Mood is a key determinant of cognitive performance in community-dwelling older adults: a cross-sectional analysis

Abstract

Identification of predictors of cognitive trajectories through the establishment of composite or single-parameter dimensional categories of cognition and mood may facilitate development of strategies to improve quality of life in the elderly. Participants (n = 487, aged 50+ years) were representative of the Portuguese population in terms of age, gender, and educational status. Cognitive and mood profiles were established using a battery of neurocognitive and psychological tests. Data were subjected to principal component analysis to identify core dimensions of cognition and mood, encompassing multiple test variables. Dimensions were correlated with age and with respect to gender, education, and occupational status. Cluster analysis was applied to isolate distinct patterns of cognitive performance and binary logistic regression models to explore interrelationships between aging, cognition, mood, and socio-demographic characteristics. Four main dimensions were identified: memory, executive function, global cognitive status, and mood. Based on these, strong and weak cognitive performers were distinguishable. Cluster analysis revealed further distinction within these two main categories into very good, good, poor, and very poor performers. Mood was the principal factor contributing to the separation between very good and good, as well as poor and very poor, performers. Clustering was also influenced by gender and education, albeit to a lesser extent; notably, however, female gender × lower educational background predicted significantly poorer cognitive performance with increasing age. Mood has a significant impact on the rate of cognitive decline in the elderly. Gender and educational level are early determinants of cognitive performance in later life.

Fig. 1

Gender relationship with age for each identified composite and single dimension. aMEM (R² linear, female, 0.126; male, 0.176), bEXEC (R² linear, female, 0.114; male, 0.155), cMMSE (R² linear, female, 0.184; male, 0.148), dGDS (R² linear, female, 0.003; male, 0.003). Males and females are represented in blue and green circles, respectively

Fig. 2

Relationship between educational level and age for each identified composite and single dimension. aMEM (R² linear, 0, 0.034; 1–2, 0.047; 3–4, 0.096; 5–8, 0.289; 9–12, 0.054; 13+ years, 0.519), bEXEC (R² linear, 0, 0.109; 1–2, 0.02; 3–4, 0.118; 5–8, 0.114; 9–12, <0.001; 13+ years, 0.002), cMMSE (R² linear, 0, 0.046; 1–2, 0.088; 3–4, 0.084; 5–8, 0.013; 9–12, 0.035; 13+ years, 0.039), dGDS (R² linear: 0, <0.001; 1–2, <0.001; 3–4, 0.004; 5–8, 0.031; 9–12, 0.079; 13+ years, 0.271). School years 0, 1–2, 3–4, 5–8, 9–12, and 13+ are shown as blue, green, red, purple, orange, and white circles, respectively

Fig. 3

Relationship between occupational status and age for each identified composite and single dimension. aMEM (R² linear, retirement, 0.093; employment, 0.164; unemployment, 0.039), bEXEC (R² linear, retirement, 0.117; employment, 0.067; unemployment, 0.003), cMMSE (R² linear retirement, 0.116; employment, 0.214; unemployment, 0.007), dGDS (R² linear, retirement, 0.001; employment, 0.042; unemployment, <0.001). Occupational statuses are shown as blue (retirement), green (employed), and red (unemployed) circles

Fig. 4

Cluster analysis. Mean performance z-scores by clusters in the MEM, EXEC, MMSE, and GDS dimensions

Fig. 5

Cluster relationship, according to age, for each identified composite and single dimension. aMEM (R² linear, C1, <0.001; C2, 0.031; C3, 0.092; C4, 0.006), bEXEC (R² linear C1, <0.001; C2, 0.037; C3, 0.073; C4, 0.089), cMMSE (R² linear, C1, 0.006; C2, 0.032; C3, 0.018; C4, 0.049), dGDS (R² linear, C1, 0.005; C2, 0.009; C3, 0.051; C4, 0.031). Total fit line is represented in black dotted line (R² linear MEM, 0.123; EXEC, 0.122; MMSE, 0.114; GDS, <0.001). Individuals in clusters C1, C2, C3, and C4 are depicted as blue, green, red, and purple circles, respectively