Detection of Longitudinal Progression of Alzheimer's Disease (AD) Based on Speech Analysis

Abstract

A method implemented by one or more computer devices includes detecting the long-term progression of Alzheimer's disease (AD) in a patient. The method includes receiving speech data including the patient's description of one or more previous or current experiences, the speech data captured at multiple moments over a period of time. The method further includes analyzing the speech data to quantify a plurality of speech variables, the plurality of speech variables including a word length variable and a particle usage variable. The method includes determining a composite score based on standardization and a substantive weight assigned to each of the plurality of quantified speech variables. Thus, the method includes detecting predicted long-term changes in the quantified speech variables based on the composite score, and further estimating the patient's AD progression based on the predicted long-term changes.

Description

[Technical Field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of U.S. Provisional Application No. 63 / 354,165, filed June 21, 2022, the entire contents of which are incorporated herein by reference.

[0002] This application relates generally to speech analysis, and more particularly to techniques for detecting the longitudinal progression of Alzheimer's disease (AD) based on speech analysis. [Background technology]

[0003] Alzheimer's disease (AD) is a progressive neurodegenerative disorder characterized by a decline in a patient's memory, speech, and cognitive skills, as well as adverse changes in the patient's mood and behavior. AD can generally be attributed to one or more identified biological changes that can occur in a patient's brain over many years, such as the excessive accumulation of amyloid-beta (Aβ) plaques and tau tangles in the patient's brain. Specifically, while Aβ and tau proteins are generally produced as part of the brain's normal function, patients diagnosed with AD may exhibit either excessive production of Aβ protein, which can accumulate as plaques around brain cells, or excessive production of tau protein, which can misfold and accumulate as tangles within brain cells.

[0004] Identifying and detecting early signs of cognitive decline in patients using less invasive and clinically burdensome techniques can help more effectively treat or prevent the progression of AD. For example, a patient's speech often contains at least some signs of a patient's cognitive decline or adverse changes over time. Furthermore, for patients with ubiquitous access to personal electronic devices suitable for capturing their voice, analysis of speech samples for acoustic and linguistic characteristics and / or content can be readily performed. Summary of the Invention

[0005]

[0003] Embodiments of the present disclosure relate to one or more computing devices, methods, and non-transitory computer-readable media that can be utilized to detect predicted longitudinal changes in quantified speech variables associated with a patient as an estimate of the patient's Alzheimer's disease (AD) progression or the AD patient's treatment response. Specifically, according to embodiments disclosed herein, one or more computing devices can utilize machine learning models (e.g., natural language processing (NLP) models, transformation-based language models, automatic speech recognition (ASR) models) to convert raw audio files of the patient's speech data captured at several moments over a period of time into text transcripts and analyze linguistic speech variables, including word length variables and particle usage variables, as well as one or more acoustic speech variables, to determine an estimate of the patient's AD progression or the patient's treatment response to which the patient's speech data corresponds.

[0006] The patient's speech data includes a recording of the patient's description of one or more previous or current experiences. In certain embodiments, one or more computing devices may analyze the text transcript to quantify at least two speech variables derived from one or more linguistic speech variables (e.g., a word length variable and a particle use variable, and optionally, a word frequency variable, a syntactic depth variable, a noun use variable, or a pronoun use variable) and one or more acoustic speech variables (e.g., one or more Mel-Frequency Cepstral Coefficient (MFCC) features). In one embodiment, the at least two quantified speech variables include a word length variable and a particle use variable. In another embodiment, the at least two quantified speech variables include a word length variable and at least one MFCC feature, including a mean of the eleventh MFCC coefficient (MFCC mean 11) variable, a variance of the first derivative of the eleventh MFCC coefficient (MFCC var 25) variable, or a variance of the first derivative of the twelfth MFCC coefficient (MFCC var 26) variable.

[0007] In certain embodiments, one or more computing devices may generate a composite score based on standardization of at least two quantified speech variables derived from one or more linguistic speech variables and / or one or more acoustic speech variables, and substantial weighting of the at least two quantified speech variables derived from one or more linguistic speech variables and / or one or more acoustic speech variables. For example, in some embodiments, at least two quantified speech variables derived from one or more linguistic speech variables (e.g., a word length variable and a particle use variable, and optionally a word frequency variable, a syntactic depth variable, a noun use variable, or a pronoun use variable) and / or one or more acoustic speech variables (e.g., one or more Mel-Frequency Cepstral Coefficient (MFCC) features) may be standardized and combined into a composite score (e.g., an equal-weighted composite score, a weighted composite score). In one embodiment, the at least two quantified speech variables utilized to generate the composite score may include a word length variable and a particle use variable. In another embodiment, the at least two quantified speech variables utilized to generate the composite score may include a word length variable and at least one of an MFCC mean 11 variable, an MFCC var 25 variable, or an MFCC var 26 variable.

[0008] In certain embodiments, one or more computing devices can then determine predicted long-term changes in the quantified voice variables based on the composite score as an estimate of the patient's AD progression or AD patient treatment response.In this way, the present technology can provide an alternative to more invasive and more clinically burdensome tests for screening AD patients over time.Indeed, by generating a composite score based on at least two quantified voice variables derived from one or more linguistic voice variables and one or more acoustic voice variables identified as indicating progressive long-term changes, the present technology can provide a quantitative estimate of the patient's AD progression or AD patient treatment response using only the patient's voice.

[0009] In certain embodiments, one or more computing devices may receive audio data including a recording of a patient's description of one or more previous or current experiences, the audio data being captured at multiple moments during the period. For example, in some embodiments, the one or more computing devices may receive the audio data by receiving an audio file including an electronic recording of the patient's voice. In one embodiment, the electronic recording of the patient's voice may include an electronic recording of one or more verbal responses of the patient to a Clinical Dementia Rating (CDR) interview. In certain embodiments, the audio data is captured on one or more dates selected from the group including the first day and approximately 0.25, 0.5, 0.75, 1, 3, 6, 9, 12, 15, 18, 21, 24, 27, 30, 33, and 36 months after the first day.

[0010] In certain embodiments, one or more computing devices may analyze the audio data to quantify a plurality of audio variables. In certain embodiments, the one or more computing devices may analyze the audio data to determine the quantified plurality of audio variables by analyzing the audio data using one or more natural language processing (NLP) machine learning models. The plurality of audio variables may include a word length variable and a particle usage variable. In some embodiments, the plurality of audio variables may further include a word frequency variable, a syntactic depth variable, a noun usage variable, or a pronoun usage variable. In certain embodiments, the plurality of audio variables may further include one or more Mel-Frequency Cepstral Coefficient (MFCC) features. For example, in some embodiments, the one or more MFCC features may include the mean of the 11th MFCC coefficient (MFCC mean 11), the variance of the first derivative of the 11th MFCC coefficient (MFCC var 25), or the variance of the first derivative of the 12th MFCC coefficient (MFCC var 26).

[0011] In certain embodiments, the one or more computing devices may then determine a composite score based at least in part on the standardization or weighting of the quantified speech variables. For example, in some embodiments, determining the composite score may include standardizing the quantified speech variables, applying equal weighting to each of the quantified speech variables, and combining the standardized and equally weighted quantified speech variables to generate a composite score. In certain embodiments, the one or more computing devices may then detect predicted long-term changes in the quantified speech variables based on the composite score. In certain embodiments, the one or more computing devices may then estimate the patient's AD progression based on the predicted long-term changes. For example, in some embodiments, estimating the AD progression based on the predicted long-term changes may include correlating the composite score with one or more clinical assessment metrics.

[0012] In certain embodiments, the one or more clinical assessment metrics may be selected from the group consisting of Mini-Mental State Examination (MMSE) score, Clinical Dementia Rating (CDR) interview, Clinical Dementia Rating-Sum of Boxes (CDR-SB) scale, Alzheimer's Disease Assessment Scale-Cognitive (ADAS-Cog) subscale test battery, Alzheimer's Disease Cooperative Study Group-Activities of Daily Living Inventory (ADCS-ADL) scale, Neuropsychiatric Symptom Assessment (NPI) scale, Neuropsychiatric Symptom Assessment-Questionnaire (NPI-Q), Alzheimer's Disease Caregiver Global Impression (CaGI) scale, Instrumental Activities of Daily Living (IADL) scale, Amsterdam Activities of Daily Living Questionnaire (A-IADL-Q), and Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) scale.

[0013] In certain embodiments, the one or more computing devices may determine whether the patient is responsive to treatment based on the estimated AD progression. In certain embodiments, the one or more computing devices may send a notification of the estimated AD progression to a computing device associated with a clinician. In certain embodiments, in response to the AD prediction, the one or more computing devices may generate a recommendation for adjusting the patient's treatment regimen. For example, in some embodiments, the treatment regimen may include a therapeutic agent comprising at least one compound selected from the group consisting of compounds against oxidative stress, anti-apoptotic compounds, metal chelators, inhibitors of DNA repair, 3-amino-1-propanesulfonic acid (3APS), 1,3-propanedisulfonate (1,3PDS), secretase activators, beta- and gamma-secretase inhibitors, tau protein, anti-tau antibodies, anti-tau agents, gene therapy agents, neurotransmitters, beta-sheet breakers, anti-inflammatory molecules, atypical antipsychotics, cholinesterase inhibitors, other drugs, and dietary supplements, symptomatic medications, psychiatric medications, and the like. The therapeutic agent may include a therapeutic agent selected from the group consisting of an oral drug, a corticosteroid, an antibiotic, an antiviral agent, an anti-tau antibody, a tau inhibitor, an anti-amyloid beta (anti-Aβ) antibody, a beta-amyloid aggregation inhibitor, a target binding therapeutic agent, an anti-BACE1 antibody, a BACE1 inhibitor, a cholinesterase inhibitor, an NMDA receptor antagonist, a monoamine depleting agent, an ergoloid mesylate, an anticholinergic antiparkinsonian agent, a dopaminergic antiparkinsonian agent, tetrabenazine, an anti-inflammatory agent, a hormone, a vitamin, a dimebolin, a homotaurine, a serotonin receptor activity modulator, an interferon, and a glucocorticoid.

[0014] In certain embodiments, the symptomatic treatment agent may be selected from the group consisting of a cholinesterase inhibitor, galantamine, rivastigmine, donepezil, an N-methyl-D-aspartate receptor antagonist, memantine, and a dietary supplement (optionally, the dietary supplement is Souvenaid®). In some embodiments, the anti-tau antibody may be selected from the group consisting of an N-terminal binding agent, a mid-domain binding agent, and a fibrillar tau binding agent. In certain embodiments, the anti-tau antibody is selected from the group consisting of semolinemab, BMS-986168, C2N-8E12, goslanemab, tiravonemab, and zagotenemab. In some embodiments, the therapeutic agent is one that specifically binds to a target, and the target may be selected from the group consisting of beta-secretase, tau, presenilin, amyloid precursor protein or a portion thereof, amyloid beta peptide or an oligomer or fibril thereof, death receptor 6 (DR6), receptor for advanced glycation end products (RAGE), parkin, and huntingtin.

[0015] In certain embodiments, the therapeutic agent may be a monoamine-depleting drug, optionally tetrabenazine. In some embodiments, the therapeutic agent may be an anticholinergic antiparkinsonian selected from the group consisting of procyclidine, diphenhydramine, trihexylphenidyl, benztropine, biperiden, and trihexyphenidyl. In some embodiments, the therapeutic agent may be a dopaminergic antiparkinsonian selected from the group consisting of entacapone, selegiline, pramipexole, bromocriptine, rotigotine, selegiline, ropinirole, rasagiline, apomorphine, carbidopa, levodopa, pergolide, tolcapone, and amantadine. In some embodiments, the therapeutic agent may be an anti-inflammatory selected from the group consisting of nonsteroidal anti-inflammatory drugs and indomethacin. In some embodiments, the therapeutic agent may be a hormone selected from the group consisting of estrogen, progesterone, and leuprolide. In some embodiments, the therapeutic agent can be a vitamin selected from the group consisting of folate and nicotinamide. In some embodiments, the therapeutic agent can be xaliproden or homotaurine, which is 3-aminopropanesulfonic acid or 3APS. [Brief explanation of the drawings]

[0016] [Figure 1] 1 illustrates an exemplary embodiment of a telehealth service environment that may be utilized to detect predicted longitudinal changes in quantified speech variables associated with a patient as an estimate of the progression of Alzheimer's disease (AD) in the patient or the patient's treatment response.

[0017] [Figure 2] FIG. 1 shows a flow diagram of a method for detecting predicted longitudinal changes in quantified speech variables, including word length and particle usage variables, associated with a patient as an estimate of the progression of AD in the patient or the treatment response of the AD patient.

[0018] [Figure 3A]FIG. 1 shows a flow diagram of a method for detecting predicted long-term changes in quantified speech variables, including word length variables and Mel-Frequency Cepstral Coefficient (MFCC) variables, associated with a patient as an estimate of the progression of AD in the patient or the treatment response of the AD patient.

[0019] [Figure 3B] FIG. 1 shows a flow diagram of a method for detecting predicted long-term changes in quantified speech variables, including particle usage variables and Mel-Frequency Cepstral Coefficient (MFCC) variables associated with a patient, as an estimate of the progression of AD in the patient or the treatment response of the AD patient.

[0020] [Figure 4] FIG. 1 shows plots illustrating the longitudinal trajectories of a patient's language and acoustic-phonetic variables, which change linearly over time.

[0021] [Figure 5] 1 shows a tabular representation of standardized effect sizes of changes from baseline to endpoint in clinical assessment scores correlated with composite scores.

[0022] [Figure 6] FIG. 1 illustrates an exemplary computing system.

[0023] [Figure 7] FIG. 7 is a diagram of an exemplary artificial intelligence (AI) architecture included as part of the exemplary computing system of FIG. DETAILED DESCRIPTION OF THE INVENTION

[0024] The present disclosure relates to one or more computing devices, methods, and non-transitory computer-readable media that can be utilized to detect predicted longitudinal changes in quantified speech variables associated with a patient as an estimate of the patient's Alzheimer's disease (AD) progression or the AD patient's treatment response. Specifically, according to embodiments disclosed herein, one or more computing devices can utilize a machine learning model (e.g., a natural language processing (NLP) model, a transformation-based language model, an automatic speech recognition (ASR) model) to convert raw audio files of the patient's speech data captured at several moments over a period of time into text transcripts and analyze linguistic speech variables, including word length variables and particle usage variables, as well as one or more acoustic speech variables, to determine an estimate of the patient's AD progression or the patient's treatment response to which the patient's speech data corresponds.

[0025] The patient's speech data includes a recording of the patient's description of one or more previous or current experiences. In certain embodiments, the one or more computing devices may analyze the text transcript to quantify at least two speech variables derived from either or both one or more linguistic speech variables (e.g., a word length variable and a particle use variable, and optionally, a word frequency variable, a syntactic depth variable, a noun use variable, or a pronoun use variable) and one or more acoustic speech variables (e.g., one or more Mel-Frequency Cepstral Coefficient (MFCC) features).

[0026] In certain embodiments, one or more computing devices may generate a composite score based on standardization of at least two quantified speech variables derived from one or more linguistic speech variables and / or one or more acoustic speech variables, and substantial weighting of the at least two quantified speech variables derived from one or more linguistic speech variables and / or one or more acoustic speech variables. For example, in some embodiments, at least two quantified speech variables derived from one or more linguistic speech variables (e.g., a word length variable and a particle use variable, and optionally a word frequency variable, a syntactic depth variable, a noun use variable, or a pronoun use variable) and / or one or more acoustic speech variables (e.g., one or more Mel-Frequency Cepstral Coefficient (MFCC) features) may be standardized and combined into a composite score (e.g., an equal-weighted composite score, a weighted composite score). In one embodiment, the at least two quantified speech variables utilized to generate the composite score may include a word length variable and a particle use variable. In another embodiment, the at least two quantified speech variables utilized to generate the composite score may include a word length variable and at least one of an MFCC mean 11 variable, an MFCC var 25 variable, or an MFCC var 26 variable.

[0027] In certain embodiments, one or more computing devices can then determine predicted long-term changes in the quantified voice variables based on the composite score as an estimate of the patient's AD progression or AD patient treatment response.In this way, the present technology can provide an alternative to more invasive and more clinically burdensome tests for screening AD patients over time.Indeed, by generating a composite score based on at least two quantified voice variables derived from one or more linguistic voice variables and one or more acoustic voice variables identified as indicating progressive long-term changes, the present technology can provide a quantitative estimate of the patient's AD progression or AD patient treatment response using only the patient's voice.

[0028] As further described herein with respect to treatment or procedure: Therapeutic agents may include neuronal transmission enhancers, psychotherapeutic drugs, acetylcholinesterase inhibitors, calcium channel blockers, biogenic amines, benzodiazepine tranquilizers, acetylcholine synthesis, storage, or release enhancers, acetylcholine postsynaptic receptor agonists, monoamine oxidase A or B inhibitors, N-methyl-D-aspartate glutamate receptor antagonists, nonsteroidal anti-inflammatory drugs, antioxidants, or serotonin receptor antagonists. In particular, therapeutic agents may include compounds against oxidative stress, anti-apoptotic compounds, metal chelators, inhibitors of DNA repair, such as pirenzepine and metabolites, 3-amino-1-propanesulfonic acid (3APS), 1,3-propanedisulfonic acid (1,3PDS), secretase activators, beta- and gamma-secretase inhibitors, tau protein, anti-tau antibodies or anti-tau agents, neurotransmitters, beta-sheet breakers, anti-inflammatory molecules, "atypical antipsychotics," such as clozapine, ziprasin, and the like. The active ingredient may comprise at least one compound selected from the group consisting of acetaminophen, risperidone, aripiprazole or olanzapine, or cholinesterase inhibitors (ChEIs), such as tacrine, rivastigmine, donepezil, and / or galantamine, as well as other drugs or nutritional supplements, such as vitamin B12, cysteine, precursors of acetylcholine, lecithin, choline, ginkgo, acetyl-L-carnitine, idebenone, propentofylline, and / or xanthine derivatives.

[0029] In some embodiments, the therapeutic agent is a tau inhibitor. Non-limiting examples of tau inhibitors include methylthioninium, LMTX (also known as leuco-methylthioninium or Trx-0237, TauRx Therapeutics Ltd.), Rember™ (methylene blue or methylthioninium chloride [MTC], Trx-0014, TauRx Therapeutics Ltd), PBT2 (Prana Biotechnology), and PTI-51-CH3 (TauPro™, ProteoTech).

[0030] In some embodiments, the therapeutic agent is an anti-tau antibody. "Anti-tau immunoglobulin," "anti-tau antibody," and "antibody that binds to tau" are used interchangeably herein and refer to an antibody that can bind to tau (e.g., human tau) with sufficient affinity so as to be useful as a diagnostic and / or therapeutic agent in targeting tau. In some embodiments, the extent to which the anti-tau antibody binds to unrelated non-tau proteins is less than about 10% of the antibody's binding to tau, as measured, for example, by radioimmunoassay (RIA). In certain embodiments, the antibody that binds to tau has an affinity of 1 μM or less, 100 nM or less, 10 nM or less, 1 nM or less, 0.1 nM or less, 0.01 nM or less, or 0.001 nM or less (e.g., 10 -8 M or less, e.g. 10 -8 M~10 -13 M, e.g. 10 -9 M~10 -13 Dissociation constant (K D ). In certain embodiments, the anti-tau antibody binds to an epitope of tau that is conserved among tau from different species. In some cases, the antibody binds to monomeric tau, oligomeric tau, and / or phosphorylated tau. In some embodiments, the anti-tau antibody binds to monomeric tau, oligomeric tau, non-phosphorylated tau, and phosphorylated tau with comparable affinities, such as affinities that differ by no more than 50-fold in vitro. In some embodiments, antibodies that bind to monomeric tau, oligomeric tau, non-phosphorylated tau, and phosphorylated tau are referred to as "pan-tau antibodies." In some embodiments, the anti-tau antibody binds to an epitope within the N-terminal region of tau, e.g., an epitope within residues 2-24, e.g., an epitope within / spanning residues 6-23. In a specific embodiment, the anti-tau antibody is semolinemab.

[0031] In some embodiments, the anti-tau antibody is one or more selected from the group consisting of different N-terminal binders, mid-domain binders, and fibrillar tau binders. Non-limiting examples of other anti-tau antibodies include BIIB092 or BMS-986168 (Biogen, Bristol-Myers Squibb), APN-mAb005 (Aprinoia Therapeutics / Samsung Biologics), BIIB076 (Biogen / Eisai), ABBV-8E12 or C2N-8E12 (AbbVie, C2N Diagnostics, LLC), WO 2012049570, WO 2014028777, WO 2014165271, WO 2014100600, WO 2015200806, antibodies disclosed in U.S. Pat. No. 8,980,270 or U.S. Pat. No. 8,980,271, E2814 (Eisai), goslanemab (Biogen), tiravonemab (Abbvie), and zagotenemab (Lilly).

[0032] In some embodiments, the therapeutic agent is an anti-tau agent. Non-limiting examples include BIIB080 (Biogen / Ionis), LY3372689 (Lilly), PNT001 (Pinteon Therapeutics), OLX-07010 (Oligomerix, Inc.), TRx-0237 / LMTX (TauRx), JNJ-63733657 (Janssen), tau siRNA (Lilly / Dicerna), and PSY-02 (Psy Therapeutics).

[0033] These include GV-971 (Green Valley), CT1812 (Cognition Therapeutics), and ATH-1017 (Athira). Pharma), COR388 (Cortexyme), Carbohydrates (Cassava), Drugs (Novo Nordisk), Cosmetics (Anavex Life). Sciences) AR1001 (AriBio) BE (KeifeRx / Life Molecular Imaging / Sun Pharma, ALZ-801 (Alzheon), AL003 (Alector / AbbVie), Lomecel-B (Longeveron), UB-311 (Vaxxinity), XPro1595 / Manufacturer (INmune Bio), NLY-01 (D&D). Biotech) Protein / PQ912(Vivoryon / Nordic / Simcere) Protein (Novartis) Protein (New Amsterdam). Pharma)、AADvac1(Axon Neuroscience)、ANVS-401 / Posiphen(Annovis Bio)、TB006(TureBinding)、BI 474121(Boehringer). Ingelheim, NuCerin (Shaperon / Kukjeon), ALZ-101 (Alzinova), NNI-362 (Neuronascent), MK-1942 (Merck), E2511 (Eisai), IGC-AD1 (India Globalization Capital), AL001 (Alzamend). Neuro) AL002 (Alzamend Neuro) AL101 (Alector / GSK) MW-151 (ImmunoChem Therapeutics, DNL-788 / SAR443820 (Denali / Sanofi), ALN-APP (Alnylam / Regeneron), E2F4DN (Tetraneuron), EmtinB (NeuroScientific Biopharma), NIT-001 (Neurostech), ACD679 (AlzeCure). Pharma)、ACD680(AlzeCure Pharma)、YDC-103(YD Global Life).and at least one compound for treating AD selected from the group consisting of: Bioscience), BMD-001 (Biorchestra), STL-101 (Stellate Therapeutics), AV-1959R (Nuravax), AV1959D (Nuravax), AV1980R (Nuravax), Duvax (Nuravax), dapanstrill (Olatec Therapeutics), LX1001 (Cornell University), BDNF (UC San Diego), ST-501 (Biogen), AMT-240 (uniQure), SOL-410 (Sola), SOL-258 (Sola), AAVhmAb, SHP-231 (Shape), SHP-232 (Shape), TEL-01 (Telocyte), and GT-0007X (Gene Therapy).

[0034] In some embodiments, the therapeutic agent is a general misfolding inhibitor, such as NPT088 (NeuroPhage Pharmaceuticals).

[0035] In some embodiments, the therapeutic agent is a neurological drug. Neurological drugs include, but are not limited to, beta-secretase, presenilin, amyloid precursor protein or a portion thereof, amyloid beta peptide or an oligomer or fibril thereof, death receptor 6 (DR6), receptor for advanced glycation end products (RAGE), parkin, and huntingtin; NMDA receptor antagonists (i.e., memantine), monoamine depleting agents (i.e., tetrabenazine); ergoloid mesylates; anticholinergic parkinsonism agents (i.e., procyclidine, diphenhydramine, trihexylphenidyl, benztropine, biperiden, and trihexyphenidyl); dopaminergic parkinsonism agents (i.e., entacapone, selegiline, pramipexole, bromocriptine, rotigotine, selegiline, ropinirole, rasagiline, apomorphine, carbidopa, levothyroxine ... anti-inflammatory agents (including but not limited to nonsteroidal anti-inflammatory drugs (i.e., indomethacin and other compounds listed above)); hormones (i.e., estrogen, progesterone and leuprolide); vitamins (i.e., folic acid and nicotinamide); dimebolins; homotaurines (i.e., 3-aminopropanesulfonic acid, 3APS); serotonin receptor activity modulators (i.e., xaliproden); interferons, and antibodies or other binding molecules (including but not limited to small molecules, peptides, aptamers, or other protein binding agents) that specifically bind to a target selected from glucocorticoids or corticosteroids. The term "corticosteroid" includes, but is not limited to, fluticasone (including fluticasone propionate (FP)), beclomethasone, budesonide, ciclesonide, mometasone, flunisolide, betamethasone, and triamcinolone. "Inhalable corticosteroid" means a corticosteroid suitable for delivery by inhalation. Exemplary inhalable corticosteroids are fluticasone, beclomethasone propionate, budenoside, mometasone furoate, ciclesonide, flunisolide, and triamcinolone acetonide.

[0036] In certain embodiments, the therapeutic agent is one or more selected from the group consisting of corticosteroids, antibiotics, antiviral agents, anti-tau antibodies, tau inhibitors, anti-amyloid beta antibodies, beta-amyloid aggregation inhibitors, anti-BACE1 antibodies, BACE1 inhibitors; therapeutic agents that specifically bind to a target; cholinesterase inhibitors; NMDA receptor antagonists; monoamine depleting agents; ergoloid mesylates; anticholinergic parkinsonism agents; dopaminergic parkinsonism agents; tetrabenazine; anti-inflammatory agents; hormones; vitamins; dimebolins; homotaurines; serotonin receptor activity modulators; interferons, and glucocorticoids.

[0037] Non-limiting examples of anti-Abeta antibodies include crenezumab, solanezumab (Lilly), bapineuzumab, aducanumab, gantenerumab, donanemab (Lilly), LY3372993 (Lilly), ACU193 (Acumen Pharmaceuticals), SHR-1707 (Hengrui USA / Atridia), ALZ-201 (Alzinova), PMN-310 (ProMIS neurosciences), and lecanemab (BAN-2401, Biogen, Eisai Co., Ltd.). Non-limiting exemplary beta-amyloid aggregation inhibitors include ELND-005 (also known as AZD-103 or scyllo-inositol), tramiprosate, and PTI-80 (Exebryl-1®, ProteoTech). Non-limiting examples of BACE inhibitors include E-2609 (Biogen, Eisai Co., Ltd.), AZD3293 (also known as LY3314814, AstraZeneca, Eli Lilly & Co.), MK-8931 (Verubecestat), and JNJ-54861911 (Janssen, Shionogi Pharma).

[0038] In some embodiments, the therapeutic agent is an "atypical antipsychotic" such as, for example, clozapine, ziprasidone, risperidone, aripiprazole, or olanzapine for the treatment of positive and negative psychotic symptoms including hallucinations, delusions, thought disorder (manifested by prominent disorganized, deviant, and erratic thinking), and bizarre or disorganized behavior, as well as anhedonia, affective flattening, blunted affect, and social withdrawal.

[0039] In some embodiments, other therapeutic agents include, for example, those described in WO 2004 / 058258 (see especially pages 16 and 17), including therapeutic drug targets (pages 36-39), alkanesulfonic acids and alkanolsulfonic acids (pages 39-51), cholinesterase inhibitors (pages 51-56), NMDA receptor antagonists (pages 56-58), estrogens (pages 58-59), nonsteroidal anti-inflammatory drugs (pages 60-61), antioxidants (pages 61-62), and peroxisome proliferator-activated receptors. (PPAR) agonists (pages 63-67), cholesterol-lowering agents (pages 68-75); amyloid inhibitors (pages 75-77), amyloidogenesis inhibitors (pages 77-78), metal chelators (pages 78-79), antipsychotic and antidepressant agents (pages 80-82), nutritional supplements (pages 83-89) and compounds that increase the availability of biologically active substances in the brain (see pages 89-93) and prodrugs (pages 93 and 94), this document being incorporated herein by reference, in particular the compounds mentioned on the pages indicated above.

[0040] As further described herein with respect to measuring the severity and progression of Alzheimer's disease: The Mini-Mental State Examination ("MMSE") is a brief clinical cognitive test commonly used to screen for dementia and other cognitive deficits (Folstein et al. J Psychiatr Res 1975;12:189-98). The MMSE provides a total score of 0 to 30. A score of 26 or less is generally considered to indicate a deficit. The lower the numerical score on the MMSE, the greater the deficit or impairment in the tested patient compared to another individual with a higher score. An increase in the MMSE score may indicate an improvement in the patient's condition, while a decrease in the MMSE score may indicate a worsening of the patient's condition. In some embodiments, a stable MMSE score may indicate a slowing, delay, or halt in the progression of AD, or the lack of the emergence of new clinical, functional, or cognitive symptoms or impairments, or an overall stabilization of the disease.

[0041] The Clinical Dementia Rating Scale ("CDR") (Morris Neurology 1993;43:2412-4) is a semi-structured interview that yields five degrees of impairment of performance in each of six categories of cognitive-based function: memory, orientation, judgment and problem-solving, social problems, home and hobbies, and personal care. The CDR was originally designed with a global score: 0 - no dementia, 0.5 - probable dementia, 1 - mild dementia, 2 - moderate dementia, 3 - severe dementia.

[0042] The complete CDR-SB score is based on the sum of the scores across all six boxes. Subscores can also be obtained for each box or component individually, for example, CDR / Memory or CDR / Judgment and Problem Solving. As used herein, a "deterioration in CDR-SB performance" or an "increase in CDR-SB score" indicates a worsening of the patient's symptoms and may reflect the progression of AD.

[0043] The term "CDR-SB" refers to the Clinical Dementia Rating-Sum of Boxes, which provides a score of 0 to 18 (O'Bryant et al., 2008, Arch Neurol 65:1091-1095). The CDR-SB score is based on semi-structured interviews with patient and caregiver informants and yields five levels of impairment in performance for each of six categories of cognitive-based functioning: memory, orientation, judgment / problem-solving, community issues, home and hobbies, and personal care. The test is administered to both the patient and caregiver, and each component (or "box") is scored on a scale of 0 to 3 (the five levels are 0, 0.5, 1, 2, and 3). The sum of the scores for the six categories is the CDR-SB score. A decrease in the CDR-SB score may indicate an improvement in the patient's symptoms, whereas an increase in the CDR-SB score may indicate a worsening of the patient's symptoms. In some embodiments, a stable CDR-SB score may indicate a slowing, delay, or halt in the progression of AD, or a lack of emergence of new clinical, functional, or cognitive symptoms or impairments, or an overall stabilization of the disease.

[0044] The Alzheimer's Disease Assessment Scale-Cognitive Subscale ("ADAS-Cog") is a frequently used measure to assess cognition in clinical trials for mild to moderate AD (Rozzini et al. Int J Geriatr Psychiatry 2007;22:1217-22; Connor and Sabbagh, J Alzheimers Dis. 2008;15:461-4; Ihl et al. Int J Geriatr Psychiatry 2012;27:15-21). The ADAS-Cog is an examiner-administered battery that assesses multiple cognitive domains, including memory, comprehension, praxis, orientation, and spontaneous speech (Rosen et al. 1984, Am J Psychiatr 141:1356-64; Mohs et al. 1997, Alzheimer Dis Assoc Disord 11(S2):S13-S21). The ADAS-Cog is a standard primary endpoint in AD treatment trials (Mani 2004, Stat Med 23:305-14). The higher the numerical score on the ADAS-Cog, the greater the deficit or impairment of the tested patient compared to another individual with a lower score. The ADAS-Cog can be used to assess whether a treatment for AD is therapeutically effective. An increase in the ADAS-Cog score indicates a worsening of the patient's symptoms, while a decrease in the ADAS-Cog score indicates an improvement in the patient's symptoms. In some embodiments, a stable ADAS-Cog score can indicate a slowing, delay, or halt in the progression of AD, or the absence of the emergence of new clinical or cognitive symptoms or impairments, or an overall stabilization of the disease.

[0045] The ADAS-Cog12 is a 70-point version of the ADAS-Cog with the addition of a 10-point delayed word recall item that assesses recall of a trained word list. The ADAS-Cog11 is another version with a range of 0 to 70. Other ADAS-Cog scales include the ADAS-Cog13 and ADAS-Cog14.

[0046] A decrease in the ADAS-Cog11 score may indicate an improvement in the patient's condition, while an increase in the ADAS-Cog11 score may indicate a worsening of the patient's condition. In some embodiments, a stable ADAS-Cog11 score may indicate a slowing, delay, or halt in the progression of AD, or a reduction in the progression of clinical or cognitive decline, or the absence of the appearance of new clinical or cognitive symptoms or impairments, or an overall stabilization of the disease.

[0047] The component subtests of the ADAS-Cog11 can be grouped into three cognitive domains: memory, language, and praxis (Verma et al. Alzheimer's Research & Therapy 2015). This "breakdown" can improve sensitivity in measuring cognitive decline, for example, when focusing on mild to moderate AD stages (Verma, 2015). Thus, the ADAS-Cog11 score can be analyzed for changes in each of the three cognitive domains: memory, language, and praxis. The memory domain value of the ADAS-Cog11 score may be referred to herein as the "ADAS-Cog11 memory domain score" or simply the "memory domain." A slowdown in memory decline may refer to a reduction in the rate of decline in memory capacity and / or function, memory retention, and / or a reduction in memory loss. A slowdown in memory decline may be evidenced, for example, by a smaller (or less negative) score on the ADAS-Cog11 memory domain.

[0048] Similarly, the language domain value of the ADAS-Cog11 score may be referred to herein as the "ADAS-Cog11 language domain score" or simply the "language domain score," and the praxis domain value of the ADAS-Cog11 score may be referred to herein as the "ADAS-Cog11 praxis domain score" or simply the "praxis domain." Praxis can refer to the planning and / or execution of simple tasks, and / or praxis can refer to the ability to conceptualize, plan, and execute complex sequences of motor movements, as well as copy drawings or three-dimensional structures and following commands.

[0049] The memory domain score is further divided into components, including scores reflecting the subject's ability to recognize and / or recall words, thereby assessing "word recognition" or "word recall" abilities. The word recognition assessment of the ADAS-Cog11 memory domain score may be referred to herein as the "ADAS-Cog11 word recognition score" or simply the "word recognition score." For example, equivalent alternative forms of subtests for word recall and word recognition can be used in sequential testing of a given patient. A slowdown in memory decline can be evidenced, for example, by a smaller (or less negative) score on the word recognition component of the ADAS-Cog11 memory domain.

[0050] The Alzheimer's Disease Cooperative Study Group-Activities of Daily Living Inventory or Alzheimer's Disease Cooperative Study Group-Activities of Daily Living Scale ("ADCS-ADL;" Galasko et al. Alzheimer Dis Assoc Disord 1997;11(Suppl 2):S33-9) is the most widely used measure for assessing functional outcome in patients with AD (Vellas et al. Lancet Neurol. 2008;7:436-50). Scores range from 0 to 78, with higher scores indicating better ADL function. The ADCS-ADL is administered to caregivers and covers both basic ADLs (e.g., eating and toileting) and more complex or instrumental ADLs (e.g., telephone use, financial management, meal preparation) (Galasko et al. Alzheimer Disease and Associated Disorders, 1997 11(Suppl 2),S33-S39).

[0051] The Neuropsychiatric Inventory ("NPI") (Cummings et al. Neurology 1994;44:2308-14) is a widely used scale that assesses the behavioral symptoms of AD, including their frequency, severity, and associated distress. Individual symptom scores range from 0 to 12, and the total NPI score ranges from 0 to 144. The NPI is administered to caregivers and refers to the patient's behavior over the past month.

[0052] The Alzheimer's Disease Caregiver Global Impression Scale ("CaGI-Alz") is a novel scale used in the clinical trials described herein and consists of four items to assess caregivers' perceptions of changes in a patient's disease severity. All items are rated on a 7-point Likert-type scale ranging from 1 (much improved since treatment initiation / previous CaGI-Alz assessment) to 7 (much worse since treatment initiation / previous CaGI-Alz assessment).

[0053] The term "IADL" refers to the Instrumental Activities of Daily Living Scale (Lawton, MP, and Brody, EM, 1969, Gerontologist 9:179-186). This scale measures the ability to perform typical daily activities such as housework, laundry, using the telephone, shopping, and preparing meals. The lower the score, the more impaired the individual is in performing activities of daily living.

[0054] Another measure that may be used is the Amsterdam Activities of Daily Living Questionnaire (A-IADL-Q).

[0055] 1 illustrates an exemplary embodiment of a telehealth service environment 100 that may be utilized to detect predicted longitudinal changes in quantified audio variables associated with a patient as an estimate of the progression of AD in the patient or the treatment response of an AD patient, according to embodiments of the present disclosure. As shown, the telehealth service environment 100 may include multiple patients 102A, 102B, 102C, and 102D, each associated with a respective electronic device 104A, 104B, 104C, and 104D, which may be suitable for enabling the multiple patients 102A, 102B, 102C, and 102D to launch and collaborate with respective telehealth applications 106A (e.g., “Telehealth App1”), 106B (e.g., “Telehealth App2”), 106C (e.g., “Telehealth App3”), and 106D (e.g., “Telehealth AppN”).

[0056] 1, each electronic device 104A, 104B, 104C, and 104D may be connected to a telehealth service platform 112 via one or more communication network(s) 110. In particular embodiments, the telehealth service platform 112 may include, for example, a cloud-based computing architecture suitable for hosting and servicing telehealth applications 106A (e.g., “Telehealth Appl”), 106B (e.g., “Telehealth App2”), 106C (e.g., “Telehealth App3”), and 106D (e.g., “Telehealth AppN”) executing on each electronic device 104A, 104B, 104C, and 104D. For example, in one embodiment, the telemedicine services platform 112 may include a Platform as a Service (PaaS) architecture, a Software as a Service (SaaS) architecture, an Infrastructure as a Service (IaaS) architecture, a Compute as a Service (CaaS) architecture, a Data as a Service (DaaS) architecture, a Database as a Service (DBaaS) architecture, or other similar cloud-based computing architecture (e.g., "X" as a Service (XaaS)).

[0057] In certain embodiments, as further illustrated by FIG. 1, the telemedicine service platform 112 may include one or more processing devices 114 (eg, servers) and one or more data stores 116 . For example, in some embodiments, one or more processing devices 114 (e.g., servers) may include one or more general-purpose processors, graphics processing units (GPUs), application specific integrated circuits (ASICs), systems-on-chips (SoCs), microcontrollers, field programmable gate arrays (FPGAs), central processing units (CPUs), application processors (APs), visual processing units (VPUs), neural processing units (NPUs), neural decision processors (NDPs), deep learning processors (DLPs), tensor processing units (TPUs), neuromorphic processing units (NPUs), or any of a variety of other processing device(s) or accelerators that may be suitable for providing processing and / or computing support for telehealth applications 106A (e.g., “Telehealth App1”), 106B (e.g., “Telehealth App2”), 106C (e.g., “Telehealth App3”), and 106D (e.g., “Telehealth AppN”). Similarly, the data store 116 may include one or more internal databases that may be utilized to store, for example, information associated with multiple patients 102A, 102B, 102C, and 102D (e.g., audio files of patient voice data 118).

[0058] In particular embodiments, as previously described, the telehealth service platform 112 may be a hosting service platform for telehealth applications 106A (e.g., “Telehealth App1”), 106B (e.g., “Telehealth App2”), 106C (e.g., “Telehealth App3”), and 106D (e.g., “Telehealth AppN”) running on respective electronic devices 104A, 104B, 104C, and 104D. For example, in some embodiments, telehealth applications 106A (e.g., “Telehealth App1”), 106B (e.g., “Telehealth App2”), 106C (e.g., “Telehealth App3”), and 106D (e.g., “Telehealth AppN”) may each include a telehealth mobile application (e.g., a mobile “app”) that may be utilized to, for example, enable multiple patients 102A, 102B, 102C, and 102D to remotely access healthcare and medical care services and / or to collaborate with one or more patient-selected clinicians (e.g., clinician 126) as part of on-demand healthcare services.

[0059] In certain embodiments, one or more of the plurality of patients 102A, 102B, 102C, and 102D may include one or more patients with AD, one or more patients suspected of having AD, and / or one or more patients prone to developing AD. Thus, as further shown in FIG. 1 , in certain embodiments, one or more of the plurality of patients 102A, 102B, 102C, and 102D may undergo an audio-based assessment utilized to detect predicted longitudinal changes in quantified audio variables associated as estimates of AD progression or AD treatment response in one or more of the plurality of patients 102A, 102B, 102C, and 102D. For example, in certain embodiments, one or more of the plurality of patients 102A, 102B, 102C, and 102D may input speech 108A, 108B, 108C, and 108D utilizing telehealth applications 106A (e.g., “Telehealth App1”), 106B (e.g., “Telehealth App2”), 106C (e.g., “Telehealth App3”), and 106D (e.g., “Telehealth AppN”) executing on respective electronic devices 104A, 104B, 104C, and 104D. For example, in some embodiments, the input speech 108A, 108B, 108C, and 108D may include, for example, electronic recordings of the speech of the plurality of patients 102A, 102B, 102C, and 102D.

[0060] In particular embodiments, the input speech 108A, 108B, 108C, 108D may be performed in response to one or more requests provided by the telehealth service platform 112 to one or more of the plurality of patients 102A, 102B, 102C, and 102D, for example, via telehealth applications 106A (e.g., “Telehealth App1”), 106B (e.g., “Telehealth App2”), 106C (e.g., “Telehealth App3”), and 106D (e.g., “Telehealth AppN”). In other embodiments, one or more of the multiple patients 102A, 102B, 102C, and 102D may utilize one or more microphones of their respective electronic devices 104A, 104B, 104C, and 104D to record input audio 108A, 108B, 108C, and 108D without first being directed via a telehealth application 106A (e.g., “Telehealth App1”), 106B (e.g., “Telehealth App2”), 106C (e.g., “Telehealth App3”), and 106D (e.g., “Telehealth AppN”).

[0061] For example, in some embodiments, as part of the audio-based assessment, the telemedicine services platform 112 may generate and provide one or more audio-based tasks to one or more of the plurality of patients 102A, 102B, 102C, and 102D that instruct them to generate and record audio through one or more microphones on the respective electronic devices 104A, 104B, 104C, and 104D. In one embodiment, the audio-based assessment may include, for example, a description of an image that may be displayed via the telemedicine applications 106A, 106B, 106C, and 106D, a reading of a book passage that may be presented via the telemedicine applications 106A, 106B, 106C, and 106D, a series of question-and-answer tasks that may be presented via the telemedicine applications 106A, 106B, 106C, and 106D, or other audio-based assessments such as a medical-grade neuropsychological speech and language assessment.

[0062] In certain embodiments, the audio-based assessment may include a series of question-and-answer tasks based on a Clinical Dementia Assessment (CDR) interview. For example, in some embodiments, the series of question-and-answer tasks performed based on the CDR interview may include, for example, a series of question-and-answer tasks regarding one or more recent daily activities of the plurality of patients 102A, 102B, 102C, and 102D, one or more work-related activities of the plurality of patients 102A, 102B, 102C, and 102D, one or more hobby-related activities of the plurality of patients 102A, 102B, 102C, and 102D, or other previous or current experiences and / or activities that a cognitively healthy patient would be expected to easily recall. In certain embodiments, the audio-based assessment may be conducted at different time instants over some given period of time. For example, in some embodiments, an audio-based assessment may be conducted on an initial date, and then on one or more dates selected from the group including, for example, approximately 0.25 months, 0.5 months, 0.75 months, 1 month, 3 months, 6 months, 9 months, 12 months, 15 months, 18 months, 21 months, 24 months, 27 months, 30 months, 33 months, and / or 36 months from the initial date.

[0063] In certain embodiments, when one or more of the plurality of patients 102A, 102B, 102C, and 102D completes the audio-based assessment, one or more of the respective electronic devices 104A, 104B, 104C, and 104D may transmit one or more audio files of the patient audio data 118 to the telemedicine service platform 112. In certain embodiments, the one or more audio files of the patient audio data 118 may be stored in one or more data stores 116 of the telemedicine service platform 112. In certain embodiments, the one or more processing devices 114 (e.g., a server) may then access the one or more audio files of the patient audio data 118, analyze the one or more audio files of the patient audio data 118, and quantify one or more audio variables utilizing the one or more audio files of the patient audio data 118. For example, in particular embodiments, one or more processing devices 114 (e.g., a server) may utilize one or more machine learning models (e.g., natural language processing (NLP) models, transformation-based language models, automatic speech recognition (ASR) models) to convert raw audio files of the patient's speech data 118 into text representations (e.g., transcripts) or other representational data, and quantify at least two speech variables derived from, for example, one or more linguistic speech variables and one or more acoustic speech variables from the patient's speech data 118. In one embodiment, the at least two quantified speech variables include a word length variable and a particle usage variable. In another embodiment, the at least two quantified speech variables include a word length variable and at least one MFCC feature, including an MFCC mean 11 variable, an MFCC var 25 variable, or an MFCC var 26 variable.

[0064] In certain embodiments, the one or more linguistic phonetic variables may include one or more of a word length phonetic variable, a particle usage phonetic variable, a word frequency phonetic variable, a syntactic depth phonetic variable, a noun usage phonetic variable, and a pronoun usage phonetic variable. For example, in certain embodiments, the word length phonetic variable (e.g., the number of letters in a word spoken by one or more of the plurality of patients 102A, 102B, 102C, and 102D) measures the length of words in characters in the speech of one or more of the plurality of patients 102A, 102B, 102C, and 102D. The particle usage phonetic variable measures the usage of different particles (e.g., prepositions used in combination with other words to form multi-word phrases, clauses, or sentences). The word frequency phonetic variable measures the average frequency (e.g., lexical richness) of words utilized by one or more of the plurality of patients 102A, 102B, 102C, and 102D, for example, based on frequency norms in standard corpora.

[0065] In certain embodiments, the syntactic depth speech variable measures the complexity of syntactic structures (e.g., phrase length, clause complexity, proportion of different syntactic structures) utilized by one or more of the plurality of patients 102A, 102B, 102C, and 102D. The noun usage speech variable measures the usage of different parts of speech, such as nouns, and the pronoun usage speech variable measures the usage of different parts of speech, such as pronouns. Similarly, in certain embodiments, the one or more acoustic speech variables may include one or more Mel-Frequency Cepstral Coefficient (MFCC) features. The one or more MFCC features may include the mean of the 11th MFCC coefficient (MFCC mean 11), the variance of the first derivative of the 11th MFCC coefficient (MFCC var 25), and the variance of the first derivative of the 12th MFCC coefficient (MFCC var 26).

[0066] In certain embodiments, the one or more processing devices 114 (e.g., a server) may generate a composite score based on standardization of at least two quantified speech variables derived from one or more linguistic speech variables and / or one or more acoustic speech variables, and substantial weighting of the at least two quantified speech variables derived from one or more linguistic speech variables and / or one or more acoustic speech variables. For example, in some embodiments, the at least two quantified speech variables derived from one or more linguistic speech variables (e.g., a word length variable and a particle use variable, and optionally a word frequency variable, a syntactic depth variable, a noun use variable, or a pronoun use variable) and / or one or more acoustic speech variables (e.g., one or more Mel-Frequency Cepstral Coefficient (MFCC) features) may be standardized, for example, by subtracting the mean value of the speech variables across multiple patients 102A, 102B, 102C, and 102D and then dividing by the standard deviation. In some embodiments, at least two quantified speech variables derived from one or both of one or more linguistic speech variables (e.g., a word length variable and a particle use variable, and optionally a word frequency variable, a syntactic depth variable, a noun use variable, or a pronoun use variable) and one or more acoustic speech variables (e.g., one or more Mel-Frequency Cepstral Coefficient (MFCC) features) may then be substantially weighted (e.g., weightings including word length variable=1.0, word frequency variable=0.9, syntactic depth variable=0.9, noun use variable=0.8, pronoun use variable=0.8, particle use variable=1.0) and combined into a composite score (e.g., an equal-weighted composite score, a weighted composite score). In one embodiment, substantial weighting may refer to weights assigned to the plurality of quantified speech variables, for example, so as not to trivialize any one of the plurality of quantified speech variables. In one embodiment, the at least two quantified speech variables utilized to generate the composite score may include a word length variable and a particle use variable.In another embodiment, the at least two quantified speech variables utilized to generate the composite score may include a word length variable and at least one of an MFCC mean 11 variable, an MFCC var 25 variable, or an MFCC var 26 variable.

[0067] For example, in some embodiments, at least two quantified phonetic variables derived from either or both one or more linguistic phonetic variables (e.g., a word length variable and a particle use variable, and optionally a word frequency variable, a syntactic depth variable, a noun use variable, or a pronoun use variable) and one or more acoustic phonetic variables (e.g., one or more Mel-Frequency Cepstral Coefficient (MFCC) features) may be combined to generate a composite score for measuring or predicting longitudinal change over time.

[0068] In particular embodiments, one or more processing devices 114 (e.g., a server) may generate a composite score by: 1) standardizing at least two quantified speech variables derived from one or both of one or more linguistic speech variables (e.g., a word length variable and a particle use variable, and optionally, a word frequency variable, a syntactic depth variable, a noun use variable, or a pronoun use variable) and one or more acoustic speech variables (e.g., one or more Mel-Frequency Cepstral Coefficients (MFCC) features); 2) applying equal weighting to each of the at least two quantified speech variables derived from one or both of one or more linguistic speech variables (e.g., weightings including: word length variable=0.9, word frequency variable=0.9, syntactic depth variable=0.9, noun use variable=0.9, pronoun use variable=0.9, particle use variable=0.9) and one or more acoustic speech variables (e.g., one or more Mel-Frequency Cepstral Coefficients (MFCC) features); and 3) combining the standardized equally weighted quantified speech variables to generate a composite score. In one embodiment, the at least two quantified speech variables utilized to generate the composite score may include a word length variable and a particle usage variable. In another embodiment, the at least two quantified speech variables utilized to generate the composite score may include a word length variable and at least one of an MFCC mean 11 variable, an MFCC var 25 variable, or an MFCC var 26 variable.

[0069] In particular embodiments, one or more processing devices 114 (e.g., a server) may then determine measured or predicted long-term changes in the quantified speech variables based on the generated composite score. For example, as further understood with respect to Figures 1-3, as shown in Figures 4 and 5, one or more quantified linguistic speech variables (e.g., a word length variable and a particle use variable, and optionally a word frequency variable, a syntactic depth variable, a noun use variable, or a pronoun use variable) and one or more quantified acoustic speech variables (e.g., one or more Mel-Frequency Cepstral Coefficient (MFCC) features) may each be associated with long-term changes, and thus the composite score may include an overall prediction of long-term changes in the quantified speech variables.

[0070] In certain embodiments, the one or more processing devices 114 (e.g., a server) may then estimate the progression of AD for one or more of the plurality of patients 102A, 102B, 102C, and 102D based on the predicted long-term changes. For example, in some embodiments, the one or more processing devices 114 (e.g., a server) may estimate the progression of AD for one or more of the plurality of patients 102A, 102B, 102C, and 102D based on the correlation of the composite score with one or more clinical assessment metrics. For example, in some embodiments, the one or more processing devices 114 (e.g., a server) may utilize one or more linear mixed models (LMMs) or one or more other similar statistical models to correlate the effect of change over time on the composite score with the effect of change over time on one or more of, for example, the MMSE score, the CDR interview, the CDR-SB score, the ADAS-Cog score, the ADCS-ADL score, the NPI score, the NPI-Q score, the CaGI score, the IADL score, the A-IADL-Q score, or the RBANS score. Then, based on the correlation of the composite score with one or more clinical assessment metrics, the one or more processing devices 114 (e.g., a server) may generate an estimate 120 of AD progression for one or more of the plurality of patients 102A, 102B, 102C, and 102D.

[0071] 1, the one or more processing devices 114 (e.g., a server) may then transmit the generated AD progression estimate 120 to a computing device 122 and present a notification or report 124 to a clinician 126, who may be associated with a corresponding one of the plurality of patients 102A, 102B, 102C, and 102D. In an embodiment, the one or more processing devices 114 (e.g., a server) may also transmit the generated AD progression estimate 120 to a corresponding one of the plurality of patients 102A, 102B, 102C, and 102D via a respective electronic device 104A, 104B, 104C, or 104D. In particular embodiments, the clinician 126 may review the notification or report 124 and communicate with a corresponding one of the plurality of patients 102A, 102B, 102C, and 102D via a respective telehealth application 106A (e.g., “Telehealth App1”), 106B (e.g., “Telehealth App2”), 106C (e.g., “Telehealth App3”), or 106D (e.g., “Telehealth AppN”) regarding the cognitive health of the corresponding one of the plurality of patients 102A, 102B, 102C, and 102D.

[0072] For example, in certain embodiments, based on the medical review and analysis of the generated AD progression estimate 120, the clinician 126 may communicate via the computing device 122 a recommendation for a treatment regimen or adjustment of a therapeutic regimen for a corresponding one of the plurality of patients 102A, 102B, 102C, and 102D. In response to receiving the clinician's 126 input via the computing device 122, the one or more processing devices 114 (e.g., a server) may communicate a recommendation for a treatment regimen or adjustment of a therapeutic regimen for a corresponding one of the plurality of patients 102A, 102B, 102C, and 102D, e.g., a compound selected from the group consisting of compounds against oxidative stress, anti-apoptotic compounds, metal chelators, inhibitors of DNA repair, 3-amino-1-propanesulfonic acid (3APS), 1,3-propanedisulfonate (1,3PDS), secretase activators, beta- and gamma-secretase inhibitors, tau protein, anti-tau antibodies, anti-tau agents, gene therapy drugs, neurotransmitters, beta-sheet breakers, anti-inflammatory molecules, atypical antipsychotics, cholinesterase inhibitors, other drugs, and dietary supplements. In some embodiments, recommendations may be generated for adjustment of a treatment regimen comprising a therapeutic agent selected from the group consisting of a therapeutic agent comprising a serotonin receptor modulator, a symptomatic agent, a neurological drug, a corticosteroid, an antibiotic, an antiviral agent, an anti-tau antibody, a tau inhibitor, an anti-amyloid beta (anti-Aβ) antibody, a beta-amyloid aggregation inhibitor, a target binding therapeutic agent, an anti-BACE1 antibody, a BACE1 inhibitor, a cholinesterase inhibitor, an NMDA receptor antagonist, a monoamine depleting agent, an ergoloid mesylate, an anticholinergic antiparkinsonian agent, a dopaminergic antiparkinsonian agent, tetrabenazine, an anti-inflammatory agent, a hormone, a vitamin, a dimebolin, a homotaurine, a serotonin receptor activity modulator, an interferon, and a glucocorticoid. In certain embodiments, one or more processing devices 114 (e.g., a server) may then transmit, via the respective electronic devices 104A, 104B, 104C, or 104D, a notification regarding a treatment regimen or a recommendation for administration of a therapeutic regimen for a corresponding one of the plurality of patients 102A, 102B, 102C, and 102D, respectively.

[0073] FIG. 2 shows a flow diagram of a method 200 for detecting predicted longitudinal changes in quantified speech variables, including word length variables and particle usage, associated with a patient as an estimate of the progression of AD in the patient or the treatment response of the AD patient, according to an embodiment of the present disclosure. Method 200 may be implemented utilizing one or more processing devices (e.g., telemedicine service platform 112 as described above in connection with FIG. 1 ), which may include hardware (e.g., a general-purpose processor, a graphics processing unit (GPU), an application specific integrated circuit (ASIC), a system-on-chip (SoC), a microcontroller, a field programmable gate array (FPGA), a central processing unit (CPU), an application processor (AP), a visual processing unit (VPU), a neural processing unit (NPU), a neural decision processor (NDP), a deep learning processor (DLP), a tensor processing unit (TPU), a neuromorphic processing unit (NPU), or any other processing device(s) that may be suitable for processing various medical profile data and / or audio data and making one or more decisions based thereon), software (e.g., instructions operating / executing on one or more processors), firmware (e.g., microcode), or some combination thereof.

[0074] Method 200 may begin at block 202 with one or more processing devices receiving audio data including a patient's description of one or more previous or current experiences, the audio data captured at multiple moments over a period of time. For example, in certain embodiments, the patient's audio data may include a recording of one or more patient responses to questions or prompts included in a Clinical Dementia Assessment (CDR) interview. Method 200 then continues at block 204, where the one or more processing devices analyze the audio data to quantify multiple audio variables, the multiple audio variables including a word length variable and a particle usage variable. For example, in certain embodiments, the word length audio variable may include a measure of the number of letters included in words spoken by one or more of the plurality of patients 102A, 102B, 102C, and 102D. Similarly, the particle usage audio variable may include a measure of the usage of different particles (e.g., a preposition used in combination with another word to form a multi-word phrase, fragment, or sentence) spoken by one or more of the plurality of patients 102A, 102B, 102C, and 102D.

[0075] Method 200 then continues at block 206, where one or more processing devices may determine a composite score based on the standardization of the quantified plurality of speech variables and the substantial weighting assigned to each of the quantified plurality of speech variables. For example, in some embodiments, at least two quantified speech variables derived from either or both one or more linguistic speech variables (e.g., a word length variable and a particle use variable, and optionally a word frequency variable, a syntactic depth variable, a noun use variable, or a pronoun use variable) and one or more acoustic speech variables (e.g., one or more Mel-Frequency Cepstral Coefficient (MFCC) features) may be standardized, substantially weighted, and combined into a composite score (e.g., an equal-weighted composite score, a weighted composite score). In one embodiment, the at least two quantified speech variables utilized to generate the composite score may include a word length variable and a particle use variable. In another embodiment, the at least two quantified speech variables utilized to generate the composite score may include a word length variable and at least one of an MFCC mean 11 variable, an MFCC var 25 variable, or an MFCC var 26 variable.

[0076] Method 200 may then continue at block 208 with one or more processing devices detecting predicted long-term changes in the quantified plurality of speech variables based on the composite score. Method 200 may then end at block 210 with one or more processing devices estimating the patient's AD progression based on the predicted long-term changes. For example, in some embodiments, the one or more processing devices may estimate the AD progression by correlating the composite score with one or more clinical assessment metrics (e.g., MMSE score, CDR interview, CDR-SB score, ADAS-Cog score, ADCS-ADL score, NPI score, NPI-Q score, CaGI score, IADL score, A-IADL-Q score, or RBANS score) based on the predicted long-term changes.

[0077] FIG. 3A shows a flow diagram of a method 300A for detecting predicted longitudinal changes in quantified speech variables, including word length variables and MFCC variables, associated with a patient as an estimate of the progression of AD in the patient or the treatment response of the AD patient, according to an embodiment of the present disclosure. Method 300A may be implemented utilizing one or more processing devices (e.g., the telemedicine service platform 112 as described above in connection with FIG. 1 ), which may include hardware (e.g., a general-purpose processor, a graphics processing unit (GPU), an application specific integrated circuit (ASIC), a system-on-chip (SoC), a microcontroller, a field programmable gate array (FPGA), a central processing unit (CPU), an application processor (AP), a visual processing unit (VPU), a neural processing unit (NPU), a neural decision processor (NDP), a deep learning processor (DLP), a tensor processing unit (TPU), a neuromorphic processing unit (NPU), or any other processing device(s) that may be suitable for processing various medical profile data and / or audio data and making one or more decisions based thereon), software (e.g., instructions operating / executing on one or more processors), firmware (e.g., microcode), or some combination thereof.

[0078] Method 300A may begin at block 302 with one or more processing devices receiving audio data including a patient's description of one or more previous or current experiences, the audio data captured at multiple moments over a period of time. For example, in certain embodiments, the patient's audio data may include a recording of one or more responses of the patient to questions or prompts included in a Clinical Dementia Assessment (CDR) interview. Method 300A then continues at block 304 with one or more processing devices analyzing the audio data to quantify multiple audio variables, the multiple audio variables including a word length variable and a Mel-Frequency Cepstral Coefficient (MFCC) audio variable. For example, in certain embodiments, the word length audio variable may include a measure of the number of letters included in words spoken by one or more of the multiple patients 102A, 102B, 102C, and 102D. Similarly, the MFCC audio variables may include the mean of the 11th MFCC coefficient (MFCC mean 11), the variance of the first derivative of the 11th MFCC coefficient (MFCC var 25), or the variance of the first derivative of the 12th MFCC coefficient (MFCC var 26).

[0079] Method 300A then continues at block 306, where one or more processing devices may determine a composite score based on the standardization of the quantified speech variables and the substantial weighting assigned to each of the quantified speech variables. For example, in some embodiments, at least two quantified speech variables derived from one or both of one or more linguistic speech variables (e.g., a word length variable and a particle use variable, and optionally a word frequency variable, a syntactic depth variable, a noun use variable, or a pronoun use variable) and one or more acoustic speech variables (e.g., one or more Mel-Frequency Cepstral Coefficient (MFCC) features) may be standardized, substantially weighted, and combined into a composite score (e.g., an equal-weighted composite score, a weighted composite score). For example, the at least two quantified speech variables utilized to generate the composite score may include a word length variable and at least one of an MFCC mean 11 variable, an MFCC var 25 variable, or an MFCC var 26 variable.

[0080] Method 300A may then continue with one or more processing devices detecting predicted long-term changes in the quantified plurality of speech variables based on the composite score at block 308. Method 300A may then end with one or more processing devices estimating the patient's AD progression based on the predicted long-term changes at block 310. For example, in some embodiments, the one or more processing devices may estimate AD progression by correlating the composite score with one or more clinical assessment metrics (e.g., MMSE score, CDR interview, CDR-SB score, ADAS-Cog score, ADCS-ADL score, NPI score, NPI-Q score, CaGI score, IADL score, A-IADL-Q score, or RBANS score) based on the predicted long-term changes.

[0081] FIG. 3B shows a flow diagram of a method 300A for detecting predicted longitudinal changes in quantified speech variables, including particle usage variables and MFCC speech variables, associated with a patient as an estimate of the progression of AD in the patient or the treatment response of the AD patient, according to an embodiment of the present disclosure. Method 300B may be implemented utilizing one or more processing devices (e.g., the computing systems and artificial intelligence architectures described below in connection with FIGS. 6 and 7 ) that may include hardware (e.g., a general-purpose processor, a graphics processing unit (GPU), an application specific integrated circuit (ASIC), a system-on-chip (SoC), a microcontroller, a field programmable gate array (FPGA), a central processing unit (CPU), an application processor (AP), a visual processing unit (VPU), a neural processing unit (NPU), a neural decision processor (NDP), a deep learning processor (DLP), a tensor processing unit (TPU), a neuromorphic processing unit (NPU), or any other processing device(s) that may be suitable for processing various medical profile data and / or audio data and making one or more decisions based thereon), software (e.g., instructions operating / executing on one or more processors), firmware (e.g., microcode), or some combination thereof.

[0082] Method 300B may begin at block 312 with one or more processing devices receiving audio data including a patient's description of one or more previous or current experiences, the audio data captured at multiple moments over a period of time. For example, in certain embodiments, the patient's audio data may include a recording of one or more responses of the patient to questions or prompts included in a Clinical Dementia Assessment (CDR) interview. Method 300B then continues at block 314, where the one or more processing devices analyze the audio data to quantify multiple audio variables, including particle usage variables and Mel-Frequency Cepstral Coefficient (MFCC) variables. For example, in certain embodiments, the particle usage audio variable may include a measure of the usage of different particles (e.g., a preposition used in combination with another word to form a multi-word phrase, fragment, or sentence). Similarly, the MFCC audio variables may include the mean of the 11th MFCC coefficient (MFCC mean 11), the variance of the first derivative of the 11th MFCC coefficient (MFCC var 25), or the variance of the first derivative of the 12th MFCC coefficient (MFCC var 26).

[0083] Method 300B then continues at block 316, where one or more processing devices may determine a composite score based on the standardization of the quantified speech variables and the substantial weighting assigned to each of the quantified speech variables. For example, in some embodiments, at least two quantified speech variables derived from one or both of one or more linguistic speech variables (e.g., a word length variable and a particle use variable, and optionally a word frequency variable, a syntactic depth variable, a noun use variable, or a pronoun use variable) and one or more acoustic speech variables (e.g., one or more Mel-Frequency Cepstral Coefficient (MFCC) features) may be standardized, substantially weighted, and combined into a composite score (e.g., an equal-weighted composite score, a weighted composite score). For example, the at least two quantified speech variables utilized to generate the composite score may include a word length variable and at least one of an MFCC mean 11 variable, an MFCC var 25 variable, or an MFCC var 26 variable.

[0084] Method 300B may then continue with one or more processing devices detecting predicted long-term changes in the quantified plurality of speech variables based on the composite score at block 318. Method 300B may then end with one or more processing devices estimating the patient's AD progression based on the predicted long-term changes at block 320. For example, in some embodiments, the one or more processing devices may estimate the AD progression by correlating the composite score with one or more clinical assessment metrics (e.g., MMSE score, CDR interview, CDR-SB score, ADAS-Cog score, ADCS-ADL score, NPI score, NPI-Q score, CaGI score, IADL score, A-IADL-Q score, or RBANS score) based on the predicted long-term changes.

[0085] 4 illustrates a plot 400 showing a longitudinal trajectory of a patient's language and acoustic speech variables that change linearly over time, according to an embodiment of the present disclosure. In certain embodiments, plots 402, 404, 406, 408, 410, and 412 may illustrate several quantified language speech variables, including, for example, a word length speech variable (e.g., plot 402), a syntactic depth speech variable (e.g., plot 404), a word frequency speech variable (e.g., plot 406), a noun usage speech variable (e.g., plot 408), a particle usage speech variable (e.g., plot 410), and a pronoun usage speech variable (e.g., plot 412).

[0086] In certain embodiments, plots 402, 404, 406, 408, 410, and 412 each show a Pearson correlation coefficient (e.g., "R," which represents a number between -1 and +1 that reflects the tendency for two random variables to be linearly related) and / or a Pearson correlation p-value (e.g., "p," which represents an indication of whether the correlation is statistically significant) for a number of quantified linguistic phonetic variables, including, for example, a word length phonetic variable (e.g., plot 402), a syntactic depth phonetic variable (e.g., plot 404), a word frequency phonetic variable (e.g., plot 406), a noun use phonetic variable (e.g., plot 408), a particle use phonetic variable (e.g., plot 410), and a pronoun use phonetic variable (e.g., plot 412), plotted against time (e.g., first day from baseline, approximately 6 months from first day, approximately 12 months from first day, and approximately 18 months from first day), respectively.

[0087] In one embodiment, the Pearson correlation p-values for each of a number of quantified language phonetic variables, including, for example, a word length phonetic variable (e.g., plot 402), a syntactic depth phonetic variable (e.g., plot 404), a word frequency phonetic variable (e.g., plot 406), a noun use phonetic variable (e.g., plot 408), a particle use phonetic variable (e.g., plot 410), and a pronoun use phonetic variable (e.g., plot 412), were each plotted against time with a statistically significant effect of time at p<0.001 (e.g., first day from baseline, approximately 6 months from first day, approximately 12 months from first day, and approximately 18 months from first day). Specifically, as shown in Figure 4, the longitudinal trajectories for word length phonetic variables (e.g., plot 402), syntactic depth phonetic variables (e.g., plot 404), word frequency phonetic variables (e.g., plot 406), noun use phonetic variables (e.g., plot 408), particle use phonetic variables (e.g., plot 410), and pronoun use phonetic variables (e.g., plot 412) each trend toward corresponding one or more patients utilizing shorter, more frequent words, simpler sentence syntax, fewer nouns, and more particles and pronouns over time, for example. Thus, plots 402, 404, 406, 408, 410, and 412 in Figure 4 illustrate correlations between language phonetic variables and longitudinal change over time (e.g., over approximately 18 months).

[0088] Similarly, in particular embodiments, plots 414, 416, and 418 may show several quantified acoustic speech variables, including, for example, the 11th MFCC coefficient (MFCC mean 11) speech variable (e.g., plot 414), the variance of the first derivative of the 11th MFCC coefficient (MFCC var 25) (e.g., plot 416), and the variance of the first derivative of the 12th MFCC coefficient (MFCC var 26) (e.g., plot 416), each plotted against time (e.g., first day from baseline, approximately 6 months from first day, approximately 12 months from first day, and approximately 18 months from first day). In one embodiment, the Pearson correlation p-values for each of a number of quantified acoustic speech variables, including, for example, the 11th MFCC coefficient (MFCC mean 11) speech variable (e.g., plot 414), the variance of the first derivative of the 11th MFCC coefficient (MFCC var 25) (e.g., plot 416), and the variance of the first derivative of the 12th MFCC coefficient (MFCC var 26) (e.g., plot 416), each plotted against time had a statistically significant effect of time at p<0.001 (e.g., first day from baseline, approximately 6 months from first day, approximately 12 months from first day, and approximately 18 months from first day). Thus, plots 414, 416, and 418 in FIG. 4 show correlations of acoustic speech variables to longitudinal change over time (e.g., greater than approximately 18 months).

[0089] 5 shows a tabular illustration 500 of standardized effect sizes of changes from baseline to endpoint in clinical assessment scores correlated with composite scores, according to an embodiment of the present disclosure. For example, in one embodiment, the composite score 502 may be generated based on several phonetic variables, including a word length phonetic variable, a word frequency phonetic variable, a syntactic depth phonetic variable, a noun use phonetic variable, a pronoun use phonetic variable, a particle use phonetic variable, a mean of the eleventh MFCC coefficient (MFCC mean 11) phonetic variable, a variance of the first derivative of the eleventh MFCC coefficient (MFCC var 25) phonetic variable, and a variance of the first derivative of the twelfth MFCC coefficient (MFCC var 26) phonetic variable. Specifically, according to an embodiment of the present disclosure, the composite score 502 may be generated by standardizing and equally weighting each of the following phonetic variables: a word length phonetic variable, a word frequency phonetic variable, a syntactic depth phonetic variable, a noun usage phonetic variable, a pronoun usage phonetic variable, a particle usage phonetic variable, the mean of the 11th MFCC coefficient (MFCC mean 11) phonetic variable, the variance of the first derivative of the 11th MFCC coefficient (MFCC var 25) phonetic variable, and the variance of the first derivative of the 12th MFCC coefficient (MFCC var 26) phonetic variable, and combining these phonetic variables into the composite score 502.

[0090] In other embodiments, the composite score 502 may be generated based on standardization of at least two phonetic variables derived from one or more linguistic phonetic variables (e.g., a word length variable and a particle use variable, and optionally a word frequency variable, a syntactic depth variable, a noun use variable, or a pronoun use variable) and one or more acoustic phonetic variables (e.g., MFCC mean 11, MFCC var 25, MFCC var 26) phonetic variables), and substantial weighting of the at least two phonetic variables derived from one or more linguistic phonetic variables and one or more acoustic phonetic variables. In some embodiments, substantial weighting may refer to weights assigned to at least two phonetic variables derived from one or more linguistic phonetic variables and one or more acoustic phonetic variables, for example, to avoid trivializing any one of the one or more linguistic phonetic variables and one or more acoustic phonetic variables. In one embodiment, the at least two quantified phonetic variables utilized to generate the composite 502 score may include a word length variable and a particle use variable. In another embodiment, the at least two quantified speech variables utilized to generate the composite score 502 may include a word length variable and at least one of an MFCC mean 11 variable, an MFCC var 25 variable, or an MFCC var 26 variable.

[0091] In certain embodiments, as shown in table diagram 500 of FIG. 5 , the generated composite score 502 (e.g., composite score=0.29) has a similar effect size for detecting longitudinal change compared to CDR-Sum of Boxes score 504 (e.g., CDR-SB=0.30), Alzheimer's Disease Cooperative Study Group-Activities of Daily Living Inventory (ADCS-ADL) score 506 (e.g., ADCS-ADL=−0.30), Mini-Mental State Examination (MMSE) score 508 (e.g., MMSE=−0.23), Alzheimer's Disease Assessment Scale-Cognitive Subscale (ADAS-Cog) score 510 (e.g., ADAS-Cog=0.22), and Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) score 512 (e.g., RBANS=−0.15). Thus, the table diagram 500 of FIG. 5 illustrates that the generated composite score 502 described herein (e.g., composite score=0.29) can be utilized to accurately detect predicted long-term changes in quantified speech variables associated with a patient as an estimate of the progression of AD in the patient or the treatment response of the AD patient.

[0092] FIG. 6 illustrates an exemplary computing system 600 that may be utilized to detect predicted longitudinal changes in quantified voice variables associated with a patient and to detect the severity and progression of AD in the patient based on the predicted longitudinal changes in the quantified voice variables, according to embodiments of the present disclosure. In certain embodiments, computing system 600 may perform one or more steps of one or more methods described or illustrated herein. In certain embodiments, computing system 600 provides functionality described or illustrated herein. In certain embodiments, software running on computing system 600 performs one or more steps of one or more methods described or illustrated herein or provides functionality described or illustrated herein. Certain embodiments include one or more portions of computing system 600. As used herein, references to a computer system may encompass computing devices, and vice versa, where appropriate. Furthermore, references to a computer system may encompass one or more computer systems, where appropriate.

[0093] The present disclosure contemplates any suitable number of computing systems 600. The present disclosure contemplates computing system 600 taking any suitable physical form. By way of example and not limitation, computing system 600 may be an embedded computer system, a system-on-chip (SOC), a single-board computer system (SBC) (e.g., a computer-on-module (COM) or system-on-module (SOM)), a desktop computer system, a laptop or notebook computer system, an interactive kiosk, a mainframe, a mesh of computer systems, a mobile phone, a personal digital assistant (PDA), a server, a tablet computer system, an augmented / virtual reality device, or a combination of two or more of these. Where appropriate, computing system 600 may include one or more computing systems 600, may be unitary or distributed, may span multiple locations, span multiple machines, span multiple data centers, or reside in a cloud, which may include one or more cloud components in one or more networks.

[0094] Where appropriate, computing system 600 may perform one or more steps of one or more methods described or shown herein without substantial spatial or temporal limitations. By way of example and not limitation, computing system 600 may perform one or more steps of one or more methods described or shown herein in real time or in batch mode. Computing system 600 may perform one or more steps of one or more methods described or shown herein at different times or in different locations, where appropriate.

[0095] In particular embodiments, computing system 600 includes a processor 602, memory 604, storage 606, an input / output (I / O) interface 608, a communication interface 610, and a bus 612. Although this disclosure describes and illustrates a particular computer system having a particular number of particular components in a particular arrangement, this disclosure contemplates any suitable computer system having any suitable number of any suitable components in any suitable arrangement. In particular embodiments, processor 602 includes hardware for executing instructions, such as instructions making up a computer program. By way of example and not limitation, to execute instructions, processor 602 may retrieve (or fetch) instructions from an internal register, an internal cache, memory 604, or storage 606, decode and execute those instructions, and then write one or more results to an internal register, an internal cache, memory 604, or storage 606. In particular embodiments, processor 602 may include one or more internal caches for data, instructions, or addresses. This disclosure contemplates processor 602 including any suitable number of any suitable internal caches, where appropriate. By way of example and not limitation, processor 602 may include one or more instruction caches, one or more data caches, and one or more translation lookaside buffers (TLBs). Instructions in an instruction cache may be copies of instructions in memory 604 or storage 606, and the instruction cache may speed up retrieval of those instructions by processor 602.

[0096] The data in the data cache may be a copy of data in memory 604 or storage 606 upon which instructions executing in processor 602 operate, the results of a previous instruction executed in processor 602 for access by a subsequent instruction executing in processor 602 or for writing to memory 604 or storage 606, or other suitable data. The data cache may speed up read or write operations by processor 602. The TLB may speed up virtual-address translation for processor 602. In particular embodiments, processor 602 may include one or more internal registers for data, instructions, or addresses. This disclosure contemplates processor 602 including any suitable number of any suitable internal registers, where appropriate. Where appropriate, processor 602 may include one or more arithmetic logic units (ALUs), be a multi-core processor, or include one or more processors 602. While this disclosure describes and illustrates a particular processor, this disclosure contemplates any suitable processor.

[0097] In particular embodiments, memory 604 includes main memory for storing instructions for processor 602 to execute or data upon which processor 602 operates. By way of example and not limitation, computing system 600 may load instructions into memory 604 from storage 606 or another source (such as, for example, another computing system 600). Processor 602 may then load the instructions from memory 604 into an internal register or internal cache. To execute instructions, processor 602 may retrieve the instructions from the internal register or internal cache and decode them. During or after execution of an instruction, processor 602 may write one or more results (which may be intermediate or final results) to an internal register or internal cache. Processor 602 may then write one or more of those results to memory 604.

[0098] In particular embodiments, processor 602 executes only instructions in one or more internal registers or internal caches or in memory 604 (as opposed to storage 606 or elsewhere) and operates only on data in one or more internal registers or internal caches or in memory 604 (as opposed to storage 606 or elsewhere). One or more memory buses (which may each include an address bus and a data bus) may connect processor 602 to memory 604. Bus 612, as described below, may include one or more memory buses. In particular embodiments, one or more memory management units (MMUs) reside between processor 602 and memory 604 and facilitate accesses to memory 604 requested by processor 602. In particular embodiments, memory 604 includes random access memory (RAM). This RAM may be volatile memory, where appropriate. Where appropriate, this RAM may be dynamic RAM (DRAM) or static RAM (SRAM). Further, where appropriate, this RAM may be single-ported or multi-ported RAM. The present disclosure contemplates any suitable RAM. Memory 604 may include, where appropriate, one or more memory devices 604. Although this disclosure describes and illustrates particular memory, this disclosure contemplates any suitable memory.

[0099] In particular embodiments, storage 606 includes mass storage for data or instructions. By way of example and not limitation, storage 606 may include a hard disk drive (HDD), a floppy disk drive, flash memory, an optical disk, a magneto-optical disk, magnetic tape, or a universal serial bus (USB) drive, or a combination of two or more of these. Storage 606 may include removable or non-removable (or fixed) media, where appropriate. Storage 606 may be internal or external to computing system 600, where appropriate. In particular embodiments, storage 606 is non-volatile solid-state memory. In particular embodiments, storage 606 includes read-only memory (ROM). Where appropriate, this ROM may be mask-programmed ROM, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), electrically alterable ROM (EAROM), or flash memory, or a combination of two or more of these. The present disclosure contemplates mass storage 606 taking any suitable physical form. Storage 606 may include, where appropriate, one or more storage control units that facilitate communication between processor 602 and storage 606. Where appropriate, storage 606 may include one or more storages 606. Although this disclosure describes and illustrates particular storage, this disclosure contemplates any suitable storage.

[0100] In particular embodiments, I / O interface 608 includes hardware, software, or both that provide one or more interfaces for communication between computing system 600 and one or more I / O devices. Computing system 600 may include one or more of these I / O devices, where appropriate. One or more of these I / O devices may enable communication between a person and computing system 600. By way of example and not limitation, an I / O device may include a keyboard, keypad, microphone, monitor, mouse, printer, scanner, speaker, still camera, stylus, tablet, touch screen, trackball, video camera, another suitable I / O device, or a combination of two or more of these. An I / O device may include one or more sensors. This disclosure contemplates any suitable I / O devices and any suitable I / O interface 606 for those I / O devices. Where appropriate, I / O interface 608 may include one or more device or software drivers that enable processor 602 to drive one or more of these I / O devices. I / O interface 608 may include, where appropriate, one or more I / O interfaces 606. Although this disclosure describes and illustrates a particular I / O interface, this disclosure contemplates any suitable I / O interface.

[0101] In particular embodiments, communication interface 610 includes hardware, software, or both that provide one or more interfaces for communication (e.g., packet-based communication) between computing system 600 and one or more other computer systems 600 or one or more networks. By way of example and not limitation, communication interface 610 may include a network interface controller (NIC) or network adapter for communicating with an Ethernet or other wire-based network, or a wireless NIC (WNIC) or wireless adapter for communicating with a wireless network, such as a WI-FI network. The present disclosure contemplates any suitable network and any suitable communication interface 610 for that network.

[0102] By way of example, and not by way of limitation, computing system 600 may communicate with one or more portions of an ad hoc network, a personal area network (PAN), a local area network (LAN), a wide area network (WAN), a metropolitan area network (MAN), or the Internet, or a combination of two or more of these. One or more portions of one or more of these networks may be wired or wireless. By way of example, computing system 600 may communicate with a wireless PAN (WPAN) (e.g., a BLUETOOTH WPAN), a Wi-Fi network, a Wi-MAX network, a cellular telephone network (e.g., a Global System for Mobile Communications (GSM) network), or other suitable wireless networks, or a combination of two or more of these. Computing system 600 may include any suitable communication interface 610 for any of these networks, where appropriate. Communication interface 610 may include one or more communication interfaces 610, where appropriate. Although this disclosure describes and illustrates a particular communication interface, this disclosure contemplates any suitable communication interface.

[0103] In particular embodiments, bus 612 includes hardware, software, or both that connects components of computing system 600 to one another. By way of example and not limitation, bus 612 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a MicroChannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI Express (PCIe) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or another suitable bus, or a combination of two or more of these. Bus 612 may include one or more buses 612, where appropriate. Although this disclosure describes and illustrates a particular bus, this disclosure contemplates any suitable bus or interconnect.

[0104] As used herein, one or more computer-readable non-transitory storage media may comprise, where appropriate, one or more semiconductor-based or other integrated circuits (ICs) (such as, for example, field programmable gate arrays (FPGAs) or application-specific ICs (ASICs)), hard disk drives (HDDs), hybrid hard drives (HHDs), optical disks, optical disk drives (ODDs), magneto-optical disks, magneto-optical drives, floppy diskettes, floppy disk drives (FDDs), magnetic tapes, solid-state drives (SSDs), RAM drives, secure digital cards or drives, any other suitable computer-readable non-transitory storage media, or any suitable combination of two or more of these. A computer-readable non-transitory storage medium may, where appropriate, be volatile, non-volatile, or a combination of volatile and non-volatile.

[0105] FIG. 7 illustrates a diagram 700 of an exemplary artificial intelligence (AI) architecture 702 (which may be included as part of computing system 600, as described above with respect to FIG. 6) that may be utilized to detect predicted long-term changes in quantified voice variables associated with a patient and to detect the severity and progression of AD in the patient based on the predicted long-term changes in the quantified voice variables, according to an embodiment of the present disclosure. In particular embodiments, the AI architecture 702 may be implemented utilizing one or more processing devices, which may include, for example, hardware (e.g., a general-purpose processor, a graphics processing unit (GPU), an application specific integrated circuit (ASIC), a system-on-chip (SoC), a microcontroller, a field programmable gate array (FPGA), a central processing unit (CPU), an application processor (AP), a visual processing unit (VPU), a neural processing unit (NPU), a neural decision processor (NDP), a deep learning processor (DLP), a tensor processing unit (TPU), a neuromorphic processing unit (NPU), and / or any other processing device(s) that may be suitable for processing various medical profile data and making one or more decisions based thereon), software (e.g., instructions operating / executing on one or more processing devices), firmware (e.g., microcode), or some combination thereof.

[0106] 7, AI architecture 702 may include machine learning (ML) algorithms and functions 704, natural language processing (NLP) algorithms and functions 706, expert systems 708, computer-based vision algorithms and functions 710, speech recognition algorithms and functions 712, planning algorithms and functions 714, and robotics algorithms and functions 716. In particular embodiments, ML algorithms and functions 704 may include any statistically-based algorithms that may be suitable for finding patterns across large amounts of data (e.g., “big data” such as genomics data, proteomics data, metabolomics data, metagenomics data, transcriptomics data, medication data, medical diagnostic data, medical procedure data, medical diagnosis data, medical symptom data, demographic data, patient lifestyle data, physical activity data, family history data, socioeconomic data, geographic environment data, etc.). For example, in particular embodiments, ML algorithms and functions 704 may include deep learning algorithms 718, supervised learning algorithms 720, and unsupervised learning algorithms 722.

[0107] In particular embodiments, the deep learning algorithm 718 may include any artificial neural network (ANN) that can be utilized to learn deep-level representations and abstractions from large amounts of data. For example, the deep learning algorithm 718 may include an ANN such as a perceptron, a multi-layer perceptron (MLP), an autoencoder (AE), a convolutional neural network (CNN), a recurrent neural network (RNN), a long short-term memory (LSTM), a grated recurrent unit (GRU), a restricted Boltzmann machine (RBM), a deep belief network (DBN), a bidirectional recurrent deep neural network (BRDNN), a generative adversarial network (GAN), and a deep Q-network, a neural autoregressive distribution estimator (NADE), an adversarial network (AN), an attention model (AM), a spiking neural network (SNN), deep reinforcement learning, etc.

[0108] In particular embodiments, supervised learning algorithm 720 may include any algorithm that can be utilized to apply what has been learned in the past to new data, e.g., using labeled examples to predict future events. For example, starting from an analysis of a known training data set, supervised learning algorithm 720 may create an inferred function to make a prediction about output values. Supervised learning algorithm 600 may also compare its output with the correct intended output and find errors in order to correct supervised learning algorithm 720 accordingly. On the other hand, unsupervised learning algorithm 722 may include any algorithm that can be applied, e.g., when the data used to train unsupervised learning algorithm 722 is neither classified nor labeled. For example, unsupervised learning algorithm 722 may study and analyze how a system can infer functions to describe hidden structure from unlabeled data.

[0109] In particular embodiments, NLP algorithms and functions 706 may include any algorithms or functions that may be suitable for automatically manipulating natural language, such as speech and / or text. For example, in some embodiments, NLP algorithms and functions 706 may include content extraction algorithms or functions 724, classification algorithms or functions 726, machine translation algorithms or functions 728, question answering (QA) algorithms or functions 730, and text generation algorithms or functions 732. In particular embodiments, content extraction algorithms or functions 724 may include means for extracting text or images from electronic documents (e.g., web pages, text editor documents, etc.) for use in other applications, for example.

[0110] In particular embodiments, classification algorithm or function 726 may include any algorithm that may utilize a supervised learning model (e.g., logistic regression, naive Bayes, stochastic gradient descent (SGD), k-nearest neighbors, decision trees, random forests, support vector machines (SVMs), etc.) to learn from and make new observations or classifications based on data input into the supervised learning model. Machine translation algorithm or function 728 may include any algorithm or function that may be suitable for automatically converting source text in one language into text in another language, for example. QA algorithm or function 730 may include any algorithm or function that may be suitable for automatically answering questions posed by humans in natural language, such as those performed by a voice-controlled personal assistant device, for example. Text generation algorithm or function 732 may include any algorithm or function that may be suitable for automatically generating natural language text.

[0111] In particular embodiments, expert system 708 may include any algorithms or functions that may be suitable for simulating the judgment and actions of a human or organization with expertise and experience in a particular field (e.g., stock trading, medicine, sports statistics, etc.). Computer-based vision algorithms and functions 710 may include any algorithms or functions that may be suitable for automatically extracting information from images (e.g., photographic images, video images). For example, computer-based vision algorithms and functions 710 may include image recognition algorithms 734 and machine vision algorithms 736. Image recognition algorithms 734 may include any algorithms that may be suitable, for example, for automatically identifying and / or classifying objects, places, people, etc. that may be contained in one or more image frames or other display data. Machine vision algorithms 736 may include any algorithms that may be suitable for enabling a computer to "see" or that may be suitable, for example, for relying on image sensor cameras with specialized optics to acquire images in order to process, analyze, and / or measure various data characteristics for decision-making purposes.

[0112] In particular embodiments, speech recognition algorithms and functions 712 may include any algorithms or functions that may be suitable for recognizing and translating spoken language into text, such as through automatic speech recognition (ASR), computer speech recognition, speech-to-text (STT) 738, or text-to-speech (TTS) 740, for computing purposes to communicate with one or more users via voice. In particular embodiments, planning algorithms and functions 714 may include any algorithms or functions that may be suitable for generating a series of actions, each of which may include its own set of preconditions to be satisfied before performing the action. Examples of AI planning may include classical planning, reduction to other problems, temporal planning, probabilistic planning, preference-based planning, conditional planning, etc. Finally, robotics algorithms and functions 716 may include any algorithms, functions, or systems that may enable one or more devices to replicate human behavior, for example, through movements, gestures, performance tasks, decision-making, emotions, etc.

[0113] As used herein, "or" is inclusive and not exclusive, unless clearly indicated otherwise or indicated otherwise by context. Thus, as used herein, "A or B" means "A, B, or both," unless clearly indicated otherwise or indicated otherwise by context. Furthermore, "and" is both conjunctive and several, unless expressly indicated otherwise or indicated otherwise by context. Thus, as used herein, "A and B" means "A and B, jointly or severally," unless clearly indicated otherwise or indicated otherwise by context.

[0114] As used herein, "automatically" and its derivatives mean "without human intervention" unless expressly indicated otherwise or indicated otherwise by context.

[0115] The embodiments disclosed herein are merely examples, and the scope of the present disclosure is not limited thereto. Embodiments according to the present disclosure are disclosed in the appended claims, particularly for methods, storage media, systems, and computer program products. Any feature recited in one claim category, e.g., a method, may also be claimed in another claim category, e.g., a system. Dependencies or references in the appended claims are chosen for formality reasons only. However, just as any combination of a claim and its features may be disclosed and claimed without regard to the dependencies recited in the appended claims, any subject matter resulting from an intentional reference to any preceding claim (e.g., multiple dependencies) may likewise be claimed. Subject matter that may be claimed includes not only combinations of features as recited in the appended claims, but also any other combinations of features within the scope of the claims, and each feature recited in a claim may be combined with any other feature or combination of features within the scope of the claim. Furthermore, any of the embodiments and features described or illustrated in this specification may be claimed in a separate claim and / or in any combination with any of the embodiments or features described or illustrated in this specification or with any of the features of the accompanying claims.

[0116] The scope of the present disclosure encompasses all changes, substitutions, variations, alterations, and modifications to the exemplary embodiments described or illustrated herein that would be understood by a person skilled in the art. The scope of the present disclosure is not limited to the exemplary embodiments described or illustrated herein. Furthermore, although the present disclosure describes and illustrates each embodiment herein as including particular components, elements, features, functions, operations, or steps, any of these embodiments may include any combination or permutation of any of the components, elements, features, functions, operations, or steps described or illustrated anywhere herein that would be understood by a person skilled in the art. Furthermore, references in the appended claims to an apparatus or system or system component adapted, arranged, enabled, configured, enabled, operative, or operating to perform a particular function encompass that apparatus, system, or component, so long as the apparatus, system, or component is so adapted, arranged, enabled, configured, enabled, operative, or operating, regardless of whether it or that particular function is activated, turned on, or unlocked. Furthermore, although this disclosure describes or illustrates particular embodiments as providing certain advantages, the particular embodiments may provide none, some, or all of these advantages.

[0117] Embodiment Among the embodiments provided are the following: 1. A method for detecting the longitudinal progression of Alzheimer's disease (AD) in a patient, comprising: receiving audio data including a patient's description of one or more previous or current experiences of the patient, the audio data being captured at multiple moments during a period of time; analyzing the speech data to quantify a plurality of speech variables, the plurality of speech variables including a word length variable and a particle use variable; determining a composite score based on standardizing the plurality of quantified speech variables and a substantive weighting assigned to each of the plurality of quantified speech variables; To detect predicted long-term changes in multiple quantified speech variables based on the composite score; and To estimate a patient's progression of AD based on predicted long-term changes A method comprising: 2. The method of embodiment 1, wherein receiving the voice data includes receiving an audio file containing an electronic recording of the patient's voice. 3. The method of any one of embodiments 1-2, wherein the electronic recording of the patient's voice comprises an electronic recording of one or more verbal responses of the patient to a Clinical Dementia Rating (CDR) interview. 4. The method of any one of embodiments 1 to 3, wherein the audio data is captured on one or more dates selected from the group including the first day and approximately 0.25, 0.5, 0.75, 1, 3, 6, 9, 12, 15, 18, 21, 24, 27, 30, 33, and 36 months after the first day. 5. The method of any one of embodiments 1 to 4, wherein the plurality of phonetic variables further comprises a word frequency variable, a syntactic depth variable, a noun use variable, or a pronoun use variable. 6. A method according to any one of embodiments 1 to 5, wherein the plurality of speech variables further comprises one or more Mel-Frequency Cepstral Coefficient (MFCC) features. 7. The method of any one of embodiments 1-6, wherein the one or more MFCC features include the mean of the 11th MFCC coefficient (MFCC mean 11), the variance of the first derivative of the 11th MFCC coefficient (MFCC var 25), or the variance of the first derivative of the 12th MFCC coefficient (MFCC var 26). 8. The method of any one of embodiments 1 to 7, wherein determining the composite score comprises: Standardizing multiple quantified speech variables; applying equal weighting to each of the quantified speech variables; and combining standardized, equally weighted, quantified speech variables to generate a composite score. 9. The method of any one of embodiments 1-8, wherein estimating the progression of AD based on predicted long-term changes comprises correlating the composite score with one or more clinical assessment metrics. 10. The method of any one of embodiments 1-9, wherein the one or more clinical assessment metrics are selected from the group consisting of Mini-Mental State Examination (MMSE) score, Clinical Dementia Rating (CDR) interview, Clinical Dementia Rating-Sum of Boxes (CDR-SB) scale, Alzheimer's Disease Assessment Scale-Cognitive (ADAS-Cog) subscale test battery, Alzheimer's Disease Cooperative Study Group-Activities of Daily Living Inventory (ADCS-ADL) scale, Neuropsychiatric Symptom Assessment (NPI) scale, Neuropsychiatric Symptom Assessment-Questionnaire (NPI-Q), Alzheimer's Disease Caregiver Global Impression (CaGI) scale, Instrumental Activities of Daily Living (IADL) scale, Amsterdam Activities of Daily Living Questionnaire (A-IADL-Q), and Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) scale. 11. The method of any one of embodiments 1-10, further comprising determining whether the patient is responsive to treatment based on the estimated progression of AD. 12. The method of any one of embodiments 1 to 11, wherein analyzing the audio data to determine a plurality of quantified audio variables comprises analyzing the audio data using one or more natural language processing (NLP) machine learning models. 13. The method of any one of embodiments 1-12, further comprising sending a notification of the estimated AD progression to a computing device associated with a clinician. 14. The method of any one of embodiments 1-13, further comprising sending a notification of the estimated AD progression to an electronic device associated with the patient. 15. The method of any one of embodiments 1 to 14, further comprising generating a recommendation for adjusting the patient's treatment regimen in response to the estimation of AD progression. 16. The treatment regimen comprises at least one compound selected from the group consisting of compounds against oxidative stress, anti-apoptotic compounds, metal chelators, inhibitors of DNA repair, 3-amino-1-propanesulfonic acid (3APS), 1,3-propanedisulfonate (1,3PDS), secretase activators, beta- and gamma-secretase inhibitors, tau protein, anti-tau antibodies, anti-tau agents, gene therapy drugs, neurotransmitters, beta-sheet breakers, anti-inflammatory molecules, atypical antipsychotics, cholinesterase inhibitors, other drugs, and dietary supplements. 16. The method of any one of embodiments 1-15, comprising a therapeutic agent selected from the group consisting of a substance, an antiviral agent, an anti-tau antibody, a tau inhibitor, an anti-amyloid beta (anti-Aβ) antibody, a beta-amyloid aggregation inhibitor, a target binding therapeutic agent, an anti-BACE1 antibody, a BACE1 inhibitor, a cholinesterase inhibitor, an NMDA receptor antagonist, a monoamine depleting agent, an ergoloid mesylate, an anticholinergic antiparkinsonian agent, a dopaminergic antiparkinsonian agent, tetrabenazine, an anti-inflammatory agent, a hormone, a vitamin, a dimebolin, a homotaurine, a serotonin receptor activity modulator, an interferon, and a glucocorticoid. 17. The method of any one of embodiments 1-16, wherein the symptomatic medication is selected from the group consisting of a cholinesterase inhibitor, galantamine, rivastigmine, donepezil, an N-methyl-D-aspartate receptor antagonist, memantine, and a dietary supplement (optionally, the dietary supplement is Souvenaid®). 18. The method of any one of embodiments 1-17, wherein the anti-Aβ antibody is selected from the group consisting of bapineuzumab, solanezumab, aducanumab, gantenerumab, crenezumab, donanemab, and lecanemab. 19. The method of any one of embodiments 1-18, wherein the anti-tau antibody is selected from the group consisting of an N-terminal binding agent, a mid-domain binding agent, and a fibrillar tau binding agent. 20. The method of any one of embodiments 1-19, wherein the anti-tau antibody is selected from the group consisting of semolinemab, BMS-986168, C2N-8E12, goslanemab, tiravonemab, and zagotenemab. 21. The method of any one of embodiments 1-20, wherein the therapeutic agent is a therapeutic agent that specifically binds to a target, and the target is selected from the group consisting of beta-secretase, tau, presenilin, amyloid precursor protein or a portion thereof, amyloid beta peptide or an oligomer or fibril thereof, death receptor 6 (DR6), receptor for advanced glycation end products (RAGE), parkin, and huntingtin. 22. The method of any one of embodiments 1-21, wherein the therapeutic agent is a monoamine-depleting drug, optionally tetrabenazine. 23. The method of any one of embodiments 1-22, wherein the therapeutic agent is an anticholinergic antiparkinsonian agent selected from the group consisting of procyclidine, diphenhydramine, trihexylphenidyl, benztropine, biperiden, and trihexyphenidyl. 24. The method of any one of embodiments 1-23, wherein the therapeutic agent is a dopaminergic antiparkinsonian selected from the group consisting of entacapone, selegiline, pramipexole, bromocriptine, rotigotine, selegiline, ropinirole, rasagiline, apomorphine, carbidopa, levodopa, pergolide, tolcapone, and amantadine. 25. The method of any one of embodiments 1-24, wherein the therapeutic agent is an anti-inflammatory agent selected from the group consisting of nonsteroidal anti-inflammatory drugs and indomethacin. 26. The method of any one of embodiments 1-25, wherein the therapeutic agent is a hormone selected from the group consisting of estrogen, progesterone, and leuprolide. 27. The method of any one of embodiments 1-26, wherein the therapeutic agent is a vitamin selected from the group consisting of folate and nicotinamide. 28. The method of any one of embodiments 1-27, wherein the therapeutic agent is 3-aminopropanesulfonic acid or 3APS, xaliproden, or homotaurine.

Claims

1. A method for detecting the long-term progression of Alzheimer's disease (AD) in patients, By one or more computing devices, Receiving audio data containing the patient's description of one or more past or current experiences, wherein the audio data is captured at multiple points in time during a period. Analyzing the audio data to quantify multiple audio variables, wherein the multiple audio variables include a word length variable and a particle usage variable. A composite score is determined based on the standardization of the quantified plurality of speech variables and the substantial weights assigned to each of the quantified plurality of speech variables. Based on the composite score, the predicted long-term changes in the quantified plurality of speech variables are detected, and Based on the predicted long-term changes, estimate the progression of AD in the patient. Methods that include...

2. The method according to claim 1, wherein receiving the voice data includes receiving an audio file containing an electronic recording of the patient's voice.

3. The method according to claim 2, wherein the electronic recording of the patient's voice includes an electronic recording of one or more oral responses of the patient to a clinical dementia assessment (CDR) interview.

4. The method according to claim 1, wherein the audio data is captured on the first day and on one or more days selected from the group including approximately 0.25, 0.5, 0.75, 1, 3, 6, 9, 12, 15, 18, 21, 24, 27, 30, 33, and 36 months after the first day.

5. The method according to claim 1, wherein the plurality of phonetic variables further include word frequency variables, syntactic depth variables, noun usage variables, or pronoun usage variables.

6. The method according to claim 1, wherein the plurality of speech variables further include one or more Mel-frequency cepstrum coefficient (MFCC) features.

7. The method according to claim 6, wherein one or more MFCC features include the mean of the 11th MFCC coefficient (MFCC mean 11), the variance of the first derivative of the 11th MFCC coefficient (MFCC var 25), or the variance of the first derivative of the 12th MFCC coefficient (MFCC var 26).

8. Determining the aforementioned composite score Standardizing the quantified plurality of speech variables; Applying equal weighting to each of the quantified speech variables; and The composite score is generated by combining the multiple standardized, equally weighted, and quantified speech variables. The method according to claim 1, including the method described in claim 1.

9. The method according to claim 1, wherein estimating the progression of AD based on the predicted long-term changes includes correlating the composite score with one or more clinical evaluation metrics.

10. The method according to claim 9, wherein the one or more clinical assessment metrics are selected from the group consisting of Mini-Mental State Examination (MMSE) scores, Clinical Dementia Assessment (CDR) interviews, Clinical Dementia Assessment-Box Sum (CDR-SB) scales, Alzheimer's Disease Assessment Scale-Cognitive (ADAS-Cog) subscale test battery, Alzheimer's Disease Collaborative Study-Activities of Daily Living Assessment (ADCS-ADL) scales, Neuropsychiatric Symptom Assessment (NPI) scales, Neuropsychiatric Symptom Assessment-Questionnaire (NPI-Q), Caregiver Global Impression (CaGI) scale, Instrumental Activities of Daily Living (IADL) scales, Amsterdam Activities of Daily Living Questionnaire (A-IADL-Q), and Repeatable Battery for Assessment of Neuropsychological State (RBANS) scales.

11. The method according to claim 1, further comprising determining whether the patient is responsive to treatment based on the estimated progression of AD.

12. The method according to claim 1, wherein analyzing the speech data to determine the plurality of quantified speech variables includes analyzing the speech data using one or more natural language processing (NLP) machine learning models.

13. The method according to claim 1, further comprising generating recommendations for adjusting the patient's treatment regimen in response to an estimate of the progression of AD.

14. The treatment regimen comprises at least one compound selected from the group consisting of compounds against oxidative stress, anti-apoptotic compounds, metal chelators, DNA repair inhibitors, 3-amino-1-propanesulfonic acid (3APS), 1,3-propanedisulfonate (1,3PDS), secretase activators, beta- and gamma-secretase inhibitors, tau proteins, anti-tau antibodies, anti-tau agents, gene therapy drugs, neurotransmitters, beta-sheet disruptors, anti-inflammatory molecules, atypical antipsychotics, cholinesterase inhibitors, other drugs, and nutritional supplements, and is a therapeutic agent, symptomatic treatment agent, neuropharmaceutical, corticosteroid. The method according to claim 13, comprising a therapeutic agent selected from the group consisting of antibiotics, antiviral agents, anti-tau antibodies, tau inhibitors, anti-amyloid beta (anti-Aβ) antibodies, beta-amyloid aggregation inhibitors, target-binding therapeutic agents, anti-BACE1 antibodies, BACE1 inhibitors, cholinesterase inhibitors, NMDA receptor antagonists, monoamine depletion agents, ergoloid mesylate, anticholinergic antiparkinsonist agents, dopaminergic antiparkinsonist agents, tetrabenazine, anti-inflammatory agents, hormones, vitamins, dimevorin, homotaurine, serotonin receptor activity modulators, interferon, and glucocorticoids.

15. The method according to claim 14, wherein the symptomatic treatment agent is selected from the group consisting of cholinesterase inhibitors, galantamine, rivastigmine, donepezil, N-methyl-D-aspartate receptor antagonists, memantine, and nutritional supplements (optionally, the nutritional supplement is Souvenaid®).

16. The method according to claim 14, wherein the anti-Aβ antibody is selected from the group consisting of bapineuzumab, solanezumab, aducanumab, gantenerumab, crenezumab, donanemab, and lecanemab.

17. The method according to claim 14, wherein the anti-tau antibody is selected from the group consisting of an N-terminal binder, an intermediate domain binder, and a fibrillary tau binder.

18. The method according to claim 17, wherein the anti-tau antibody is selected from the group consisting of semolinemab, BMS-986168, C2N-8E12, goslanemab, tilabonemab, and zagotenemab.

19. A system for detecting the long-term progression of Alzheimer's disease (AD) in a patient, the system comprising one or more computing devices, the computing devices are One or more non-temporary computer-readable storage media containing instructions, The system comprises one or more processors connected to one or more storage media, wherein the one or more processors execute the instructions, Receiving audio data which includes the patient's description of one or more past or current experiences, and which is captured at multiple points in time during a certain period, The audio data is analyzed to quantify multiple phonetic variables, including word length variables and particle usage variables. A composite score is determined based on the standardization of the quantified plurality of speech variables and the substantial weights assigned to each of the quantified plurality of speech variables. Based on the composite score, predictable long-term changes in the quantified plurality of speech variables are detected, and Based on the predicted long-term changes, the progression of AD in the patient is estimated. A system that is configured in such a way.

20. A non-temporary computer-readable medium containing instructions, wherein when the instructions are executed by one or more processors of one or more computing devices, the one or more processors Receiving audio data which includes the patient's description of one or more past or current experiences, and which is captured at multiple points in time during a certain period, Analyzing the audio data in order to quantify multiple phonetic variables, including word length variables and particle usage variables, The composite score is determined based on the standardization of the quantified plurality of speech variables and the substantial weights assigned to each of the quantified plurality of speech variables. Based on the composite score, predict long-term changes in the quantified plurality of speech variables are detected. A non-temporary computer-readable medium for estimating the progression of AD in the patient based on the predicted long-term changes.

Images