Performance of Normal-Hearing Children on a New Working Memory Span Task1

This paper reports memory span data from 45 monolingual, normal-hearing children ranging in age from 3;5 to 5;11 years. A newly implemented methodology (here termed the “memory span game”) required participants to reproduce sequences of lights, or sequences of lights accompanied by sounds, by pressing appropriately matched buttons on a response box. Auditory stimuli used in the light-plus-sound condition included digitnames and nonsense syllables. The effects of auditory stimulus type, and of presenting the target sequence as lights-plus-sound vs. lights-only, were examined. Digit-spans obtained using the memory span game were compared to spans from the same children obtained using the conventional digit memory span sub-test of the Wechsler Intelligence Scale for Children III, in order to assess the degree to which the processing demands of these two tasks overlap. In addition, expanding on previously reported findings in the literature concerning a relationship between working memory span and vocabulary development, we obtained measures of each child’s receptive and expressive vocabulary as well as nonword repetition skills using two standardized tests. Results show that children improve as a function of age on the new memory game task at rates very similar to that obtained for the WISC forward digit span task. We found that children younger than 3;5 were generally unable to meet the demands of the memory span game task. Differences between the group means between the three different memory span game conditions were not significant; however, correlational analyses showed clear differences between the visual-spatial/lights-only condition vs. the auditory plus visual-spatial condition in how predictive each score was of vocabulary measures. Scores obtained from the auditory plus visual-spatial condition of the memory game task accounted for approximately 12-16% of the variance in vocabulary scores with the effects of age partialled out, while the lights-only/silent condition scores only predicted approximately 2% of this variance. Results from the nonword repetition measures show that the phonetic similarity of the nonsense words to familiar words influenced the strength of this task’s relationship with vocabulary measures, but not with working memory span measures such as WISC digit-span and memory span game scores (although the previously reported small positive correlations between these measures were replicated). These results have both theoretical and clinical implications for the use of this reproductive memory span task with hearing-impaired children who may have expressive language problems or delays. The findings are also relevant to the development of phonological representations in short-term memory and spoken word recognition abilities in normalhearing children.