Revisitando modelos teóricos e métodos de avaliação das funções executivas a partir de análise de variáveis latentes

Abstract

The goal of this dissertation was to contribute to the current literature on the underlying structure of the executive functions (EF) through a combination of theoretical and empirical approaches. The findings produced by these studies have ample clinical applicability to áreas raning from neuropsychology to psychiatry. The project contains two empirical studies, titled “Factor analysis of cognitive tasks in healthy adults and patients with mood disorders: the challenges of finding a universal model of executive functions” and “Cognitive abilities underlying performance on the Modified Card Sorting Test: novel and traditional scores”, respectively. The studies followed a retrospective design, and involved a sample of approximately 100 adults with no mood disorders, 80 patients with major depressive disorder (MDD), 45 individuals with bipolar disorder type II (BDII) and 40 participants with bipolar disorder type I (BDI). The overall aim of this project and of each individual study was to revisit existing theoretical models of EF through the statistical analysis of assessment tasks and batteries within a clinical context. The goals of the first study included the identification of underlying factors in a neuropsychological assessment battery; the comparison of the fator structure across homogeneous (i.e. healthy adults, patients with a single mood disorder) and heterogeneous populations (i.e. mixed sample of adults with and without mood disorders); and the comparison between the results of these analysis and the current interpretation of EF scores in widely used assessment tasks. The analyses revealed an underlying structure with three to six factors depending on the population studied. Robust working memory and visuoverbal processing speed factors were identified across all populations, while the majority of other assessment measures loaded on to a single factor, or were inconsistently associated with different variables in each analysis. These findings highlight the importance of studying the underlying variables of EF tasks in different populations, including healthy individuals and patients with mental illnesses. There is also a need to refine the most widely used measures in neuropsychological assessment so as to provide a more specific evaluation of the EF. In light of these findings, the aim of the second study was to identify the cognitive functions underlying traditional and novel scores on the Modified Card Sorting Test. A significant association was observed beteween performance on this test and measure sof working memory and inhibition. Novel scores may complement assessments using this instrument, allowing for the evaluation of diferente cognitive functions and providing more detailed results. In conclusion, this dissertation project illustrates the important contributions of a hybrid empirical and clinical approach to neuropsychological assessment and research, in order to investigate and validate theoretical models and assessment tasks.