Sleep restriction can attenuate prioritization benefits on declarative memory consolidation

Abstract

As chronic sleep restriction is a widespread problem among adolescents, the present study investigated the effects of a 1-week sleep restriction (SR) versus control period on the consolidation of long-term memory for prose passages. We also determined whether the benefit of prioritization on memory is modulated by adequate sleep occurring during consolidation. Fifty-six healthy adolescents (25 male, aged 15-19 years) were instructed to remember a prose passage in which half of the content was highlighted (prioritized), and were told that they would receive an additional bonus for remembering highlighted content. Following an initial free recall test, participants underwent a 7-night period in which they received either a 5-h (SR) or 9-h (control) nightly sleep opportunity, monitored by polysomnography on selected nights. Free recall of the passage was tested at the end of the sleep manipulation period (1 week after encoding), and again 6 weeks after encoding. Recall of highlighted content was superior to that of non-highlighted content at all three time-points (initial, 1 week, 6 weeks). This beneficial effect of prioritization on memory was stronger 1 week relative to a few minutes after encoding for the control, but not the SR group. N3 duration was similar in the control and SR groups. Overall, the present study shows that the benefits of prioritization on memory are enhanced over time, requiring time and sleep to unfold fully. Partial sleep deprivation (i.e. 5-h nocturnal sleep opportunity) may attenuate such benefits, but this may be offset by preservation of N3 sleep duration.

Keywords: adolescents; memory; preferential consolidation; reward; sleep; sleep loss.

Figure 1

This protocol started with three baseline nights [B1–B3; 9 h time in bed (TIB)], followed by seven nights of sleep manipulation (M1–M7) with 9 h TIB for the control group and 5 h TIB for the sleep restriction (SR) group. Both groups had 9 h TIB during recovery nights (R1–R3). Encoding (study) took place on the evening of the third baseline night. Retrieval took place before and after the sleep manipulation period (test 1, test 2) and 6 weeks after encoding (test 3; not depicted). Asterisks indicate nights that participants’ sleep was monitored by polysomnography. Sleep–wake patterns were monitored continuously by actigraphy except during night B1, when all Actiwatches were charged.

Figure 2

Free recall performance of the control and sleep restriction (SR) groups. This bar graph shows the mean percentage of highlighted (HL) and non‐highlighted (nHL) idea units that were recalled correctly a few minutes after learning (initial), after the sleep opportunity manipulation period (1 week), and 6 weeks after encoding (6 weeks). Error bars reflect standard error of the mean.

Figure 3

Prioritization benefit on memory. For all participants [control and sleep restriction (SR) groups combined] and for the control group alone, the benefit of prioritization (percentage of highlighted idea units correctly recalled minus that of non‐highlighted idea units) on memory was strengthened after the manipulation period relative to minutes after encoding. However, this did not hold true for the SR group. Error bars reflect standard error of the mean.

Figure 4

Sleep characteristics, separated by group [control, sleep restriction (SR)], averaged across manipulation nights 1, 4, and 7. Error bars reflect standard error of the mean.

Figure 5

Relation between sleep and memory performance at the post‐manipulation (a–b) and delayed (c–h) time points. Regression lines indicate the linear associations of the duration (averaged across manipulation nights M1, M4, and M7) of N2 (a–e) and rapid eye movement (REM) sleep (f–h) with memory for total content [i.e. highlighted (HL) and non‐highlighted (nHL) content; a,c,f], HL content (d,g), and nHL content (b,e,h). Data for the control and sleep restriction (SR) groups are illustrated in blue diamonds and red squares, respectively. All graphs depict significant correlations between sleep and memory for the SR, but not control, groups.