Individual differences in working memory capacity and their effect on speech processing

Though tests of working memory (WM) correlate with scales of language development, it is unclear how WM capacity relates to spoken-language processing. However, Gilbert et al. (2014) have shown that listeners perceptually chunk speech in temporal groups (TGs) and that the span these TGs influences memory of heard items. Assuming that WM capacity links to this processing of speech in groups, listeners with the highest WM spans would be better at recalling items from long TGs. To examine this, we presented two sets of stimuli (utterances and sequences of meaningless syllables) containing long TGs. After each stimuli, listeners had to determine if a target item was previously heard. An analysis using GLME models showed that correct recognition memory of items heard in utterances was significantly better for listeners with high WM spans than for listeners with smaller spans. The effect was marginally significant for sequences of nonsense syllables.