Interestingly, midfrontal theta power could tell errors apart (Hit < Miss, FA < CR) regardless of whether images were learned before. Therefore, we proposed that our brain detects memory-based errors with ERN, and general recognition errors with midfrontal theta power. 3/3

First, error-related negativity (ERN) could tell whether our brain made a recognition error or not: but only for images that were learned (i.e. hits versus misses). It couldn't tell false alarms and correct rejections apart even though FAs were also errors. 2/3

In Study 3, we kept the presentation time at 250ms but gave participants a longer ISI, and we observed that the WM-LTM relationship as we saw in Study 1. This suggested that WM matters to LTM if people get more time to form spatio-temporal binding (during ISI). 4/5

However, in Study 2, if we speed up the presentation rate of the recognition memory task (250ms/image), WM stopped predicting recognition memory differences. This suggests that WM affects the formation of spatio-temporal binding of items formed only with slower presentation rates. 3/5

In Study 1, we found that working memory correlated with simple recognition memory when the images were presented slowly (3 seconds per image). The correlation holds even we regressed out source memory differences, suggesting that WM generally predicts LTM abilities. 2/5

Tonight 5-7 PM I’ll be presenting my poster at CNS @cogneuronews.bsky.social C46: “Repetition learning produces stronger and faster recollection during recognition” Don’t miss it if you’re interested in EEG and visual long-term memory! #CNS2025

Across 5 experiments, we found that working memory and attentional control (WMAC) ability continued to predict long-term memory (LTM) performance even after participants showed significant learning. Huge thanks to my advisor Dr. Ed Vogel!

New paper out now in JEP:G.

"Individual differences in working memory and attentional control continue to predict memory performance despite extensive learning."

psycnet.apa.org/doi/10.1037/xg…