The unity and diversity of the executive functions: sociodemographic and cultural effects

• Autor(a) principal: Segura, Isis Angélica [UNIFESP]
• Data de Publicação: 2024
• Tipo de documento: Tese
• Idioma: eng
• Título da fonte: Repositório Institucional da UNIFESP
• Texto Completo: https://hdl.handle.net/11600/72319
• topic: Executive Functions; Adolescence; Age; Socioeconomic Factors; Culture; Cross-country; Lower-level Cognitive Processes

Executive Functions (EF) are a set of cognitive abilities that regulate behavior to achieve goals people have in mind. One widely accepted theory of EF is the unity and diversity framework, which posits that there is a pattern of intercorrelation (unity) among three (of many) EF domains (inhibition, shifting, updating) which are also separable (diversity). The exact factor structure of EF remains controversial, particularly during adolescence, when these domains become distinguishable. This controversy may also stem from various issues explored here in four studies: 1) psychometric issues; 2) sample differences in socioeconomic status (SES), culture and age; 3) inconsistent control of lower-level cognitive processes (LLP) that are also involved in EF tasks. Objectives: Study 1 and 2 determined the best EF factor structure and invariance to SES and age in early adolescents only from Iran (Study 1) plus a Brazilian sample (Study 2) in which invariance for country was also tested. Study 3 investigated the effect of LLP in the Iranian adolescent sample and three other databanks from other publications. Study 4 explored the role of LLP on adolescents’ age-related EF improvement using the Iranian and another adolescent dataset. Methods: Study 1 and 2: adolescents from Iran and Brazil (total sample: 739; 407 Iranians; 9-15 years) completed two tasks of each EF domain using a socioculturally adaptable EF test battery (Free Research Executive Evaluation). The best Confirmatory Factor Analysis EF model (out of seven tested ones) was assessed for invariance across age and SES (Multiple indicators Multiple causes), and also country (Multiple-group confirmatory factor analysis). Study 3: examined the impact of LLP on EF model in four reanalyzed datasets (two with adolescents and two with adults) (> 180 participants per sample). Using Structural Equation Modeling, a LLP latent factor reflecting shared variance of control conditions from EF tasks was regressed onto the three intercorrelated EF factors only using executive indicators (not executive cost measures). Study 4: with two adolescent datasets, the LLP factor was tested as a mediator of the direct age effects on the EF factors. Results: the three-correlated factor structure fit the data well for all tested datasets. This model was mostly invariant across age and SES (Study 1 and 2) and country (Study 2). All three EF traits improved with age, while SES had minimal positive effects on shifting and updating. The reanalyzed datasets (Study 3) also supported the three-intercorrelated EF factor using only executive indicators. LLP strongly predicted all three EF and altered EF intercorrelations (unity), which became mostly non-significant. LLP factor mediated age effects (Study 4) in both adolescent samples. Conclusion: EF differentiation is evident since early adolescence. This model is also invariant to age and SES, as well as to culture with socioculturally adaptable tests (Study 1 and 2). In spite of this, LLP explains most of EF performance, EF unity (Study 3) and the improvement of age across adolescence (Study 4). These results suggest the need to reappraise the EF unity and diversity framework.