The assumptions about relations between sleep and memory have changed considerably since Moruzzi and Magoun (Reference Moruzzi and Magoun1949) proposed that sleep reinstates the function of synapses. About a decade ago, the biological significance of dreams was downplayed by some authors, who called them “spandrels of sleep” (e.g., Flanagan Reference Flanagan2000). Llewellyn's target article confirms a new impetus in sleep research, suggesting that Freud's wish to develop a scientific theory for “the interpretation of dreams” might not be a “lost dream” (Freud Reference Freud1900). By emphasizing elaborative encoding (as opposed to consolidation) and a major role of rapid eye movement (REM) sleep (as opposed to non-REM [NREM] sleep or slow-wave sleep) in the formation of episodic memories, Llewellyn is partly in opposition to other frameworks (Diekelmann et al. Reference Diekelmann, Wilhelm and Born2009; Stickgold & Walker Reference Stickgold and Walker2007). Recently, for instance, Hahn et al. (Reference Hahn, McFarland, Berberich, Sakmann and Mehta2012) demonstrated that entorhinal layer III may be a mediator for memory consolidation during slow-wave sleep. Although Llewellyn reviews an impressive amount of data, she neither elaborates on several established findings nor comments on the association between REM sleep, dreaming, and episodic memory in special populations, such as children.
Young children seem to have at least the same amount of REM sleep as adults (Davis et al. Reference Davis, Parker and Montgomery2004; Siegel Reference Siegel2011). However, developmental data show that the episodic memory system emerges only later in life (Fivush Reference Fivush2011; Willoughby et al. Reference Willoughby, Desrocher, Levine and Rovet2012). Furthermore, motor-perceptual skills – which probably are acquired very early in life – also seem to bear relationships with both NREM and REM sleep (Rauchs et al. Reference Rauchs, Desgranges, Foret and Eustache2005). One question that subsequently arises is why infants (and many mammalian species) have so much REM sleep but do not have episodic memory (Tulving Reference Tulving, Terrace and Metcalfe2005). Children's dream recall was found to increase significantly after age 7, correlating with mental imagery ability and visuospatial skills (Nir & Tononi 2010) rather than other intellectual capacities. These abilities are subserved by neural substrates that comprise extrahippocampal areas.
Llewellyn puts special weight on the relationship between REM-sleep dreaming, hippocampus, and episodic memory. Her focus on the hippocampus is justified, mirroring that of other researchers. However, during the first years of life, the hippocampal formation – one of the crucial brain regions for the formation (or “indexing,” as Llewellyn writes [sect. 4.2.5, para. 5]) of episodic memories – is still immature (Josselyn & Frankland Reference Josselyn and Frankland2012), whereas semantic memories can already be formed via parahippocampal areas (e.g., Vargha-Khadem et al. Reference Vargha-Khadem, Gadian, Watkins, Connelly, Van Paesschen and Mishkin1997). Parahippocampal cortex, which receives a diverse gamut of synthesized sensory-specific in addition to multimodal cortical information (Nieuwenhuys et al. Reference Nieuwenhuys, Voogd and van Hujzen2008), was also shown to be engaged in REM sleep (as Llewellyn also remarks). Activation of the parahippocampal region was furthermore found in relation to place perception (Kanwisher Reference Kanwisher2010) and processing of scenes and landmarks (Piefke et al. Reference Piefke, Weiss, Markowitsch and Fink2005). The amygdala is also activated during REM sleep and plays roles in different phases of episodic mnemonic memory processing (encoding, consolidation, and retrieval) (Ally et al. Reference Ally, Hussey and Donahue2012; Markowitsch & Staniloiu Reference Markowitsch and Staniloiu2011a).
Paraphrasing Llewellyn, we ought to say that the term episodic memory is rarely precisely defined before it is used. This is also the case in target article. The various ways the term is currently employed in the literature constitute a source of confusion for empirical data interpretation. Llewellyn herself points to data “difficult to reconcile” (sect 7, para. 2), implying that the relation between REM sleep and episodic memory may depend on the conceptualization of episodic memory and testing paradigms employed. Llewellyn's definition of episodic memory – “enduring memories of personally significant past events with contextual place and time underpinnings (Squire Reference Squire1987; Tulving Reference Tulving1983; Reference Tulving, Squire, Weinberger and Squire1991)” (sect. 1.1, para. 2) – is itself outdated (Markowitsch & Staniloiu Reference Markowitsch and Staniloiu2011b; Reference Markowitsch and Staniloiu2012; Tulving Reference Tulving, Terrace and Metcalfe2005). As Szpunar and McDermott (Reference Szpunar, McDermott, Sweat, Menzel and Sweat2008) remark, episodic memory has been an “evolving concept.” Tulving, who coined the construct episodic memory, used to emphasize the refinements in its conceptualization by giving a date for his most recent definition (e.g., Schacter & Tulving Reference Schacter, Tulving, Schacter and Tulving1994). Whereas a few decades ago the term episodic could be applied to describe memory for laboratory stimuli with a specific embedding in time and place, currently the episodic memory system is viewed as equivalent to the episodic autobiographical memory system. In 2005, Tulving's definition of episodic memory comprised 17 lines, ending with this sentence: “The essence of episodic memory lies in the conjunction of three concepts – self, autonoetic awareness, and subjective time” (p. 9). This definition deviates considerably from previous ones and from those still in use in some literature – for example, “What–Where–When” tasks (sect. 7, para. 2), to which Llewellyn refers with respect to the findings of Rauchs et al. (Reference Rauchs, Bertran, Guillery-Girard, Desgranges, Kerrouche, Denise, Foret and Eustache2004). It has been observed that specific details, pertaining to “what, where, when,” may be produced by some individuals without true first-person reexperiencing (Markowitsch et al. Reference Markowitsch, Fink, Thöne, Kessler and Heiss1997) that is essential for strictly episodic memories (events) (Eustache & Desgranges Reference Eustache and Desgranges2008; Viard et al. Reference Viard, Desgranges, Eustache and Piolino2012) . On the other hand, there are individuals who judge their recalled material with a high confidence as representing true reexperiencing while delivering few episodic details (Levine et al. Reference Levine, Svoboda, Turner, Mandic and Mackey2009). This may reflect impoverished narrative abilities or that having several similar episodic events may reduce the memory strength for them, leading to semanticization (Cermak Reference Cermak, Squire and Butters1984).
We concur with Llewellyn's reservation about using Verbal Paired Associates tests to uncover the relationship between REM sleep and episodic memory because they do not match the current understanding of episodic autobiographical memory. Furthermore, LePort and colleagues (Reference LePort, Mattfeld, Dickinson-Anson, Fallon, Stark, Kruggel, Cahill and McGaugh2012) found that individuals with highly superior autobiographical memory did not perform significantly better on the Verbal Paired Associates test than did controls. Since 2005, several authors have increased the sophistication of their testing paradigms for episodic autobiographical memory (e.g., Levine et al. Reference Levine, Svoboda, Turner, Mandic and Mackey2009). It seems therefore plausible that by employing testing paradigms that tap into the more recent definition by Tulving (Reference Tulving, Terrace and Metcalfe2005), the contribution of slow-wave sleep to episodic memory will be viewed from a different perspective.
In conclusion, the link between REM sleep and strict episodic memory (as defined currently) is far from being unequivocal. Also, the contribution of NREM sleep to strict episodic(-autobiographical) memory needs further exploration. How different stages of sleep support the likely interactive processing within different memory systems remains a topic for future research (Dew & Cabeza Reference Dew and Cabeza2011).
The assumptions about relations between sleep and memory have changed considerably since Moruzzi and Magoun (Reference Moruzzi and Magoun1949) proposed that sleep reinstates the function of synapses. About a decade ago, the biological significance of dreams was downplayed by some authors, who called them “spandrels of sleep” (e.g., Flanagan Reference Flanagan2000). Llewellyn's target article confirms a new impetus in sleep research, suggesting that Freud's wish to develop a scientific theory for “the interpretation of dreams” might not be a “lost dream” (Freud Reference Freud1900). By emphasizing elaborative encoding (as opposed to consolidation) and a major role of rapid eye movement (REM) sleep (as opposed to non-REM [NREM] sleep or slow-wave sleep) in the formation of episodic memories, Llewellyn is partly in opposition to other frameworks (Diekelmann et al. Reference Diekelmann, Wilhelm and Born2009; Stickgold & Walker Reference Stickgold and Walker2007). Recently, for instance, Hahn et al. (Reference Hahn, McFarland, Berberich, Sakmann and Mehta2012) demonstrated that entorhinal layer III may be a mediator for memory consolidation during slow-wave sleep. Although Llewellyn reviews an impressive amount of data, she neither elaborates on several established findings nor comments on the association between REM sleep, dreaming, and episodic memory in special populations, such as children.
Young children seem to have at least the same amount of REM sleep as adults (Davis et al. Reference Davis, Parker and Montgomery2004; Siegel Reference Siegel2011). However, developmental data show that the episodic memory system emerges only later in life (Fivush Reference Fivush2011; Willoughby et al. Reference Willoughby, Desrocher, Levine and Rovet2012). Furthermore, motor-perceptual skills – which probably are acquired very early in life – also seem to bear relationships with both NREM and REM sleep (Rauchs et al. Reference Rauchs, Desgranges, Foret and Eustache2005). One question that subsequently arises is why infants (and many mammalian species) have so much REM sleep but do not have episodic memory (Tulving Reference Tulving, Terrace and Metcalfe2005). Children's dream recall was found to increase significantly after age 7, correlating with mental imagery ability and visuospatial skills (Nir & Tononi 2010) rather than other intellectual capacities. These abilities are subserved by neural substrates that comprise extrahippocampal areas.
Llewellyn puts special weight on the relationship between REM-sleep dreaming, hippocampus, and episodic memory. Her focus on the hippocampus is justified, mirroring that of other researchers. However, during the first years of life, the hippocampal formation – one of the crucial brain regions for the formation (or “indexing,” as Llewellyn writes [sect. 4.2.5, para. 5]) of episodic memories – is still immature (Josselyn & Frankland Reference Josselyn and Frankland2012), whereas semantic memories can already be formed via parahippocampal areas (e.g., Vargha-Khadem et al. Reference Vargha-Khadem, Gadian, Watkins, Connelly, Van Paesschen and Mishkin1997). Parahippocampal cortex, which receives a diverse gamut of synthesized sensory-specific in addition to multimodal cortical information (Nieuwenhuys et al. Reference Nieuwenhuys, Voogd and van Hujzen2008), was also shown to be engaged in REM sleep (as Llewellyn also remarks). Activation of the parahippocampal region was furthermore found in relation to place perception (Kanwisher Reference Kanwisher2010) and processing of scenes and landmarks (Piefke et al. Reference Piefke, Weiss, Markowitsch and Fink2005). The amygdala is also activated during REM sleep and plays roles in different phases of episodic mnemonic memory processing (encoding, consolidation, and retrieval) (Ally et al. Reference Ally, Hussey and Donahue2012; Markowitsch & Staniloiu Reference Markowitsch and Staniloiu2011a).
Paraphrasing Llewellyn, we ought to say that the term episodic memory is rarely precisely defined before it is used. This is also the case in target article. The various ways the term is currently employed in the literature constitute a source of confusion for empirical data interpretation. Llewellyn herself points to data “difficult to reconcile” (sect 7, para. 2), implying that the relation between REM sleep and episodic memory may depend on the conceptualization of episodic memory and testing paradigms employed. Llewellyn's definition of episodic memory – “enduring memories of personally significant past events with contextual place and time underpinnings (Squire Reference Squire1987; Tulving Reference Tulving1983; Reference Tulving, Squire, Weinberger and Squire1991)” (sect. 1.1, para. 2) – is itself outdated (Markowitsch & Staniloiu Reference Markowitsch and Staniloiu2011b; Reference Markowitsch and Staniloiu2012; Tulving Reference Tulving, Terrace and Metcalfe2005). As Szpunar and McDermott (Reference Szpunar, McDermott, Sweat, Menzel and Sweat2008) remark, episodic memory has been an “evolving concept.” Tulving, who coined the construct episodic memory, used to emphasize the refinements in its conceptualization by giving a date for his most recent definition (e.g., Schacter & Tulving Reference Schacter, Tulving, Schacter and Tulving1994). Whereas a few decades ago the term episodic could be applied to describe memory for laboratory stimuli with a specific embedding in time and place, currently the episodic memory system is viewed as equivalent to the episodic autobiographical memory system. In 2005, Tulving's definition of episodic memory comprised 17 lines, ending with this sentence: “The essence of episodic memory lies in the conjunction of three concepts – self, autonoetic awareness, and subjective time” (p. 9). This definition deviates considerably from previous ones and from those still in use in some literature – for example, “What–Where–When” tasks (sect. 7, para. 2), to which Llewellyn refers with respect to the findings of Rauchs et al. (Reference Rauchs, Bertran, Guillery-Girard, Desgranges, Kerrouche, Denise, Foret and Eustache2004). It has been observed that specific details, pertaining to “what, where, when,” may be produced by some individuals without true first-person reexperiencing (Markowitsch et al. Reference Markowitsch, Fink, Thöne, Kessler and Heiss1997) that is essential for strictly episodic memories (events) (Eustache & Desgranges Reference Eustache and Desgranges2008; Viard et al. Reference Viard, Desgranges, Eustache and Piolino2012) . On the other hand, there are individuals who judge their recalled material with a high confidence as representing true reexperiencing while delivering few episodic details (Levine et al. Reference Levine, Svoboda, Turner, Mandic and Mackey2009). This may reflect impoverished narrative abilities or that having several similar episodic events may reduce the memory strength for them, leading to semanticization (Cermak Reference Cermak, Squire and Butters1984).
We concur with Llewellyn's reservation about using Verbal Paired Associates tests to uncover the relationship between REM sleep and episodic memory because they do not match the current understanding of episodic autobiographical memory. Furthermore, LePort and colleagues (Reference LePort, Mattfeld, Dickinson-Anson, Fallon, Stark, Kruggel, Cahill and McGaugh2012) found that individuals with highly superior autobiographical memory did not perform significantly better on the Verbal Paired Associates test than did controls. Since 2005, several authors have increased the sophistication of their testing paradigms for episodic autobiographical memory (e.g., Levine et al. Reference Levine, Svoboda, Turner, Mandic and Mackey2009). It seems therefore plausible that by employing testing paradigms that tap into the more recent definition by Tulving (Reference Tulving, Terrace and Metcalfe2005), the contribution of slow-wave sleep to episodic memory will be viewed from a different perspective.
In conclusion, the link between REM sleep and strict episodic memory (as defined currently) is far from being unequivocal. Also, the contribution of NREM sleep to strict episodic(-autobiographical) memory needs further exploration. How different stages of sleep support the likely interactive processing within different memory systems remains a topic for future research (Dew & Cabeza Reference Dew and Cabeza2011).