Declarative Memory Predicts Phonological Processing Abilities in Adulthood

Abstract

Individual differences in phonological processing abilities have often been attributed to perceptual factors, rather than to factors relating to learning and memory. Here, we consider the contribution of individual differences in declarative and procedural memory to phonological processing performance in adulthood. We examined the phonological processing, declarative memory, and procedural memory abilities of 79 native English-speaking young adults with typical language and reading abilities. Declarative memory was assessed with a recognition memory task of real and made-up objects. Procedural memory was assessed with a serial reaction time task. For both tasks, learning was assessed shortly after encoding, and again after a 12-h, overnight delay. We regressed phonological processing ability with memory performance on both days. We found that declarative memory, but not procedural memory, was highly predictive of phonological processing abilities. Specifically, declarative memory scores obtained shortly after learning were associated with non-word repetition performance, whereas declarative memory scores obtained after the overnight delay were associated with phonological awareness. Procedural memory was not associated with either of the phonological processing measures. We discuss these findings in the context of adult participants with mature phonological systems. We examine possible implications for the relationship between declarative memory and phonological processing in adulthood.

Keywords: declarative memory; learning; nonword repetition; phonological awareness; phonological processing; procedural memory.

FIGURE 1

Effect plots for Session 2 memory abilities on phonological awareness (PA). Figure depicts the relative predictive relationships between PA and declarative memory (DM) and PA and procedural memory (PM), assessed after an overnight delay, and controlled for site. Values for all variables are scaled according to the proximity-to-maximum method (Moeller, 2015). Graphs were generated using the package “effects” in R (Fox et al., 2016). Shaded regions represent the pointwise confidence band for the predictor effects, based on standard errors calculated from the covariance matrix of the fitted regression coefficients (Fox and Weisberg, 2018).

FIGURE 2

Effect plots for Session 1 memory abilities on nonword repetition (NWR). Figure depicts the relative predictive relationships between NWR and DM assessed shortly after learning, and NWR and PM, and controlled for site. Values for all variables are scaled according to the proximity-to-maximum method (Moeller, 2015). As in Figure 1, graphs were generated using the package “effects” in R (Fox et al., 2016), and shaded regions represent the pointwise confidence band for the predictor effects.