The benefit of offline sleep and wake for novel object recognition

Abstract

How do we segment and recognize novel objects? When explicit cues from motion and color are available, object boundary detection is relatively easy. However, under conditions of deep camouflage, in which objects share the same image cues as their background, the visual system must reassign new functional roles to existing image statistics in order to group continuities for detection and segmentation of object boundaries. This bootstrapped learning process is stimulus dependent and requires extensive task-specific training. Using a between-subject design, we tested participants on their ability to segment and recognize novel objects after a consolidation period of sleep or wake. We found a specific role for rapid eye movement (REM, n = 43) sleep in context-invariant novel object learning, and that REM sleep as well as a period of active wake (AW, n = 35) increased segmentation of context-specific object learning compared to a period of quiet wake (QW, n = 38; p = .007 and p = .017, respectively). Performance in the non-REM nap group (n = 32) was not different from the other groups. The REM sleep enhancement effect was especially robust for the top performing quartile of subjects, or "super learners" (p = .037). Together, these results suggest that the construction and generalization of novel representations through bootstrapped learning may benefit from REM sleep, and more specific object learning may also benefit from AW. We discuss these results in the context of shared electrophysiological and neurochemical features of AW and REM sleep, which are distinct from QW and non-REM sleep.

REM sleep rescues learning from interference

“…One possibility is that the improvements observed in the NREM and REM groups were not due to any active sleep consolidation processes in a specific stage, but rather a passive reduction in information input, similar to that experienced during QW. Other studies have found similar learning profiles between QW and NREM sleep (McDevitt, Rowe, Brady, Duggan, & Mednick, 2014). QW and NREM sleep share some neurophysiological characteristics that may make both brain states conducive to consolidation.…”

The Consolidation and Transformation of Memory

“…Behavioral studies likewise provide a mixed picture. REM sleep seems to benefit particularly memory that is not dependent on cortico-hippocampal circuitry, including procedural skills, object recognition memory, and amygdalamediated cued fear conditioning (Karni et al, 1994;Plihal and Born, 1997;McDevitt et al, 2014;Popa et al, 2010; also see below), as well as stimuli conforming with preexisting schemas (Durrant et al, 2015). These observations could indeed be parsimoniously explained by assuming a stabilization effect mediated via local, synaptic consolidation.…”

Caffeine and modafinil given during 48h sleep deprivation modulate object recognition memory and synaptic proteins in the hippocampus of the rat

“…Novel objection recognition test (NORT) is a well-established method to study recognition memory function [1,2].Both rapid eye movement (REM) sleep and slow wave sleep (SWS) increase scores of object recognition task [3]. SD impairs recognition memory as presented by a deficit in object exploration, discrimination and recognition (familiar and novel objectas well as their respective location) [4].…”