Input and output modes modulate phonological and semantic contributions to immediate serial recall: Evidence from a brain-damaged patient

Abstract

Psycholinguistic models of short-term retention suggest that performance at verbal short-term memory (STM) tasks relies on the activation of phonological, lexical, and semantic representations, the relative impact of each depending on task variables. This was tested in normal individuals and in I.R., a brain-damaged patient with a phonological deficit. In Experiment 1, the effect of phonological and semantic similarity was assessed under different presentation formats (words, pictures) and recall modes (oral, picture pointing, and picture pointing among distractors). In Experiment 2, effects were compared using reproduction and reconstruction responses. When words were used at input, controls showed robust phonological similarity effects irrespective of response mode. In contrast, I.R. showed a reliable semantic effect. However, both studies indicated that when response mode promoted order recuperation (reconstruction and picture pointing modes), I.R. showed a typical phonological similarity effect with no semantic contribution. The data support current psycholinguistic views suggesting that the short-term retention of verbal items depends on the temporary activation of word representations. In healthy controls, presentation mode appears to modulate the role of those representations but in I.R., it was the output condition—particularly whether order was or was not required—that was found to be crucial with respect to the appearance of semantic or phonological effects. This supports the important role that order information plays in short-term memory tasks.

Keywords:

• Short-term memory
• Immediate serial recall
• Phonological representations
• Semantic representations
• Order memory
• Interactive activation model

This work was supported by studentships from the Natural Sciences and Engineering Research Council of Canada (NSERC) and the Canadian Institutes of Health Research (CIHR) to Véronique Chassé and by a grant from NSERC and a chercheur-boursier fellowship from the Fonds de la Recherche en Santé du Québec (FRSQ) to Sylvie Belleville. We would like to thank I.R. for her assiduous collaboration and Bernard Bouchard for assistance in constructing and editing the stimuli.

Notes

¹ A visual digit span of 4 is considered abnormal when compared to French norms provided by Belleville, Châtelois, Fontaine, and Peretz (2002), which show average digit spans of 6.78 (1.06) in aged controls and 7.11 (1.02) in young participants using visually presented material.