Abstract
Electromyographic studies of subvocal speech are evaluated with two different aims: (1)a methodological review of controls for nonspeech artifact and of data analysis techniques, and (2) a review of substantive findings and suggestions for future research. A highly effective procedure, involving manipulation of phoneme content in certain trials, has been developed to control nonspeech artifact generated by the speech musculature. The commonly used data analysis technique of amplitude measurement of the single highest polygraph pen deflection is inadequate. Techniques which take frequency into account, such as analog computer integration of voltage values, are more sensitive. Even greater sensitivity is needed to isolate individual words so that rehearsal strategies can be investigated. It has been clearly established that preschool children as young as age 4 engage in spontaneous subvocalization that is related to recall in boys but not in girls.

This publication has 35 references indexed in Scilit: