Contralateral delay activity as a marker of visual working memory capacity: a multi-site registered replication

The contralateral delay activity (CDA) is a widely used electrophysiological marker of visual working memory (VWM), yet recent work has questioned whether typical sample sizes in CDA studies are sufficient to robustly detect set size effects and brain-behavior correlations. As part of the #EEGManyLabs initiative, the present multi-site replication study aimed to rigorously test replicability of the key findings of Vogel and Machizawa (2004) using a large sample of 304 participants across 10 laboratories and a preregistered analysis plan. We replicated the expected contralateral-ipsilateral asymmetry and observed increases in CDA amplitude from set size 2 to 4 and from set size 2 to 6. In contrast, the hypothesized positive correlation between the CDA increase from set size 2 to 4 and individual VWM capacity was not replicated in the preregistered meta-analytic correlation. Across different pipelines and statistical analyses, the meta-analytic correlation estimate was small (r = 0.15) and substantially attenuated relative to the original effect size in Vogel and Machizawa (2004) study (r = 0.78). To contextualize these findings, we applied a funnel-plot diagnostic combining published effects with the #EEGManyLabs data, indicating small-study inflation and publication bias. Taken together, our results indicate that reports of strong correlations between CDA amplitude and VWM capacity may have been overestimated, in part because statistically significant findings were selectively reported. Our results highlight the importance of open science practices, including well-powered, preregistered studies with transparent data and analysis pipelines, in order to characterize the magnitude and robustness of individual-difference associations in psychophysiology.