Effects of learnt landmarks and sleep on spatial working memory

• Published in: The Lancet (British edition) Vol. 387; p. S58
• Main Authors: Katshu, Mohammad Zia Ul Haq, Dr, d'Avossa, Giovanni, MD
• Format: Journal Article
• Language: English
• Published: London Elsevier Ltd 25.02.2016; Elsevier Limited
• Subjects: Internal Medicine
• ISSN: 0140-6736; 1474-547X
• DOI: 10.1016/S0140-6736(16)00445-1

Abstract Background Sleep consolidates spatial memory through reactivation of patterns of neural activity associated with spatial navigation. Recently, post-training sleep has been shown to improve spatial working memory (sWM). Whether improvement in sWM reflects a general post-sleep facilitation of learning, or is more specific and based on consolidation of landmarks, is not known. To tease these effects apart, we assessed, in a mixed-design study, the effect of learnt landmarks and sleep on sWM. Methods Healthy participants were randomly assigned by computer generated sequence to morning, afternoon, or night groups, indicating timing of first session, to assess effect of sleep on sWM. Each participant completed three sessions at approximately 8 h intervals while maintaining regular sleep–wake schedules. The first session included learning a landmark and a fixed-memory item whose effect was assessed by comparing prelearning and postlearning sWM tasks over the three sessions. Group level effects for accuracy and precision of sWM were tested with mixed ANOVAs. Findings We randomised 36 people (12 in each group). During the learning phase in all participants, targets in the proximity of the landmark ( F =4·564, p<0·0001) and fixed-memory item ( F =3·219, p<0·0001) were recalled more precisely, confirming that participants learnt the landmark and the fixed-memory item. Both landmark ( F =4·795, p<0·0001) and fixed-memory item ( F =1·984, p=0·005) also improved recall accuracy of nearby targets during the learning session. In the postlearning phase, improved recall accuracy ( F =1·771, p<0·0001), but not precision, was observed for targets in the vicinity of the landmark. There was no improvement in recall precision or accuracy for targets closer to fixed-memory item. This improvement in recall accuracy was lost if there was no postlearning sleep period. No overall improvement in spatial recall precision or accuracy across sessions or after sleep was observed. Interpretation Learnt landmarks facilitate encoding of novel visual information in sWM. Crucially, our study has shown that the improvement in sWM observed after sleep reflects the consolidation of these learnt landmarks rather than procedural learning. Funding None.