Dementia and cognitive impairment associated with aging are a major medical and social concern. Neuropsychological testing is a key element in the diagnostic procedures of Mild Cognitive Impairment (MCI), but has presently a limited value in the prediction of progression to dementia. We advance the hypothesis that newer statistical classification methods derived from data mining and machine learning methods like Neural Networks, Support Vector Machines and Random Forests can improve accuracy, sensitivity and specificity of predictions obtained from neuropsychological testing. Seven non parametric classifiers derived from data mining methods (Multilayer Perceptrons Neural Networks, Radial Basis Function Neural Networks, Support Vector Machines, CART, CHAID and QUEST Classification Trees and Random Forests) were compared to three traditional classifiers (Linear Discriminant Analysis, Quadratic Discriminant Analysis and Logistic Regression) in terms of overall classification accuracy, specificity, sensitivity, Area under the ROC curve and Press'Q. Model predictors were 10 neuropsychological tests currently used in the diagnosis of dementia. Statistical distributions of classification parameters obtained from a 5-fold cross-validation were compared using the Friedman's nonparametric test.

Recent studies show that most of the oxidative changes found in Alzheimer's disease (AD) are already present in mild cognitive impairment (MCI) patients. The question arises as to whether oxidative stress has a role in the progression of MCI to AD. W

Background: Mild cognitive impairment (MCI) is considered to be an early stage of a neurodegenerative disorder, particularly Alzheimer’s disease, and the clinical diagnosis requires the objective demonstration of cognitive deficits. The aim of the present study was to evaluate the predictive value of MCI for the conversion to dementia when using four different verbal memory tests (Logical Memory, LM; California Verbal Learning Test, CVLT; Verbal Paired-Associate Learning, VPAL; and Digit Span, DS) in the MCI criteria. Methods: Participants were consecutive patients with subjective cognitive complaints who performed a comprehensive neuropsychological evaluation and were not demented, observed in a memory clinic setting. Results: At baseline, 272 non-demented patients reporting subjective cognitive complaints were included. During the follow-up time (3.0 ± 1.9 years), 58 patients converted to dementia and 214 did not. Statistically significant differences between the converters and non-converters were present in LM, VPAL, and CVLT. A multivariate Cox regression analysis combining the four memory tests revealed that only the CVLT test remained significant as a predictor of conversion to dementia. Non-demented patients with cognitive complaints diagnosed as having MCI according to abnormal (<1.5 SD) learning in the CVLT test had a 3.61 higher risk of becoming demented during the follow-up. Conclusion: The verbal memory assessment using the CVLT should be preferred in the diagnostic criteria of MCI for a more accurate prediction of conversion to dementia.

Predicting progression from a stage of Mild Cognitive Impairment to dementia is a major pursuit in current research. It is broadly accepted that cognition declines with a continuum between MCI and dementia. As such, cohorts of MCI patients are usually heterogeneous, containing patients at different stages of the neurodegenerative process. This hampers the prognostic task. Nevertheless, when learning prognostic models, most studies use the entire cohort of MCI patients regardless of their disease stages. In this paper, we propose a Time Windows approach to predict conversion to dementia, learning with patients stratified using time windows, thus fine-tuning the prognosis regarding the time to conversion. In the proposed Time Windows approach, we grouped patients based on the clinical information of whether they converted (converter MCI) or remained MCI (stable MCI) within a specific time window. We tested time windows of 2, 3, 4 and 5 years. We developed a prognostic model for each time window using clinical and neuropsychological data and compared this approach with the commonly used in the literature, where all patients are used to learn the models, named as First Last approach. This enables to move from the traditional question “Will a MCI patient convert to dementia somewhere in the future” to the question “Will a MCI patient convert to dementia in a specific time window”. The proposed Time Windows approach outperformed the First Last approach. The results showed that we can predict conversion to dementia as early as 5 years before the event with an AUC of 0.88 in the cross-validation set and 0.76 in an independent validation set. Prognostic models using time windows have higher performance when predicting progression from MCI to dementia, when compared to the prognostic approach commonly used in the literature. Furthermore, the proposed Time Windows approach is more relevant from a clinical point of view, predicting conversion within a temporal interval rather than sometime in the future and allowing clinicians to timely adjust treatments and clinical appointments.

The use of neuropsychological tests to detect cognitive decline in the initial phases of Alzheimer's disease (AD) has faced significant limitations, namely the fact that most cohort studies of conversion to dementia had relatively short follow-up periods. The aim of the present study is to assess the predictive value for future conversion to dementia of a comprehensive neuropsychological battery applied to a cohort of non-demented patients followed-up for 5 years. Participants (n = 250) were selected from the Cognitive Complaints Cohort (CCC) having cognitive complaints, assessment with a comprehensive neuropsychological battery, and a follow-up period of 5 years (unless patients have converted to dementia earlier). During the follow-up period (2.6 ± 1.8 years for converters and 6.1 ± 2.1 years for non-converters), 162 patients (64.8%) progressed to dementia (mostly AD), and 88 (35.2%) did not. A Linear Discriminant Analysis (LDA) model constituted by Digit Span backward, Semantic Fluency, Logical Memory (immediate recall), and Forgetting Index significantly discriminated converters from non-converters (λ Wilks = 0.64; χ(2) (4) = 81.95; p < 0.001; RCanonical = 0.60). Logical Memory (immediate recall) was the strongest predictor with a standardized canonical discriminant function coefficient of 0.70. The LDA classificatory model showed good sensitivity, specificity and accuracy values (78.8%, 79.9% and 78.6%, respectively) of the neuropsychological tests to predict long-term conversion to dementia. The present results show that it is possible to predict, on the basis of the initial clinical and neuropsychological evaluation, whether non-demented patients with cognitive complaints will probably convert to dementia, or remain stable, at a reasonably long and clinically relevant term.