The integrative memory model is comprised of six interacting memory systems. There is a relational representation core system central to recollection, and an entity representation core system central to familiarity. These connect to a context representation core system, a self-referential system, and a connectivity hub. The productions of these systems interface with a unitary attribution system reminiscent of the evaluation process in Whittlesea's (Reference Whittlesea and Medin1997) Selective Construction and Preservation of Experiences (SCAPE) account of memory. SCAPE's attribution system is where cued memory traces, stimulus structure, task, and context interact in what Leboe-McGowan (Reference Leboe-McGowan2019) dubbed the “hokey pokey” (in reference to the popular children's participation dance) to yield subjective experiences and reports of recollection or familiarity.
We are thrilled to see a neurocognitive model that incorporates an attributional process informed by Whittlesea's work. And we agree with Bastin et al. that the melding of the core and attributional systems is the most novel aspect of the integrative memory model. However, the “hokey pokey” of multiple systems in the integrative memory model, particularly the inclusion of separate systems for recollection and familiarity, is antithetical to Whittlesea's unitary memory system approach. In Whittlesea's (Reference Whittlesea and Medin1997) words, “Memory is fundamentally very simple. Human performance derives its complexity not from the architecture or processing of memory but from the variety of tasks, stimulus structures, and contexts to which memory is exposed” (p. 260). Here, we side with Whittlesea, and suggest that a drawback of the integrative memory model's complexity is that it may be difficult to determine whether recollection or familiarity arise from productions within the model's core systems or evaluations within the attribution system. Given the alignment of integrative memory model systems with distinct brain architectures, perhaps researchers can leverage brain imaging and connectivity analyses to justify this complexity. Regardless of whether the integrative memory model's complexity proves justifiable, however, we suggest that it currently has at least two major blind-spots: false memories and development. We consider each in turn.
False memories
Memory is a fundamentally reconstructive process; therefore, a truly integrative model of memory must explain how false memories arise. Rates of false memory phenomena, including the misinformation effect, can exceed 50% of participants (e.g., Loftus et al. Reference Loftus, Miller and Burns1978). Similarly, high levels of false memories occur in the Deese–Roediger–McDermott effect, in which people falsely recollect a non-studied word (e.g., sleep) that is the top semantic associate of a list of presented words (e.g., bed, rest, tired, etc.) (Deese Reference Deese1959; Roediger & McDermott Reference Roediger and McDermott1995). Finally, rates of rich false memories, in which people come to believe entire events that never happened, can reach 50% of participants (e.g., Scoboria et al. Reference Scoboria, Wade, Lindsay, Azad, Strange, Ost and Hyman2017). Some of these false memories merely feel familiar. The integrative memory model describes how familiarity-based retrieval can produce correct or false recognition (see Bastin et al.’s Fig. 2). However, a sizable proportion of these and other types of false memories are experienced as recollected. Although the integrative memory model provides a detailed description of recollection-based true memories (see their Fig. 3), it does not currently address the processes underlying recollection-based false memories.
From both theoretical and applied perspectives, it is important to understand how and when false recollection versus false familiarity arise, and how true and false memories differ (Bernstein & Loftus Reference Bernstein and Loftus2009). Instead of tackling false memories, Bastin et al. focus on the model's ability to explain memory impairments, including amnesia, frontal lesions, and especially the various stages of Alzheimer's disease. Damage to the integrative memory model systems may explain deficits in recollection or familiarity, but how does it explain whether someone experiences false memories as recollected versus familiar? Do false recollections arise within the core systems or the attribution system? How could we tell? Perhaps the aforementioned brain imaging and connectivity analyses can be used to answer these questions. We suggest that studying the conditions that predict whether individuals with amnesia, frontal lesions, and Alzheimer's disease will experience a false memory as recollected versus familiar would enhance the model's contribution.
Development
Bastin et al. thoroughly review the neuro-atypical memory literature, and detail how lesions and other neural insults impair recollection and/or familiarity. They suggest that longitudinal study of memory impairments would have great utility. We agree. However, we suggest that the authors have overlooked an important complementary approach – namely, the study of how recollection and familiarity develop and shift across the lifespan. Childhood and old age involve dramatic structural and functional changes to brain and behavior. Therefore, we believe that it would be informative to consider developmental patterns in recollection and familiarity across the lifespan. In the case of false memory, the integrative memory model might be informed by considering how the likelihood of different memory errors shifts in childhood and adulthood. In memory implantation studies, experimenters use suggestive techniques to lead participants to remember having experienced certain event details or entire events that never occurred (see Loftus Reference Loftus2018). From our reading of the lifespan developmental literature on false memory, misinformation-based and rich false memories tend to follow a U-shaped development: These false memories are more frequent in childhood and older adulthood than in younger adulthood (see also Brainerd & Reyna Reference Brainerd and Reyna2005; Frenda et al. Reference Frenda, Nichols and Loftus2011). Conversely, the Deese–Roediger–McDermott illusion increases linearly from childhood to older adulthood (e.g., Brainerd et al. Reference Brainerd, Reyna and Ceci2008; Gallo Reference Gallo2010). How might these different data patterns relate to the core versus attributional systems in the integrative memory model? We welcome Bastin et al.’s insights on how developmental patterns in false memory illusions might constrain or validate their model. Indeed, we feel it would be informative to consider the development of recollection and familiarity processes for both true and false memories.
In sum, Bastin et al. should justify the integrative memory model's substantial complexity by addressing how that complexity contributes to our understanding of (1) different types of false memory phenomena (particularly false recollection), and (2) the development of recollection and familiarity for true and false memories across the lifespan. By incorporating these missing elements, we feel the integrative memory model would be more integrative and thus better live up to its name.
You have
Access
The integrative memory model is comprised of six interacting memory systems. There is a relational representation core system central to recollection, and an entity representation core system central to familiarity. These connect to a context representation core system, a self-referential system, and a connectivity hub. The productions of these systems interface with a unitary attribution system reminiscent of the evaluation process in Whittlesea's (Reference Whittlesea and Medin1997) Selective Construction and Preservation of Experiences (SCAPE) account of memory. SCAPE's attribution system is where cued memory traces, stimulus structure, task, and context interact in what Leboe-McGowan (Reference Leboe-McGowan2019) dubbed the “hokey pokey” (in reference to the popular children's participation dance) to yield subjective experiences and reports of recollection or familiarity.
We are thrilled to see a neurocognitive model that incorporates an attributional process informed by Whittlesea's work. And we agree with Bastin et al. that the melding of the core and attributional systems is the most novel aspect of the integrative memory model. However, the “hokey pokey” of multiple systems in the integrative memory model, particularly the inclusion of separate systems for recollection and familiarity, is antithetical to Whittlesea's unitary memory system approach. In Whittlesea's (Reference Whittlesea and Medin1997) words, “Memory is fundamentally very simple. Human performance derives its complexity not from the architecture or processing of memory but from the variety of tasks, stimulus structures, and contexts to which memory is exposed” (p. 260). Here, we side with Whittlesea, and suggest that a drawback of the integrative memory model's complexity is that it may be difficult to determine whether recollection or familiarity arise from productions within the model's core systems or evaluations within the attribution system. Given the alignment of integrative memory model systems with distinct brain architectures, perhaps researchers can leverage brain imaging and connectivity analyses to justify this complexity. Regardless of whether the integrative memory model's complexity proves justifiable, however, we suggest that it currently has at least two major blind-spots: false memories and development. We consider each in turn.
False memories
Memory is a fundamentally reconstructive process; therefore, a truly integrative model of memory must explain how false memories arise. Rates of false memory phenomena, including the misinformation effect, can exceed 50% of participants (e.g., Loftus et al. Reference Loftus, Miller and Burns1978). Similarly, high levels of false memories occur in the Deese–Roediger–McDermott effect, in which people falsely recollect a non-studied word (e.g., sleep) that is the top semantic associate of a list of presented words (e.g., bed, rest, tired, etc.) (Deese Reference Deese1959; Roediger & McDermott Reference Roediger and McDermott1995). Finally, rates of rich false memories, in which people come to believe entire events that never happened, can reach 50% of participants (e.g., Scoboria et al. Reference Scoboria, Wade, Lindsay, Azad, Strange, Ost and Hyman2017). Some of these false memories merely feel familiar. The integrative memory model describes how familiarity-based retrieval can produce correct or false recognition (see Bastin et al.’s Fig. 2). However, a sizable proportion of these and other types of false memories are experienced as recollected. Although the integrative memory model provides a detailed description of recollection-based true memories (see their Fig. 3), it does not currently address the processes underlying recollection-based false memories.
From both theoretical and applied perspectives, it is important to understand how and when false recollection versus false familiarity arise, and how true and false memories differ (Bernstein & Loftus Reference Bernstein and Loftus2009). Instead of tackling false memories, Bastin et al. focus on the model's ability to explain memory impairments, including amnesia, frontal lesions, and especially the various stages of Alzheimer's disease. Damage to the integrative memory model systems may explain deficits in recollection or familiarity, but how does it explain whether someone experiences false memories as recollected versus familiar? Do false recollections arise within the core systems or the attribution system? How could we tell? Perhaps the aforementioned brain imaging and connectivity analyses can be used to answer these questions. We suggest that studying the conditions that predict whether individuals with amnesia, frontal lesions, and Alzheimer's disease will experience a false memory as recollected versus familiar would enhance the model's contribution.
Development
Bastin et al. thoroughly review the neuro-atypical memory literature, and detail how lesions and other neural insults impair recollection and/or familiarity. They suggest that longitudinal study of memory impairments would have great utility. We agree. However, we suggest that the authors have overlooked an important complementary approach – namely, the study of how recollection and familiarity develop and shift across the lifespan. Childhood and old age involve dramatic structural and functional changes to brain and behavior. Therefore, we believe that it would be informative to consider developmental patterns in recollection and familiarity across the lifespan. In the case of false memory, the integrative memory model might be informed by considering how the likelihood of different memory errors shifts in childhood and adulthood. In memory implantation studies, experimenters use suggestive techniques to lead participants to remember having experienced certain event details or entire events that never occurred (see Loftus Reference Loftus2018). From our reading of the lifespan developmental literature on false memory, misinformation-based and rich false memories tend to follow a U-shaped development: These false memories are more frequent in childhood and older adulthood than in younger adulthood (see also Brainerd & Reyna Reference Brainerd and Reyna2005; Frenda et al. Reference Frenda, Nichols and Loftus2011). Conversely, the Deese–Roediger–McDermott illusion increases linearly from childhood to older adulthood (e.g., Brainerd et al. Reference Brainerd, Reyna and Ceci2008; Gallo Reference Gallo2010). How might these different data patterns relate to the core versus attributional systems in the integrative memory model? We welcome Bastin et al.’s insights on how developmental patterns in false memory illusions might constrain or validate their model. Indeed, we feel it would be informative to consider the development of recollection and familiarity processes for both true and false memories.
In sum, Bastin et al. should justify the integrative memory model's substantial complexity by addressing how that complexity contributes to our understanding of (1) different types of false memory phenomena (particularly false recollection), and (2) the development of recollection and familiarity for true and false memories across the lifespan. By incorporating these missing elements, we feel the integrative memory model would be more integrative and thus better live up to its name.