Training versus engagement as paths to cognitive enrichment with aging

Abstract

While a training model of cognitive intervention targets the improvement of particular skills through instruction and practice, an engagement model is based on the idea that being embedded in an intellectually and socially complex environment can impact cognition, perhaps even broadly, without explicit instruction. We contrasted these 2 models of cognitive enrichment by randomly assigning healthy older adults to a home-based inductive reasoning training program, a team-based competitive program in creative problem solving, or a wait-list control. As predicted, those in the training condition showed selective improvement in inductive reasoning. Those in the engagement condition, on the other hand, showed selective improvement in divergent thinking, a key ability exercised in creative problem solving. On average, then, both groups appeared to show ability-specific effects. However, moderators of change differed somewhat for those in the engagement and training interventions. Generally, those who started either intervention with a more positive cognitive profile showed more cognitive growth, suggesting that cognitive resources enabled individuals to take advantage of environmental enrichment. Only in the engagement condition did initial levels of openness and social network size moderate intervention effects on cognition, suggesting that comfort with novelty and an ability to manage social resources may be additional factors contributing to the capacity to take advantage of the environmental complexity associated with engagement. Collectively, these findings suggest that training and engagement models may offer alternative routes to cognitive resilience in late life.

Figure 1

CONSORT diagram charting the flow of participants from through the study. (Note that “retained” refers to participant status in the intervention groups, regardless of whether the individual returned for testing; because our analysis as intent-to-treat, the number tested could be greater than the number retained in the program.)

Figure 2

Pre- to posttest change in manifest variables and latent constructs. Error bars represent standard errors. Change in latent constructs is highlighted with slightly larger symbols.

Figure 3

Model-estimated change in Divergent Thinking in the Engagement intervention as a function of age, cognitive status (MoCA), verbal ability, Openness, Need for Cognition, and Social Network Size.

Figure 4

Model-estimated change in Inductive Reasoning in the Training intervention as a function of cognitive status (MoCA) and verbal ability.