Does working memory training have to be adaptive?

von Bastian, Claudia C.; Eschen, Anne · 2016 · Crossref

DOI: 10.1007/s00426-015-0655-z

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates whether adaptive working memory (WM) training—where task difficulty is automatically adjusted to individual performance—is necessary for effective cognitive training, or if exposure to varying levels of task difficulty is sufficient. Previous research suggesting adaptive training is superior often confounded adaptivity with variability, as control groups typically practiced only at the easiest difficulty level. To disentangle these factors, von Bastian and Eschen (2016) conducted a randomized controlled trial with 130 young adults, comparing three WM training procedures: adaptive (difficulty adjusted to performance), randomized (difficulty varied randomly independent of performance), and self-selected (participants chose difficulty changes). An active control group completed trivia quizzes with adaptive difficulty. Participants underwent an intensive regimen of 20 training sessions over four weeks, each lasting approximately 30–45 minutes. Training involved three complex span tasks (numerical, verbal, and figural-spatial) designed to measure WM capacity. The study employed a pretest-posttest design using a comprehensive battery of assessments to measure practice effects, intermediate transfer to structurally dissimilar WM tasks (word-position binding, Brown-Peterson, and memory updating), and far transfer to reasoning tasks (Raven’s Advanced Progressive Matrices, letter sets, locations, nonsense syllogisms, and diagramming relationships). The design was double-blind, and compliance was monitored via automated data uploads. The results demonstrated large performance increases in the trained WM tasks for all three WM training groups. However, no significant transfer effects were observed for either intermediate transfer to untrained WM tasks or far transfer to reasoning tasks. Crucially, there were no significant differences between the adaptive, randomized, and self-selected training conditions regarding training gains or transfer outcomes. While the randomized group practiced at lower average difficulty levels than the adaptive and self-selected groups, this did not result in superior or inferior outcomes. The active control group also showed improvements in their trivia tasks but no transfer to cognitive measures. The findings challenge the assumption that adaptive adjustment of task difficulty is a critical component for inducing cognitive plasticity or transfer effects. Instead, the study suggests that exposure to varying levels of task difficulty is sufficient to induce training gains in WM tasks. Since neither training nor transfer effects were modulated by the specific training procedure, the authors conclude that the variability of demands, rather than the adaptive alignment to individual performance, drives the observed improvements. This implies that simpler, non-adaptive training protocols may be equally effective for improving performance on trained tasks, though they do not appear to facilitate broader transfer to reasoning or untrained WM tasks in this sample.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• simulator training transfer