Working memory training improves episodic memory in older people: transfer based on controlled retrieval processes

Introduction: The results of working memory (WM) training to improve episodic memory in older people are inconsistent, potentially because transfer depends on overlap in cognitive resources between trained and tested tasks. This study tested whether performance on an episodic memory test improves only when it requires controlled retrieval from secondary memory (recollection), akin to the processes engaged during WM training. Method: Fifty-five adults over 60 years (27 experimental group, EG; 28 control group, CG) completed pre- and post-tests including episodic memory (verbal and visuospatial recognition) and WM span tasks (reading, digit, spatial location). The EG completed adaptive complex span WM training; the CG completed non-adaptive perceptual speed tasks. Results: ANOVAs showed greater improvement in recollection (exclusion/local conditions) for the EG than the CG in both verbal recognition (p = 0.023, with schooling covariate; marginal without covariate) and visuospatial recognition (p = 0.014). Discussion: Findings support that transfer from WM training to episodic memory occurs when tasks share controlled retrieval processes. Complex span training enhanced recognition only when it relied on recollection (controlled search in secondary memory). Identifying shared cognitive resources is important for designing effective cognitive interventions for older adults.