Abstract
Sleep is shown to facilitate learning on a variety of tasks that require procedural memory. Evidence of replay during sleep of waking experience in rat hippocampus (Ji, 2007) has been proposed as a mechanism of the sleep-dependent learning process in humans (Stickgold, 2007). However, behavioral data in humans that support the notion of improved hippocampal-related memory due to sleep has been inconsistent (Frank, 2007). We asked whether sleep improved hippocampal-dependent implicit memory in a contextual cueing (CC) task (Chun & Jiang, 1998; Chun & Phelps, 1999). Sixty-nine subjects were tested twice on the contextual cueing task (Chun &Jiang, 2003). Subjects searched for a target (T) amongst distractors (L). In session one, half of the item configurations are repeated throughout the session (OLD) and half contain random item configurations (NEW). Session two tested three conditions 1) OLD configurations from session 1, 2) NEW configurations, 3) and a new set of 12 repeated configurations (NEW/OLD). Subjects were assigned to a Nap, Nocturnal Sleep, or Rest condition between the two sessions. We found an overall CC effect for session one (i.e. faster RT for OLD configurations compared with NEW), with no group interaction. The GroupXSessionXCondition ANOVA found no differences between groups, indicating that improvement between sessions was similar in subjects who slept and in those who did not sleep. Post-hoc tests showed no differences in session two between any of the groups in their retention of session one OLD or acquisition of session two NEW/OLD. All groups were also at chance in recognition of the OLD configurations. These results are further evidence of a lack of involvement of sleep in learning involving the hippocampus. Thus, the proposed mechanism of hippocampal replay as a general model for sleep-dependent learning may need to be reconsidered.
Dr. Sara C. Mednick's K01 MH080992-01, Dr Yuhong Jiang's MH071788.