Computational techniques for diagnosing and treating language-related disorders

Abstract

The present disclosure relates to computing techniques for diagnosing and treating speech-related disorders. Such techniques implement computer-generated diagnoses and computer-generated treatments delivered to at least one computing device via a network. The diagnoses and treatments are customized for each patient through a comprehensive analysis of the patient's production and reception errors obtained from the patient via the network, along with a set of correct responses at each stage of evaluation and treatment.

Description

[0001] This application is a divisional application. The application date of the original application is October 29, 2014, the application number is 201480065682.2, and the name of the invention is “Computational technology for diagnosing and treating language-related disorders”.

[0002] CROSS-REFERENCE TO RELATED APPLICATIONS

[0003] This application claims priority to U.S. Provisional Patent Application Serial No. 61 / 898,052, filed October 31, 2013, which is hereby incorporated by reference in its entirety for all purposes. Technical Field

[0004] The present disclosure relates generally to computing. More particularly, the present disclosure relates to computing techniques for diagnosing and treating language-related disorders.

[0005] background

[0006] Where in the present disclosure, a document, statute and / or item of knowledge is referenced and / or discussed, such reference and / or discussion is not an admission that said document, statute and / or item of knowledge and / or any combination thereof was publicly available, known to the public, part of the common general knowledge and / or otherwise constituted prior art under applicable statutory provisions at the priority date; and / or was known to be relevant in an attempt to solve any problem with which the present disclosure is concerned. In addition, nothing is disclaimed.

[0007] Language is the basis of human spirit and communicative function to a great extent. Therefore, the obstacle of a part of language expression can bring extensive or significant adverse effects. Some popular examples of this obstacle include language-related disorders, such as dyslexia, specific language impairment (SLI), auditory processing disorder (APD), aphasia, or other. For example, although dyslexia is usually regarded as dyslexia, individuals with this disease often also experience many other difficulties. Among these difficulties are problems in speech pronunciation, attention or memory. Therefore, some of these individuals may often work hard in school, especially at great personal cost, because dyslexia affects many basic language and cognitive functions. Others often drop out of school and experience self-esteem or other psychosocial problems. However, although these language barriers are common, many experts (such as teachers or therapeutic experts) are not trained accordingly.

[0008] This problematic state of affairs is further complicated by the fact that some language-related disorders, such as dyslexia, cover a wide range of deficits. Therefore, it remains questionable whether these disorders are even useful as components of research and assessment. In addition, many existing diagnostic tests for language-related disorders are not designed to be implemented to a broad group, despite allowing for self-pacing and customization based on the deficits of each individual user. Even worse, many providers simply stop at diagnosis and do not go on to recommend treatment to address even a large portion of the deficits found based on the diagnosis. At best, some providers who link treatment to diagnosis use the assessment results to select pre-set treatment modules that are intended to be performed by all users at a certain level of ability.

[0009] Although interventions can be useful in treating these disorders, at present, the state of interventional treatments for dyslexia is discouraging. For example, some dyslexia interventions in secondary schools have produced disappointing results. In addition, some popular reading instruction methods (such as the Orton-Gillingham method or methods based on Orton-Gillingham) have insufficient scientific evidence of efficacy, either partially or completely.

[0010] Brief Overview

[0011] The present disclosure at least partially solves at least one of the above contents. However, the present disclosure may prove to be useful in other technical fields. Therefore, the claims should not be interpreted as necessarily limited to solving any one of the above contents.

[0012] A system of one or more computers may be configured to perform specific operations or actions by installing software, firmware, hardware, or a combination thereof on the system, which, when in operation, causes the system to perform the operations or actions. One or more computer programs may be configured to perform specific operations or actions by containing instructions that, when executed by a data processing device, cause the device to perform the operations or actions.

[0013] An example embodiment of the present disclosure includes a method, which includes: diagnosing a language-related disorder via the following operations: obtaining a first set of criteria via a first computer, wherein the first set of criteria is based on a first analysis of a patient data structure relative to a master data structure, wherein the patient data structure includes a set of actual patient task responses, wherein the master data structure includes a set of cell generation data related to multiple patients and a set of predicted patient task responses; storing a first result in the patient data structure via the first computer, wherein the first result is received from a second computer, wherein the first result is based on the first computer selecting a first diagnostic shell based on the first set of criteria, generating a first diagnostic cell based on the first diagnostic shell and the cell generation data set, and communicating the first diagnostic cell to the second computer; obtaining a second set of criteria via the first computer, wherein the second set of criteria is based on a second analysis of the patient data structure including the first result relative to the master data structure; and determining via the first computer whether to generate a second diagnostic cell and whether to select at least one of a second diagnostic shell, wherein the second diagnostic cell is based on the first diagnostic shell, and wherein the first diagnostic shell and the second diagnostic shell are different in task type.

[0014] An example embodiment of the present disclosure includes a system comprising: a first computer that facilitates diagnosis of language-related disorders by: obtaining a first set of criteria, wherein the first set of criteria is based on a first analysis of a patient data structure relative to a master data structure, wherein the patient data structure includes a set of actual patient task responses, wherein the master data structure includes a set of cell generation data related to multiple patients and a set of predicted patient task responses; storing a first result in the patient data structure, wherein the first result is received from a second computer, wherein the first result is based on the first computer selecting a first diagnostic shell based on the first set of criteria, generating a first diagnostic cell based on the first diagnostic shell and the cell generation data set, and communicating the first diagnostic cell to the second computer; obtaining a second set of criteria, wherein the second set of criteria is based on a second analysis of the patient data structure containing the first result relative to the master data structure; determining at least one of whether to generate a second diagnostic cell and whether to select a second diagnostic shell, wherein the second diagnostic cell is based on the first diagnostic shell, and wherein the first diagnostic shell and the second diagnostic shell differ in task type.

[0015] An example implementation of the present disclosure includes a non-transitory computer-readable storage medium storing a set of instructions for execution via a hardware processor, wherein the set of instructions instructs the hardware processor to implement a method, the method comprising: diagnosing a language-related disorder via the following operations: obtaining a first set of criteria via a first computer, wherein the first set of criteria is based on a first analysis of a patient data structure relative to a master data structure, wherein the patient data structure includes a set of actual patient task responses, wherein the master data structure includes a set of cell generation data related to multiple patients and a set of predicted patient task responses; storing a first result in the patient data structure via the first computer, wherein the first result is received from a second computer, wherein the first result is based on the first computer selecting a first diagnostic shell based on the first set of criteria, generating a first diagnostic cell based on the first diagnostic shell and the cell generation data set, and communicating the first diagnostic cell to the second computer; obtaining a second set of criteria via the first computer, wherein the second set of criteria is based on a second analysis of the patient data structure including the first result relative to the master data structure; and determining via the first computer whether to generate a second diagnostic cell and whether to select at least one of a second diagnostic shell, wherein the second diagnostic cell is based on the first diagnostic shell, and wherein the first diagnostic shell and the second diagnostic shell are different in task type. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings illustrate example embodiments of the present disclosure. These drawings should not be construed as necessarily limiting the present disclosure. The same numbers and / or similar numbering schemes may refer to the same and / or similar elements throughout the text.

[0017] Figure 1 A schematic diagram illustrating an example implementation of a computer network model according to the present disclosure.

[0018] Figure 2 A schematic diagram illustrating an example implementation of a computer network architecture according to the present disclosure.

[0019] Figure 3 A schematic diagram illustrating an example implementation of a computer network diagram according to the present disclosure.

[0020] Figure 4 A schematic diagram illustrating an example implementation of a computer according to the present disclosure.

[0021] Figure 5 A flow chart illustrating an example implementation of a generative model-based diagnostic process according to the present disclosure.

[0022] Figure 6 A flow chart illustrating an example embodiment of a generative model-based treatment process according to the present disclosure.

[0023] Figure 7 A flow chart illustrating an example implementation of a diagnostic process according to the present disclosure.

[0024] Figure 8 A flow chart illustrating an example embodiment of a treatment process according to the present disclosure.

[0025] Fig. 9 Diagrams showing example embodiments of diagnostic and therapeutic procedures according to the present disclosure.

[0026] Fig.10 A diagram showing an example implementation of phoneme recognition diagnostic shells and cells for consonants according to the present disclosure.

[0027] Fig.11 A diagram showing an example embodiment of a phoneme recognition diagnostic shell and cell for vowels according to the present disclosure.

[0028] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0029] The present disclosure is now described more fully with reference to the accompanying drawings, in which example embodiments of the present disclosure are shown. However, the present disclosure may be implemented in many different forms and should not be construed as necessarily limited to the example embodiments disclosed herein. Rather, these example embodiments are provided so that the present disclosure is detailed and complete, and fully conveys the concepts of the present disclosure to those skilled in the art.

[0030] Features described with reference to certain example embodiments may be combined and subcombined in and / or with various other example embodiments. Moreover, different aspects and / or elements of the example embodiments as disclosed herein may also be combined and subcombined in a similar manner. In addition, some example embodiments (whether individually and / or collectively) may be components of a larger system, where other procedures may take precedence over and / or otherwise modify their application. In addition, as disclosed herein, many steps may be required before, after, and / or in parallel with the example embodiments. It should be noted that at least any and / or all methods and / or processes as disclosed herein may be performed in any manner, at least in part, via at least one entity.

[0031] The terms used herein may imply an action or inaction, direct or indirect, complete or partial, temporary or permanent. For example, when an element is referred to as being "on," "connected," or "coupled" to another element, then the element may be directly on, directly connected, or directly coupled to the other element, and / or there may be intervening elements, including indirect and / or direct variations. In contrast, when an element is referred to as being "directly connected" or "directly coupled" to another element, there are no intervening elements.

[0032] Although the terms "first", "second", etc. may be used herein to describe various elements, components, regions, layers and / or parts, these elements, components, regions, layers and / or parts are not necessarily limited by these terms. These terms are used to distinguish one element, component, region, layer or part from another element, component, region, layer or part. Therefore, without departing from the teachings of the present disclosure, the first element, component, region, layer or part discussed below may be referred to as a second element, component, region, layer or part.

[0033] The terms used herein are used to describe specific example embodiments and are not intended to necessarily limit the present disclosure. As used herein, the singular forms "a", "an" and "said" are intended to also include plural forms, unless the context clearly indicates otherwise. When used in this specification, the terms "include" and / or "comprise" specify the presence of stated features, integers, steps, operations, elements and / or components, but do not exclude the presence and / or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0034] Example embodiments of the present disclosure are described herein with reference to illustrations of idealized embodiments (and intermediate structures) of the present disclosure. As such, variations in the shapes of the illustrations as a result, for example, of manufacturing techniques and / or tolerances, are to be expected.

[0035] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those skilled in the art to which the present disclosure belongs. Terms (such as those defined in commonly used dictionaries) should be interpreted as having a meaning consistent with their meaning in the context of the relevant art, and should not be interpreted in an idealized and / or overly formal sense, unless explicitly defined as such herein.

[0036] In addition, relative terms such as "below," "lower," "above," and "upper" may be used herein to describe the relationship of one element to another element as illustrated in the accompanying drawings. These relative terms are intended to encompass different orientations of the illustrated technology in addition to the orientation depicted in the accompanying drawings. For example, if the device in the accompanying drawings is flipped, then the elements described as being on the "lower" side of the other elements will be oriented on the "upper" side of the other elements. Similarly, if the device in one of the drawings is flipped, then the elements described as being "below" or "beneath" the other elements will be oriented as being "above" the other elements. Therefore, the example terms "below" and "lower" may encompass both upper and lower orientations.

[0037] As used herein, the term "about" and / or "substantially" refers to a variation of + / - 10% relative to a nominal value / item. This variation is always included in any given value / item provided herein, whether or not this variation is specifically referred to that value / item.

[0038] If any disclosure is incorporated herein by reference and these disclosures conflict with the present disclosure in part and / or in full, the present disclosure controls with respect to the conflict and / or the broader disclosure and / or the broader definition of terms. If these disclosures conflict with each other in part and / or in full, the disclosure with the later date controls with respect to the conflict.

[0039] In some embodiments, the present disclosure implements a computing technology for providing personalized diagnosis and treatment to patients with language-related disorders. The technology implements computer-generated diagnosis and computer-generated treatment delivered to at least one computing device via a network. Diagnosis and treatment are customized for each patient by a comprehensive analysis of the patient's production and reception errors obtained from the patient via the network and a set of correct responses under each evaluation and treatment stage. The technology performs this error analysis by matching the patient's response with a preset matrix of all possible correct and incorrect responses predicted for the patient population involved. This error analysis enables the technology to provide individual-specific diagnosis and treatment cells, which efficiently and comprehensively target specific language processing defects that are the basis of the patient's defects. Each of the cells is a test and / or practice unit that focuses on one aspect of language and / or language function. The technology further implements a database that stores the patient's correct responses and ungrammatical production, based on which a learning analysis computer algorithm or similar method is used to monitor and improve the effectiveness of the technology for correcting language processing defects. Since the present disclosure relates to language processing, the scope of the present disclosure further extends beyond language structure to the communication and / or cognitive functions of language services. As scientific understanding of language-related disorders and impairments improves, the present disclosure may also be used as a preventive program for populations identified as being at risk. It should be noted that while language-related disorders include at least one of dyslexia, specific language impairment, auditory processing disorder, and aphasia, other language-related disorders are also included.

[0040] In some embodiments, the present disclosure enables the use of the patient's diagnostic results to generate customized treatment content for the patient. Through the use of computerized diagnosis and treatment programs, when delivering network services to multiple remote devices, the present disclosure is relatively effective and cost-effective in contacting a large number of patients with obstacles that affect language processing (such as speech disorders, dyslexia, aphasia, or other obstacles known to those skilled in the art). Through the use of cloud computing and / or other network technologies with comparable advantages, the present disclosure provides convenience as a computerized program available to multiple mobile users at any time or any day, thereby strengthening the effectiveness of the program. The present disclosure enables running programs and saving at least some patient data on remote servers, thereby expanding the number of remote devices that can be adopted by patients.

[0041] The present disclosure is intended to diagnose, evaluate and treat language-related disorders from several unique aspects. First, in some embodiments, the diagnosis based on the present disclosure is clear, because the diagnosis is based at least in part on a determination model, thereby identifying the set of specific problem areas in each patient's language function and structure. Each assessment level confirms and verifies the analysis of the previous level. Currently, there is no comparable clear diagnostic test for known dyslexia, because the patient's dyslexia is judged by a test set covering a wide range of spoken language and cognitive abilities, because dyslexia is regarded as an unexpected abnormality of an individual whose IQ (IQ) is within the normal range and functions fully in other aspects. Therefore, in order to diagnose dyslexia, psychologists can currently implement the Wechsler Intelligence Scale for Children-IV Integrated (WISC-IV), Wechsler Individual Achievement Test-III (WIAT-III), Boston Naming Test, Menyuk Syntactic Comprehension Test, Wide Range Memory and Learning Assessment-II (WRAML-II), Peabody Picture Vocabulary Test-IV, and neuropsychological assessment and testing with executive function. These tests allow certified experts to implement time-consuming and therefore costly for most families. Furthermore, diagnoses based on this test are usually probabilistic, based on normative data. Although some sectors of the scientific community currently state that IQ tests are unnecessary for diagnosing dyslexia, the field still lacks reliable diagnostic instruments.

[0042] Second, in some embodiments, the present disclosure enables the establishment of personalized assessments and treatments. The present disclosure enables clearer diagnosis and efficient and effective treatment because the disclosed technology analyzes and / or solves the reception and production errors of each patient at each stage of their training. The program individually customizes diagnostic and treatment cells to solve the basic language processing problems of each user. For example, if a patient makes an error regarding the pronunciation of / θ / ("th") during a phoneme recognition test during the diagnostic stage, then words with this / θ / pronunciation are included in a word segmentation test during further diagnostic testing to confirm the problem and obtain finer details of their processing defects. These details may include the pronunciation environment in which these problems with / θ / occur. The pronunciation environment is the phonetic context surrounding a particular phoneme, such as the pronunciations adjacent to the phoneme and / or the position of the phoneme in a word, such as the initial position, the middle position, and / or the final position. For example, a patient may have difficulty with the phoneme / θ / in the final position, such as tee / . th , but there is no problem with the phoneme / θ / in word-initial position, such as think. In contrast, currently, many clinicians use preset diagnostic tests that are uniformly applied to all patients. Furthermore, currently used methods that claim to be "individual-based" simply progress the user to the next preset test and / or train at a higher or equal difficulty based on previous performance. These preset linear programs cannot "fine-tune" their training methods because they cannot analyze the error patterns generated by the user during their task performance, nor can they use the analysis results to develop subsequent treatments, which the present disclosure achieves. Furthermore, the present disclosure achieves "future applicability" by generating new treatment cells to make corrections based thereon, i.e., responding to newly discovered treatment problems at any stage of treatment. More specifically, diagnosis, assessment and / or treatment are tightly integrated in the present disclosure in a non-linear generative manner, which is important because language-related disorders tend to occur continuously with a wide range of individual differences.

[0043] Third, in some embodiments, the present disclosure enables correction rather than compensation of language processing defects. More specifically, the treatment of dyslexia can generally be divided into treatment with the aid of a correction method and treatment with the aid of a compensation method. Most providers focus on compensatory techniques and therefore generally accept dyslexia as a lifelong defect. In fact, some experts label dyslexia as a lifelong disease. The few providers who have attempted to correct the problem have not yet succeeded in doing so in a manner that is replicable for the population involved. Although some cognitive scientists retain hope, this goal remains difficult to achieve in this field. In addition, some scientists have concluded that correction programs (such as Scientific Learning Corporation's Fast English Implantation) have failed to achieve this goal. One reason for this gap between hope and reality is clear in the context of the present disclosure: correction programs (such as Fast English Implantation) cannot directly solve the language processing problems of most individuals because the correction programs are based on linear modules with preset modules. In contrast, the present disclosure is nonlinear and responds to the performance of each user at each training stage. In addition, these existing methods artificially modify speech input and use non-speech pronunciation to focus on processing speed. In contrast, the present disclosure does not digitally alter speech signals in a way that the input no longer sounds like natural speech. Natural speech is the spoken production of native speakers without speech impairment and is produced spontaneously in a natural environment during human interaction. However, certain prosodic features of natural speech may be exaggerated during moments of excitement or interactions with young children. In fact, the present disclosure may use live and / or recorded speech of a speaker that exaggerates the prosody (i.e., length, pitch, stress) of natural speech as input to help the user hear the input clearly. The present disclosure not only focuses on auditory processing speed, but also analyzes the user's production and reception of all components of language and also addresses most, if not all, other basic issues, including vocabulary (word) representation and retrieval.

[0044] Fourth, in some embodiments, the present disclosure enables comprehensive correction of language processing problems for each user. One reason why such language-related disorders (such as dyslexia) continue to exist is that existing diagnosis, assessment, and treatment methods are insensitive to multiple aspects of each individual's language deficit. Most current methods record the patient's correct response at most and discard incorrect responses. In contrast, the present disclosure enables analysis of the patient's incorrect response to find out the error pattern that identifies a specific difficult area. The present technology uses these error patterns to establish a model of the patient's language reception and production ability to serve as a roadmap for the treatment of only and all problems that are specific (or unique) to and correct for each person. In the present disclosure, error analysis is used to compare the patient's actual response with a target set of correct responses. For example, error analysis can reveal that the patient has difficulty processing articles (the, a, an) based on the patient's omission and incorrect substitution in the context of needing substitution. This persistent error analysis throughout the entire treatment phase allows the computer program to continuously update treatment and assessment, and to revise as needed. These unique strengths of the program are made possible by the program's predictive features, which are built upon a vast knowledge base of aggregated oral behaviors and outputs of both patient populations with these language processing disorders and typical populations. Thus, in the absence of an effective, standardized approach to reading intervention, the present disclosure enables a personalized, data-driven approach.

[0045] Fifth, in some embodiments, the present disclosure enables the application of learning analysis algorithms and other intelligent data processing techniques and / or artificial intelligence (AI) techniques, such as IBM Watson, Apple Siri, Google Now, or Microsoft Cortana, to at least one database of stored patient responses to continuously improve program performance. The present disclosure implements data distillation of accumulated information to discover (for example) the types of test items that are frequently generated and the evaluation units that are repeated because the first attempt failed. The program uses this type of stored information to focus resources on improving the content of the most frequently used test types and enhancing the effectiveness of certain treatment cells, as described herein. This information is stored in a learning analysis database, as described herein. For example, because the program can identify which specific functional areas of the patient's brain are affected, this information, as collected in this database from a wide group of users, can be used to predict the progression path of each new patient, plan their treatment schedule, and / or estimate the duration of treatment.

[0046] In some embodiments, the present disclosure implements a computer program that includes a diagnostic phase and a therapeutic phase, both of which are based on data hosted in at least one computerized database in any manner. The program interfaces with the at least one database via a network. It should be noted that at least one of the diagnostic phase and the therapeutic phase can occur in whole or in part without the intervention (whether direct or indirect) of a clinician with a language-related disorder.

[0047] The diagnostic phase that can be performed at least in part via hardware modules and / or software modules includes the deployment of multiple diagnostic shells and multiple diagnostic cells. In some embodiments, at least one of the diagnostic shells can be implemented via at least one of an instruction set, function, program, call, routine, subroutine, vector, algorithm, heuristic, parameter, criterion, applet, library, operation, command, module, data structure (for example, matrix, queue, array, stack, deck of cards, linked list, table, tree or other), category, object, node, flag, alphanumeric value, symbolic value, hash, file, driver, software application and / or any combination thereof and / or equivalent. Each in the diagnostic shell is a program for serving as a specified activity of a language function test. This program does not contain any test items. Instead, the test is delivered via a diagnostic cell, and the diagnostic cell is generated via inserting a test item into the diagnostic shell. More specifically, each diagnostic shell is a test type, and each diagnostic cell is an actual specific test. In some embodiments, at least one of the diagnosis cell can be implemented via at least one of an instruction set, a function, a program, a call, a routine, a subroutine, a vector, an algorithm, a heuristic, a parameter, a criterion, a small program, a library, an operation, a command, a module, a data structure (e.g., a matrix, a queue, an array, a stack, a deck of cards, a linked list, a table, a tree or other), a category, an object, a node, a flag, an alphanumeric value, a symbolic value, a hash, a file, a driver, a software application and / or any combination thereof and / or an equivalent. Technology disclosed herein enables generating diagnostic tests for each patient by placing content in a diagnosis shell. Therefore each test generated is a diagnosis cell. For example, if the diagnosis shell relates to vocabulary (word) retrieval, then program generation relates to the first diagnosis cell of a noun and relates to the second diagnosis cell of a verb. Program generates a diagnosis cell via retrieving a specific item and / or a specific item set from a first data structure (such as a main matrix). It should be noted that other types of data structures can also be used, such as queues, stacks, decks of cards, linked lists, tables, trees or other. Additionally, it should be noted that the first data structure may be at least one of indexed and searchable.

[0048] Some embodiments of the present disclosure include a pronunciation-symbol matching diagnostic shell. For example, this shell involves the patient entering symbols (such as letters and / or characters) such as via a keyboard (whether physical and / or virtual) coupled to a computer running the program and / or such as via selection from a set of options displayed on a display coupled to a computer running the program, by matching the pronunciation heard by the patient in an audio recording output by the program (such as by typing h for the pronunciation / h / heard). The program then matches the patient's response with a set of correct responses stored in a data structure (such as a master matrix), and provides a total of the patient's correct responses and provides other feedback when necessary. The program then matches the patient's incorrect response with a set of predicted errors as stored in a data structure (such as a master matrix) to generate new diagnostic cells and / or treatment cells.

[0049] Some embodiments of the present disclosure include a vocabulary-accessible diagnostic shell. For example, this shell involves a patient listing words according to specified criteria, such as words beginning with "B", by speaking into a microphone coupled to a computer running the program within a set time constraint or by typing via a keyboard (whether physical and / or virtual) coupled to a computer running the program. The program then matches the patient's response with a set of correct responses in a data structure (such as a master matrix), and provides a total of the patient's correct responses and provides other feedback when necessary. The program then classifies the patient's incorrect responses into a set of categories (such as pronunciation-based or phonetic errors, meaning-based or semantic errors), and matches incorrect responses with a set of predicted errors of related categories as stored in a data structure (such as a master matrix) to generate new diagnostic cells and / or treatment cells.

[0050] Some embodiments of the present disclosure include a vocabulary retrieval diagnostic shell. For example, this shell involves the patient naming the object by speaking to a microphone coupled to a computer running the program or selectively activating / clicking the name from a list / screen of options displayed on the display when the image of the object is displayed on a display coupled to the computer. The program then matches the patient's response with a set of stored correct responses in a data structure (such as a master matrix), and provides a total of the patient's correct responses and provides other feedback when necessary. The program then classifies the patient's incorrect responses into a set of categories (such as pronunciation-based or phonetic errors, meaning-based or semantic errors), and matches incorrect responses with a set of predicted errors of related categories as stored in a data structure (such as a master matrix) to generate new diagnostic cells and / or treatment cells.

[0051] Some embodiments of the present disclosure include a syllable-divided diagnostic shell. For example, this shell involves patient splitting whether it is outputted by hearing via a speaker coupled to a computer running the program and / or outputted by vision via a display coupled to a computer running the program. This word splitting occurs via a number of individual syllables (such as un-der-stand) in a word entered in order, such as via typing into a keyboard (whether physical or virtual) coupled to a computer running the program and / or speaking to a microphone coupled to a computer running the program. The program then matches the patient's response with a stored set of correct responses such as stored in a data structure such as a master matrix, and provides a total of the patient's correct responses and provides other feedback when needed. The program then matches incorrect responses with a set of prediction errors such as stored in a data structure such as a master matrix to generate new diagnostic cells and / or treatment cells.

[0052] Some embodiments of the present disclosure include a word segmentation diagnosis shell. For example, this shell involves the patient splitting words displayed on a display coupled to a computer running the program. This word splitting occurs by speaking the individual segments of the word (such as / k lIk / for click) in order to a microphone coupled to a computer running the program. The program then matches the patient's response with a stored set of correct responses as stored in a data structure (such as a master matrix), and provides a total of the patient's correct responses and provides other feedback when needed. The program then matches incorrect responses with a set of prediction errors as stored in a data structure (such as a master matrix) to generate new diagnostic cells and / or treatment cells.

[0053] Some embodiments of the present disclosure include a word recognition diagnostic shell. For example, this shell involves the patient listening to the recorded pronunciation or words and / or looking at the symbols or words displayed on the display of the computer coupled to the running program. The patient selects these words from the article displayed on the display via highlighting (such as via an input device, such as a keyboard or touch pad) or clicking on any of the pronunciation and / or symbols printed or with the matching read out (such as via an input device, such as a mouse). The program then matches the patient's response with a set of correct responses such as stored in a data structure (such as a main matrix), and provides a total of the patient's correct responses and provides other feedback when needed. The program then matches incorrect responses with a set of predicted errors such as stored in a data structure (such as a main matrix) to generate new diagnostic cells and / or treatment cells.

[0054] Some embodiments of the present disclosure include a phoneme recognition diagnostic shell. For example, this shell involves the patient selecting any word containing a certain pronunciation from a string of output words, either visually (such as via a display coupled to a computer running the program) and / or auditorily (such as via a speaker coupled to a computer running the program). This selection is performed by clicking a button on a display (whether physical or virtual) and / or repeating the selected word into a microphone. The program then matches the patient's response with a set of correct responses as stored in a data structure (such as a master matrix), and provides a summary of the patient's correct responses and other feedback when necessary. The program then matches the patient's incorrect response with a set of prediction errors as stored in a data structure (such as a master matrix) to generate new diagnostic cells and / or treatment cells.

[0055] Some embodiments of the present disclosure include a rhyming diagnosis shell. For example, this shell involves the patient selecting any words (reaction words) that rhyme with the words (stimulus words) output. The words output are output visually (such as via a display coupled to a computer running the program) and / or auditorily (such as via a speaker coupled to a computer running the program). This selection is performed by clicking a button on the display (whether physical or virtual) when the reaction word appears or by speaking a word that rhymes with the words (stimulus words) output to a microphone. The program then matches the patient's response with a set of correct responses such as stored in a data structure (such as a master matrix), and provides a total of the patient's correct responses and provides other feedback when needed. The program then matches incorrect responses with a set of prediction errors such as stored in a data structure (such as a master matrix) to generate new diagnosis cells and / or treatment cells.

[0056] Some embodiments of the present disclosure include a morpheme recognition diagnostic shell. For example, this shell involves the patient highlighting and / or typing on a keyboard coupled to a computer running the program with an input device (such as a mouse) coupled to a computer running the program affixes found in any word displayed on a display coupled to a computer running the program. The program then matches the patient's response with a set of correct responses as stored in a data structure (such as a master matrix), and provides a total of the patient's correct responses and provides other feedback when necessary. The program then matches incorrect responses with a set of predicted errors as stored in a data structure (such as a master matrix) to generate new diagnostic and / or therapeutic cells.

[0057] Some embodiments of the present disclosure include a rapid naming diagnosis shell. For example, this shell involves the patient reading out the words flashing on the display of the computer coupled to the running program in a microphone coupled to the computer running the program under the time constraints of the setting or at an accelerated speed. The words can be displayed in any way, such as one letter or some letters from left to right at a time in different colors or forms. The program then matches the patient's response with a group of correct responses such as stored in a data structure (such as a master matrix), and provides the total of the patient's correct responses at the target speed and provides other feedback when needed. The program then matches incorrect responses with a set of prediction errors such as stored in a data structure (such as a master matrix) to generate new diagnosis cells and / or treatment cells.

[0058] Some embodiments of the present disclosure include a fast processing diagnostic shell. For example, this shell involves the patient performing one or more of the tests in at least another diagnostic shell under a set time constraint or at an accelerated speed. The program then matches the patient's response with a set of correct responses as stored in a data structure (such as a master matrix), and provides a total of the patient's correct responses at the target speed and provides other feedback when necessary. The program then matches incorrect responses with a set of predicted errors as stored in a data structure (such as a master matrix) to generate new diagnostic cells and / or treatment cells.

[0059] Some embodiments of the present disclosure include a reading fluency diagnostic shell. For example, this shell involves a patient reading an article displayed on a display coupled to a computer running the program into a microphone coupled to the computer running the program. The program then uses voice or speech recognition software or an online assistant to identify and record reading errors and provide feedback when necessary. The online assistant can be contacted via at least one method (such as a phone call, a telecommunications conference session, a chat, or other). The program then classifies any captured reading errors into a set of categories (such as pronunciation-based or phonetic errors, meaning-based or semantic errors), and matches the patient's errors with a set of predicted errors as stored in a data structure (such as a master matrix) to generate new diagnostic cells and / or treatment cells.

[0060] It should be noted that any cell test or cell task can be performed through a new form or device (such as a touch screen feature, clicker, or application). In addition, it should be noted that any cell test or cell task can be repeated in a cell. In addition, it should be noted that any cell test or cell task can be designed as an age-appropriate interactive game.

[0061] The treatment phase, which may be performed at least in part via a hardware module and / or a software module (whether distinct from or as a diagnostic module), includes the deployment of a plurality of treatment shells and a plurality of treatment cells. In some embodiments, at least one of the treatment shells may be implemented via at least one of an instruction set, function, program, call, routine, subroutine, vector, algorithm, heuristic, parameter, criterion, applet, library, operation, command, module, data structure (e.g., matrix, queue, array, stack, deck, linked list, table, tree, or other), class, object, node, flag, alphanumeric value, symbolic value, hash, file, driver, software application, and / or any combination thereof and / or equivalents. It should be noted that these embodiments may be the same and / or different from at least one of the treatment shells. The treatment cells are generated from the treatment shells in a manner similar to the generation of the diagnostic cells from the diagnostic shells. In some embodiments, at least one of the therapeutic cells may be implemented via at least one of an instruction set, a function, a program, a call, a routine, a subroutine, a vector, an algorithm, a heuristic, a parameter, a criterion, a small program, a library, an operation, a command, a module, a data structure (e.g., a matrix, a queue, an array, a stack, a deck of cards, a linked list, a table, a tree or other), a class, an object, a node, a flag, an alphanumeric value, a symbolic value, a hash, a file, a driver, a software application, and / or any combination thereof and / or an equivalent. It should be noted that these embodiments may be the same and / or different from at least one of the therapeutic cells. In addition, each therapeutic cell contains a training unit and an evaluation unit. The training unit allows practice with new exercises, while the evaluation unit assesses the performance of previously or currently covered aspects. However, in other embodiments, the therapeutic cell is generated from the therapeutic shell in a different manner than the diagnostic cell is generated from the diagnostic shell. Moreover, in some embodiments, at least one of the training units may be implemented via at least one of an instruction set, a function, a program, a call, a routine, a subroutine, a vector, an algorithm, a heuristic, a parameter, a criterion, a small program, a library, an operation, a command, a module, a data structure (e.g., a matrix, a queue, an array, a stack, a deck, a linked list, a table, a tree, or other), a class, an object, a node, a flag, an alphanumeric value, a symbolic value, a hash, a file, a driver, a software application, and / or any combination thereof and / or an equivalent. Moreover, in some embodiments, at least one of the evaluation units may be implemented via at least one of an instruction set, a function, a program, a call, a routine, a subroutine, a vector, an algorithm, a heuristic, a parameter, a criterion, a small program, a library, an operation, a command, a module, a data structure (e.g., a matrix, a queue, an array, a stack, a deck, a linked list, a table, a tree, or other), a class, an object, a node, a flag, an alphanumeric value, a symbolic value, a hash, a file, a driver, a software application, and / or any combination thereof and / or an equivalent. It should be noted that these implementations may be the same and / or different than at least one of the training units.

[0062] Some embodiments of the present disclosure include a phoneme discrimination treatment shell. For example, this shell involves a computer display showing minimal pairs, such as pit and bit. An audio recording is played one word at a time in random order at a specified speed. The patient selects / clicks on the word as it is read before the record plays the next word, or presses on the arrow keys representing each word. The program then matches the patient's response with a set of correct responses stored in a first data structure (such as a master matrix) and provides a total of the patient's correct responses and other feedback as needed. The program then matches the total of the patient's correct responses with specified criteria in the first data structure (such as a master matrix) to generate a new treatment cell. It should be noted that as part of the treatment phase, the patient can go through the same phoneme discrimination cell multiple times at increasing speeds with the same minimal pairs to improve their auditory processing speed.

[0063] Some embodiments of the present disclosure include a rapid word recognition treatment shell. For example, this shell involves a computer display showing minimal opposites, such as pit and bit. One of the two words is highlighted one at a time in a random order at a specified speed. The patient reads the highlighted word by speaking it into a microphone while the word is highlighted before the next word is illuminated. The program then matches the patient's response with a set of correct responses stored in a first data structure (such as a master matrix) and provides a total of the patient's correct responses and other feedback when necessary. The program matches the total of the patient's correct responses with specified criteria in a first data structure (such as a master matrix) to generate a new treatment cell. It should be noted that as part of the treatment phase, the patient can go through the same rapid word recognition cell multiple times at increasing speeds with the same minimal opposites to improve his reading speed.

[0064] Some embodiments of the present disclosure include a word amplification therapy shell. For example, this shell involves a program playing an audio file and / or video file of a speaker lengthening and amplifying the prosodic features of a word (such as the / br / segment in brick or the intonation contour of variety). Alternatively, this amplification is performed through an online assistant function and / or animation. The patient records his imitation of the amplified form by speaking into a microphone or by using a camera contained in and / or coupled to the computer with or without body gestures (such as using a chopping motion to indicate a syllable break or waving his hand high or low to signal pitch). The program matches the patient's recording with a stored set of parameters in a first data structure (such as a master matrix) and provides corrective or evaluative feedback when necessary, such as further lengthening the / br / sound group to form two different segments. The patient answers a series of questions about his or her pronunciation of the indicated word, such as "Where is the tip of your tongue when you say / l / in help?" The program then matches the patient's responses to a set of correct responses stored in a first data structure (such as a master matrix) and provides a summary of the patient's correct responses and other feedback as needed. The program then matches the summary of the patient's correct responses to the specified criteria in the first data structure (such as a master matrix) to generate a new treatment cell.

[0065] In the training unit, in some embodiments, each treatment cell focuses on only one specific language processing problem identified in the patient's response during diagnosis and / or treatment. Each cell is designed to correct only one such problem through sufficient practice followed by evaluation to confirm that the deficit has been satisfactorily addressed. For example, if the patient has difficulty handling speech with / The program generates a word containing this / Phonemes and syllables (such as / i / 、 ), real words, and / or nonsense words (such as s ure、lu sh or sh um) and sentences containing this pronunciation (such as This is surely the best sh Each list is then placed in a separate shell that specifies a different task. For example, in one such therapy cell, the patient can imitate a sentence containing / each time he hears a prompt in the form of recording the pronunciation of this phoneme. syllables or words to practice accurately making this sound. Alternatively, it can be required to draw this dental fricative (sibilant) sound within the specified duration of the prompt indication on the display of the computer coupled to the running program. In another embodiment, once the baseline in fundamental frequency has been set for the patient's specific voice, the program uses speech or speech recognition software to give the patient real-time feedback on the accuracy of each of its verbal reactions. For example, the patient's speech signal is represented by a spectrogram (sound wave pattern) converted by the program into a visual clue on the display, thereby indicating the distance between its generation and the target form when the patient tries to approach the target. In another treatment cell, the user can activate / click the word with this pronunciation in the list displayed on the display, such as via an input device (e.g., a keyboard, whether physical or virtual, a touchpad, a mouse, a clicker, a joystick or a touch screen). The tasks in the treatment cell can reflect the tasks performed for the diagnostic test. Other treatment cells can cover other reception and / or generation difficulties. The task range can range from paying attention to speech features to the full text. Overall, a cell group may cover complex tasks, such as reading, while individual cells within the group may focus on spelling rules and vocabulary acquisition (vocabulary building).

[0066] In the evaluation unit, in some embodiments, after sufficient practice in the training unit of the treatment cell, the patient progresses to the evaluation unit of the same cell. The patient performs the same task with a test item similar or identical to the associated training unit. The user must pass this evaluation before proceeding to the next treatment cell. The passing score is pre-specified in a data structure (such as a master matrix) for each evaluation unit. Correct and incorrect reactions are recorded in a second data structure (such as a patient matrix) and are used to generate new treatment cells by computer when necessary with reference to the first data structure (such as a master matrix) described herein. It should be noted that other types of data structures can also be used, such as queues, stacks, card decks, linked lists, tables, trees or other. In addition, it should be noted that the first data structure and the second data structure can be a subset of a parent data structure and / or a parent data structure. The first data structure and the second data structure can be stored in a separate computerized database (whether remote or local to each other), or stored in a computerized database. In addition, it should be noted that the second data structure can be at least one of indexed and searchable.

[0067] In another embodiment, the assessment unit may be separate from the training unit, or one or more assessment units may accompany the training unit in the cell or vice versa, or the assessment may be incorporated into the training phase itself. The cell may also contain other types of units, such as practical units, such as watching videos or webinars, and different formats such as described herein. In yet another embodiment, when the patient is in the treatment phase, the program may perform further diagnostic tests if necessary before the patient resumes his or her treatment.

[0068] In yet another embodiment, the disclosed technology may employ interactive games, such as to facilitate at least one of the diagnostic phase and the therapeutic phase. More specifically, the therapeutic cell may be in the form of an age-appropriate interactive game, wherein the game is a shell that can accommodate different content suitable for each user, and the content follows an error analysis of the user's response. For example, a pronunciation therapy cell as described herein may use at least one audio-visual clue in the game to prompt the patient to approach a target pronunciation or word. Various game rewards or penalties may be included to encourage the patient to achieve a specified goal. It should be noted that these rewards that may be operated based on a loyalty system may be redeemable, such as redeemed for prizes, cash, goods, services, airline miles, additional treatment sessions, personal diagnosis and / or treatment sessions, and the like.

[0069] The first data structure (such as a master matrix) contains a list set of responses to the predicted tests and tasks assigned and the information required to generate diagnostic cells and / or treatment cells. As described herein, the first data structure can be implemented in a computerized database in whole or in part. In other embodiments, the first data structure is implemented in a plurality of databases (such as a relational database or a non-relational database) that are remotely and / or locally hosted and / or similar to each other or different in any way (directly and / or indirectly). The first data structure is remotely hosted / resides (such as) on a server computer rather than on a patient's computer as described herein, thereby allowing end users (such as patients) to have more flexibility, while providing more efficient content updates because of this resident / hosting. However, in other embodiments, the first data structure can be hosted / resides on a patient's computer in whole or in part, directly and / or indirectly locally. This configuration can be implemented in whole or in part, directly and / or indirectly from a central data repository (such as a computerized database) periodic update (such as weekly or monthly). For example, in this configuration, a first data structure hosted / residing locally in whole or in part on the patient's computer is at least a partial copy of a first data structure hosted / residing remotely from the patient's computer.

[0070] In some embodiments, the first data structure contains the reaction of most (if not all) predictions and such as when the first data structure is a database, data is fed to the computer, as described herein. The computer is configured to guide / generate cell content and shell based on this data feed. The first data structure contains a list of reactions predicted by a given population with language processing problems when performing tasks based on the diagnostic cell and / or therapeutic cell according to the program. Then, when necessary, the computer first classifies incorrect reactions into several categories, such as speech, semantics, morphology or vocabulary, and subdivides (such as morphology> affix> prefix) with visual depiction (such as graphics, for example, including mathematical objects of vertices and / or arcs) for storage in the first data structure. The first data structure may also contain phonemes, dictionaries (vocabulary), phrases and / or sentence patterns and other components of the target language required for generating diagnostic cells and / or therapeutic cells. For example, the target language can be any natural language (i.e., the native language spoken by any group of people in any world region).

[0071] For example, if we refer to Fig. 9 As shown and described, each stimulus item in a diagnostic cell and / or treatment cell is represented as a node on a graph, where the connected vertices at the next lower level represent most, if not all, of the possible correct and incorrect responses predicted to be made by the population involved (such as a patient with aphasia). The program matches the patient's incorrect response with the same prediction error (node) in the relevant graph, and proceeds down the graph to locate the appropriate set of test and / or practice items to generate the next diagnostic cell and / or treatment cell. Thus, for Fig. 9In the example illustrated in, the stimulus item in the word segmentation diagnostic test is the word be. In this test, the patient must separate the word into its individual pronunciation. The corresponding node for this stimulus item in the first data structure (such as the main matrix) is connected to its correct response ( / bi / ) and the incorrect response of prediction. Incorrect response comprises the understanding of the process of the principle and rule of natural language, linguistics, as the basis of language acquisition or development, and the process of governing abnormal language (that is, the speaker outside the typical population, such as the language of the individual with communication disorder) and most (if not all) possible examples predicted. In the case of the stimulus word be, the incorrect response of prediction comprises the example that the user omits the consonant / b / or replaces this consonant or vowel / i / . At least the knowledge in the domain set identified above helps to predict, that is, if / b / is replaced, the phoneme that is likely to be replaced because of voice similarity is / p / or / d / . Graphics also allow other possible replacements. In addition, the specific incorrect response of the patient (such as no / b / ) causes a specific computer operation (generating a word with / b / ). The generated words can be real words from the dictionary, or nonsense words with the desired pronunciation combination. When the patient's incorrect response is a substitution error (such as / p / for / b / ), the computer generates at least one pair of words with a critical contrast, such as the minimal contrast pit. With bit ). If the patient's response contains more than one error, most (if not all) of the affected nodes on the graph are activated, which means that the patient must practice with more than one word list during the treatment phase. When this word list is placed in the listen and repeat treatment shell, the patient performs the required task by completing each word in this list one by one. In addition, since the complete test usually contains several stimulus items, this test can produce more than one incorrect response from the patient. Therefore, several graphs can be activated from one test output in the first data structure (such as the main matrix). For example, a rapid naming diagnostic test can produce errors involving several words. In this case, a computer-implemented priority sorting algorithm determines which test or practice item set to present to the patient next.

[0072] The second data structure (such as a patient matrix) is uniquely associated with each patient. As described herein, the second data structure can be implemented in a computerized database in whole or in part. In other embodiments, the second data structure is implemented in multiple databases (such as a relational database or a non-relational database) that are remotely and / or locally hosted and / or similar to each other or different in any way (directly and / or indirectly). The second data structure is remotely hosted / resides (such as) on a server computer rather than on a patient's computer as described herein, thereby allowing end users (such as patients) to have more flexibility, while providing more efficient content updates because of this resident / hosting. However, it should be noted that the server computer can host the first data structure and the second data structure, or the first data structure and the second data structure are hosted / reside on different server computers, as described herein. In addition, it should be noted that in other embodiments, the second data structure can be hosted / reside on a patient's computer in whole or in part, directly and / or indirectly locally. This configuration can be implemented in whole or in part, directly and / or indirectly from a central data repository (such as a computerized database) periodic update (such as weekly or monthly). For example, in this configuration, the second data structure that is hosted / resides locally on the patient's computer in whole or in part is an at least partial copy of a second data structure that is hosted / resides remotely from the patient's computer. In further embodiments, the first data structure and the second data structure are one data structure that may be hosted in whole or in part in any manner described herein with reference to at least one of the first data structure and the second data structure individually.

[0073] The second data structure stores patient personal information, such as user identification (ID), name, domicile address, background (whether personal, medical, social, ethnic, racial or other), and its response to diagnostic and assessment tests. In addition, the second data structure can be dynamically and automatically updated via a computer (whether via a service requester segment as described herein or via a service provider segment as described herein). This update can occur after each new response from the patient is filtered through the first data structure via a computer. In addition, the second data structure contains a list of errors and correct responses generated by a specific patient when performing tasks via a diagnostic cell and / or a treatment cell in the program. The patient's errors are classified into several categories via a computer, such as phonetic, semantic, morphological or lexical, and are graphically subdivided (such as morphology> affix> prefix) in the manner of a first data structure (such as a main matrix).

[0074] When the incorrect response of the patient is located on a node on one of the graphs in the first data structure (such as the main matrix), the connected vertices at the next lower level of this graph and the computer-implemented priority sorting algorithm determine the next diagnostic cell and / or treatment cell for this patient. The second data structure (such as the patient matrix) is continuously updated during the entire treatment process when the patient makes new errors and / or provides all correct responses to previously identified problems. The second data structure (such as the patient matrix) is used to track the progress of the patient, and the patient can access his progress report in real time in a suitable format (such as a spreadsheet document, a visual depiction document, a word processor document, or any combination thereof). The relevant information in the second data structure (such as the patient matrix) is sent in whole or in part to the learning analysis database for the performance of the analysis and monitoring program.

[0075] The technology is further implemented via a database storing the patient's correct responses and ungrammatical productions, based on which a learning analysis computer algorithm or similar method is used to monitor and improve the effectiveness of the technology for correcting language processing deficiencies. For example, the effectiveness of the program is enhanced by refining the stored patient data using learning analysis or similar methods (such as via a second data structure), the second data structure contains a lot of useful information, such as patient demographics, general error types (e.g., phonetic errors, morphological errors, lexical errors, semantic errors, or syntactic errors), success rates of different treatment cells, and the like. Efficient data management and retrieval of individual and group information by specified criteria yields useful insights for research and treatment enhancement. In addition, the learning analysis data may be stored in whole or in part in a computerized database as described herein with respect to at least one of the first data structure and the second data structure, or in another computerized database. In other embodiments, the learning analysis data is implemented in any manner (directly and / or indirectly) in multiple databases (such as relational or non-relational databases) that are similar or different from each other and are hosted remotely and / or locally to each other. The learning analytics data is remotely hosted / resides (such as) on a server computer rather than on a patient's computer as described herein, thereby allowing end users (such as patients) more flexibility while providing more efficient content updates due to this residency / hosting. However, it should be noted that the server computer may host the learning analytics data, the first data structure, and the second data structure, or the learning analytics data, the first data structure, and the second data structure may be hosted / reside on different server computers, as described herein. In addition, it should be noted that in other embodiments, the learning analytics data may be hosted / reside on the patient's computer locally, in whole or in part, directly and / or indirectly. This configuration may enable periodic updates (such as weekly or monthly) from a central data repository (such as a computerized database) in whole or in part, directly and / or indirectly. For example, in this configuration, the learning analytics data that is hosted / resides locally on the patient's computer in whole or in part is at least a partial copy of the learning analytics data that is remotely hosted / resides from the patient's computer. In further embodiments, the learning analytics data, the first data structure, and the second data structure are one data structure that can be hosted in whole or in part in any manner individually described herein with reference to at least one of the learning analytics data, the first data structure, and the second data structure.

[0076] A learning analysis database may be used in a back-end computer infrastructure that collects and stores most, if not all, results of patient performance history containing game play. A computer-implemented learning analysis algorithm running on the back-end infrastructure analyzes these results to determine themes such as general performance issues or errors, frequency of calling online assistants, frequency of game types and / or tasks performed, game types or tasks that are prone to repeated failures, time played by time of day and day of the week, duration of consecutive play, intervals between play, frequency of patient requests for assistant assistance, and the like. Based on these algorithmic analyses, changes to a first data structure (such as a master matrix) may be implemented to help enhance the speed and / or efficiency of patient progress. These changes may be implemented with the aid of a quick restart of the server computer to bring the new changes to all patients immediately. Alternatively, these changes may be implemented in real time on-site without restarting the server computer.

[0077] The data stored in the learning analysis database can be analyzed by computer algorithms running on the back-end computer infrastructure or viewed by system administrators by theme or diagram to quickly identify trends and / or troubleshoot. This algorithm can be designed to remotely alert system administrators of trouble spots via messages (such as email, text, chat, sound, vibration, visual cues, or other), thereby allowing for relatively rapid or real-time changes to at least one of the service provider or service requester techniques described herein. Similar mechanisms can be deployed to identify specific patients who need immediate or additional attention.

[0078] In some embodiments, the disclosed technology implements a computer-implemented error analysis algorithm. More specifically, the program executes the error analysis algorithm by processing the test output relative to a pre-specified predicted patient response set through a first data structure. For example, for a question requiring a patient to identify affixes in the word hunter, the predicted responses include 1) -er (correct), 2) hunt (incorrect), 3) none (incorrect), 4) hunter (incorrect), or 5) other (incorrect). Therefore, the error analysis algorithm needs to use the knowledge of the structure and processing (principles and rules) of natural language to analyze the learner's errors when the problem occurs to identify error patterns and track down the source of the error. As described herein, this error analysis algorithm is used to compare the patient's actual response with a target set of correct responses. For example, the error analysis generated by the patient can reveal that it is consistently unable to recognize common affixes, such as -er (hunter), -or (editor) and -tion (action). In this case, the results of the analysis algorithm are based at least in part on the word formation process (morphology); but this error analysis can cover any component of the language (phonetics, semantics, vocabulary, syntax) and can be based on any defined dimension. It should be noted that the error analysis algorithm can be run on a back-end computer infrastructure. In addition, it should be noted that the error analysis algorithm is adopted from the field of second language acquisition.

[0079] In some embodiments, the disclosed technology implements a computer-implemented prioritization algorithm. More specifically, weighted values ​​are assigned to most (if not all) categories and / or items in a first data structure (such as a master matrix) based on language frequency of occurrence, communicative function, significance of the impact of its omission or ungrammaticality, and the like. This value determines which diagnostic or therapeutic item or cell will be presented next. The prioritization algorithm will select the article the over the adjective unimportant because the latter occurs more frequently than the former. It should be noted that the prioritization algorithm can be run on a back-end computer infrastructure.

[0080] In some embodiments, the disclosed technology implements an online assistant function. More specifically, if the patient fails the evaluation test of the cell after several (such as two) attempts, or if the patient produces a reaction that is not predicted by the content of the first data structure, then the online assistant can intervene to help the patient proceed to the next stage. For example, this online assistant presents the form of an automatic program with advanced speech recognition, human-computer interaction and / or communication capabilities. Alternatively, this online assistant presents the form of communication with a real human operator such as in a call center. For example, this communication may include telephone communication, telephone conference, chat or personal visit, such as when the disclosed technology is implemented in an office-laboratory environment.

[0081] In some embodiments, for ease of access, the present disclosure implements patient diagnosis and / or patient treatment from an Internet-based website or online portal and runs a program on a cloud server. However, this delivery method can be supplemented and / or supplemented in whole or in part via a mobile application as described herein. A convenient feature of this deployment is that it is relatively simple to maintain the patient's profile and allow easy access to its diagnosis and / or treatment program through the network. It should be noted that different applications can be developed for different treatment shells (tasks). In addition, it should be noted that this deployment can allow temporary stopping and recovery of at least one of diagnosis and treatment. Therefore, this functionality can be used within a period of time (such as within a few months).

[0082] In some embodiments, the present disclosure implements multiple forms. More specifically, the diagnosis and / or treatment cells can use any one or combination of other forms including video, audio, text, graphics, animation, web conferencing, or new technologies. The patient can use electronic and / or physical support during speech pronunciation practice, such as front camera or rear camera, and other future technologies. The program can also use voice and / or speech recognition software to train and receive and analyze input from the patient. Biometric monitoring and feedback can be incorporated to increase the sensitivity of the program to the patient's performance and reaction, and thus increase the effectiveness of the program. This can be particularly useful for people with attention deficits. The sensing system can be installed with a program to collect biometric data, which can include eye gaze (to monitor concentration), pulse and blink rate (to monitor stress and mental fatigue), and lip movement (to monitor pronunciation). These sensors can provide biometric feedback to the program and to the patient. When the patient receives this information as instant feedback, the program-especially when delivered unobtrusively on a wearable or other portable device-can become seamlessly integrated into its daily life. In this embodiment, the program can provide real-time corrections to the patient's speech errors as they are made during their daily activities.

[0083] In some embodiments, the present disclosure implements a performance test / patient's performance history. More specifically, the program not only stores the patient's correct and incorrect responses in a second data structure (such as a patient matrix), but also displays their scores from all previously completed diagnostic and treatment cells in an easily accessible format. The patient's progress report can be continuously updated as they complete each cell. The program can display a comparison of the patient's performance history with the planned specified development path. This visual display encourages the patient to keep moving toward his or her specified goal - satisfactory removal of all identified language processing problems.

[0084] In some embodiments, the present disclosure implements a reward system. More specifically, diagnostic tests, assessment tests, and / or training practices can be delivered as age-appropriate interactive games. The patient's performance in each section can be scored and converted into reward points, which the patient can exchange for virtual or physical objects, multiplayer games or such earned time, or other forms of incentives to motivate the patient to complete his training satisfactorily. It should be noted that these rewards that can be operated based on a loyalty system can be convertible, such as converted into prizes, cash, goods, services, airline miles, additional treatment sessions, personal diagnosis and / or treatment sessions, etc.

[0085] Figure 1 A schematic diagram of an example implementation of a computer network model according to the present disclosure is shown. The computer network model 100 includes a network 102, a server 104, and a client 106. This distributed operation model distributes tasks / workloads between the server 104 that provides resources / services and the client 106 that requests resources / services. The server 104 and the client 106 illustrate different computers / applications, but in other embodiments, the server 104 and the client 106 reside in one system / application. In addition, in some embodiments, the model 100 requires that a large amount of resources be allocated to a small number of computers (such as the server 104), wherein the complexity of the client 106 depends on the amount of computation downloaded to the computer of the amount, i.e., more computation downloaded from the client 106 to the server 104 results in a lighter client 106, such as being more dependent on network sources and less dependent on local computing resources.

[0086] The network 102 includes a plurality of nodes, such as a collection of computers and / or other hardware interconnected via a plurality of communication channels, which allows sharing of resources and / or information. This interconnection may be direct and / or indirect. The network 102 may be wired and / or wireless. The network 102 may allow encrypted and / or unencrypted communication via short and / or long distances. The network 102 may operate via at least one network protocol, such as Ethernet, transmission control protocol (TCP) / Internet protocol (IP), etc. The network 102 may have any scale, such as a personal area network, a local area network, a home area network, a storage area network, a campus area network, a backbone network, a metropolitan area network, a wide area network, an enterprise private network, a virtual private network, a virtual network, a satellite network, a computer cloud network, the Internet, a cellular network, etc. The network 102 may be and / or may include an intranet and / or an extranet. The network 102 may be and / or may include the Internet. The network 102 may include and / or allow for communication with other networks, which may be subnetworks and / or different networks, the same and / or different from the network 102. The network 102 may include hardware, such as computers, network interface cards, repeaters, hubs, bridges, switches, extenders, and / or hardware-based and / or software-based firewalls. The network 102 may be operated directly and / or indirectly by and / or on behalf of one and / or more entities, regardless of any relationship to the contents of the present disclosure.

[0087] The server 104 may be hardware-based and / or software-based. The server 104 is a server computer and / or is directly and / or indirectly hosted on a server computer (whether fixed or mobile), such as a public telephone booth, a workstation, a transportation vehicle (whether land, sea or air), a desktop computer, a laptop computer, a tablet, a mobile phone, a mainframe computer, a supercomputer, a server farm, etc. The server computer may be touchscreen-enabled and / or non-touchscreen. The server computer may include and / or may be part of another computer system and / or a cloud computing network. The server computer may run any type of operating system (OS), such as And / or other. The server computer may include and / or may be directly and / or indirectly coupled to an input device, such as a mouse, keyboard, front and / or rear camera, accelerometer, touch screen, biometric reader, clicker and / or microphone. The server computer may include and / or may be directly and / or indirectly coupled to an output device, such as a display, speaker, headset, joystick, video game controller, vibrator and / or printer. In some embodiments, the input device and output device may be implemented in one unit. The server computer may include circuits for determining global positioning such as via a global positioning system (GPS), a signal triangulation system, etc. The server computer may be equipped with a near field communication (NFC) circuit. The server computer may directly and / or indirectly host, run and / or be coupled to a database (such as a relational database or a non-relational database), which may directly and / or indirectly feed data to the server 104.

[0088] The server 104 communicates with the network 102 via a server computer, such as directly and / or indirectly, selectively and / or non-selectively, encrypted and / or unencrypted, wired and / or wirelessly, via contact and / or contactless. This communication can be via a software application, a software module, a mobile application, a browser, a browser extension, an OS, and / or any combination thereof. For example, this communication can be via a public framework / application programming interface (API), such as Hypertext Transfer Protocol Secure (HTTPS).

[0089] The client 106 may be hardware-based and / or software-based. The client 106 is a patient computer and / or is directly and / or indirectly hosted on a patient computer (whether fixed or mobile), such as a terminal, a public telephone booth, a workstation, a transportation vehicle (whether land, sea or air), a desktop computer, a laptop computer, a tablet, a mobile phone, a mainframe computer, a supercomputer, a server farm, etc. The patient computer may be touch-screen capable and / or non-touch-screen. The patient computer may include and / or may be part of another computer system and / or a cloud computing network. The patient computer may run any type of OS, such as And / or other. The patient computer may include and / or may be coupled to input devices such as a mouse, keyboard, front and / or rear cameras, accelerometers, touch screens, biometric readers, clickers and / or microphones, and / or output devices such as displays, speakers, headsets, joysticks, video game controllers, vibrators and / or printers. In some embodiments, the input devices and output devices may be implemented in one unit. The patient computer may include circuits for global positioning determination such as via GPS, signal triangulation systems, etc. The patient computer may be equipped with NFC circuits. The patient computer may directly and / or indirectly host, run and / or be coupled to a database (such as a relational database or a non-relational database), which may directly and / or indirectly feed data to the patient 106.

[0090] The client 106 communicates with the network 102 via the patient computer, such as directly and / or indirectly, selectively and / or non-selectively, encrypted and / or unencrypted, wired and / or wirelessly, via contact and / or contactless. This communication can be via a software application, a software module, a mobile application, a browser, a browser extension, an OS, and / or any combination thereof. For example, this communication can be via a public framework / API, such as HTTPS.

[0091] In other embodiments, the server 104 and the client 106 may also communicate with each other, such as when hosted in one system or when in local proximity to each other, such as via a short-range wireless communication protocol (such as infrared or ) communicate directly with each other. This direct communication can be selective and / or non-selective, encrypted and / or unencrypted, wired and / or wireless, via contact and / or contactless. Since many of the clients 106 can initiate sessions with the server 104 relatively simultaneously, in some embodiments, the server 104 employs load balancing techniques and / or failover techniques to obtain operational efficiency, continuity and / or redundancy.

[0092] It should be noted that other computing models are possible. For example, these models may include decentralized computing, such as peer-to-peer (P2P), e.g. Or distributed computing, such as is done via a computer cluster, where a collection of networked computers cooperate so that the computers appear to function as a single system.

[0093] Figure 2A schematic diagram of an example implementation of a computer network architecture according to the present disclosure is shown. Computer network architecture 200 includes a network 202 that communicates with a service provider segment and with a service requester segment. The service provider segment includes a server computer 204 and a database 206. The service requester segment includes a workstation computer 208, a tablet computer 210, a desktop computer 212, a laptop computer 214, and a mobile phone 216. Architecture 200 operates according to model 100. However, in other embodiments, architecture 200 operates according to other computing models as described herein (such as direct communication, decentralized computing, distributed computing, and / or any combination thereof). Network 202 operates according to network 102. However, in other embodiments, network 202 operates according to other network types as described herein.

[0094] It should be noted that the service provider segment may include one or more server computers 204 and / or one or more databases 206, which are structurally and / or functionally the same and / or different from each other, are communicatively coupled to each other and / or are non-communicatively coupled to each other (such as directly and / or indirectly, wired and / or wirelessly, selectively and / or non-selectively, encrypted and / or unencrypted, via contact and / or non-contact), are synchronized and / or asynchronous, controlled via a single entity and / or via multiple entities, regardless of any relationship to the contents of the present disclosure. Likewise, it should be noted that the service requester segment may include less than five and / or more than five computers 208, 210, 212, 214, 216 that are structurally and / or functionally identical and / or different from one another, communicatively coupled to one another and / or non-communicatively coupled to one another (such as directly and / or indirectly, wired and / or wirelessly, selectively and / or non-selectively, encrypted and / or unencrypted, via contact and / or contactless), synchronized and / or non-synchronized, controlled via a single entity and / or via multiple entities, regardless of any relationship to the contents of the present disclosure.

[0095] The computer 204 communicates with the network 202, such as directly and / or indirectly, wired and / or wirelessly, selectively and / or non-selectively, encrypted and / or unencrypted, via contact and / or contactless, synchronously and / or asynchronously. The computer 204 facilitates this communication via a hardware unit (such as a hardware component of the computer 204, for example, a network card). However, in other embodiments, the computer 204 facilitates this communication via a software unit (such as a software application, a software module, a mobile application, a browser, a browser extension, an OS, and / or any combination thereof). For example, this communication can be via a public framework / API, such as HTTPS. Due to the size of the service requester segment, the computer 204 adopts load balancing technology and / or failover technology to obtain operational efficiency, continuity and / or redundancy.

[0096] The computer 204 is operatively coupled to the database 206, such that the computer 204 communicates with the database 206, such as directly and / or indirectly, wired and / or wirelessly, selectively and / or non-selectively, encrypted and / or unencrypted. The computer 204 facilitates this communication via a hardware unit (such as a hardware component of the computer 204, for example, a network card). However, in other embodiments, the computer 204 facilitates this communication via a software unit (such as a software application, a software module, a mobile application, a browser, a browser extension, an OS, and / or any combination thereof). For example, this communication can be via a common framework / API, such as via a database management system (DBMS) hosted on the computer 204 (such as or other suitable system). Furthermore, it should be noted that the computer 204 can host the database 206 locally and / or access the database 206 remotely. Alternatively, the computer 204 and the database 206 can be in one location but implemented differently. In addition, it should be noted that the computer 204 can host and / or be operatively coupled to one or more databases 206, such as directly and / or indirectly, wired and / or wirelessly, selectively and / or non-selectively, encrypted and / or unencrypted, via contact and / or contactless, synchronously and / or asynchronously. Furthermore, it should be noted that the database 206 can be hosted on more than one computer 204, such as directly and / or indirectly, wired and / or wirelessly, selectively and / or non-selectively, encrypted and / or unencrypted, via contact and / or contactless, synchronously and / or asynchronously.

[0097] Database 206 includes an organized data set. The data may be of any type, whether a basic type (such as Boolean and / or character), a composite type (such as an array and / or a union), and / or an abstract data type (such as a list, a queue, a deck, a stack, a string, a tree, and / or a graph). The data may be organized by any structure, such as a linear structure (such as an array, an image, a table, a matrix, a vector, and / or a list), a tree structure (such as a tree, a pagoda, a tree pile, a heap, and / or a dictionary tree), a hash structure (such as a table, a list, and / or a filter), a graphical structure (such as a graph, a matrix, a stack, and / or a diagram), and / or any combination of any of them. The organized data set may contain content, such as patient information, shell information for a language-related disorder, cell information for a language-related disorder, patient matrix information, master matrix information, analysis information, and / or other related information. Database 206 is accessed via computer 204 (such as via a DBMS running on computer 206). Database 206 is a relational database, but other database models are possible, such as post-relational. It should be noted that although the computer 204 and the database 206 are located distinct from each other, in other embodiments, the computer 204 hosts the database 206. It should be noted that the computer 204 and the database 206 are operated via a single actor, but in other embodiments, the computer 204 and the database 206 are operated via different actors. Additionally, it should be noted that the database 206 can be in communication with the network 202, such that the computer 204 communicates with the database 206 via the network 202.

[0098] The workstation computer 208, tablet computer 210, desktop computer 212, laptop computer 214, and mobile phone 216 communicate with the network 202, such as directly and / or indirectly, wired and / or wirelessly, selectively and / or non-selectively, encrypted and / or unencrypted, synchronously and / or asynchronously, on demand and / or non-on demand. The workstation computer 208, tablet computer 210, desktop computer 212, laptop computer 214, and mobile phone 216 facilitate this communication in any combination via hardware units (such as hardware components of the workstation computer 208, tablet 210, desktop computer 212, laptop computer 214, and mobile phone 216, for example, transceivers and / or network cards). In other embodiments, the workstation computer 208, tablet computer 210, desktop computer 212, laptop computer 214, and mobile phone 216 facilitate this communication via software units (such as software applications, software modules, mobile applications, browsers, browser extensions, OS, and / or any combination thereof). For example, this communication may be via a public framework / API, such as HTTPS. Additionally, it should be noted that other types of service requesters are possible, such as stand-alone cameras, automated teller machines (ATMs), cryptocurrency mining machines, public telephone kiosks, terminals, wearable computers (such as eyeglass computers), implanted computers, or other suitable computing devices.

[0099] It should be noted that at least two of the workstation computer 208, tablet computer 210, desktop computer 212, laptop computer 214, and mobile phone 216 can communicate in the same manner and / or in different manners in parallel and / or non-parallel via network 202. Additionally, it should be noted that the workstation computer 208, tablet computer 210, desktop computer 212, laptop computer 214, and mobile phone 216 are operated via different actors, but in other embodiments, at least two of the workstation computer 208, tablet computer 210, desktop computer 212, laptop computer 214, and mobile phone 216 are operated via a single actor.

[0100] The service provider segment serves data to the service requester segment via the network 202. This service may be via push technology and / or pull technology. For example, the push technology enables request initiation via the service provider segment (such as via the computer 204). Thus, information that may be updated periodically may be pushed to the service requester segment via the computer 204 (such as via synchronous conferencing, messaging, and / or file distribution). Also, for example, the pull technology enables request initiation via the service requester segment (such as via the mobile phone 216). Thus, information may be pulled from the service provider segment via the mobile phone 216 (such as via web browsing and / or web feeds).

[0101] In one mode of operation, language-related disorder diagnostic data and / or treatment data based on the diagnostic data are provided to the service requester segment via the service provider segment via the network 202. For example, the computer 204 feeds the diagnostic data and / or treatment data from the database 206 as needed to the mobile phone 216 operated by the language-related disorder patient. The computer 204 receives patient responses from the mobile phone 216 and dynamically and repeatedly processes these responses to obtain a more refined diagnosis and / or treatment. The operator of the computer 204 and / or the database 206 can control the manner in which this feeding occurs (such as via patient subscription), and / or update the diagnostic data and / or treatment data, such as based on data dynamically and repeatedly obtained from other language-related disorder patients.

[0102] Figure 3 A schematic diagram of an example implementation of a computer network diagram according to the present disclosure is shown. Computer network diagram 300 includes network 302, computer system 304, operator computer 306, and patient computer 308. Network 302 operates according to network 202, but other network types are possible, as described herein. The service provider segment includes a system 304 such as for network-based language-related disorder diagnosis and / or treatment, which is used as a network-based telemedicine service. System 304 communicates with network 302, as described herein. Operator computer 308 (such as at least one of workstation computer 208, tablet computer 210, desktop computer 212, laptop computer 214, and mobile phone 216) can communicate with system 304, such as for controlling the manner in which feeds occur (such as via patient subscription), and / or such as updating diagnostic data and / or treatment data based on data dynamically obtained from other language-related disorder patients repeatedly. The service requester segment includes a patient computer 308, such as at least one of a workstation computer 208, a tablet computer 210, a desktop computer 212, a laptop computer 214, and a mobile phone 216. In another embodiment, the operator computer 306 and the patient computer 308 are a single computer. The operator computer 306 and the patient computer 308 communicate with the network 302 as described herein. In addition, the operator computer 306 can be configured to provide online assistant functions, in whole or in part, to the patient computer 308, directly or indirectly (such as via a telephone call, teleconferencing, chat, or other).

[0103] Figure 4A schematic diagram of an example embodiment of a computer according to the present disclosure is shown. The computer 400 includes a processor 402, a memory 404 operatively coupled to the processor 402, a network communication unit 406 operatively coupled to the processor 402, a camera 408 operatively coupled to the processor 402, a display 410 operatively coupled to the processor 402, a speaker 412 operatively coupled to the processor 402, a geo-positioning unit 414 operatively coupled to the processor 402, a graphics unit 416 operatively coupled to the processor 402, and a microphone 418 operatively coupled to the processor 402. The computer 400 includes a power supply 420 that powers the processor 402, the memory 404, the network communication unit 406, the camera 408, the display 410, the speaker 412, the geo-positioning unit 414, the graphics unit 416, and the microphone 418. Although at least two of the processor 402, memory 404, network communication unit 406, camera 408, display 410, speaker 412, geo-positioning unit 414, graphics unit 416, microphone 418, and power supply 420 are implemented in one unit, at least one of the processor 402, memory 404, network communication unit 406, camera 408, display 410, speaker 412, geo-positioning unit 414, graphics unit 416, microphone 418, and power supply 420 may be operatively coupled to the computer 400 when being independent, such as locally or remotely, directly or indirectly. Additionally, in other embodiments, the computer 400 lacks at least one of the network communication unit 406, camera 408, display 410, speaker 412, geo-positioning unit 414, graphics unit 416, and microphone 418. It should be noted that the computer 400 may include other units, whether input units and / or output units, such as a biometric reader, a clicker, a vibrator, a printer, and the like.

[0104] Processor 402 includes a hardware processor, such as a multi-core processor. For example, processor 402 includes a central processing unit (CPU).

[0105] Memory 404 includes computer-readable storage media, which may be non-transitory. The media stores a plurality of computer-readable instructions, such as software applications, for execution via processor 402. The instructions instruct processor 402 to facilitate execution of the method for diagnosing and / or treating speech-related disorders as described herein. Some examples of memory 404 include volatile storage units, such as random access memory (RAM), or non-volatile storage units, such as hard disk drives or read-only memories (ROM). For example, memory 404 includes flash memory. Memory 404 communicates with processor 402 by wire. Moreover, for example, memory 402 stores a plurality of computer-readable instructions, such as a plurality of instruction sets, for operating at least one of network communication unit 406, camera 408, display 410, speaker 412, geolocation unit 414, graphics unit 416, microphone 418 or other input and / or output units.

[0106] The network communication unit 406 includes a network interface controller for wired or wireless, direct or indirect computer network communication. For example, the network communication unit 406 includes hardware for computer networking communication based on at least one standard selected from the set of the Institute of Electrical and Electronics Engineers (IEEE) 802 standards (such as the IEEE 802.11 standard). For example, the network communication unit 406 includes a wireless network card operating according to the IEEE 802.11 (g) standard. The network communication unit 406 communicates with the processor 402 by wire.

[0107] The camera 408 includes a lens for image capture (such as photos and / or videos). The camera 408 stores the captured visual information on the memory 404, which may be in a compressed format or an uncompressed format. The camera 408 may allow image display on the display 410, such as before, during and / or after image capture. The camera 408 may include a flash illumination unit. The camera 408 may allow optical or software-based zooming. The camera 408 communicates with the processor 402 by wire. The camera 408 may also be remotely coupled to the processor 402, such as wirelessly.

[0108] The display 410 includes an area for displaying visual and / or tactile information. The display 410 includes at least one of an electronic visual display, a flat panel display, a liquid crystal display (LCD), and a stereoscopic display. For example, the display 410 includes a computer monitor with touch functionality. The display 410 communicates with the processor 402 by wire. The display 410 may also be remotely coupled to the processor 402, such as wirelessly.

[0109] Speaker 412 includes a loudspeaker, such as an electroacoustic transducer that provides sound in response to an electrical audio signal input. Speaker 412 is, for example, a dynamic speaker. Speaker 412 communicates with processor 402 by wire. Speaker 412 may also be remotely coupled to processor 402, such as wirelessly.

[0110] The geolocation unit 414 includes a GPS receiver. The geolocation unit 414 communicates with the processor 402. It should be noted that other types of geolocation are possible, such as triangulation via cell site signals. The geolocation unit 414 may also be remotely coupled to the processor 402, such as wirelessly.

[0111] The graphics unit 416 includes a graphics processing unit (GPU) for image processing. The graphics unit 416 is a graphics-dedicated unit, but in other embodiments, the processor 402 is integrated with the graphics unit 416. For example, the graphics unit 416 includes a video card. The graphics unit 416 communicates with the processing unit 402 by wire.

[0112] Microphone 418 comprises an acoustic-electric transducer / sensor that operates to convert sound in the air into an electrical signal for subsequent use (such as output via speaker 412). Microphone 418 can be based on electromagnetic induction, based on capacitance change, or based on piezoelectricity. Microphone 418 can be coupled to a preamplifier upstream of an audio power amplifier. For example, microphone 418 is a dynamic microphone. Microphone 418 can also be remotely coupled to processor 402 (such as wirelessly).

[0113] The power supply 420 supplies power to the computer 400. The power supply 420 includes at least one of an onboard rechargeable battery, such as a lithium-ion battery, and an onboard renewable energy source, such as a photovoltaic cell, a wind turbine, and / or a water turbine. It should be noted that this power supply may be via mains electricity, such as via a power cable.

[0114] It should be noted that the computer 400 may also include and / or be operatively coupled to at least one input device, such as a computer keyboard, a computer mouse, a touchpad, a clicker, a scanner, a fax, a biometric reader, a pointer, or other suitable input device. Likewise, the computer 400 may include and / or be operatively coupled to at least one output device, such as a printer, a projector, or other suitable output device. In addition, at least one of the computer 204, the workstation computer 208, the tablet 210, the desktop computer 212, the laptop computer 214, and the mobile phone 216 may be constructed according to the schematic diagram of the computer 400.

[0115] Figure 5A flowchart of an example implementation of a diagnostic process based on a generative model according to the present disclosure is shown. A process such as a computer implemented via at least one of a service provider segment and a service requester segment uses multiple diagnostic shells 502, 506, 514. The process further uses multiple diagnostic cells 504, 508, 510, 512, 516. It should be noted that any number of shells or cells can be used in any combination. The process on which the program operation can be based uses patient information in a first data structure (such as a master matrix) and a second data structure (such as a patient matrix) to select a first diagnostic shell 502 (Dx1, represented by a dotted rectangle) for activation. Each diagnostic shell specifies a specific task to be performed by the patient, such as matching pronunciation with letters or symbols. The program generates at least one content-specific test cell 504 (diagnostic cell Dx1y) from a diagnostic shell 502 (Dx1) by inserting selected information from a first data structure (such as a master matrix) into the shell 502. 1-n ). It should be noted that this computerized generation can be parallel, simultaneous, real-time, concurrent, batch-processed, multi-threaded, or any negative thereof, such as non-parallel, non-simultaneous, and the like. For example, if the diagnostic shell Dx1 specifies the task of naming objects shown on a computer display, then particular diagnostic cells Dx1y1, Dx1y3, Dx1y5, and the like may involve each naming a different class of objects, such as household objects, tools, or things whose names begin with the letter "B". Cells 504 are implemented to the patient in an order that is optimal (such as most effective) for this particular patient based on a specified set of criteria in a first data structure (such as a master matrix). Cells 504 may be implemented in a non-sequential order because the present disclosure does not necessarily specify a preset sequence of diagnoses and treatments.

[0116] Additionally, Dx1y1 is followed by Dx1y3 and Dx1y5. This does not necessarily mean that Dx1y1, Dx1y3, and Dx1y5 are predefined cells with predefined content provided in a predefined order. Rather, these cells are generated after analyzing the results for each patient. Thus, the process may deploy a different set of diagnostic cells, or a different sequence of those same sets of cells, or skip this shell entirely, based on this same diagnostic shell for another patient, depending on at least one underlying language problem identified.

[0117] In addition, the process enables the generation of diagnostic cells 504 (Dx1y1, Dx1y3, Dx1y 5) , but the patient undergoes these tests sequentially in an order dynamically set via a first data structure (such as a master matrix). Patient responses based on cells 504 enable automatic selection of diagnostic shells 506 (Dx 2)to generate the next test in cell 508 (Dx2y1). The patient's response to cell 508 (Dx2y1) results in the automatic generation of yet another cell 510 (Dx2y2) for further testing. The patient's results for cell 510 (Dx2y2) then enable the process to automatically return to selecting shell 502 again for further more refined testing, but this time retrieving new content from the first data structure (such as the main matrix) to generate diagnostic cell 512 (Dx1y4). Based on its results, the patient then proceeds to the next diagnostic shell 514 (Dx3) and then cell 516 (Dx3y2) in an order dynamically set via the first data structure (such as the main matrix).

[0118] Figure 6 A flowchart showing an example implementation of a generative model-based treatment process according to the present disclosure. Each shell in the treatment phase contains a training unit (Tx 1-n ) and evaluation units (Ex 1-n ). The process that program execution may be based on first uses the patient's performance history from the diagnosis phase as contained in a second data structure (such as a patient matrix) to generate treatment cells for treatment based at least in part on the treatment shells via comparing, analyzing or filtering the second data structure (such as a patient matrix) relative to the first data structure (such as a master matrix). As the information in the second data structure (such as a patient matrix) is continuously updated, the patient's performance with respect to earlier treatment cells then determines the subsequent cells to be implemented. The cells in the treatment phase are generated from their corresponding shells in the same or similar manner as in the diagnosis phase. It should be noted that any number of shells or cells may be used.

[0119] like Figure 6 As shown, the process captures a portion of a treatment phase for a patient. This process can be implemented by a computer via at least one of a service provider segment and a service requester segment. At one stage of the process described, such as via Figures 1 to 4The computer architecture in implements a treatment cell 602 (Tx1y3Ex1y3). In order to proceed to the next cell, the patient has to pass the assessment contained in this cell 602 according to the criteria specified by the first data structure (such as the main matrix). If the patient passes, the process continues to generate the next cell 606 (Tx3y2Ex3y2) from the shell Tx3Ex3 by retrieving the selected content from the first data structure (such as the main matrix) again. However, if the patient cannot pass the assessment of the cell 602 (Tx1y3Ex1y3), the process generates a practice cell (Px1y1) in accordance with box 604 to help the patient build the skills required to pass. Repeated failures require more practice cells (Px1y3) or other treatment cells (Tx4y1Ex4y1). Other treatment cells (Tx4y1Ex4y1) can be deployed to help this patient build the skills required to pass the initial treatment cell (Tx1y3Ex1y3). If the patient is still unable to pass the same cell Tx1y3Ex1y3 after several (such as two) attempts, then the online assistant function as described in this article is invoked.

[0120] The patient proceeds to the generated treatment cell 606 (Tx3y2Ex3y2) upon successfully performing the task based on cell 602 as evaluated relative to a second data structure (such as a patient matrix). Similarly, if the patient passes, the patient proceeds to the next cell generated, namely cell 608. However, if the patient fails, a similar process occurs as with the loop involving cell 602 (Tx1y3Ex1y3). However, different failure loops for different cells may require different types of cells and different presentation sequences thereof. For example, in Figure 6, for the first failed loop involving cell 602 (Tx1y3Ex1y3), two practice cells (Px1y1, Px1y3) are implemented before another treatment cell (Tx4y1Ex4y1). Although the process may deploy similar or different types of practice cells to assist the patient depending on the identified problem, for the second failed loop involving cell 606 (Tx3y2Ex3y2), the treatment cell (Tx4y2Ex4y2) is implemented before the practice cell in accordance with box 610. However, cell 608 is implemented immediately after successfully passing the task based on cell 606. In some cases, it is appropriate to provide at least some answers before the patient attempts the same assessment for the last time before proceeding to the next treatment cell. However, in other cases, it is appropriate to provide an instructional video or other assistance after a failed attempt. In addition, as the process continues, the patient's responses are recorded in a second data structure (such as a patient matrix) as the patient passes through new treatment cells. Based on this record, the process enables modification and enhancement of the patient's treatment, continuously updating the type and order of cells to be activated next based on information from a first data structure, such as a master matrix. At least one of the diagnostic phase and the therapeutic phase is thus based on a reactive generative model that is used to establish personalized diagnostic and therapeutic cells for each patient.

[0121] Figure 7 Flowchart of an example embodiment of a diagnostic process according to the present disclosure is shown. The process includes at least a plurality of blocks 702 to 726. In addition, the process may be performed in a sequential numerical order and / or a non-sequential numerical order. The process is performed as described herein (such as Figures 1 to 4 The process may be performed by a computing architecture (such as a computer system 304) in a patient computer 308. The process may be performed, facilitated for performance, and / or assisted in such performance by at least one actor (such as a service provider) either domestically and / or internationally. For example, the process may be performed by a computer system 304 interfaced with a patient computer 308.

[0122] In block 702 , the system 304 obtains first filtered data, such as via comparing, analyzing, or filtering a second data structure (such as a patient matrix) relative to a first data structure (such as a master matrix).

[0123] In block 704 , the system 304 selects a diagnostic shell based at least in part on the first filtered data from block 702 .

[0124] In block 706 , system 304 generates a diagnostic cell based at least in part on the diagnostic shell from block 704 .

[0125] In block 708, the system 304 executes the diagnostic cell selected in block 706. This execution may include interfacing with the patient computer 308, such as via network communications with the patient computer 308. The patient computer 308 executes the diagnostic cell to receive patient input.

[0126] In block 710, the system 304 stores the results of the diagnostic cells in a second data structure, such as a patient matrix. This storage, which may be dynamic, is based at least in part on receiving at least some of the patient inputs from the patient computer 308 in real time or in a delayed manner.

[0127] In block 712, the system 304 stores the statistics in an analysis data structure, such as a computerized database. This storage, which may be dynamic, is based at least in part on obtaining the statistics from the results stored in the second data structure. Alternatively, the statistics may be obtained based at least in part on receiving at least some of the patient inputs from the patient computer 308 in real time or in a delayed manner.

[0128] In block 714 , the system 304 obtains second filtered data, such as via comparing, analyzing, or filtering the second data structure (such as the patient matrix) as updated in block 710 relative to the first data structure (such as the master matrix).

[0129] In block 716, system 304 analyzes patient performance based at least in part on the second filtered data from block 714. This analysis enables a determination as to whether another diagnostic cell from the same shell as block 704 needs to be used. Additionally, this analysis is based at least in part on factors as described herein.

[0130] In block 718, the system 304 determines whether the generation of the next diagnostic cell should occur. If so, the process continues to block 724. Otherwise, the process continues to block 720.

[0131] In block 720, the system 304 determines whether selection of the next diagnostic shell should occur. If so, the process continues to block 726. Otherwise, the process continues to block 722.

[0132] In block 722, the system 304 determines that the patient diagnosis is complete. This determination may include an output to the patient computer 308. For example, this output may be visual, auditory, vibrational, or other.

[0133] In block 724, the system 304 generates a next diagnostic cell based at least in part on the diagnostic shell from block 704. It should be noted that the contents of this cell are generated based on the first data structure, such as the primary matrix.

[0134] In block 726, the system 304 selects the next diagnostic shell for activation. It should be noted that this process goes through as many cycles as necessary until information from the first data structure and the second data structure indicates that the diagnostic test phase for this patient is complete, wherein the process may optionally continue to the treatment phase. In addition, it should be noted that since most of the steps are performed via the system 304, the patient computer 308 is lighter on resources for other background tasks. It should be noted that the diagnostic phase can be paused or resumed at any time via the patient.

[0135] Figure 8 802 to 826. In addition, the process can be performed in a sequential numerical order and / or a non-sequential numerical order. The process is performed as described herein (such as Figures 1 to 4 The process may be performed by a computing architecture (such as a computer system 304) in a patient computer 308. The process may be performed, facilitated for performance, and / or assisted in such performance by at least one actor (such as a service provider) either domestically and / or internationally. For example, the process may be performed by a computer system 304 interfaced with a patient computer 308.

[0136] In block 802 , the system 304 obtains first filtered data, such as via comparing, analyzing, or filtering a second data structure (such as a patient matrix) relative to a first data structure (such as a master matrix).

[0137] In block 804 , system 304 selects a treatment shell based at least in part on the first filtered data from block 802 .

[0138] In block 806 , system 304 generates treatment cells based at least in part on the treatment shell from block 804 .

[0139] In block 808, the system 304 executes the therapy cell selected in block 806. This execution may include interfacing with the patient computer 308, such as via network communication with the patient computer 308. The patient computer 308 executes the therapy cell to receive patient input. For example, the therapy cell may be Tx i y j Ex i y j .

[0140] In block 810, the system 304 stores the results of the treatment cell in a second data structure, such as a patient matrix. This storage, which may be dynamic, is based at least in part on receiving at least some of the patient inputs from the patient computer 308 in real time or in a delayed manner.

[0141] In block 812, the system 304 stores the statistics in an analysis data structure, such as a computerized database. This storage, which may be dynamic, is based at least in part on obtaining the statistics from the results stored in the second data structure. Alternatively, the statistics may be obtained based at least in part on receiving at least some of the patient inputs from the patient computer 308 in real time or in a delayed manner.

[0142] In block 814 , the system 304 obtains second filtered data, such as via comparing, analyzing, or filtering the second data structure (such as the patient matrix) as updated in block 810 relative to the first data structure (such as the master matrix).

[0143] In block 816, system 304 analyzes patient performance based at least in part on the second filtered data from block 814. This analysis enables a determination as to whether another treatment cell from the same shell from block 804 needs to be used. Additionally, this analysis is based at least in part on factors as described herein. It should be noted that the new and previous information input into the second data structure is compared, analyzed, or filtered by the first data structure to determine whether the same shell Tx from block 804 needs to be used. i Ex i Another therapeutic cell.

[0144] In block 818, the system 304 determines whether the generation of the next treatment cell should occur. If so, the process continues to block 820. Otherwise, the process continues to block 824.

[0145] In block 824, the system 304 determines whether selection of the next treatment shell should occur. If so, the process continues to block 822. Otherwise, the process continues to block 826.

[0146] In block 820, system 304 generates a next treatment cell based at least in part on the treatment shell from block 804. Note that the contents of this cell are generated based on a first data structure, such as a master matrix.

[0147] In block 822, the system 304 selects the next treatment shell for activation. It should be noted that this process goes through as many cycles as necessary until information from the first data structure and the second data structure indicates that the treatment phase for this patient is complete. In addition, it should be noted that since most of the steps are performed via the system 304, the patient computer 308 is lighter on resources for other background tasks.

[0148] In block 826, the system 304 determines that the patient treatment is complete for at least one treatment session or one treatment action. This determination may include an output to the patient computer 308. For example, this output may be visual, auditory, vibratory, or other. It should be noted that the treatment session may be paused or resumed by the patient at any time.

[0149] Fig. 9 The diagram of the example embodiment of diagnosis and treatment according to the present disclosure is shown.The stimulus item in the word segmentation diagnostic test is the word be.In this test, the patient must separate the word into its individual pronunciation.The corresponding node for this stimulus item in the first data structure (such as the main matrix) is connected to its correct response ( / bi / ) and the incorrect response of prediction.Incorrect response comprises the knowledge of the process of the principle and rule of natural language, linguistics, as the basis of language acquisition or development, and the process of governing abnormal language (that is, the speaker outside the typical population, such as the language of the individual with communication disorder) and most (if not all) possible examples predicted.In the case of the stimulus word be, the incorrect response of prediction comprises the example that the user omits the consonant / b / or replaces this consonant or vowel / i / .The knowledge in the domain set identified above at least helps to predict, assuming that if / b / is replaced, then the phoneme that is likely to be replaced because of the phonetic similarity is / p / or / d / .Graphics also allow other possible substitutions.In addition, the specific incorrect response of the patient (such as no / b / ) leads to a specific computer operation (generating a word with / b / ). The generated words can be real words from the dictionary, or nonsense words with the desired pronunciation combination. When the patient's incorrect response is a substitution error (such as / p / for / b / ), the computer generates at least one pair of words with a critical contrast, such as the minimal contrast pit. With bit If the patient's response contains more than one error, most (if not all) of the affected nodes on the graph are activated, which means that the patient must practice with more than one word list during the treatment phase. When this word list is placed in the listen and repeat treatment shell, the patient performs the required task by completing each word in this list one by one. In addition, since the complete test usually contains several stimulus items, this test can produce more than one incorrect response from the patient. Therefore, several graphics can be activated from one test output in the first data structure (such as the main matrix). For example, a rapid naming diagnostic test can produce errors involving several words. In this case, a computer-implemented priority sorting algorithm determines which test or practice item set is presented to the patient next.

[0150] Fig.10A diagram showing an example implementation of a phoneme recognition diagnostic shell and cell for consonants according to the present disclosure. This method is implemented via the techniques described herein.

[0151] In one mode of operation, such as based on Figures 1 to 4 The computational architecture of the diagnostic phase is constructed to provide a comprehensive profile of the patient's ability to process linguistic information as a speaker, listener, reader, and writer. The infrastructure is designed to allow the patient to proceed in an orderly manner through most, if not all, of the language components that are important for playing these roles, with built-in mechanisms to confirm the accuracy of the assessment. For example, the process begins with a phoneme recognition diagnostic shell. The first phoneme recognition diagnostic cell generated (cell A) tests the patient's ability to distinguish between / p / and / b / in word position, such as Fig.10 As shown. The phonemes / p / and / b / are distinguished from each other by the voiced phonetic feature. Words distinguished by only one phonetic feature are symmetrically paired as minimal pairs, such as staple and stable. When implementing the phoneme recognition diagnostic cell A with the help of the / p / - / b / contrast, the patient hears strings of words containing / p / or / b / in random order, such as staple, clamber, or flappy. The patient is asked to pick out only the words with / p / . If the patient does not make any errors in this cell, the process generates a / pr / - / br / contrast in the word-initial position (such as pr im or br im). If the patient again makes no error, the process proceeds in sequence to the next phoneme contrast, / t / - / d / . If the patient makes an error in cell B, the process generates cells D through F to test their ability to distinguish / p / - / b / in word-initial and word-final positions and to recognize at least one previous error in word-medial position.

[0152] If the patient makes an error in cell A, the process generates the next cell C, which again has the / p / - / b / contrast in the word-medial position, to confirm the error in cell A. If the patient does not make an error in cell C, the process enables the generation of cell G to test its ability to distinguish / pr / - / br / in the word-initial position. Cells B and G both test the same / pr / - / br / contrast in the word-initial position, but are marked differently to indicate the different paths the patient takes through the process and to emphasize the fact that the process generates each cell with different content, because if the patient makes an error in cell C, the process first generates cells H and I containing / p / - / b / contrasts in the word-initial and word-final positions before generating cell J with the / pr / - / br / contrast in the word-initial position. One reason for presenting the word-medial contrast before the word-initial is because it is expected that the word-medial is more difficult to detect than the word-initial. Therefore, if the patient is able to perform the more difficult task first, the easier task does not have to be implemented.

[0153] Other contrasts related to the phoneme / p / , such as / p / - / t / ( p ip, t ip)、 / p / - / f / ( p it, f it)、 / p / - / k / ( p ick, k ick) and / p / Consonant clusters in initial and final positions (such as / pl / - / bl / ( pl ume、 bl oom)、 / pr / - / tr / ( pr ick, tr ick)、 / spl / - / sl / ( spl ay, sl ay)) can be similarly incorporated into the diagnostic phase via more diagnostic cells.

[0154] It should be noted that the phoneme recognition task can be incorporated with a measure of processing speed. In this case, the rapid processing diagnostic shell is used. The phoneme recognition task remains as described above, but now the patient performs it under practical constraints. For example, the patient detects phoneme contrasts auditorily at the normal rapid speed of natural speech. If the patient cannot detect most contrasts at this speed, then those same tasks are given at a slower speed (such as 60-80 words per minute). The obvious difference between their performance at normal speed and slow speed will point to a processing speed problem.

[0155] The process is further uniquely structured to ensure efficiency, thereby revealing the order of presentation of important diagnostic cells. The more subtle contrasts are tested before the more obvious contrasts. For example, only if the patient does not pass / o / - Comparison (h o le、h a ll) about the contrast between / o / and / au / (h o le、h ow l) Testing patients, since patients who can detect the former are likely to also be able to detect the latter, and therefore do not have to test the easier comparison.

[0156] Most, if not all, of the results from the diagnostic cells are recorded in a second data structure (such as a patient matrix) for later generation of treatment cells. This part of the diagnostic phase covers most, if not all, phonemes and phoneme groups in the language. At the end of this diagnosis, the process produces a basic, if not complete, speech profile for the patient that specifies which phonemes and phoneme groups are problematic in which articulation environments.

[0157] Diagnostic cells can also be incorporated into the treatment phase. For example, more complex consonant clusters (such as / spr / - / skr / ) can be tested at the end of a treatment phase for / p / to see if there are still residual problems with this phoneme.

[0158] The process can proceed from the phoneme recognition diagnostic shell to the pronunciation-symbol matching diagnostic shell. The process can use the results of the phoneme recognition diagnostic cells to generate cells for pronunciation-symbol matching. For example, if the patient makes an error in the phoneme recognition shell regarding the / v / - / w / comparison, the process can generate a pronunciation-symbol matching cell that requires the patient to enter (such as via typing) letters that match the recorded pronunciations / v / and / w / . Thus, the pronunciation-symbol matching cell can be used to confirm the results of the previously implemented phoneme recognition cells.

[0159] The process can be carried out from the phoneme recognition diagnostic shell to the syllable division diagnostic shell and to the word segmentation diagnostic shell separately. Any word containing these specific phonemes can be generated in the syllable division diagnostic cell according to the pronunciation that is found to be problematic for the patient from the earlier phoneme recognition test. For example, if the patient makes an error in the phoneme recognition cell involving / p / in the middle and final positions of the word, the syllable division diagnostic cell can include the word harpsichord to see if the patient retains the / p / pronunciation when splitting this word into its syllables. If the patient omits / p / , the subsequent word segmentation diagnostic cell can include more words with / p / in these positions, such as mishaps and capsized. Then ask the patient to separate these words into individual pronunciations, thereby giving the finer details of his speech ability. When performing these tasks in the syllable division and word segmentation shells, new errors can trigger the program to run the phoneme recognition cell again for the phonemes that were not initially found to be problematic in certain pronunciation environments.

[0160] Other types of speech tests can also be performed based on phoneme recognition (including rhyme). In particular, when the patient makes mistakes with vowels in the phoneme recognition shell, the process can generate rhyme diagnostic cells, such as when the patient cannot detect the previously compared vowels in the phoneme recognition shell. - / i / (such as b ee r and b ee ) when patients are asked to produce Rhyming words. To perform this task, they can record their words using a microphone, type in, or select rhyming word pairs.

[0161] The diagnostic phase can proceed from pronunciation-symbol matching to word-level tests, such as vocabulary acquisition and vocabulary retrieval. If the results of the phoneme recognition and pronunciation-symbol matching diagnostics suggest that the patient has difficulty processing the phoneme / b / in certain pronunciation environments, then the process can generate a vocabulary acquisition diagnostic cell, which requires the patient to produce as many words as possible with / b / pronunciation in the initial, middle or final position of the word by recording or typing within a set time. The process first scores the number of words that the patient can produce within the time limit to determine whether there is a problem in learning words with / b / itself. The process then matches its production with the stored words to further detect the problem and obtain additional details about its processing difficulties. For example, if she misspells blubber as blummer, then this error confirms that she has difficulty with the / b / - / m / contrast in the middle position of the word. However, if she spells bubble correctly in the same task, then this result will suggest that her problem may be limited to the / b / - / m / contrast in the middle position of the word when the word ends with / r / . Simply put, follow-up diagnostic tests during the process can confirm and generate finer details about the patient's treatment issue.

[0162] The vocabulary retrieval diagnostic shell can be after or before the vocabulary acquisition shell. When after the vocabulary acquisition shell, the vocabulary retrieval diagnostic cell can be used to track in more detail a portion of the patient's mental lexicon (the network of stored words in the brain) that has been adversely affected by processing difficulties at the phonological level. For example, the vocabulary retrieval diagnostic cell may ask the patient to name the objects pear, rail, and crest based on the images on the screen. If the patient records her answers as pail, wail, and quest, then these results suggest not only that she has difficulty with the phoneme / r / in all positions, but also that the words she has stored with / r / are ungrammatical because of interference from / l / and / w / .

[0163] The patient can proceed from these word-level diagnostic cells to others, such as morpheme recognition. For example, the process generates a morpheme recognition diagnostic cell that requires the patient to type a word with the suffix –er based on her previous errors in phoneme recognition, syllable segmentation, and / or word segmentation regarding the phoneme / r / .

[0164] Another important word-level test is rapid naming. For example, the process generates rapid naming cells that require the patient to read the words displayed on the display as they appear. If the / i / - / I / vowel was found to be problematic in the previous diagnostic cells, the words used may involve / i / - / I / vowels, such as beat and bit. When incorporated into the rapid processing shell, the program requires the patient to quickly and automatically recognize words at the speed necessary for fluent reading.

[0165] Other diagnostic shells may involve sentence level and paragraph level testing as described in the sample diagnostic shell section. Moreover, shells may involve reading or writing. In addition, some diagnostic cells are sensitive to the patient's age, while others are not. For example, it is expected that most (if not all) patients have acquired a complete list of the phonemes of a language. However, at least one word used in the cell may be selected based on age.

[0166] During the treatment phase, the treatment shell can be used as the diagnostic shell and vice versa.Three treatment shells that are novel to the art are described below: Phoneme Discrimination, Rapid Word Recognition, and Word Amplification as implemented herein.

[0167] Regarding the treatment cell, problems found via the phoneme recognition shell in the diagnostic phase can lead to the program activating the phoneme discrimination treatment shell. In the phoneme discrimination shell, the smallest pair of words (e.g., pleats, bleats) is used for the phoneme contrast of interest ( / p / - / b / ). The two words are shown on the display. The audio recording plays the two words randomly at a set speed when the patient selects the word to be read at that point in time by clicking or other input components. The patient can start at a slow speed of 60 words per minute (wpm) and proceed to faster and faster speeds to more than 100wpm. One goal is to approach the speed required for auditory recognition of speech and automatic recognition of fluently read words.

[0168] In the fast word recognition shell, the minimal pair of words is used for the phoneme contrast of interest, just as in the phoneme discrimination shell. The two words are shown on the display. At this point, the patient speaks the word highlighted by the display. Again, the patient can start at a slow speed and progress to speeds in excess of 100 wpm. The fast word recognition shell can be implemented independently of the phoneme discrimination shell, or it can be placed after the phoneme discrimination shell when the patient fails to progress to faster speeds in the phoneme discrimination shell. In the second scenario, the fast word recognition shell shows an improvement over the speed of the phoneme discrimination shell.

[0169] In the word amplification shell, the patient is taught to pay attention to the prosodic features of a word or phrase, such as intonation, stress patterns, and vowel sound quality. For example, if a patient registers a question about the / pr / consonant cluster in the phoneme recognition shell during diagnosis, the process can generate a word amplification cell that plays a video or audio file of a person speaking, and the person speaks elongated segments and exaggerates the intonation contours of the word, especially the initial / pr / segment. This can be similar to mother talk, a language with exaggerated prosodic patterns used by caregivers to babies. The patient is asked to imitate this amplification pattern by recording and playing back his production or by answering a series of questions about the position and movement of his pronunciation (such as her lips, tongue, jaw, etc.) or by pointing out various auditory features (such as pitch and sound transitions).

[0170] A sequence of activating treatment shells depends on the problems identified in the diagnostic phase and the new problems found in the treatment phase. For illustration, if the patient shows a problem about the / p / - / b / contrast in the initial position in the diagnostic phase, the process activates the pronunciation treatment cell so that the unit allows the patient to observe, practice, record and review its pronunciation of / p / and / b / phonemes from the model. Next, the process presents a word discrimination cell to it, and the unit requires it to select the correct word (or the image associated with the word) when the audio recording plays a phrase or sentence. This cell can play five phrases or sentences. Her correct and incorrect responses are input into the second data structure (such as the patient matrix) for later analysis. Next, the process activates the phoneme discrimination cell with five minimum pairs of / p / - / b / contrasts involving the initial position. If there is no error in the task execution as described above for this type of cell, the patient proceeds one by one and proceeds to higher and higher speeds. However, if the patient encounters problems when proceeding in this way, the process can run other cells (such as fast word recognition involving the same minimum contrast) before returning to phoneme discrimination to see if the patient can overcome the obstacle for the second time. Other cells can be called after other attempts fail, such as additional pronunciation therapy cells that provide more detailed clues, or online assistants can be called at this time. After this, the process activates the spelling pattern cell for / p / , and the unit involves teaching the patient to recognize when p is pronounced as / f / (ph). In addition, the process trains and tests the patient in the fast naming of words containing the letter p ( / p / or / f / pronunciation) and pronunciation / b / . The process can use the mistakes made previously in word discrimination and elsewhere to generate words in the fast naming cell. The patient can slowly start by 1.0 seconds per word, and proceed to faster and faster speeds until close to 0.3 seconds per word. It should be noted that the student who is expected to develop typically in grade 4 and above realizes 120 to 180 correct words per minute. Once the patient achieves this goal, the training phase for the / p / - / b / contrast in word-initial position is complete, and the patient proceeds to the next shell or cell. If the patient instead shows problems with the / p / - / b / contrast in word-medial position during the diagnostic phase, then the syllable division therapy cell is activated. This is because the syllable division therapy cell can help train the patient to pay attention to the phonetic features in the middle of the word, such as focusing on flabby. with flappy The difference between.

[0171] Fig.11 A diagram showing an example implementation of a phoneme recognition diagnostic shell and cell for vowels according to the present disclosure. This method is implemented via the techniques described herein. Phoneme recognition diagnostic cell U Test patient distinguishes / ε / - (l e d、l ad) ability. If the patient does not make an error, then the process generates a cell V to test / ε / - Contrast (f ai r or f u r). If the patient makes an error in cell U, the process generates cell W to confirm the error about / ε / - The problem of contrast and the generation of cells X to test the new contrast / ε / - / ∧ / (b e t or b u tt). If the patient does not make an error in cell X, then she proceeds to the next cell. But if she makes an error in cell X, then the process generates cell Y to test / ε / - contrast.

[0172] Additionally, it should be noted that any end-user-based techniques disclosed herein (whether patient-based or operator-based) can be employed via any graphical user interface (whether monochrome, grayscale, or color). For example, such an interface may include structural elements such as windows, menus, icons, control elements / widgets, or tabs. Also, for example, such an interface may include interactive elements such as a cursor, a selection element, or an adjustment handle.

[0173] In addition, in some embodiments, the computer 400 may be configured to automatically detect the patient's geographic location, such as via a geographic location unit 414. Based on this detection, the computer 400 may be configured to at least partially output at least one of the diagnostic stage (whether in part or in whole) and the treatment stage (whether in part or in whole) based on the language or dialect associated with the geographic location. For example, if the computer 400 detects that the patient is in France, the computer 400 at least partially outputs at least one of the diagnostic stage (whether in part or in whole) and the treatment stage (whether in part or in whole) based on France (whether in part or in whole). It should be noted that the relevant cell data, shell data, game data, etc. can be automatically (whether in part or in whole) converted and / or can be obtained to be converted in advance (whether in part or in whole) via the computer 400. It should be noted that these actions can be performed in any way via the computer system 304 and / or the patient's computer 308. It should be noted in addition that the patient can, such as for travel purposes, disregard (whether in part or in whole) and / or disable (whether in part or in whole) this feature.

[0174] Additionally, in some embodiments, the computer 400 can be configured to automatically detect patient speech as inputted audibly into the microphone 418, such as via speech or voice recognition software. Then, based on this detection, the computer 400 outputs, at least in part, at least one of a diagnostic stage (whether in part or in whole) and a treatment based on the speech (whether in part or in whole).

[0175] Furthermore, in some embodiments, the computer 400 may be configured to at least partially employ the NFC circuit to interact with another NFC circuit during at least a portion of the output of at least one of a diagnostic phase (whether partial or total) and a treatment phase (whether partial or total). For example, the patient may be instructed to move the computer 400 based on a task, such as by waving or touching another object (such as an NFC cell). Upon such movement, the NFC circuit may indicate to the computer 400 whether the task was performed, and if so, whether the task was performed appropriately.

[0176] In addition, in some embodiments, the computer 400 (such as the mobile phone 216) can be configured to run a software application, which at least partially implements at least some of the technologies disclosed herein, such as for the diagnostic phase and / or the treatment phase (whether in whole or in part). The software application can be silently run in the background of the computer 400. The software application can be configured so that when the patient starts her day and talks to others (such as at work, at school, at home, etc.), the computer 400 listens to the patient's response via the microphone 418. This listening can be when the patient is using the computer 400 (such as via implementing a phone call or a telecommunications conference session) and / or the patient is not using the computer 400 (such as when the computer 400 is placed on the table). The application can be configured to automatically filter background noise and / or voices other than the patient. In addition, for the purpose of privacy and / or data security, the application can also be configured to automatically delete background noise and / or voices other than the patient as preset in advance. However, for any language spoken by the patient and recorded via the application, the system 304 can be configured to process this information, such as upon receipt from the computer 400, and utilize this information for providing a diagnostic phase and / or a therapeutic phase (whether in whole or in part), such as via a diagnostic cell and / or a therapeutic cell.

[0177] Furthermore, in some embodiments, the computer 400 may be implantable (such as a hearing aid) or wearable (such as with an optical head mounted display (OHMD)). This computer 400 may include a vibrator for bone conduction (such as for listening to sounds), or a similar device for pronunciation analysis. For example, the vibrator may provide a vibration output to the jawbone so that the jawbone conducts the sound to the patient's inner ear. Alternatively, the computer 400 may be configured to provide information about the patient's pronunciation of a particular sound by monitoring features such as the vibration of the patient's vocal cords and the airflow pattern in the patient's mouth. Thus, the computer 400 may be configured to provide at least partial output of at least one of the diagnostic phase (whether in part or in whole) and the treatment phase (whether in part or in whole) even in at least one of the language selected as via the geolocation unit 414 based on the geographic location of the computer 400.

[0178] In addition, in some embodiments, the computer 400 may be configured for facial coding to classify the patient's facial movements via their facial appearance during at least one of the diagnostic phase (whether partial or total) and the treatment phase (whether partial or total), such as for emotion determination and pronunciation training or speech therapy. This classification can be used as an iterative feedback loop to enhance at least one of the diagnostic phase (whether partial or total) and the treatment phase (whether partial or total). For example, the computer 400 may detect the patient's face in at least one of the diagnostic phase (whether partial or total) and the treatment phase (whether partial or total), extract geometric features of the patient's face, generate a temporal profile of each patient's facial movement, and then provide this profile, such as to the computer system 304 (whether to the first data structure and / or the second data structure) to serve as an iterative feedback loop to enhance at least one of the diagnostic phase (whether partial or total) and the treatment phase (whether partial or total).

[0179] In addition, in some embodiments, computer 400 can be implantable or wearable. This computer 400 can identify and map the problem area of ​​the patient's brain for language processing, which is determined by at least one of the diagnostic stage and the treatment stage. As an implantable or wearable device, computer 400 can monitor the activity of nerve cells and use brain signals to correlate the physical area of ​​the brain with the functional area, which is mapped by the program by at least one of the diagnostic stage and the treatment stage. This computer 400 can then stimulate the brain circuit through the treatment stage by corrective action. This computer 400 can monitor the changes of the brain circuit as determined by brain activation and the patient output from the treatment cell. For example, this monitoring of the activated area in the brain can indicate that the patient is close to achieving the target pronunciation during its approach during treatment. In this case, the program strengthens its treatment at this moment so that it passes this goal and strengthens its training to ensure stable acquisition.

[0180] In addition, in some embodiments, the computer 400 may be configured for second language learning. In this case, a first data structure (such as a master matrix) will be configured to predict the errors that a second language learner is likely to make for the language or dialect. Most of the treatment shells will remain, especially when the learner's goal is to be fluent like a native speaker. New diagnostic and treatment shells will cover and / or be updated to cover more components of grammar, sentence structure, and pragmatics (language usage context), especially when at least some of the information is obtained from other patients using the computer system 304.

[0181] In addition, in some embodiments, computer 400 can be configured for accent modification, wherein the patient receives speech therapy to change his local accent. For example, a person may decide to reduce a regional or stigmatized accent for professional reasons. In this case, her treatment will focus more on phoneme recognition and discrimination and word amplification, and less on fast word recognition. Appropriate treatment cells will be generated by the diagnostic shell selected for this program. That is, the diagnostic shell will first cover the pronunciation at the segment (pronunciation) and prosody (speech) levels. In other embodiments, the patient can be a machine that produces human-like speech during task execution, such as for machine learning purposes. In these examples, computer 400 can be configured to provide correction training so that the patient-machine is close to natural speech at the segment, prosody and syntactic levels.

[0182] In addition, applicants conducted research on therapeutic methods implemented in humans, which formed the basis of this disclosure. The experimental group of this study adopted a flexible data-driven approach that used a detailed ongoing linguistic cognitive profile of each student to generate personalized training exercises. The control group used an Orton-Gillingham-based approach. Comparison of segmentation skills between the experimental group and the control group statistically showed a significant transfer effect in the former and a slight decrease in the latter. For the experimental group, combined training of phonemic awareness and rapid automatic naming also resulted in stronger performance in word recognition.

[0183] More specifically, the study compared a personalized approach based on the methods disclosed herein with a standardized approach for struggling readers in middle schools. The Wilson Reading System, based on the Orton-Gillingham principle, was used as a standardized approach in the control group. The personalized approach used in the experimental group is a new form of intervention that uses creative problem solving to solve specific reading and language processing problems identified in each student. Referred to herein as reactive intervention, this approach requires the formation of a detailed linguistic cognitive profile for each student in order to customize training. The research design follows the guidelines recommended by the National Reading Panel (2000) and the Institute of Educational Sciences (2013) for efficacy research. The results show that there is a significant difference in the impact of segmentation skills between the two approaches, thereby providing further confirmation and explanation of the weaknesses of existing intervention programs.

[0184] The study was conducted at a public junior high school with an enrollment of over 1,400 and approximately 260 students in special education. The junior high school is located in a predominantly middle to upper middle income area of ​​New York State. The study was conducted in conjunction with an on-site after-school reading program that ran from March to June 2014. The program was organized by the research team and provided to the participants as a community service.

[0185] This study meets the National Reading Panel's (2000) guidelines for well-designed experiments and the Institute of Education Sciences' (2013) guidelines, and the experiment provides strong evidence for cause. The National Reading Panel's (2000) meta-analysis of reading research concluded that effect sizes are larger when children receive focused and explicit instruction on no more than one or two phonemic awareness (PA) skills. Furthermore, programs that focus on both letters and phonemes are more effective than programs that cover only one. In this study, the experimental group learned IPA (International Phonetic Association) symbols in addition to letters and phonemes. As the only PA skill directly taught, the experimental group practiced identifying individual sounds in words (phoneme separation). However, the end-of-program assessment tested another PA skill segmentation to determine transfer.

[0186] The review of the National Reading Panel also found that students were more effective when taught in small groups than in classroom or one-to-one environments. The optimal total instruction time was found to be five to 18 hours. In the current study, group sizes ranged from an average of seven to 10 students per session. The instruction time for each session was an average of 35 minutes, totaling 8.25 hours. To ensure intrinsic effectiveness, participants were randomly assigned to two groups as described below. The two groups were equivalent on key factors (described below), as confirmed by chi-square tests. Students who received Wilson intervention during school hours were asked to withdraw from the project and were not included in the statistical analysis.

[0187] The National Reading Panel (2000) found that transfer effects were larger when studies used experimenter-designed tests to measure reading improvement because standardized tests may be less sensitive in detecting changes in the skills being investigated. The pretest and posttest used in this study were developed by the authors to measure segmentation ability. A final recommendation (follow-up posttests at some time after the intervention was completed to assess the long-term effects of the training) could not be achieved in this study because the program ended at the end of the school year.

[0188] Participants were recommended by teachers at their school as students for special education who would benefit from an after-school reading program. Twenty-six students participated in the after-school program: 15 from the 6th grade and 11 from the 7th grade. In order to assign participants to the control and experimental groups, the students were first divided by grade. Students in each grade were randomly assigned to the control or experimental group. Of the 26 students who participated in the program, 18 attended after-school tutoring until the end of the school year and were included in the control study. The 6th and 7th grade students were evenly distributed in two groups with nine students in each group (six in the 6th grade and three in the 7th grade).

[0189] All students in the control study were from two-parent families. Of the 18 students, five were girls (three in the control group and two in the experimental group) and 13 were boys (six in the control group and seven in the experimental group); and seven were minority students (four in the control group and three in the experimental group). Five students in the control group and six students in the experimental group received speech / language therapy services. The participants in the study all scored at Level 1 or Level 2 on the New York State 4th grade English Language and Literature (ELA) test. Level 1 is considered "failure", where student performance indicates a lack of understanding of the English language knowledge and skills expected for this grade (New York State Education Department, 2012). Level 2 performance indicates partial understanding. Above Level 2 are Level 3 ("reaching proficiency standards") and Level 4 ("exceeding proficiency standards"). Only one student in the control group and one student in the experimental group could obtain their 5th grade ELA scores, and both of them were at Level 2. Two of the students in the control group had ELA scores that were not available. IQ scores were not considered because these struggling readers may have language processing problems, which could affect their IQ scores. A series of chi-square tests revealed no significant differences between the control and experimental groups in terms of gender, minority status, participation in speech / language therapy, grade, and two-parent / single-parent families. Two students in the control group and three students in the experimental group were overage for their grade, suggesting that they had repeated a grade. This factor was also found to be statistically insignificant between the two groups.

[0190] The extracurricular tutorials for the control and experimental groups were held every Tuesday from March to June 2014 for 45 minutes, except for a one-week mid-term break. The instruction time was 35 minutes, after deducting 10 minutes for adaptation and preparation for the last bus. Each group was taught by a teacher in a separate classroom (see below for control and experimental groups). Each teacher was supported by a teaching assistant for 75% of the tutorial time. The teacher kept an attendance log, while his teaching assistant recorded the participation procedures for each student in his group (see below for achieving fidelity). Both teachers assigned homework totaling no more than one hour a week.

[0191] The teacher who taught the control group was hired through a selection process conducted by the school, which screened applicants among its own special education teachers. The instructor selected had nine years of teaching experience and a Wilson Level 1 certificate.

[0192] The control group followed the Wilson Reading System closely. The class used the Third Edition of the Wilson Reading System's Instructor Manual, Rules Notebook, and Student Reader One and Two. As explained on the company's website, the Wilson Reading System is a step-by-step program designed to teach encoding and decoding skills, progressing from monosyllabic words to multisyllabic words. It requires students to type the pronunciation of words as they learn segmentation (Wilson Language Training Corp., 2010).

[0193] During the program, the control group worked on closed syllables and exceptions, digraphs (two letter combinations that form a single sound, such as ee), welds (e.g., all, ing), synthesis and segmentation, vowels (a, e, i, o, u), and consonants. The teacher introduced new concepts in a way that allowed for practice and review. Class activities were guided by the Wilson Method and included identifying welds (b all ), combining letters (which represent sounds) into words (r-am=ram), fitting given words into sentences, identifying phrase boundaries in written sentences, completing words with given letters, and reading simple passages. Teachers distribute packets of information from the Wilson curriculum weekly for at-home review.

[0194] The teacher of the experimental group is one of the inventors of this application. The inventor has more than 20 years of experience in college teaching, but none of them are non-higher. The inventor has worked with students with reading disabilities in a one-to-one environment for more than eight years, but has not been trained in any standardized method.

[0195] The experimental intervention involved responding directly to the language problems found in each student. Therefore, the form and pace of intervention in the experimental group varied for each student depending on his or her difficulties and progress. In order to implement this responsive intervention, an initial profile of the students' language abilities and weaknesses was collected through a series of short tests designed by the authors to identify each student's problem areas. The tests were administered one-on-one in a quiet classroom before the start of the project and took no more than 15 minutes per student. The tests covered basic phonetic knowledge. Students were asked to identify vowels in words ( / i / in beat), participles ( / fremd / framed), and split words into syllables (ad-van-tage).

[0196] The initial assessment was followed by further weekly observations and recording of the students' receptive and production errors. These data sources were used to generate a personal profile containing information about each student's phonemes (consonants and vowels), speech processing (e.g., neutering of unstressed vowels to ), orthography (spelling patterns), morphological structure (word formation), and morphophonemic processing (e.g., devoicing of past tense morphemes in base words ending in voiceless consonants, such as The students' linguistic-cognitive profiles included problems with both reception (listening and reading) and production (slapped). Because the project lasted only three months, the experimental classes did not cover larger linguistic structures such as vocabulary (word) collocations, phrases, sentences, and paragraphs. It is generally accepted in the art that phonological, orthographic, and morphological knowledge are key components of reading development (Berninger, Abbott, Nagy, & Carlisle, 2010) and should be taught directly (Torgesen, 2004).

[0197] Each student's cumulative profile was used to generate new exercises for the next week specifically for him or her. Four students in the experimental group had pronunciation problems and difficulty controlling airflow and volume. These students began the project with pronunciation exercises, practicing with props when needed. For example, they used a handheld mirror to check their lip movements (lip expansion or rounding) and a lollipop to feel the position of their tongue during vowel production. Regardless of individual pace, the experimental group as a whole generally progressed from pronunciation, auditory discrimination of phonemes, and representing pronunciation in the form of phonetic symbols to learning phoneme-grapheme (sound-letter) mappings and spelling patterns.

[0198] The exercises in the second half of the project focused on the rapid naming of words, which contained previously learned spelling patterns. Although rapid automatic naming (RAN) has been used for assessment and research for more than three decades (Denckla & Rudel, 1976), the actual nature and role of rapid naming in reading difficulties remain unclear (see Elliott & Grigorenko's research review, 2014). RAN was used for a different purpose in this study: to capture students' errors in identifying spelling patterns and use the collected information to form future exercises. RAN traditionally involves letters, numbers, and object names, but in the version of RAN used in this study, participants read aloud single words that flashed on the screen at a specified time using MS PowerPoint. Each word list contained an average of 40 words. Most of the words were monosyllabic, such as rut, or disyllabic, such as roaster. No more than three of the words in each list contained more than two syllables, such as happiness. The scorer recorded the correct and incorrect reading of the test words, which were then used to generate new exercises for subsequent weeks. For example, when a student mispronounces sitting as seeing, subsequent practice consists of distinguishing between the spelling patterns of the vowels / ai / (-ight, -ite, -ie, etc.) and / I / (i). Instruction in the class is similarly individualized. A significant amount of class time is spent on one-on-one discussions, collaborative work, and group work. Each student receives a weekly packet of information tailored to his or her specific concerns about the language. It is important to note that segmentation skills are not taught directly and explicitly to the class. The students' weekly packets similarly do not contain segmentation practice. This is to see if a transfer effect occurs.

[0199] For efficacy research purposes, fidelity of implementation was considered consistent with the IES (Institute of Educational Sciences, 2013) Common Guidelines for Educational Research and Development. Fidelity of implementation of the Wilson Method in the control group was monitored through classroom observations, student attendance and engagement, class time use, instructional strategies, and a combination of instructional materials and content. The control group teacher prepared lesson plans by completely filling out the Wilson Reading System lesson plan form, spelling out the specific words that would be taught. The research team confirmed that she adhered closely to the lesson plans, with the only modifications being to carry over the materials into the second week when needed. Additionally, the control group’s teaching assistant, who was affiliated with the author’s university rather than the participating schools, served as an observer for the control group. Observations of the teaching assistants were conducted at least once every other week in writing or through verbal communication. The teaching assistants’ reports confirmed that the teacher closely followed the Wilson Reading System lesson plan form. --Presents language structures systematically and cumulatively; reinforces concepts through multiple sensory channels; uses questioning techniques and gives feedback on students' mistakes. In addition, as an experienced instructor, the teacher varies class activities and pace to keep students engaged.

[0200] The teaching assistants in both classes rated the students in their groups on how well they paid attention and participated in the classroom, based on attendance, attention to the task at hand, responses to the teacher's questions, and distractions (e.g., cell phone use). Based on a four-point rating system (good, fair, average, poor), the control group received the following ratings: good—3, fair—4, average—1, poor—1. The experimental group received the following ratings: good—2, fair—3, average—2, poor—2.

[0201] The experimental classes followed a highly fluid and flexible arrangement that did not allow for easy monitoring of the fidelity of the implementation. Nevertheless, the following information was collected weekly from the training period: student attendance and participation, notes on discussions with each student in the group, and distribution of personalized weekly information packets for each student. The collected data showed that each student present received at least one one-to-one discussion from the teacher or teaching assistant each period. Each student present also received a personalized information packet each week. Both classes met and ended at the same time on the same day in the same building.

[0202] The same segmentation test was implemented one week before the start of the extracurricular program (pretest) and the last week of the program (posttest). The segmentation test consists of a list of 10 common words, which contain some monosyllabic words (such as praise) and disyllabic words (such as flower). Seven of the words appear as the 5000 most frequently used words (Davies & Gardner, 2010). The test was implemented individually for each student in a quiet room. The student's oral production was transcribed into IPA phonetic symbols by person, where the sound transitions (intervals between pronunciations) were pointed out. When necessary, the students were asked to segment the words again to confirm the transcription. The same test was implemented for all students in the two groups in the same way. The scores were assigned as follows: 0—the test word was read as a single unit; the student read the wrong pronunciation; the pronunciation was in the wrong position in the word; the student said the letter instead of the pronunciation (for example, "double U" W instead of / w / pronunciation); or the student gave up trying to segment the word. 1—a single phoneme or phoneme group was given in the correct position in the word. (Even in this case where the student does not fully segment the word, points are still given for the phoneme cluster; a comparison of her score to the total score will indicate that her segmentation is incomplete.)

[0203] Near the end of the project, all participants were tested for word recognition speed. Thirty-six words were placed one by one in an MS PowerPoint slide (Calibri 44-point font) and set to display at 0.3 seconds (i.e., 200 words per minute). All words were in the top 1000 most frequently used words in modern American English, such as anything and least (Davies & Gardner, 2010). Three blank slides separated every 10 words to allow students to pause between rapid readings of the words flashing on the screen. The test was conducted individually in a quiet area where other students were not present. Only the reading of the exact form of the word was accepted. For example, if a student reads respond as respond, no points were given.

[0204] Administer the segmentation test according to the test manual containing precise instructions given verbally to the students. Give examples that illustrate segmentation (“For example, ‘cat’ is ”). Ask each student to repeat the test word to confirm hearing the correct word before segmenting the word into individual sounds (phonemes). Give students as much time as they need to complete each item. The tester is friendly but does not provide any feedback for each test item and response. For the RAN test, the tester checks with students to make sure they are ready to read the words on the screen. There are ten blank slides before the first word to give students time to prepare. The tester uses the word list in the order the slides are presented to check for correct reading.

[0205] Some of the inventors of the present application acted as testers and had been practicing for more than one year prior to this study using the same protocol of administering the same segmented test to other students and scoring. To determine inter-rater reliability, all segmented pre-tests were scored independently by two authors, and 33.3% of the segmented post-tests were scored by a second assessor. The inter-rater reliability for the pre-test was 99.4% and the inter-rater reliability for the post-test was 99.7%. Inter-rater reliability was not monitored for the RAN test because scoring only involved the correct reading of the test words on the checklist.

[0206] A series of analyses of covariance (ANCOVA) were run to examine the effect of intervention on segmentation skills. Both ELA and segmentation pretest scores were entered as covariates. Table 1 shows the unadjusted and adjusted means for these analyses. The ANCOVA for the segmentation test revealed a significant effect of intervention, F(1,12)=15.11, p=.002, η 2=.557 (see Table 2). Students in the experimental group performed better on this assessment than students in the control group. The split scores of the two students in the control group were not included in the calculation because their ELA scores were not available. However, their performance on the pre-test and post-test fell within the score range of others in their control group (pre-test: 20 and 20 of 59; post-test: 13 and 17 of 59).

[0207] Students in both groups began the project with similar unadjusted mean scores (25 vs. 26 out of 59), but differed significantly by the end (45 vs. 21). Four members of the control group actually showed a substantial decrease in scores, three of which decreased by more than 20.0%. In contrast, five of the nine students in the experimental group showed an increase of more than 33.9%. At the beginning of the project, both groups were generally only able to divide words into syllables rather than phonemes, as seen in their pre-test scores. After the intervention, all students in the experimental group were able to segment most of the test words into individual phonemes. The only student in the experimental group who could not score above 40 out of 59 on the post-test missed a month of school in the middle of the project.

[0208] RAN is part of the experimental group's reaction intervention project. As previously noted, RAN is used to detect weaknesses in the application of spelling rules at the speed required for fluent reading. At the beginning of the intervention, all students in the experimental group made mistakes even at a slower speed than 60 words per minute (wpm), and most of them could not perform the task under the situation of more than 120wpm. It is expected that the students who develop typically in grade 4 and above will achieve 120 to 180 correct words per minute (cwpm) (Shaywitz, 2003, 277 pages). The project was carried out for two months, and the students in the experimental group were able to perform at a speed between 120 and 200wpm. For example, a student misread the / u / pronunciation in roosting, toot, rooted and noose in an earlier RAN assessment. After further practice, he was able to correctly read the words tooting, drooping, loose and croon at 200wpm in subsequent RAN tests.

[0209] Near the end of the program, both groups were given the RAN test to see if the combined PA and speed training had improved the experimental group's accuracy in word recognition. Each student in the experimental group read an average of 44.4% of the test words correctly at 200 wpm, compared to only 31.0% in the control group. More importantly, five of the nine students in the experimental group scored above 50.0%, compared to only one in the control group.

[0210] Table 1

[0211] Descriptive statistics of unadjusted and adjusted means for the split-post-test by group

[0212]

[0213] a. The covariates present in the model were assessed at the following values: 4th grade ELA score = 626.88, split pretest score = 26.2500.

[0214] Table 2

[0215] Post-segmentation test: test for between-subject effects

[0216]

[0217] bR squared = .704 (adjusted R squared = .630)

[0218] c. Calculated using α = .05

[0219] In some embodiments, various functions or actions may occur at a given location and / or in conjunction with the operation of one or more devices or systems. In some embodiments, a portion of a given function or action may be performed at a first device or location, and the remainder of the function or action may be performed at one or more additional devices or locations.

[0220] In some embodiments, the device or system includes at least one processor and a memory storing instructions that, when executed by the at least one processor, cause the device or system to perform one or more method actions as described herein. In some embodiments, the memory stores data, such as one or more structures, metadata, rows, tags, blocks, strings, or other suitable data organizations.

[0221] As will be appreciated by those skilled in the art, aspects of the present disclosure may be implemented as a system, method, or computer program product. Thus, aspects of the present disclosure may take the form of a complete hardware implementation, a complete software implementation (including firmware, resident software, microcode, etc.), or as an implementation combining software and hardware aspects, which may generally be referred to herein as a "circuit," "module," or "system." In addition, aspects of the present disclosure may take the form of a computer program product implemented in one or more computer-readable media having computer-readable program code implemented thereon.

[0222] Any combination of one or more computer-readable media may be utilized. Computer-readable media may be computer-readable signal media or computer-readable storage media. Computer-readable storage media may be, for example, but not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or apparatuses, or any suitable combination of the foregoing. More specific examples of computer-readable storage media (a non-exhaustive list) would include the following: an electrical connection with one or more wires, a portable computer floppy disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer-readable storage medium may be any tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or apparatus.

[0223] Computer readable signal media may include a propagated data signal in which a computer readable program code is implemented, for example, in baseband or as part of a carrier wave. This propagated signal may take any of a variety of forms, including, but not limited to, electromagnetic form, optical form, or any suitable combination thereof. Computer readable signal media may be any computer readable medium that is not a computer readable storage medium and that can convey, propagate, or transmit a program used by or in conjunction with an instruction execution system, device, or apparatus. The program code implemented on the computer readable medium may be transmitted using any appropriate media, including, but not limited to, wireless, wired, fiber optic cable, radio frequency (RF), etc., or any suitable combination of the above.

[0224] Computer program code for performing operations of aspects of the present disclosure may be written in any combination of one or more programming languages, including object-oriented programming languages ​​(such as Java, Smalltalk, C++ or the like) and conventional procedural programming languages ​​(such as the "C" programming language or similar programming languages). The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0225] The corresponding structures, materials, actions and equivalents of all members or steps plus function elements in the claims below are intended to include any structure, material or action for combining other required elements to perform functions as specifically required. The description of the present disclosure has been presented for the purpose of illustration and description, but is not intended to be exhaustive or limited to the disclosed form. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the present disclosure. The embodiments are selected and described to best explain the principles and practical applications of the present disclosure, and to enable those skilled in the art to understand that the various embodiments of the present disclosure have various modifications, which are suitable for the intended specific use.

[0226] The figures depicted herein are illustrative. There may be many variations in the figures or steps (or operations) described therein without departing from the spirit of the present disclosure. For example, steps may be performed in different orders or steps may be added, deleted or modified. All of these variations are considered part of the present disclosure. It should be understood that various improvements and enhancements may be made by those skilled in the art now and in the future, and the improvements and enhancements are within the scope of the appended claims.

Claims

1. A method of treating, the method comprising: include: A user request is received by a server in real time over a network from a client, the client communicating with a camera and operated by a user, the user having a face with two eyes; In response to the user request, the server hosts a secure session with the client in real time over the network; And during said secure session: selecting, by the server, a first shell containing a first task type in a first graph based on a first comparison of a first matrix specific to the user and a second matrix not specific to the user; selecting, by the server, a first cell in a second graph that includes a first audio task or a first image task based on the first task type and the first comparison; The server enables the first audio task or the first image task from the first cell to serve the client through the network; The server instructs the client to activate the camera through the network, so that the camera acquires a video describing the face with the two eyes, and outputs the first audio task or the first image task through the client; In response to outputting the first audio task or the first image task via the client, receiving, by the server through the network, user input and the video from the client; The server writes the user input into the first matrix; identifying, by the server, the two eyes in the video; determining, by the server, an eye gaze value or a blink rate based on the two eyes identified in the video; Writing the eye gaze value or the blink rate into the first matrix by the server; selecting, by the server, the first shell in the first graph or a second shell containing a second task type based on the user input in the first matrix and a second comparison of the eye gaze values ​​or the blink rates with the second matrix, wherein the second task type is different from the first task type, wherein the second comparison is subsequent to the first comparison; selecting, by the server, the first cell or a second cell containing a second audio task or a second image task in the second graph based on the first task type or the second task type and the second comparison, wherein the second audio task or the second image task is different from the first audio task or the first image task, respectively; generating, by the server, a correction or evaluation feedback message including the second audio or the second image; and The correction or evaluation feedback message is served to the client by the server so that the correction or evaluation feedback message is output to the user via the client.

2. The method according to claim 1, in, The user has lips, and further comprises: During said secure session: The server instructs the client to activate the camera through the network, so that the camera acquires the video describing the face with the two eyes and the two lips, and outputs the first audio task or the first image task via the client; recognizing, by the server, the lips in the video; determining, by the server, a lip motion value based on the lips identified in the video; The server writes the lip movement value into the first matrix; The first shell in the first graph or the second shell containing the second task type is selected by the server based on the second comparison of the user input, the lip movement value and the eye gaze value or the blink rate in the first matrix with the second matrix.

3. The method according to claim 2, further comprising: include: During said secure session: The first shell in the first graph or the second shell containing the second task type is selected by the server based on the second comparison of the user input, the lip movement values, and the eye gaze values ​​in the first matrix with the second matrix.

4. The method according to claim 2, further comprising: include: During said secure session: The first shell in the first graph or the second shell containing the second task type is selected by the server based on the second comparison of the user input, the lip movement values, and the blink rate in the first matrix with the second matrix.

5. The method according to claim 2, further comprising: include: During said secure session: The first shell in the first graph or the second shell containing the second task type is selected by the server based on the second comparison of the user input, the lip movement value, the eye gaze value and the blink rate in the first matrix with the second matrix.

6. The method according to claim 1, further comprising: include: During said secure session: The correction or evaluation feedback message is served to the client by the server so that the correction or evaluation feedback message is output to the user in an unobtrusive manner.

7. The method according to claim 1, in, The client is wearable.

8. The method according to claim 7, in, The client is a glasses unit or an optical head mounted display (OHMD) having the camera.

9. The method according to claim 1, in, The client is a phone or tablet hosting the camera, wherein the camera is a front-facing camera.

10. The method according to claim 1, in, The client runs a browser, and the secure session occurs through the browser.

11. The method according to claim 1, in, The client communicates with a vibrator, the vibrator being configured for bone conduction with the user, and further comprising: the server causing the correction or evaluation feedback message to be served to the client so that the correction or evaluation feedback message is output to the user via the client, and the client instructs the vibrator to output the correction or evaluation feedback message via the bone conduction.

12. The method according to claim 1, further comprising: include: During said secure session: Recognizing, by the server, the face in the video; determining, by the server, an emotion of the user based on the face recognized in the video; The server writes the emotion into the first matrix; The first shell in the first graph or the second shell containing the second task type is selected by the server based on the user input, the emotion, and the second comparison of the eye gaze value or the blink rate in the first matrix with the second matrix.

13. The method according to claim 1, further comprising: include: During said secure session: Recognizing, by the server, the face in the video; Extracting, by the server, geometric features of the face; generating, by the server, a temporal profile for each of a plurality of motions of the face in the video; The temporal profile is written by the server into the first matrix such that an iterative feedback loop is formed during the first comparison or the second comparison.

14. The method according to claim 1, further comprising: include: During said secure session: The first shell in the first graphic or the second shell containing the second task type is selected by the server based on the user input in the first matrix and the second comparison of the eye gaze values ​​with the second matrix.

15. The method according to claim 1, further comprising: include: During said secure session: The first shell in the first graph or a second shell containing the second task type is selected by the server based on the user input in the first matrix and the second comparison of the blink rate with the second matrix.

16. The method according to claim 1, further comprising: include: During said secure session: The first shell in the first graph or a second shell containing the second task type is selected by the server based on the second comparison of the user input, the eye gaze values, and the blink rate in the first matrix with the second matrix.

17. The method according to claim 1, further comprising: include: During said secure session: receiving, by the server, a pulse of the user from the client in response to outputting the first audio task or the first image task via the client, wherein the pulse is measured while the first audio task or the first image task is output via the client; The server writes the pulse into the first matrix; The first shell in the first graph or the second shell containing the second task type is selected by the server based on the second comparison of the user input, the pulse, and the eye gaze value or the blink rate in the first matrix with the second matrix.

18. The method according to claim 1, in, The client is in communication with a video game controller, wherein the user input is via the video game controller.

19. The method according to claim 1, in, The correction or evaluation feedback message including the second audio or the second image has information on symptoms of a language-related disorder, thereby diagnosing the user.

20. The method according to claim 1, in, The server (a) instructs the client to activate the camera so that the camera acquires the video describing the user's body posture and simultaneously outputs the first audio task or the first image task via the client, (b) recognizes the body posture in the video, and (c) selects the first shell or the second shell based on the body posture recognized in the video.

21. The method according to claim 1, in, The server (a) receives a biometric data set from the client, the client communicating with a biometric reader monitoring the user while acquiring the video, and (b) selects the first shell or the second shell based on the biometric data set.

Images