An Automated Approach for the Detection of Alzheimer's Disease From Resting State Electroencephalography.

Early detection is crucial to control the progression of Alzheimer's disease and to postpone intellectual decline. Most current detection techniques are costly, inaccessible, or invasive. Furthermore, they require laborious analysis, what delays the start of medical treatment. To overcome this, researchers have recently investigated AD detection based on electroencephalography, a non-invasive neurophysiology technique, and machine learning algorithms. However, these approaches typically rely on manual procedures such as visual inspection, that requires additional personnel for the analysis, or on cumbersome EEG acquisition systems. In this paper, we performed a preliminary evaluation of a fully-automated approach for AD detection based on a commercial EEG acquisition system and an automated classification pipeline. For this purpose, we recorded the resting state brain activity of 26 participants from three groups: mild AD, mild cognitive impairment (MCI-non-AD), and healthy controls. First, we applied automated data-driven algorithms to reject EEG artifacts. Then, we obtained spectral, complexity, and entropy features from the preprocessed EEG segments. Finally, we assessed two binary classification problems: mild AD vs. controls, and MCI-non-AD vs. controls, through leave-one-subject-out cross-validation. The preliminary results that we obtained are comparable to the best reported in literature, what suggests that AD detection could be automatically detected through automated processing and commercial EEG systems. This is promising, since it may potentially contribute to reducing costs related to AD screening, and to shortening detection times, what may help to advance medical treatment.

A self-driven approach for multi-class discrimination in Alzheimer's disease based on wearable EEG.

Early detection is critical to control Alzheimer's disease (AD) progression and postpone cognitive decline. Traditional medical procedures such as magnetic resonance imaging are costly, involve long waiting lists, and require complex analysis. Alternatively, for the past years, researchers have successfully evaluated AD detection approaches based on machine learning and electroencephalography (EEG). Nonetheless, these approaches frequently rely upon manual processing or involve non-portable EEG hardware. These aspects are suboptimal regarding automated diagnosis, since they require additional personnel and hinder portability. In this work, we report the preliminary evaluation of a self-driven AD multi-class discrimination approach based on a commercial EEG acquisition system using sixteen channels. For this purpose, we recorded the EEG of three groups of participants: mild AD, mild cognitive impairment (MCI) non-AD, and controls, and we implemented a self-driven analysis pipeline to discriminate the three groups. First, we applied automated artifact rejection algorithms to the EEG recordings. Then, we extracted power, entropy, and complexity features from the preprocessed epochs. Finally, we evaluated a multi-class classification problem using a multi-layer perceptron through leave-one-subject-out cross-validation. The preliminary results that we obtained are comparable to the best in literature (0.88 F1-score), what suggests that AD can potentially be detected through a self-driven approach based on commercial EEG and machine learning. We believe this work and further research could contribute to opening the door for the detection of AD in a single consultation session, therefore reducing the costs associated to AD screening and potentially advancing medical treatment.

"Don't know" sign: description and evaluation of its diagnostic accuracy for cognitive impairment.

OBJECTIVES: Patients in neurology clinics are sometimes not aware of the reason for the consultation, and we have called this circumstance the "Don't know" sign (DKS). Our objective was to define this new sign and its modalities and to evaluate its prevalence and its diagnostic accuracy for cognitive impairment (CI) in comparison to other observation-based signs. DESIGN, SETTING, AND PARTICIPANTS: A cross-sectional prospective study included all new outpatients evaluated by the authors at neurology consultation. MEASUREMENTS: We recorded observation-based signs. The Global Deterioration Scale (GDS) was used to assess the cognitive status of patients, based on clinical history, caregiver interview, and cognitive test results. We analyzed the prevalence and the diagnostic accuracy for CI of DKS, "head turning sign," "attending with," verbal repetition, and combinations, calculating sensitivity (Se), specificity (Sp), positive predictive value (PPV), and negative predictive value (NPV). RESULTS: We enrolled 673 consecutive patients (62% female) with a mean ± SD age of 59.3 ± 20.2 years. DKS was positive in 94 patients (14%) and was strongly associated with GDS score. DKS had a Se of 0.41, Sp of 0.98, PPV of 0.89, and NPV of 0.79 for CI diagnosis. The presence of at least two positive observation signs yielded a Se of 0.50, Sp of 0.97, PPV of 0.86, and NPV of 0.81. CONCLUSIONS: DKS is frequently observed in neurology outpatients. It has low sensitivity but high specificity and PPV for CI diagnosis. It does not require additional consultation time, and its use can be recommended in combination with other observation-based signs.

PET-Amyloid After Inconclusive Cerebrospinal Fluid Biomarkers in Clinical Practice. Is it Necessary to Duplicate Procedures?

INTRODUCTION: In the absence of a gold standard for in vivo Alzheimer disease (AD) diagnosis, AD biomarkers such as cerebrospinal fluid biomarkers (CSF-B) and PET-Amyloid are considered diagnostically useful in clinical practice guidelines and have consensual appropriate use criteria (AUC). However, little evidence has been published on their utilization in the clinical setting or on approaches to mismatched results. The objective of this work was to evaluate the use of AD biomarkers in clinical practice, focusing on the implementation of PET-Amyloid in cases of inconclusive CSF-B. METHODS: This naturalistic, ambispective case series included patients fulfilling AUC for CSF-B and PET-Amyloid whose CSF-B results were non-diagnostic (target population), analyzing the diagnostic certainty, the treatment approach, and the relationship between CSF-B and PET-Amyloid results. RESULTS: Out of 2373 eligible patients, AD biomarkers were studied in 417 (17.6%), most frequently due to cognitive impairment in under 65-year-olds, using CSF-B in 311 patients and PET-Amyloid in 150. CSF-B results were non-diagnostic for 44 patients (52.3% male; aged 60.9±6.6 years), who then underwent PET-Amyloid study, which was positive in 31. A 'k' coefficient of 0.108 was obtained between CSF-B and PET-amyloid (54.5% concordance). In multivariate regression analysis, Aß42 was the only significant predictor (p= 0.018) of a positive PET-Amyloid result. In the target population, PETAmyloid increased diagnostic confidence by 53.7% (p <0.001) and modified the therapeutic approach in 36.4% of cases. CONCLUSION: These findings support the duplication of AD biomarkers and demonstrate that the implementation of PET-Amyloid provides an early and certain diagnosis to guide appropriate treatment.

Discriminative validity of an abbreviated Semantic Verbal Fluency Test.

Semantic verbal fluency (SVF) is one of the most widely used tests for cognitive assessment due to its diagnostic utility (DU). OBJECTIVE: our objective is to evaluate the DU to detect cognitive impairment (CI) of a short version of the SVF applied in 30 seconds (SVF1-30). METHODS: a prospective sample of consecutive patients evaluated in a Neurology Unit between December 2016 and December 2017 were assessed with the Global Deterioration Scale (GDS), 30-second and 60-second SVF tests (animals), and the Fototest, which includes a fluency task of people's names. The DU for CI was evaluated by the area under the ROC curve and effect size ("d" Cohen). RESULTS: the study included 1012 patients (256 with CI, 395 with dementia). SVF1-30 shows a good correlation with GDS stage. The DU of SVF1-30 is identical to that of the classical version, applied in 60 seconds, (SVFtotal) for CI (0.89 ± 0.01; p > 0.50), and shows no significant difference for dementia (0.85 ± 0.01 vs. 0.86 ± 0.01, p > 0.15). DISCUSSION: the DU of SVF1-30 is similar to that of the SVFtotal, allowing a reduction in examination time with no loss of discriminative capacity.

A fluência verbal semântica (SVF) é um dos testes mais utilizados na avaliação cognitiva devido à sua utilidade diagnóstica (UD). OBJETIVO: Nosso objetivo foi o de avaliar o DU de uma versão abreviada do SVF aplicado em 30 segundos (SVF1-30) para a detecção do comprometimento cognitivo (CC). MÉTODOS: Amostra prospectiva de pacientes avaliados em uma Unidade de Neurologia entre dezembro de 2016 e dezembro de 2017. Global Deterioration Scale (GDS), um teste de SVF (animais), registrando os resultados em 30 e 60 segundos e Fototest, que inclui uma tarefa de fluência de nomes de pessoas foram aplicadas. A UD para CC foi avaliada pela área sob a curva ROC e o tamanho do efeito ("d" Cohen). RESULTADOS: foram incluídos 1012 sujeitos (256 CC e 395 demência). O SVF1-30 mostrou uma boa correlação com o estágio GDS. A UD de SVF1-30 é idêntico ao da versão clássica (SVFtotal) para CC (0,89 ± 0,01; p > 0,50) e sem diferença significativa para demência (0,85 ± 0,01 vs. 0,86 ± 0,01; p > 0,15). DISCUSSÃO: a UD do SVF1-30 é similar ao SVFtotal, o que permite diminuir o tempo de exploração sem perder a capacidade discriminativa.

Discriminative validity of an abbreviated Semantic Verbal Fluency Test / Validade discriminativa de um teste abreviado de Fluência Verbal Semântica

ABSTRACT. Semantic verbal fluency (SVF) is one of the most widely used tests for cognitive assessment due to its diagnostic utility (DU). Objective: our objective is to evaluate the DU to detect cognitive impairment (CI) of a short version of the SVF applied in 30 seconds (SVF1-30). Methods: a prospective sample of consecutive patients evaluated in a Neurology Unit between December 2016 and December 2017 were assessed with the Global Deterioration Scale (GDS), 30-second and 60-second SVF tests (animals), and the Fototest, which includes a fluency task of people's names. The DU for CI was evaluated by the area under the ROC curve and effect size ("d" Cohen). Results: the study included 1012 patients (256 with CI, 395 with dementia). SVF1-30 shows a good correlation with GDS stage. The DU of SVF1-30 is identical to that of the classical version, applied in 60 seconds, (SVFtotal) for CI (0.89 ± 0.01; p > 0.50), and shows no significant difference for dementia (0.85 ± 0.01 vs. 0.86 ± 0.01, p > 0.15). Discussion: the DU of SVF1-30 is similar to that of the SVFtotal, allowing a reduction in examination time with no loss of discriminative capacity.

RESUMO. A fluência verbal semântica (SVF) é um dos testes mais utilizados na avaliação cognitiva devido à sua utilidade diagnóstica (UD). Objetivo: Nosso objetivo foi o de avaliar o DU de uma versão abreviada do SVF aplicado em 30 segundos (SVF1-30) para a detecção do comprometimento cognitivo (CC). Métodos: Amostra prospectiva de pacientes avaliados em uma Unidade de Neurologia entre dezembro de 2016 e dezembro de 2017. Global Deterioration Scale (GDS), um teste de SVF (animais), registrando os resultados em 30 e 60 segundos e Fototest, que inclui uma tarefa de fluência de nomes de pessoas foram aplicadas. A UD para CC foi avaliada pela área sob a curva ROC e o tamanho do efeito ("d" Cohen). Resultados: foram incluídos 1012 sujeitos (256 CC e 395 demência). O SVF1-30 mostrou uma boa correlação com o estágio GDS. A UD de SVF1-30 é idêntico ao da versão clássica (SVFtotal) para CC (0,89 ± 0,01; p > 0,50) e sem diferença significativa para demência (0,85 ± 0,01 vs. 0,86 ± 0,01; p > 0,15). Discussão: a UD do SVF1-30 é similar ao SVFtotal, o que permite diminuir o tempo de exploração sem perder a capacidade discriminativa.

Diagnostic accuracy, effectiveness and cost for cognitive impairment and dementia screening of three short cognitive tests applicable to illiterates.

BACKGROUND: Illiteracy, a universal problem, limits the utilization of the most widely used short cognitive tests. Our objective was to assess and compare the effectiveness and cost for cognitive impairment (CI) and dementia (DEM) screening of three short cognitive tests applicable to illiterates. METHODS: Phase III diagnostic test evaluation study was performed during one year in four Primary Care centers, prospectively including individuals with suspicion of CI or DEM. All underwent the Eurotest, Memory Alteration Test (M@T), and Phototest, applied in a balanced manner. Clinical, functional, and cognitive studies were independently performed in a blinded fashion in a Cognitive Behavioral Neurology Unit, and the gold standard diagnosis was established by consensus of expert neurologists on the basis of these results. Effectiveness of tests was assessed as the proportion of correct diagnoses (diagnostic accuracy [DA]) and the kappa index of concordance (k) with respect to gold standard diagnoses. Costs were based on public prices at the time and hospital accounts. RESULTS: The study included 139 individuals: 47 with DEM, 36 with CI, and 56 without CI. No significant differences in effectiveness were found among the tests. For DEM screening: Eurotest (k = 0.71 [0.59-0.83], DA = 0.87 [0.80-0.92]), M@T (k = 0.72 [0.60-0.84], DA = 0.87 [0.80-0.92]), Phototest (k = 0.70 [0.57-0.82], DA = 0.86 [0.79-0.91]). For CI screening: Eurotest (k = 0.67 [0.55-0.79]; DA = 0.83 [0.76-0.89]), M@T (k = 0.52 [0.37-0.67]; DA = 0.80 [0.72-0.86]), Phototest (k = 0.59 [0.46-0.72]; DA = 0.79 [0.71-0.86]). There were no differences in the cost of DEM screening, but the cost of CI screening was significantly higher with M@T (330.7 ± 177.1 €, mean ± sd) than with Eurotest (294.1 ± 195.0 €) or Phototest (296.0 ± 196. 5 €). Application time was shorter with Phototest (2.8 ± 0.8 min) than with Eurotest (7.1 ± 1.8 min) or M@T (6.8 ± 2.2 min). CONCLUSIONS: Eurotest, M@T, and Phototest are equally effective. Eurotest and Phototest are both less expensive options but Phototest is the most efficient, requiring the shortest application time.

Effectiveness and costs of phototest in dementia and cognitive impairment screening.

BACKGROUND: To assess and compare the effectiveness and costs of Phototest, Mini Mental State Examination (MMSE), and Memory Impairment Screen (MIS) to screen for dementia (DEM) and cognitive impairment (CI). METHODS: A phase III study was conducted over one year in consecutive patients with suspicion of CI or DEM at four Primary Care (PC) centers. After undergoing all screening tests at the PC center, participants were extensively evaluated by researchers blinded to screening test results in a Cognitive-Behavioral Neurology Unit (CBNU). The gold standard diagnosis was established by consensus of expert neurologists. Effectiveness was assessed by the proportion of correct diagnoses (diagnostic accuracy [DA]) and by the kappa index of concordance between test results and gold standard diagnoses. Costs were based on public prices and hospital accounts. RESULTS: The study included 140 subjects (48 with DEM, 37 with CI without DEM, and 55 without CI). The MIS could not be applied to 23 illiterate subjects (16.4%). For DEM, the maximum effectiveness of the MMSE was obtained with different cutoff points as a function of educational level [k = 0.31 (95% Confidence interval [95%CI], 0.19-0.43), DA = 0.60 (95%CI, 0.52-0.68)], and that of the MIS with a cutoff of 3/4 [k = 0.63 (95%CI, 0.48-0.78), DA = 0.83 (95%CI, 0.80-0.92)]. Effectiveness of the Phototest [k = 0.71 (95%CI, 0.59-0.83), DA = 0.87 (95%CI, 0.80-0.92)] was similar to that of the MIS and higher than that of the MMSE. Costs were higher with MMSE (275.9 ± 193.3€ [mean ± sd euros]) than with Phototest (208.2 ± 196.8€) or MIS (201.3 ± 193.4€), whose costs did not significantly differ. For CI, the effectiveness did not significantly differ between MIS [k = 0.59 (95%CI, 0.45-0.74), DA = 0.79 (95%CI, 0.64-0.97)] and Phototest [k = 0.58 (95%CI, 0.45-0.74), DA = 0.78 (95%CI, 0.64-0.95)] and was lowest for the MMSE [k = 0.27 (95%CI, 0.09-0.45), DA = 0.69 (95%CI, 0.56-0.84)]. Costs were higher for MMSE (393.4 ± 121.8€) than for Phototest (287.0 ± 197.4€) or MIS (300.1 ± 165.6€), whose costs did not significantly differ. CONCLUSION: MMSE is not an effective instrument in our setting. For both DEM and CI, the Phototest and MIS are more effective and less costly, with no difference between them. However, MIS could not be applied to the appreciable percentage of our population who were illiterate.