It has been recently proposed that pregnant women would perform memory tasks by focusing more on item-specific processes and less on relational processing, compared to post-partum women (Mickes, Wixted, Shapiro & Scarff, ). The present cross-sectional study tested this hypothesis by directly manipulating the type of encoding employed in the study phase. Pregnant, post-partum and control women either rated the pleasantness of word meaning (which induced item-specific elaboration) or named the semantic category to which they belonged (which induced relational elaboration). Memory for the encoded words was later tested in free recall (which emphasizes relational processing) and in recognition (which emphasizes item-specific processing). In line with Mickes et al.'s () conclusions, pregnant women in the item-specific condition performed worse than post-partum women in the relational condition in free recall, but not in recognition. However, compared to the other two groups, pregnant women also exhibited lower recognition accuracy in the item-specific condition. Overall, these results confirm that pregnant women rely on relational encoding less than post-partum women, but additionally suggest that the former group might use item-specific processes less efficiently than post-partum and control women.

Previous studies indicate that pregnancy is associated with a mild impairment in verbal recall tasks, but the relative roles of item-specific and relational encoding processes have not been fully elucidated. The present study was specifically aimed at determining whether pregnancy had a differential impact on the use of these 2 types of encoding processes. The performance of pregnant, postpartum, and control women was assessed in the category-cued recall (Experiment 1: explicit memory) and category exemplar generation (Experiment 2: implicit memory) tasks, by simultaneously varying the number of exemplars presented during the study phase (1 exemplar for small categories, 6 exemplars for large categories). Control women showed the expected crossed double dissociation: Performance in category-cued recall was better for small than for large categories (confirming that this task relied on the use of item-specific encoding processes), whereas repetition priming in category exemplar generation was higher for large than for small categories (confirming that this task relied on the use of relational encoding processes). For pregnant and postpartum women, no significant differences between small and large categories were detected in the category-cued recall task; furthermore, in the category exemplar generation task, these 2 groups exhibited greater priming for small category exemplars. Overall, our data suggest that pregnancy might involve complex changes in the use of item-specific and relational encoding processes.