Social incentives improve deliberative but not procedural learning in older adults

Abstract

Age-related deficits are seen across tasks where learning depends on asocial feedback processing, however plasticity has been observed in some of the same tasks in social contexts suggesting a novel way to attenuate deficits. Socioemotional selectivity theory suggests this plasticity is due to a deliberative motivational shift toward achieving well-being with age (positivity effect) that reverses when executive processes are limited (negativity effect). The present study examined the interaction of feedback valence (positive, negative) and social salience (emotional face feedback - happy; angry, asocial point feedback - gain; loss) on learning in a deliberative task that challenges executive processes and a procedural task that does not. We predict that angry face feedback will improve learning in a deliberative task when executive function is challenged. We tested two competing hypotheses regarding the interactive effects of deliberative emotional biases on automatic feedback processing: (1) If deliberative emotion regulation and automatic feedback are interactive we expect happy face feedback to improve learning and angry face feedback to impair learning in older adults because cognitive control is available. (2) If deliberative emotion regulation and automatic feedback are not interactive we predict that emotional face feedback will not improve procedural learning regardless of valence. Results demonstrate that older adults show persistent deficits relative to younger adults during procedural category learning suggesting that deliberative emotional biases do not interact with automatic feedback processing. Interestingly, a subgroup of older adults identified as potentially using deliberative strategies tended to learn as well as younger adults with angry relative to happy feedback, matching the pattern observed in the deliberative task. Results suggest that deliberative emotional biases can improve deliberative learning, but have no effect on procedural learning.

Keywords: age; automatic; deliberative; feedback; learning; procedural; social.

FIGURE 1

An adapted schematic of the task from Gorlick et al. (2013). This task directly manipulated task demands on cognitive control. Behavior suggests that learning with social feedback depends on a complex three-way interaction between task demands (high cognitive control demand (top): three dimensional stimuli each with four possible values, low cognitive control demand (bottom): two dimensional stimuli each with two possible values), age (younger, older) and social feedback valence (positive happy face (left), negative angry face (right)).

FIGURE 2

Example surface features (houses, vehicles) with an example of category membership for the rule-based (A,C; rule = color) and information integration (B,D; rule = shape (window, body), number (cloud, wheel) accessory (landscape, tool); see Materials and Methods) tasks. Features disambiguating category membership are highlighted in green below each column in green. (E) Screen shots for the social (top) and asocial (bottom) feedback manipulations. Within each of these condition the valence of the feedback was either positive (left; increase happiness or approach the goal) or negative (right; decrease anger or avoid increasing the goal).

FIGURE 3

The number of trials needed to learn the rule for older and younger adults given (A) rule-based (RB) and information integration (II) category structures in the happy- and angry-face-feedback conditions (B) within the information integration condition by strategy (II-II = information integration strategy; II-UD = unidimensional strategy) and (C) the rule-based and information integration tasks in the point gain and point loss feedback conditions (II = information integration structure; RB = unidimensional structure). SE bars are included.