The role of subvocal speech in memory : An EMG study of preschool children



Journal Title

Journal ISSN

Volume Title



The relationship of subvocal speech to recall on a memory task was investigated for 54 four- and five-year-old nursery school children. Electromyographic (EMG) recordings were obtained from the lip and chin speech muscles and a control nonspeech facial muscle. Total EMG activity was then calculated by integrating the power for frequencies from 10 to 500 hz. The memory task consisted of 10 trials, each containing three successively-presented black and white line drawings of familiar objects. The 15-second stimulus presentation period (SPP) was followed by a 15-second delay period (DP), after which Ss were to orally recall the pictures in the order of presentation. During the SPP and the DP, Ss were not permitted to visibly move their lips or to audibly vocalize. The names of the pictures in five of the trials contained labial phonemes which involved a maximum of lip movement when silently pronounced. The other five trials contained nonlabial phonemes involving a minimum of lip movement. The total EMG activity in the nonlabial trials was used as a conservative control (superior to the commonly-used resting baseline) for nonspeech artifact (e.g., lip licking, grimacing) in labial trials. The recall difficulty in labial and nonlabial trials was virtually identical, supporting the assumption that nonspeech artifact was randomly distributed over both sets of trials. The nonlabial sum was therefore subtracted from the sum of the labial trials for each S and, if the difference was positive, subvocal speech was considered to have occurred. When these difference scores were analyzed the following results were obtained: (1) highly significant amounts of subvocal speech were found across all Ss during both SPP and DP; (2) during the SPP no difference in amount of subvocal speech occurred between a group of younger and of older Ss, or between a group of less bright and of brighter Ss; (3) during the DP both older, and brighter, Ss subvocalized significantly more than did their counterparts; (4) this difference was not attributable to differences in either amount of subvocal speech or total raw power, since these were similar in the SPP and the DP; (5) thus, the subvocal speech activity of both older and brighter Ss had increased during the DP compared to their SPP levels, while that of younger and of less bright Ss had decreased. The following results were obtained with respect to the relationship between subvocalization and recall: (1) sub-vocal speech and total recall during SPP and DP were significantly correlated for the two speech muscles but not for the nonspeech facial control muscle; (2) the correlation for boys was much higher than for girls, particularly during the DP; this sex difference was not attributable to differences in age, IQ, total amount of subvocal speech or recall; (3) the Ss who subvocalized the most obtained significantly higher recall scores than did those Ss who subvocalized the least. Serial position (SP) analyses of recall order indicated that: (1) for the one-item responses, younger-less bright Ss and low subvocalizers recalled third SP items most often, whereas their counterparts recalled first and third SP items most often; (2) for the two-item responses, younger-less bright Ss both recalled a far smaller percentage of items in the first and second SP's, and made far more transposition errors (particularly with the second and third SP's) than did older-brighter Ss. Results are discussed in terms of Vygotsky's theory of inner speech, and the mediation- and production-deficiency hypotheses. Relevant findings of adult and child EMG and short-term memory studies are compared and contrasted with the present findings.



Memory, Verbal learning