and do not limit the scope of use or functionality of any hardware, software, embedded logic component, or a combination of two or more such components with particular implementations.

[0144] Computer system 800 may include one or more processors 801, a memory 803, and a storage 808 that communicate with each other, and with other components, via a bus 840. The bus 840 may also link a display 832, one or more input devices 833 (which may, for example, include a keypad, a keyboard, a mouse, a stylus, etc.), one or more output devices 834, one or more storage devices 835, and various tangible storage media 836. All of these elements may interface directly or via one or more interfaces or adaptors to the bus 840. For instance, the various tangible storage media 836 can interface with the bus 840 via storage medium interface 826. Computer system 800 may have any suitable physical form, including but not limited to one or more integrated circuits (Ics), printed circuit boards (PCBs), mobile handheld devices (such as mobile telephones or PDAs), laptop or notebook computers, distributed computer systems, computing grids, or servers.

[0145] Computer system 800 includes one or more processors 807 (e.g., central processing units (CPUs), general purpose graphics processing units (GPGPUs), or quantum processing units (QPUs)) that carry out functions. Processors 801 optionally contains a cache memory unit 802for temporary local storage of instructions, data, or computer addresses. Processors 801 are configured to assist in execution of computer readable instructions. Computer system 800 may provide functionality for the components depicted in FIG. 8 as a result of the processors 801 executing non-transitory, processor-executable instructions embodied in one or more tangible computer-readable storage media, such as memory 803, storage 808, storage devices 835, or storage medium 836. The computer-readable media may store software that implements particular operations, and processors 801 may execute the software. Memory 803 may read the software from one or more other computer-readable media (such as mass storage devices 835, 836) or from one or more other sources through a suitable interface, such as network interface 820. The software may cause processors 801 to carry out one or more processes or one or more operations of one or more processes described or illustrated herein. Carrying out such processes or operations may include defining data structures stored in memory 803 and modifying the data structures as directed by the software.

[0146] The memory 803 may include various components (e.g., machine readable media) including, but not limited to, a random access memory component (e.g., RAM 804) (e.g., static RAM (SRAM), dynamic RAM (DRAM), ferroelectric random access memory (FRAM), phasechange random access memory (PRAM), etc.), a read-only memory component (e.g., ROM 805), and any combinations thereof. ROM 805 may act to communicate data and instructions unidirectionally to processors 801, and RAM 804 may act to communicate data and instructions bidirectionally with processors 801. ROM 805 and RAM 804 may include any suitable tangible computer-readable media described below. In one example, a basic input / output system 806 (BIOS), including basic routines that help to transfer information between elements within computer system 800, such as during start-up, may be stored in the memory 803.

[0147] Fixed storage 808 is connected bidirectionally to processors 801, optionally through storage control unit 807. Fixed storage 808 provides additional data storage capacity and may also include any suitable tangible computer-readable media. Storage 808 may be used to store operating system 809, executables 810, data 811, applications 812 (application programs), and the like. Storage 808 can also include an optical disk drive, a solid-state memory device (e.g., flash-based systems), or a combination of any of the above. Information in storage 808 may, in appropriate cases, be incorporated as virtual memory in memory 803.

[0148] In one example, storage devices 835 may be removably interfaced with computer system 800 (e.g., via an external port connector (not shown)) via a storage device interface 825. Particularly, storage devices 835 and an associated machine-readable medium may provide nonvolatile or volatile storage of machine-readable instructions, data structures, program modules, or other data for the computer system 800. In one example, software may reside, completely orpartially, within a machine-readable medium on storage devices 835. In another example, software may reside, completely or partially, within processors 801.

[0149] Bus 840 connects a wide variety of subsystems. Herein, reference to a bus may encompass one or more digital signal lines serving a common function, where appropriate. Bus 840 may be any of several types of bus structures including, but not limited to, a memory bus, a memory controller, a peripheral bus, a local bus, and any combinations thereof, using any of a variety of bus architectures. As an example and not by way of limitation, such architectures include an Industry Standard Architecture (ISA) bus, an Enhanced ISA (EISA) bus, a Micro Channel Architecture (MCA) bus, a Video Electronics Standards Association local bus (VLB), a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, an Accelerated Graphics Port (AGP) bus, HyperTransport (HTX) bus, serial advanced technology attachment (SATA) bus, and any combinations thereof.

[0150] Computer system 800 may also include an input device 833. In one example, a user of computer system 800 may enter commands or other information into computer system 800 via input devices 833. Examples of an input devices 833 include, but are not limited to, an alphanumeric input device (e.g., a keyboard), a pointing device (e.g., a mouse or touchpad), a touchpad, a touch screen, a multi-touch screen, a joystick, a stylus, a gamepad, an audio input device (e.g., a microphone, a voice response system, etc.), an optical scanner, a video or still image capture device (e.g., a camera), and any combinations thereof. In some cases, the input device is a Kinect, Leap Motion, or the like. Input devices 833 may be interfaced to bus 840 via any of a variety of input interfaces 823 (e.g., input interface 823) including, but not limited to, serial, parallel, game port, USB, FIREWIRE, THUNDERBOLT, or any combination of the above.

[0151] In some cases, when computer system 800 is connected to network 830, computer system 800 may communicate with other devices, specifically mobile devices and enterprise systems, distributed computing systems, cloud storage systems, cloud computing systems, and the like, connected to network 830. Communications to and from computer system 800 may be sent through network interface 820. For example, network interface 820 may receive incoming communications (such as requests or responses from other devices) in the form of one or more packets (such as Internet Protocol (IP) packets) from network 830, and computer system 800 may store the incoming communications in memory 803 for processing. Computer system 800 may similarly store outgoing communications (such as requests or responses to other devices) in the form of one or more packets in memory 803 and communicated to network 830 from network interface 820. Processors 801 may access these communication packets stored in memory 803 for processing.

[0152] Examples of the network interface 820 include, but are not limited to, a network interface card, a modem, and any combination thereof. Examples of a network 830 or network segment 830 include, but are not limited to, a distributed computing system, a cloud computing system, a wide area network (WAN) (e.g., the Internet, an enterprise network), a local area network (LAN) (e.g., a network associated with an office, a building, a campus or other relatively small geographic space), a telephone network, a direct connection between two computing devices, a peer-to-peer network, and any combinations thereof. A network, such as network 830, may employ a wired or a wireless mode of communication. In general, any network topology may be used.

[0153] Information and data can be displayed through a display 832. Examples of a display 832 include, but are not limited to, a cathode ray tube (CRT), a liquid crystal display (LCD), a thin film transistor liquid crystal display (TFT-LCD), an organic liquid crystal display (OLED) such as a passive-matrix OLED (PMOLED) or active-matrix OLED (AMOLED) display, a plasma display, and any combinations thereof. The display 832 can interface to the processors 801, memory 803, and fixed storage 808, as well as other devices, such as input devices 833, via the bus 840. The display 832 is linked to the bus 840 via a video interface 822, and transport of data between the display 832 and the bus 840 can be controlled via the graphics control 821. In some cases, the display is a video projector. In some cases, the display is a head-mounted display (HMD) such as a VR headset. In further cases, suitable VR headsets include, by way of nonlimiting examples, HTC Vive, Oculus Rift, Samsung Gear VR, Microsoft HoloLens, Razer OSVR, FOVE VR, Zeiss VR One, Avegant Glyph, Freefly VR headset, and the like. In still further cases, the display is a combination of devices such as those disclosed herein.

[0154] In addition to a display 832, computer system 800 may include one or more other peripheral output devices 834 including, but not limited to, an audio speaker, a printer, a storage device, and any combinations thereof. Such peripheral output devices may be connected to the bus 840 via an output interface 824. Examples of an output interface 824 include, but are not limited to, a serial port, a parallel connection, a USB port, a FIREWIRE port, a THUNDERBOLT port, and any combinations thereof.

[0155] In addition or as an alternative, the computer system 800 may provide functionality as a result of logic hardwired or otherwise embodied in a circuit, which may operate in place of or together with software to execute one or more processes or one or more operations of one or more processes described or illustrated herein. Reference to software in this disclosure may encompass logic, and reference to logic may encompass software. Moreover, reference to a computer-readable medium may encompass a circuit (such as an IC) storing software forexecution, a circuit embodying logic for execution, or both, where appropriate. The present disclosure encompasses any suitable combination of hardware, software, or both.

[0156] Various illustrative logical blocks, modules, circuits, and algorithm operations described in connection with the examples disclosed herein may be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and operations have been described above generally in terms of their functionality.

[0157] The various illustrative logical blocks, modules, and circuits described in connection with the examples disclosed herein may be implemented or performed with a general purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions. A general purpose processor may be a microprocessor, but in the alternative, the processor may be any processor, controller, microcontroller, or state machine. A processor may also be implemented as a combination of computing devices, e.g., a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration.

[0158] The operations of a method, a technique, or an algorithm described in connection with the examples disclosed herein may be embodied directly in hardware, in a software module executed by one or more processors, or in a combination of the two. A software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An example storage medium may be coupled to the processor such the processor can read information from, and write information to, the storage medium. In the alternative, the storage medium may be integral to the processor. The processor and the storage medium may reside in an ASIC. The ASIC may reside in a user terminal. In the alternative, the processor and the storage medium may reside as discrete components in a user terminal.

[0159] In accordance with the description herein, suitable computing devices include, by way of non-limiting examples, server computers, desktop computers, laptop computers, notebook computers, sub-notebook computers, netbook computers, netpad computers, set-top computers, media streaming devices, handheld computers, Internet appliances, mobile smartphones, tablet computers, personal digital assistants, video game consoles, and vehicles. Select televisions, video players, and digital music players with optional computer network connectivity may be suitable for use in the system. Suitable tablet computers, in various cases, include those with booklet, slate, and convertible configurations.

[0160] In some cases, the computing device includes an operating system configured to perform executable instructions. The operating system is, for example, software, including programs and data, which manages the device’s hardware and provides services for execution of applications. Suitable server operating systems may include, by way of non-limiting examples, FreeBSD, OpenBSD, NetBSD®, Linux, Apple® Mac OS X Server®, Oracle® Solaris®, Windows Server®, and Novell® NetWare®. Suitable personal computer operating systems may include, by way of non-limiting examples, Microsoft® Windows®, Apple® Mac OS X®, UNIX®, and UNIX-like operating systems such as GNU / Linux®. In some cases, the operating system is provided by cloud computing. Suitable mobile smartphone operating systems may include, by way of nonlimiting examples, Nokia® Symbian® OS, Apple® iOS®, Research In Motion® BlackBerry OS®, Google® Android®, Microsoft® Windows Phone® OS, Microsoft® Windows Mobile® OS, Linux®, and Palm® WebOS®.

[0161] In some cases, the systems, the methods, the computer-readable media, and the techniques the methods, the computer-readable media, and the techniques disclosed herein include one or more non-transitory computer readable storage media encoded with a program including instructions executable by the operating system of an optionally networked computing device. In further cases, a computer readable storage medium is a tangible component of a computing device. In still further cases, a computer readable storage medium is optionally removable from a computing device. In some cases, a computer readable storage medium includes, by way of non-limiting examples, CD-ROMs, DVDs, flash memory devices, solid state memory, magnetic disk drives, magnetic tape drives, optical disk drives, distributed computing systems including cloud computing systems and services, and the like. In some cases, the program and instructions are permanently, substantially permanently, semi -permanently, or non- transitorily encoded on the media.

[0162] In some cases, the systems, the methods, the computer-readable media, and the techniques the methods, the computer-readable media, and the techniques disclosed herein include at least one computer program, or use of the same. A computer program includes a sequence of instructions, executable by one or more processors of the computing device’s CPU, written to perform a specified task. Computer readable instructions may be implemented as program modules, such as functions, objects, APIs, computing data structures, and the like, that perform particular tasks or implement particular abstract data types. A computer program may be written in various versions of various languages.

[0163] The functionality of the computer readable instructions may be combined or distributed in various ways across various environments. In some cases, a computer program comprises one sequence of instructions. In some cases, a computer program comprises a plurality of sequencesof instructions. In some cases, a computer program is provided from one location. In some cases, a computer program is provided from a plurality of locations. In some cases, a computer program includes one or more software modules. In some cases, a computer program includes, in part or in whole, one or more web applications, one or more mobile applications, one or more standalone applications, one or more web browser plug-ins, extensions, add-ins, or add-ons, or combinations thereof.

[0164] In some cases, a computer program includes a web application. A web application, in various cases, may utilize one or more software frameworks and one or more database systems. In some cases, a web application is created upon a software framework such as Microsoft® .NET or Ruby on Rails (RoR). In some cases, a web application utilizes one or more database systems including, by way of non -limiting examples, relational, non-relational, object oriented, associative, XML, and document oriented database systems. In further cases, suitable relational database systems include, by way of non-limiting examples, Microsoft® SQL Server, mySQL™, and Oracle®. In some cases, a webapplication may be built using Angular on TypeScript. A web application, in some cases, may be written in one or more versions of one or more languages. A web application may be written in one or more markup languages, presentation definition languages, client-side scripting languages, server-side coding languages, database query languages, or combinations thereof. In some cases, a web application is written to some extent in a markup language such as Hypertext Markup Language (HTML), Extensible Hypertext Markup Language (XHTML), or extensible Markup Language (XML). In some cases, a web application is written to some extent in a presentation definition language such as Cascading Style Sheets (CSS). In some cases, a web application is written to some extent in a client-side scripting language such as Asynchronous JavaScript and XML (AJAX), Flash® ActionScript, JavaScript, or Silverlight®. In some cases, a web application is written to some extent in a server-side coding language such as Active Server Pages (ASP), ColdFusion®, Perl, Java™, JavaServer Pages (JSP), Hypertext Preprocessor (PHP), Python™, Ruby, Tel, Smalltalk, WebDNA®, or Groovy. In some cases, a web application is written to some extent in a database query language such as Structured Query Language (SQL). In some cases, a web application integrates enterprise server products such as IBM® Lotus Domino®. In some cases, a web application includes a media player element. In some cases, a media player element utilizes one or more of many suitable multimedia technologies including, by way of non-limiting examples, Adobe® Flash®, HTML 5, Apple® QuickTime®, Microsoft® Silverlight®, Java™, and Unity®.

[0165] In some cases, a computer program includes a mobile application provided to a mobile computing device. In some cases, the mobile application is provided to a mobile computingdevice at the time it is manufactured. In other cases, the mobile application is provided to a mobile computing device via the computer network disclosed herein.

[0166] In view of the disclosure provided herein, a mobile application may be created using hardware, languages, and development environments known to the art. In some cases, mobile applications are written in several languages. Suitable programming languages may include, by way of non-limiting examples, C, C++, C#, Objective-C, Java™, JavaScript, Pascal, Object Pascal, Python™, Ruby, VB.NET, WML, and XHTML / HTML with or without CSS, or combinations thereof.

[0167] Suitable mobile application development environments are available from several sources. Commercially available development environments include, by way of non-limiting examples, AirplaySDK, alcheMo, Appcelerator®, Celsius, Bedrock, Flash Lite, .NET Compact Framework, Rhomobile, Swift, Kotlin, and WorkLight Mobile Platform. Other development environments are available without cost including, by way of non-limiting examples, Lazarus, MobiFlex, MoSync, and PhoneGap. Also, mobile device manufacturers distribute software developer kits including, by way of non-limiting examples, iPhone and iPad (iOS) SDK, Android™ SDK, BlackBerry® SDK, BREW SDK, Palm® OS SDK, Symbian SDK, webOS SDK, and Windows® Mobile SDK.

[0168] Several commercial forums may be available for distribution of mobile applications including, by way of non-limiting examples, Apple® App Store, Google® Play, Chrome WebStore, BlackBerry® App World, App Store for Palm devices, App Catalog for webOS, Windows® Marketplace for Mobile, Ovi Store for Nokia® devices, and Samsung® Apps.

[0169] In some cases, a computer program includes a standalone application, which is a program that is run as an independent computer process, not an add-on to an existing process, e.g., not a plug-in. Standalone applications may be compiled. A compiler may be a computer programs that transforms source code written in a programming language into binary object code such as assembly language or machine code. Suitable compiled programming languages include, by way of non-limiting examples, C, C++, Objective-C, COBOL, Delphi, Eiffel, Java™, Lisp, Python™, Visual Basic, and VB .NET, or combinations thereof. Compilation is often performed, at least in part, to create an executable program. In some cases, a computer program includes one or more executable complied applications.

[0170] In some cases, the computer program includes a web browser plug-in (e.g., extension, etc.). In computing, a plug-in is one or more software components that add specific functionality to a larger software application. Makers of software applications support plug-ins to enable third- party developers to create abilities which extend an application, to support easily adding new features, and to reduce the size of an application. When supported, plug-ins enable customizingthe functionality of a software application. For example, plug-ins are commonly used in web browsers to play video, generate interactivity, scan for viruses, and display particular file types. Web browser plug-ins may include Adobe® Flash® Player, Microsoft® Silverlight®, and Apple® QuickTime®. In some cases, the toolbar comprises one or more web browser extensions, add-ins, or add-ons. In some cases, the toolbar comprises one or more explorer bars, tool bands, or desk bands.

[0171] Several plug-in frameworks may be available that enable development of plug-ins in various programming languages, including, by way of non-limiting examples, C++, Delphi, Java™, PHP, Python™, and VB .NET, or combinations thereof.

[0172] Web browsers (also called Internet browsers) are software applications, designed for use with network-connected computing devices, for retrieving, presenting, and traversing information resources on the World Wide Web. Suitable web browsers include, by way of non-limiting examples, Microsoft® Internet Explorer®, Mozilla® Firefox®, Google® Chrome, Apple® Safari®, Opera Software® Opera®, and KDE Konqueror. In some cases, the web browser is a mobile web browser. Mobile web browsers (also called microbrowsers, mini-browsers, and wireless browsers) are designed for use on mobile computing devices including, by way of non-limiting examples, handheld computers, tablet computers, netbook computers, subnotebook computers, smartphones, music players, personal digital assistants (PDAs), and handheld video game systems. Suitable mobile web browsers include, by way of non-limiting examples, Google® Android® browser, RIM BlackBerry® Browser, Apple® Safari®, Palm® Blazer, Palm® WebOS® Browser, Mozilla® Firefox® for mobile, Microsoft® Internet Explorer® Mobile, Amazon® Kindle® Basic Web, Nokia® Browser, Opera Software® Opera® Mobile, and Sony® PSP™ browser.

[0173] In some cases, the systems, the methods, the computer-readable media, and the techniques disclosed herein include software, server, or database modules, or use of the same. Software modules may be created by techniques using machines, software, and languages. The software modules disclosed herein are implemented in a multitude of ways. In some cases, a software module comprises a file, a section of code, a programming object, a programming structure, a distributed computing resource, a cloud computing resource, or combinations thereof. In some cases, a software module comprises a plurality of files, a plurality of sections of code, a plurality of programming objects, a plurality of programming structures, a plurality of distributed computing resources, a plurality of cloud computing resources, or combinations thereof. In some cases, the one or more software modules comprise, by way of non-limiting examples, a web application, a mobile application, a standalone application, and a distributed or cloud computing application. In some cases, software modules are in one computer program or application. Insome cases, software modules are in more than one computer program or application. In some cases, software modules are hosted on one machine. In some cases, software modules are hosted on more than one machine. In some cases, software modules are hosted on a distributed computing platform such as a cloud computing platform. In some cases, software modules are hosted on one or more machines in one location. In some cases, software modules are hosted on one or more machines in more than one location.

[0174] In some cases, the systems, the methods, the computer-readable media, and the techniques disclosed herein include one or more databases, or use of the same. In some cases, various databases may be suitable for storage and retrieval of various types of data (e.g., encrypted, unencrypted, etc.) one or more of which may be historical, present, or future data or information. In some cases, suitable databases include, by way of non-limiting examples, relational databases, non-relational databases, object oriented databases, object databases, entityrelationship model databases, associative databases, XML databases, document oriented databases, and graph databases. Further non-limiting examples include SQL, PostgreSQL, MySQL, Oracle, DB2, Sybase, and MongoDB. In some cases, a database is Internet-based. In further cases, a database is web-based. In still further cases, a database is cloud computing-based. In a particular case, a database is a distributed database. In other cases, a database is based on one or more local computer storage devices.Examples of Patient Scorecards

[0175] Examples of scorecards (e.g., clinical rating scales) are shown in FIGs. 9A-9C, where FIG. 9 A shows a scorecard for the Young Mania Rating Scale, FIG. 9B shows a scorecard for the Brief Psychiatric Rating Scale, and FIG. 9C shows a scorecard for the Patient Health Questionnaire-9. Note that while the scorecards of FIGs. 9A-9C are illustrated as unfilled, in practice, the systems, the methods, the computer-readable media, and the techniques disclosed herein may be used to automatically fill out at least a portion of a scorecard. For example, the systems, the methods, the computer-readable media, and the techniques may automatically fill out multiple scorecards for a single patient, thereby enabling evaluation for multiple mental conditions simultaneously. Unlike healthcare providers who are typically trained in only evaluating a patient for a single or limited range of mental conditions, advantageously, the systems, the methods, the computer-readable media, and the techniques disclosed herein may automatically evaluate a patient for a wide range of mental conditions via filling out multiple scorecards, thereby streamlining and improving patient evaluation processes.

[0176] FIGs. 10A-10B show an example of a generated scorecard (e.g., clinical rating scale) for the Young Mania Rating Scale (YMRS). The scorecard of FIGs. 10A-10B were generated using the systems, the methods, the computer-readable media, and the techniques disclosed herein.

[0177] FIG. 10A shows a bar chart for YMRS Score Components, plotting (in descending vertical order): “YMRS Total” having a score of 24, “Top Variable 1 : 6 Speech (Rate and Amount)” having a score of 6, “Top Variable 2: 8 Thought Content” having a score of 4, and “Top Variable 3: 1 Elevated Mood” having a score of 2. The top three variables of the bar chart are reflected in the “Top eYMRS Scores” section above the bar chart. Higher values for the variables indicate more severe symptom scores. Accordingly, “Top Variable 1 : 6 Speech (Rate and Amount)” is the most severe. “Top Variable 2: 8 Thought Content” is also severe, but less severe than “Top Variable 1 : 6 Speech (Rate and Amount).” Finally, “Top Variable 3: 1 Elevated Mood” is moderate.

[0178] FIG. 10B shows more detail of the data presented in FIG. 10 A. FIG. 10B shows ten variables analyzed in the YMRS: elevated mood, increased motor activity / energy, sexual interest, sleep, irritability, speech (rate and amount), language / thought disorder, thought content, disruptive / aggressive behavior, and insight. For each variable, the systems, the methods, the computer-readable media, and the techniques disclosed herein provide a calculated level, scored from 0 (not detected) to 6 (most severe). Further, for each variable, comments are also provided to present additional insight along with the level. These comments are generated by machine learning models based at least in part on patient speech analysis. These comments may be generated by the machine learning model further based at least in part on health data of the patient (e.g., from wearable health sensors). The total score “eYMRS total” of FIG. 10A may be the sum of each of the levels for the ten variables. As illustrated in FIG. 10A, a total score of 24 may indicate the patient is potentially unwell.

[0179] Similar to the YMRS scorecard of FIGs. 10A-10B, FIGs. 10C-10F show an example of a generated scorecard (e.g., clinical rating scale) for Positive and Negative Syndrome Scale (PANSS). Similarly, the generated PANSS scorecard includes a first page having summaries of top scores and bar charts to visually plot PANSS Score Components. Also similarly, the subsequent pages of the generated PANSS scorecard includes more detailed data including level scores and generated comments for each of 30 different variables (e.g., delusions, conceptual disorganization, hallucinatory behavior, excitement, grandiosity, etc.).

[0180] Advantageously, the YMRS scorecard of FIGs. 10A-10B and the PANSS scorecard of FIGs. 10C-10F may be generated by the systems, the methods, the computer-readable media, and the techniques disclosed herein may be formatted to be similar to (e.g., at least approximately the same as) scorecards healthcare providers generate. This may help make the scorecards more readable to healthcare providers accustomed to reading and preparing scorecards in certain formats. Moreover, the YMRS scorecard of FIGs. 10A-10B and the PANSS scorecard of FIGs. 10C-10F may be prepared by the systems, the methods, the computer-readable media, and thetechniques disclosed herein simultaneously and independently from one another, solving the specific challenge of confirmation bias. Therefore, a healthcare provider can view the two scorecards to help make a more informed and less biased diagnosis.

[0181] In some cases, the systems, the methods, the computer-readable media, and the techniques disclosed herein may generate 1 scorecard (e.g., clinical rating scale) to 50 independent scorecards, each scorecard corresponding to a different condition. In some cases, the systems, the methods, the computer-readable media, and the techniques disclosed herein may generate 1 scorecard to 2 independent scorecards, 1 scorecard to 3 independent scorecards, 1 scorecard to 4 independent scorecards, 1 scorecard to 5 independent scorecards, 1 scorecard to 6 independent scorecards, 1 scorecard to 7 independent scorecards, 1 scorecard to 8 independent scorecards, 1 scorecard to 9 independent scorecards, 1 scorecard to about 10 independent scorecards, 1 scorecard to about 20 independent scorecards, 1 scorecard to about 50 independent scorecards, 2 independent scorecards to 3 independent scorecards, 2 independent scorecards to 4 independent scorecards, 2 independent scorecards to 5 independent scorecards, 2 independent scorecards to 6 independent scorecards, 2 independent scorecards to 7 independent scorecards, 2 independent scorecards to 8 independent scorecards, 2 independent scorecards to 9 independent scorecards, 2 independent scorecards to about 10 independent scorecards, 2 independent scorecards to about 20 independent scorecards, 2 independent scorecards to about 50 independent scorecards, 3 independent scorecards to 4 independent scorecards, 3 independent scorecards to 5 independent scorecards, 3 independent scorecards to 6 independent scorecards, 3 independent scorecards to 7 independent scorecards, 3 independent scorecards to 8 independent scorecards, 3 independent scorecards to 9 independent scorecards, 3 independent scorecards to about 10 independent scorecards, 3 independent scorecards to about 20 independent scorecards, 3 independent scorecards to about 50 independent scorecards, 4 independent scorecards to 5 independent scorecards, 4 independent scorecards to 6 independent scorecards, 4 independent scorecards to 7 independent scorecards, 4 independent scorecards to 8 independent scorecards, 4 independent scorecards to 9 independent scorecards, 4 independent scorecards to about 10 independent scorecards, 4 independent scorecards to about 20 independent scorecards, 4 independent scorecards to about 50 independent scorecards, 5 independent scorecards to 6 independent scorecards, 5 independent scorecards to 7 independent scorecards, 5 independent scorecards to 8 independent scorecards, 5 independent scorecards to 9 independent scorecards, 5 independent scorecards to about 10 independent scorecards, 5 independent scorecards to about 20 independent scorecards, 5 independent scorecards to about 50 independent scorecards, 6 independent scorecards to 7 independent scorecards, 6 independent scorecards to 8 independent scorecards, 6 independent scorecards to 9 independent scorecards, 6 independent scorecards toabout 10 independent scorecards, 6 independent scorecards to about 20 independent scorecards, 6 independent scorecards to about 50 independent scorecards, 7 independent scorecards to 8 independent scorecards, 7 independent scorecards to 9 independent scorecards, 7 independent scorecards to about 10 independent scorecards, 7 independent scorecards to about 20 independent scorecards, 7 independent scorecards to about 50 independent scorecards, 8 independent scorecards to 9 independent scorecards, 8 independent scorecards to about 10 independent scorecards, 8 independent scorecards to about 20 independent scorecards, 8 independent scorecards to about 50 independent scorecards, 9 independent scorecards to about 10 independent scorecards, 9 independent scorecards to about 20 independent scorecards, 9 independent scorecards to about 50 independent scorecards, about 10 independent scorecards to about 20 independent scorecards, about 10 independent scorecards to about 50 independent scorecards, or about 20 independent scorecards to about 50 independent scorecards. In some cases, the systems, the methods, the computer-readable media, and the techniques disclosed herein may generate 1 scorecard, 2 independent scorecards, 3 independent scorecards, 4 independent scorecards, 5 independent scorecards, 6 independent scorecards, 7 independent scorecards, 8 independent scorecards, 9 independent scorecards, about 10 independent scorecards, about 20 independent scorecards, or about 50 independent scorecards. In some cases, the systems, the methods, the computer-readable media, and the techniques disclosed herein may generate at least 1 scorecard, 2 independent scorecards, 3 independent scorecards, 4 independent scorecards, 5 independent scorecards, 6 independent scorecards, 7 independent scorecards, 8 independent scorecards, 9 independent scorecards, about 10 independent scorecards, or about 20 independent scorecards. In some cases, the systems, the methods, the computer-readable media, and the techniques disclosed herein may generate at most 2 independent scorecards, 3 independent scorecards, 4 independent scorecards, 5 independent scorecards, 6 independent scorecards, 7 independent scorecards, 8 independent scorecards, 9 independent scorecards, about 10 independent scorecards, about 20 independent scorecards, or about 50 independent scorecards.Examples of Performance Data

[0182] FIG. 11 shows an example of performance data of generating scorecards. Specifically, the performance data is shown to compare the performance of the systems, the methods, the computer-readable media, and the techniques disclosed herein against experts in performing The Brief Psychiatric Rating Scale (BPRS) rating scale. The BPRS scorecard may be used by a healthcare provider to measure psychiatric symptoms such as depression, anxiety, hallucinations and unusual behavior. The BPRS scorecard is one of the oldest, most widely used scales to measure psychotic symptoms and was first published in 1962. Accordingly, healthcare providers are well-versed in this scorecard and the similarity of the predictions by the methods and thesystems the methods, the computer-readable media, and the techniques disclosed herein serve as a useful benchmark for performance.

[0183] The systems, the methods, the computer-readable media, and the techniques disclosed herein evaluated 236 patients for each of the 18 variables of the BPRS scorecard. Each patient was also evaluated for each of the 18 variables by one / two / etc. healthcare providers (e.g., physicians, researchers, etc.) during an interview. For each of the 18 variables, the patient is scored from 0 (not detected) to 7 (most severe). The accepted range of scientific uncertainty in a healthcare provider’s score is ± 1. Therefore, if the systems, the methods, the computer-readable media, and the techniques disclosed herein can score within ± 1 of a healthcare provider for a variable, this score is considered to be essentially in agreement, within the uncertainty.

[0184] The performance data of FIG. 11 show that the systems, the methods, the computer- readable media, and the techniques disclosed herein are in agreement, within the uncertainty, for each variable for the majority of patients. For example, for the variable of somatic concern, the systems, the methods, the computer-readable media, and the techniques disclosed herein predicted exactly the same score as the healthcare provider for 54% of patients; one point more than the healthcare provider for 19% of the patients; and one point lower than the healthcare provider for 15% of the patients - yielding agreement, within the uncertainty, 88% of the time. On average, across each of the 18 variables, the systems, the methods, the computer-readable media, and the techniques disclosed herein were in agreement with the healthcare providers, within the uncertainty, about 87% of the time.Certain Definitions and Additional Considerations

[0185] Unless otherwise defined, all technical terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present subject matter belongs.

[0186] As used in this specification and the appended claims, the terms “artificial intelligence,” “artificial intelligence techniques,” “artificial intelligence operation,” and “artificial intelligence algorithm” generally refer to any system or computational procedure that may take one or more actions to enhance or maximize a chance of achieving a goal. An example of such a goal is to mathematically or computationally model the probabilistic relationship between an input data (e.g., voice or biometric data) and an outcome like a mental condition detection (e.g., via generation of one or more scorecards). The term “artificial intelligence” may include “generative modeling,” “deep learning” (DL), “machine learning”, or “reinforcement learning” (RL). As used in this specification and the appended claims, the terms “machine learning,” “machine learning techniques,” “machine learning operation,” and “machine learning model” generally refer to anysystem or analytical or statistical procedure that may progressively improve computer performance of a task.

[0187] As used in this specification and the appended claims, “some embodiments,” “further embodiments,” or “a particular embodiment,” means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. Thus, the appearances of the phrase “in some embodiments,” or “in further embodiments,” or “in a particular embodiment” in various places throughout this specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0188] As used in this specification and the appended claims, when the term “at least,” “greater than,” or “greater than or equal to” precedes the first numerical value in a series of two or more numerical values, the term “at least,” “greater than” or “greater than or equal to” applies to each of the numerical values in that series of numerical values. For example, greater than or equal to 1, 2, or 3 is equivalent to greater than or equal to 1, greater than or equal to 2, or greater than or equal to 3.

[0189] As used in this specification and the appended claims, when the term “no more than,” “less than,” or “less than or equal to” precedes the first numerical value in a series of two or more numerical values, the term “no more than,” “less than,” or “less than or equal to” applies to each of the numerical values in that series of numerical values. For example, less than or equal to 3, 2, or 1 is equivalent to less than or equal to 3, less than or equal to 2, or less than or equal to 1.

[0190] As used in this specification, “or” is intended to mean an “inclusive or” or what is also known as a “logical OR,” wherein when used as a logic statement, the expression “A or B” is true if either A or B is true, or if both A and B are true, and when used as a list of elements, the expression “A, B, or C” is intended to include all combinations of the elements recited in the expression, for example, any of the elements selected from the group consisting of A, B, C, (A, B), (A, C), (B, C), and (A, B, C); and so on if additional elements are listed. As such, any reference to “or” herein is intended to encompass “and / or” unless otherwise stated.

[0191] As used in this specification and the appended claims, the indefinite articles “a” or “an,” and the corresponding associated definite articles “the” or “said,” are each intended to mean one or more unless otherwise stated, implied, or physically impossible. Yet further, it should be understood that the expressions “at least one of A and B, etc.,” “at least one of A or B, etc.,” “selected from A and B, etc.” and “selected from A or B, etc.” are each intended to mean either any recited element individually or any combination of two or more elements, for example, any of the elements from the group consisting of “A,” “B,” and “A AND B together,” etc.

[0192] As used in this specification and the appended claims “about” or “approximately” may mean within an acceptable error range for the value, which will depend in part on how the value is measured or determined, e.g., the limitations of the measurement system. For example, “about” may mean within 1 or more than 1 standard deviation, per the practice in the art. Alternatively, “about” may mean a range of up to 20%, up to 10%, up to 5%, or up to 1% of a given value. Where values are described in the application and claims, unless otherwise stated the term “about” meaning within an acceptable error range for the particular value may be assumed.

[0193] While preferred embodiments of the present invention have been shown and disclosed herein, it will be obvious to those skilled in the art that such embodiments are provided by way of example only. It is not intended that the invention be limited by the specific examples provided within the specification. While the invention has been described with reference to the aforementioned specification, the descriptions and illustrations of the embodiments herein are not meant to be construed in a limiting sense. Numerous variations, changes, and substitutions will now occur to those skilled in the art without departing from the invention. Furthermore, it shall be understood that all aspects of the invention are not limited to the specific depictions, configurations or relative proportions set forth herein which depend upon a variety of conditions and variables. It should be understood that various alternatives to the embodiments of the invention disclosed herein may be employed in practicing the invention. It is therefore contemplated that the invention shall also cover any such alternatives, modifications, variations, or equivalents. It is intended that the following claims define the scope of the invention and that methods and structures within the scope of these claims and their equivalents be covered thereby.

[0194] It should be noted that various illustrative or suggested ranges set forth herein are specific to their example embodiments and are not intended to limit the scope or range of disclosed technologies, but, again, merely provide example ranges for frequency, amplitudes, etc. associated with their respective embodiments or use cases. Where values are described as ranges, it will be understood that such disclosure includes the disclosure of all possible sub-ranges within such ranges, as well as specific numerical values that fall within such ranges irrespective of whether a specific numerical value or specific sub-range is expressly stated.

[0195] It should be understood that, unless a term is expressly defined in this patent, there is no intent to limit the meaning of that term, either expressly or by implication, beyond its plain or ordinary meaning, and such term should not be interpreted to be limited in scope based at least in part on any statement made in any section of this patent (other than the language of the claims). To the extent that any term recited in the claims at the end of this patent is referred to in this patent in a manner consistent with a single meaning, that is done for sake of clarity only so as tonot confuse the reader, and it is not intended that such claim term be limited, by implication or otherwise, to that single meaning.

[0196] Throughout this specification, plural instances may implement components, operations, or structures described as a single instance. Although individual operations of one or more methods are illustrated and described as separate operations, one or more of the individual operations may be performed concurrently, and nothing requires that the operations be performed in the order illustrated. Structures and functionality presented as separate components in example configurations may be implemented as a combined structure or component. Similarly, structures and functionality presented as a single component may be implemented as separate components. These and other variations, modifications, additions, and improvements fall within the scope of the subject matter herein.

[0197] Additionally, certain embodiments are disclosed herein as including logic or a number of routines, subroutines, applications, or instructions. These may constitute either software (e.g., code embodied on a machine-readable medium) or hardware. In hardware, the routines, etc., are tangible units capable of performing certain operations and may be configured or arranged in a certain manner. In example embodiments, one or more computer systems (e.g., a standalone, client or server computer system) or one or more hardware modules of a computer system (e.g., a processor or a group of processors) may be configured by software (e.g., an application or application portion) as a hardware module that operates to perform certain operations as disclosed herein.

[0198] In various embodiments, a hardware module may be implemented mechanically or electronically. For example, a hardware module may comprise dedicated circuitry or logic that is permanently configured (e.g., as a special-purpose processor, such as a field programmable gate array (FPGA) or an application-specific integrated circuit (ASIC) to perform certain operations. A hardware module may also comprise programmable logic or circuitry (e.g., as encompassed within a general-purpose processor or other programmable processor) that is temporarily configured by software to perform certain operations. It will be appreciated that the decision to implement a hardware module mechanically, in dedicated and permanently configured circuitry, or in temporarily configured circuitry (e.g., configured by software) may be driven by cost and time considerations.

[0199] Accordingly, hardware modules may encompass a tangible entity, be that an entity that is physically constructed, permanently configured (e.g., hardwired), or temporarily configured (e.g., programmed) to operate in a certain manner or to perform certain operations disclosed herein. Considering embodiments in which hardware modules are temporarily configured (e.g., programmed), each of the hardware modules need not be configured or instantiated at any oneinstance in time. For example, where the hardware modules comprise a general -purpose processor configured using software, the general-purpose processor may be configured as respective different hardware modules at different times. Software may accordingly configure processor, for example, to constitute a particular hardware module at one instance of time and to constitute a different hardware module at a different instance of time.

[0200] Hardware modules may provide information to, and receive information from, other hardware modules. Accordingly, the described hardware modules may be regarded as being communicatively coupled. Where multiple of such hardware modules exist contemporaneously, communications may be achieved through signal transmission (e.g., over appropriate circuits and buses) that connect the hardware modules. In embodiments in which multiple hardware modules are configured or instantiated at different times, communications between such hardware modules may be achieved, for example, through the storage and retrieval of information in memory structures to which the multiple hardware modules have access. For example, one hardware module may perform an operation and store the output of that operation in a memory device to which it is communicatively coupled. A further hardware module may then, at a later time, access the memory device to retrieve and process the stored output. Hardware modules may also initiate communications with input or output devices, and may operate on a resource (e.g., a collection of information). Elements that are described as being coupled and or connected may refer to two or more elements that may be (e.g., direct physical contact) or may not be (e.g., electrically connected, communicatively coupled, etc.) in direct contact with each other, but yet still cooperate or interact with each other.

[0201] The various operations of example methods disclosed herein may be performed, at least partially, by one or more processors that are temporarily configured (e.g., by software) or permanently configured to perform the relevant operations. Whether temporarily or permanently configured, such processors may constitute processor-implemented modules that operate to perform one or more operations or functions. The modules referred to herein may, in some example embodiments, comprise processor-implemented modules.

[0202] Similarly, the methods or routines disclosed herein may be at least partially processor- implemented. For example, at least some of the operations of a method may be performed by one or more processors or processor-implemented hardware modules. The performance of certain operations may be distributed among the one or more processors, not only residing within a single machine, but deployed across a number of machines. In some example embodiments, the processor or processors may be located in a single location (e.g., within a home environment, an office environment or as a server farm), while in other embodiments the processors may be distributed across a number of locations.

[0203] The performance of certain operations may be distributed among the one or more processors, not only residing within a single machine, but deployed across a number of machines. In some example embodiments, the one or more processors or processor-implemented modules may be located in a single geographic location (e.g., within a home environment, an office environment, or a server farm). In other example embodiments, the one or more processors or processor-implemented modules may be distributed across a number of geographic locations.

[0204] It will be understood that, although the terms first, second, etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first element may be termed a second element, and, similarly, a second element may be termed a first element, without departing from the scope of the present disclosure.

Claims

CLAIMSWHAT IS CLAIMED IS:

1. A computer-implemented method for assessing mental condition of a patient, comprising:(a) obtaining, from a first source, first input data corresponding to said patient;(b) obtaining, from a second source, second input data corresponding to said patient;(c) analyzing said first input data and said second input data using one or more machine learning models to generate a plurality of patient scores for at least two mental conditions for said patient;(d) generating at least two patient scorecards for said patient based at least in part on said plurality of patient scores, wherein each patient scorecard of said at least two patient scorecards corresponds to at least one of said at least two mental conditions, and wherein each patient scorecard of said at least two patient scorecards is at least partially complete; and(e) outputting or storing said at least two patient scorecards.

2. The computer-implemented method of claim 1, wherein said first input data comprises vocal data of said patient and said second input data comprises wearable biometric data of said patient.

3. The computer-implemented method of claim 1, wherein said first input data voice tone data of said patient and said second input data comprises vocal language data of said patient.

4. The computer-implemented method of any one of the preceding claims, wherein said first source is the same device as said second source.

5. The computer-implemented method of any one of claims 1-3, wherein said first source is a different device than said second source.

6. The computer-implemented method of either claim 4 or claim 5, wherein said first input data comprises vocal data of said patient.

7. The computer-implemented method of claim 6, further comprising: preprocessing said vocal data.

8. The computer-implemented method of claim 7, wherein preprocessing said vocal data comprises: filtering third-party voices from said vocal data, thereby generating filtered vocal data; and generating a transcription based at least in part on said filtered vocal data.

9. The computer-implemented method of either claim 7 or 8, wherein preprocessing said vocal data comprises: timestamping said vocal data.

10. The computer-implemented method of any one of claims 7-9, wherein preprocessing said vocal data comprises: decrypting said vocal data.

11. The computer-implemented method of any one of claims 6-10, wherein said first source comprises a user device, and said method further comprises: causing said user device to prompt said user to record said vocal data via an audio sensor of said user device, wherein either an audio file or a video file comprises said vocal data.

12. The computer-implemented method of any one of claims 6-11, wherein analyzing said first input data comprises one or both of: analyzing voice tone of said vocal data for sentiment markers, wherein said sentiment markers correspond to at least one patient score of said plurality of patient scores, or analyzing language of said vocal data for words and meaning, wherein said words and meaning correspond to at least one patient score of said plurality of patient scores.

13. The computer-implemented method of any one of the preceding claims, wherein said second input data comprises wearable biometric data of said patient.

14. The computer-implemented method of claim 13, wherein said wearable biometric data comprises one or more of: sleep data, activity data, routine data, temperature data, or photographic data.

15. The computer-implemented method of claim 14, wherein said wearable biometric data comprises said photographic data and said method further comprises:prompting said patient to capture a facial image, wherein said photographic data comprises said facial image.

16. The computer-implemented method of any one of claims 13-15, wherein said second source comprises one or more of: a smartwatch, a smartphone, smart glasses, a fitness tracker, or a sleep tracker.

17. The computer-implemented method of any one of claims 13-16, further comprising: preprocessing said wearable biometric data.

18. The computer-implemented method of claim 17, wherein preprocessing said wearable biometric data comprises: timestamping said wearable biometric data.

19. The computer-implemented method of any one of claims 13-18, wherein analyzing said second input data comprises: analyzing said wearable biometric data for patient routine, wherein said patient routine corresponds to at least one patient score of said plurality of patient scores.

20. The computer-implemented method of any one of the preceding claims, wherein said one or more machine learning models comprises one or more neural networks.

21. The computer-implemented method of claim 20, wherein said one or more neural network comprise one or more large language model.

22. The computer-implemented method of claim 21, wherein said one or more large language models use a transformer architecture.

23. The computer-implemented method of any one of claims 20-22, wherein said one or more machine learning models are trained at least in part by: obtaining a set of training data for a plurality of training patients, wherein said training data comprises (i) one or both of training vocal data or training wearable biometric data for said plurality of training patients and (ii) a plurality of training patient scores for said plurality of training patients;classifying said set of training data into a plurality of classified subsets that each correspond to a different training patient score of said plurality of training patient scores or a different range of training patient scores of said plurality of training patient scores; and generating said one or more machine learning models using said plurality of classified subsets.

24. The computer-implemented method of any one of claims 20-23, wherein said one or more machine learning models comprises at least two machine learning models, and wherein each of said at least two machine learning models corresponds to each of said at least two mental conditions.

25. The computer-implemented method of any one of claims 20-23, wherein said one or more machine learning models consists of one machine learning model, wherein said one machine learning model corresponds to each of said at least two mental conditions.

26. The computer-implemented method of any one of the preceding claims, wherein said plurality of patient scores comprise at least three different patient scores.

27. The computer-implemented method of claim 26, wherein said plurality of patient scores comprise at least four different patient scores.

28. The computer-implemented method of claim 27, wherein said plurality of patient scores comprise at least five different patient scores.

29. The computer-implemented method of any one of claims 26-28, wherein said plurality of patient scores comprise a scale of discrete ratings.

30. The computer-implemented method of any one of the preceding claims, wherein said at least two mental conditions comprise bipolar disorder.

31. The computer-implemented method of any one of the preceding claims, wherein said at least two mental conditions comprise schizophrenia.

32. The computer-implemented method of any one of the preceding claims, wherein said at least two mental conditions comprise two or more of: mania, depression, post-traumatic stressdisorder, premenstrual dysphoric disorder, suicidal ideation, psychosis, dementia, attention- deficit / hyperactivity disorder, anti-social behavior, bullying, anxiety, stress, self-analysis, personality, communication styles, or leadership.

33. The computer-implemented method of any one of the preceding claims, wherein said at least two patient scorecards comprise two or more of: a Young Mania Rating Scale scorecard, a Bipolar Depression Rating Scale scorecard, a Positive and Negative Syndrome Scale scorecard, a Patient Health Questionnaire-9 scorecard, a Brief Psychiatric Rating Scale scorecard, or a Myers- Briggs scorecard.

34. The computer-implemented method of any one of the preceding claims, wherein each patient scorecard of said at least two patient scorecards is fully complete.

35. The computer-implemented method of any one of the preceding claims, wherein outputting or storing said at least two patient scorecards comprises: causing said at least two patient scorecards to be presented to one or more of: said patient, a caregiver of said patient, or a family member of said patient.

36. The computer-implemented method of any one of the preceding claims, wherein said patient is undiagnosed with both of said at least two mental conditions.

37. The computer-implemented method of any one of the preceding claims, wherein said second input data comprises remote sensing data.

38. The computer-implemented method of any one of the preceding claims, further comprising, after a period of time elapsed since performing one or more of operations (a)-(e):(f) obtaining, from said first source, updated first input data corresponding to said patient;(g) obtaining, from said second source, updated second input data corresponding to said patient;(h) analyzing said updated first input data and said updated second input data using said one or more machine learning models to generate a plurality of updated patient scores for said at least two mental conditions for said patient;(i) updating said at least two patient scorecards for said patient based at least in part on said plurality of updated patient scores, thereby generating at least two updated patient scorecards for said patient; and(j) outputting or storing said at least two updated patient scorecards.

39. The computer-implemented method of claim 38, wherein said period of time is at least about 24 hours.

40. The computer-implemented method of claim 39, wherein said period of time is at least about 72 hours.

41. The computer-implemented method of claim 40, wherein said period of time is at least about one week.

42. The computer-implemented method of claim 41, wherein said period of time is at least about one month.

43. A computer system for assessing mental condition of a patient, comprising: one or more processors; and one or more memories storing computer-executable instructions that, when executed, cause the one or more processors to perform any one of claims 1-42.

44. One or more non-transitory computer-readable media comprising computer-executable instructions that, when executed by at least one processor, cause the at least one processor to perform any one of claims 1-42.

Images