Measuring working memory capacity using repetition priming of words

Temporal and item capacity limits (i.e., 2-seconds and 3 +/- 1 items, respectively) have been proposed to account for limitations on working memory performance. Such limitations have implications for tasks that depend on active cognitive processing (e.g., reading comprehension, dual task performance, and reasoning ability). Theoretical accounts of Working Memory (WM) capacity rely on decay of information held in a rehearsal mechanism (the phonological loop) or a limitation of memory activation due to the limited capacity of attention. Decay of memory activation has also been used to explain decline in repetition priming in short-term Implicit Memory tasks. Decay of repetition priming in a naming task was used in the current study to obtain an indirect measure of WM capacities. Participants viewed lists of sequentially presented words followed by a final to-be-read aloud word. On priming trials the to-be-read final word was presented somewhere within the list. The serial position of the prime was varied within subjects and across trials. Also, the inter-stimulus interval between list words was varied between subjects. It was expected that reaction times (RTs) to reading primed words, compared to non-primed words, would provide a measure of decay of memory activation an indirect measure of item and temporal WM capacity limits. The results indicate that repetition priming decays rapidly from Lag 0 (immediate priming) to Lag 2 then increases to Lags 3 or 4 after which priming decays across Lags. The initial rapid decrease then increase in repetition priming is attributed to the influence of inhibitory processes, which dissipates after about Lag 4. This pattern of repetition priming decay is discussed with reference to its implications for an item limit on working memory capacity. Specifically, an activation limit alone cannot account for the observed pattern of repetition priming; another process such as inhibition is needed to justify an item WM limit on repetition priming. The study's results provided no support for a 2-second temporal limit. It may be that intentional maintenance of list information or phonological coding of list items is necessary to detect the temporal limit associated with rehearsal of information in the phonological loop.