RESUMO
The Positive Mental Health Questionnaire (PMHQ) has been validated across various populations but has displayed diverse psychometric structures depending on the procedures used. The original version of the PMHQ includes 39 items organized into 6 factors, although there are reports that indicate a reduced structure of between 1 and 4 factors. The aim of this study was to assess the psychometric properties of the PMHQ with 1, 4 and 6 factors. A total of 360 healthcare workers aged 23 to 77 (M = 37.06; SD = 10.79) participated. Construct validity was assessed through confirmatory factor analysis using weighted root mean square residual. The original 6-factor (χ2/df: 3.40; RMSEA: 0.085; CFI: 0.913; TLI: 0.906) and a reduced 4-factor (χ2/df: 2.90; RMSEA: 0.072; CFI: 0.931; TLI: 0.926) structure showed acceptable fit. The fit of the 1-factor model was unacceptable. The internal consistency was evaluated through McDonald's ω, and it was acceptable for 4 of 6 factors of the original structure and for 3 of 4 factors of the reduced structure. In conclusion, these findings suggest that the 6-factor and 4-factor models are valid for measuring positive mental health. However, issues with internal consistency must be investigated.
RESUMO
Abstract There is a growing interest to understand the neural functions and substrates of complex cognitive processes related to Obesity (OB). Artificial Intelligence (AI) is being applied, specifically the perceptron model of Artificial Neural Networks (ANN) in non-communicable chronic diseases, to identify with greater certainty the connective factors (synaptic networks) between the input variables and the output variables associated. Objective Identify the synaptic weights of the ANN whose input variables are the executive functions (EF) and healthy lifestyles as predictors of Body Fat Percentage (BFP) in a group of adult subjects with different levels of BFP. Methods It was an exploratory, quantitative, cross-sectional, comparative, convenience, and explanatory research. The Neuropsychological Battery (BANFE-2) and the Overeating Questionnaire (OQ) were administered to 40 participants aged between 18-38 years. BFP was measured using a RENPHO ES-24M Smart Body Composition Scale. The perceptron ANN model with ten trials was applied with a multilayer-perceptron. Results The ANN showed that the sensory variables with greater synaptic weight for BFP were Stroop A and B Errors and Successes of BANFE-2, and OQ scales Rationalizations and Healthy Habits. Conclusions ANN proved to be important in the simultaneous analysis of neuropsychological and healthy lifestyle data for the analysis of OB prevention and treatment by identifying the variables that are closely related. These findings open the door for the use of non-linear analysis models, which allow the identification of relationships of different weights, between input and output variables, to more effectively direct interventions to modify obesity habits.
Resumen Existe un interés creciente por comprender las funciones neuronales y sustratos cognitivos complejos relacionados con la obesidad. Se está aplicando Inteligencia Artificial, en concreto el modelo perceptrón de Redes Neuronales Artificiales en enfermedades crónicas no transmisibles, para identificar con mayor certeza los factores de conexión (redes sinápticas) entre las variables de entrada y las variables de salida. Objetivo Identificar pesos sinápticos de la RNA cuyas variables de entrada fueron las funciones ejecutivas y los estilos de vida saludable, como predictores del Porcentaje de Grasa Corporal en un grupo de sujetos adultos con diferentes niveles del Porcentaje de Grasa. Métodos se trató de una investigación exploratoria, cuantitativa, transversal, comparativa, de conveniencia y explicativa. Se administró la Batería Neuropsicológica (BANFE-2) y el Cuestionario de Sobreingesta (OQ), a 40 participantes con edades comprendidas entre los 18-38 años. El porcentaje de grasa se midió con una báscula de composición corporal (RENPHO ES-24M). El modelo redes neuronales de perceptrón, se ejecutó con diez ensayos. Resultados El modelo de Red Neuronal mostró que las variables sensoriales con mayor peso sináptico para el porcentaje de grasa, fueron Errores Stroop A y B y Aciertos de BANFE-2, y Racionalizaciones de las escalas OQ y Hábitos Saludables. Conclusiones las redes neuronales artificiales demostró ser importante en el análisis simultáneo de datos neuropsicológicos y de estilo de vida saludable para el análisis de prevención y tratamiento de la obesidad, al identificar las variables que están estrechamente relacionadas. Estos hallazgos abren la puerta al uso de modelos de análisis no lineales, que permiten identificar relaciones de diferente peso, entre variables de entrada y salida, más eficientes que los modelos lineales.