Elsevier

Neurobiology of Aging

Volume 46, October 2016, Pages 76-83
Neurobiology of Aging

Regular article
Age differences in hippocampal activation during gist-based false recognition

https://doi.org/10.1016/j.neurobiolaging.2016.06.014Get rights and content

Abstract

Age-related increases in reliance on gist-based processes can cause increased false recognition. Understanding the neural basis for this increase helps to elucidate a mechanism underlying this vulnerability in memory. We assessed age differences in gist-based false memory by increasing image set size at encoding, thereby increasing the rate of false alarms. False alarms during a recognition test elicited increased hippocampal activity for older adults as compared to younger adults for the small set sizes, whereas the age groups had similar hippocampal activation for items associated with larger set sizes. Interestingly, younger adults had stronger connectivity between the hippocampus and posterior temporal regions relative to older adults during false alarms for items associated with large versus small set sizes. With increased gist, younger adults might rely more on additional processes (e.g., semantic associations) during recognition than older adults. Parametric modulation revealed that younger adults had increased anterior cingulate activity than older adults with decreasing set size, perhaps indicating difficulty in using monitoring processes in error-prone situations.

Introduction

Prior research has shown that older adults have difficulty remembering previous events as well as younger adults (Light, 1991). This age-related difference is likely the result of deficits in recollection for specific contextual details, whereas familiarity-based recognition remains intact with age (e.g., Spencer and Raz, 1995, Yonelinas, 2002). Extant work suggests that encoding engenders 2 types of memory traces: verbatim and gist traces (Brainerd and Reyna, 1990). Verbatim traces contain the distinctive details of an event, whereas gist traces retain the general meaning of the event without specific perceptual details (Brainerd and Reyna, 1990). False memories typically occur when the gist trace is retrieved, and no verbatim trace is accessible. Consistent with this notion, age-related increases in false memories are most robust when old and new information share common semantic or perceptual characteristics (Balota et al., 1999, Kensinger and Schacter, 1999, Koutstaal and Schacter, 1997, Norman and Schacter, 1997).

Although reliance on gist processing has been shown to vary by task demands among younger adults (Loftus et al., 1995, Roediger and McDermott, 1995, Schacter, 1999, Schacter, 2008), aging increases the use of gist, resulting in poor memory in older adults when recollection requires remembering specific details (Brainerd and Reyna, 2002, Tun et al., 1998). Identifying mechanisms underlying age differences in false recognition that result from gist-based processing is important for understanding how to reduce vulnerability to such errors, especially given a shift toward greater reliance on gist-based memory processes with age (Koutstaal and Schacter, 1997). The present study employed neuroimaging to elucidate a neural mechanism for age differences in false memory at retrieval that results from gist-based processing.

Extant neuroimaging research has revealed age differences in activation relative to verbatim versus gist retrieval. Regions including the hippocampus, early visual cortex, lateral parietal cortex, occipitoparietal cortex, and rhinal cortex are involved at retrieval, although their roles depend on the processes required, namely recollection versus familiarity. At retrieval, these regions associated with recollection (e.g., the hippocampus, early visual cortex, and lateral parietal cortex) show age-related deficits in activation (Ally et al., 2008, Daselaar et al., 2006, Duarte et al., 2008). However, regions implicated in familiarity (e.g., rhinal cortex, occipitoparietal cortex) show intact or often enhanced functioning with age (Daselaar et al., 2006, Dennis et al., 2008b, Duarte et al., 2010). This work evidences that older adults typically show significant reductions in true recollection (Bastin and Van der Linden, 2003, Davidson and Glisky, 2002), whereas familiarity processes remain preserved throughout the lifespan (Bastin and Van der Linden, 2003, Naveh-Benjamin, 2000).

In memory, hippocampal activation contributes to binding details during encoding and reconstructing them during retrieval (Yassa et al., 2011). In younger adults, the hippocampus can be differentially activated on a within-subject basis by varying the amount of to-be-remembered information. For example, set size of information at encoding (i.e., manipulating the number of object exemplars at encoding to evoke more gist-based processing) is positively associated with both false alarm rates and increased hippocampal activity at retrieval in younger adults (Gutchess and Schacter, 2012). Increased hippocampal activity for larger set sizes suggests that features shared by target and lure items elicit similar reconstructive processes that ultimately lead to false memories.

Among older adults, the role of the hippocampus in gist processing at retrieval may be responsible for eliciting errors in several ways. First, older adults may have difficulty monitoring retrieval attempts (e.g., Fandakova et al., 2013a, Fandakova et al., 2013b). In this case, the hippocampus may fail to access the correct information, possibly due to a failure in binding the details together (Naveh-Benjamin, 2000). Here, decreased hippocampal activity would reflect deficient processing in which original encoding information is not properly reactivated. However, increased hippocampal engagement may lead to reactivation failure if this engagement corresponds with the number of reactivated features. In this case, older adults may retrieve more information by activating irrelevant features, ultimately eliciting more incorrect bindings.

Second, hippocampal activation may also play a role in false recognition with regard to gist processing through pattern separation. Aging diminishes the capacity to separate new information from similar inputs that could cause interference (pattern separation), leading to an increase in the retrieval of previous information from a partial cue (pattern completion) (Yassa and Stark, 2011). True recollection requires memories to be unique enough from other interfering information, relying on pattern separation (Norman, 2010). The more interference from overlapping events that needs to be overcome, the more separation becomes critical for recollection to occur (Yassa and Stark, 2011). Prior work in both rodents (e.g., Wilson et al., 2006) and humans (e.g., Stark et al., 2010, Yassa et al., 2010) suggests that false memories elicited through deficient pattern separation occurs specifically in the dentate gyrus and is coupled with pattern completion in older adults. In this case, the hippocampus responds to the novelty of information, with decreased activation suggesting impaired pattern separation abilities (Duncan et al., 2012, Fandakova et al., 2013a, Kumaran and Maguire, 2006). With regard to the present work, we hypothesized that decreased hippocampal activation would occur concomitantly with increased false alarms, due to age-related differences in the ability to differentiate target from novel stimuli. This may potentially result from deficits in pattern separation ability.

In the present study, we analyzed younger and older adults' neural activity when recognizing novel exemplars taken from small, medium, or large set sizes of objects drawn from the same category (e.g., bicycles). This manipulation varied the level of gist associated with each category. Younger adults employ reconstructive processes as shown with increases in hippocampal activation during gist-related false recognition (Gutchess and Schacter, 2012). Given declines in hippocampal activation with age (Fandakova et al., 2013a, Naveh-Benjamin, 2000), we anticipated reduced hippocampal activity in younger versus older adults, suggestive of a shift toward gist-based processing, and induced through the use of large versus small set sizes. However, although we expected hippocampal engagement to change with age due to gist, this could be reflected through several patterns of neural activity. Whether hippocampal activation tracks the level of gist to a lesser degree for older versus younger adults, or does not respond to manipulation of gist at all in older adults, is an open question addressed through this work.

Due to age-related differences in the ability to differentiate target from novel stimuli, older adults may instead draw on other processes to help memory performance. One possibility is that given age-related deficits in pattern separation (Duncan et al., 2012, Fandakova et al., 2013a, Kumaran and Maguire, 2006), older adults may recruit regions implicated in semantic processes during retrieval (Dennis et al., 2007, Dennis et al., 2008a). We thereby predicted that older versus younger adults would activate more lateral temporal regions associated with semantic processing for large versus small set sizes. These predictions coalesce with patterns found in prior aging research at retrieval using both verbal (Dennis et al., 2008b) and visual (Koutstaal and Schacter, 1997) stimuli. The present study extends this work by further manipulating gist in a graded fashion for visual stimuli sharing perceptual as well as semantic properties to investigate how gist reliance influences age differences in the neural correlates of false memory at retrieval.

Beyond assessing age differences in activity related to gist-based processing and consequent false memory in isolated regions, we tested how hippocampal connectivity changes with age given different levels of gist contributing to false memories. Age-related decreases in connectivity between the hippocampus and posterior temporal regions underlie memory deficits in aging, and specifically, increases in false memories (Dennis et al., 2008b). Despite reduced connectivity between the hippocampus with posterior regions, however, enhanced hippocampal connectivity with frontal regions at encoding predicts better memory performance for older adults (Daselaar et al., 2006, Dennis et al., 2008a). Hippocampal-frontal cortex connectivity suggests that older adults compensate for deficits in posterior connectivity by relying on frontal top-down modulation (Daselaar et al., 2006, Dennis et al., 2008a, Grady et al., 2003). However, little research has considered age-related functional connectivity changes between the hippocampus and these regions during false recognition for items that vary in gist. For large versus small set sizes, we predicted that younger adults would have greater connectivity relative to older adults from the hippocampus to regions involved in perceptual processes. Although older adults may exhibit increased activation in inferior temporal regions during false recognition (Daselaar et al., 2006), the connectivity between these regions and the hippocampus may be weakened with age (Dennis et al., 2008b). Concomitant with age-related reductions in posterior connectivity, we predicted that older adults would have greater connectivity with frontal regions, a pattern suggested to reflect strategic shifts to support task performance (Davis et al., 2008). Further we predicted that connectivity would be greater for large relative to small set sizes due to the heavy reliance on gist-based processing.

Finally, we conducted a parametric modulation analysis to capitalize on our graded levels of gist, allowing us to assess age-related changes in neural activity corresponding to increasing or decreasing levels of gist that are associated with varying levels of false memory. Beyond the hippocampus, we identified lateral temporal cortex as a candidate region whose activity may be modulated by the extent of gist-based processing, given its involvement in semantic processing. Extant work has shown age differences in reliance on semantic processes with increased activation in the lateral temporal cortex for older adults at retrieval (Davis et al., 2008, Dennis et al., 2008a). We predicted that with increasing set size (i.e., small, medium, and large), older adults would show increased reliance on semantic processes, reflected in increased lateral temporal cortex activation. These analyses can help distinguish the potential roles of the hippocampus and lateral temporal cortex in gist-based memory errors.

Section snippets

Participants

Sixteen younger (mean age = 24.13 years, SD = 4.57; 8 male; age range = 19–33 years) and 16 older (mean age = 71.81 years, SD = 5.58; 7 male; age range = 61–80 years) adults participated. An additional 7 younger adults and 3 older adults were excluded due to failure to respond to large numbers of trials (>40% no responses; 1 younger adult, 1 older adult), failure to follow instructions (1 younger adult), too few false alarms in at least 1 condition (<6 items; 5 younger adults), or excessive

Behavioral data

We assessed age differences in false recognition by entering false alarm rates into a 2 (age: young, old) × 3 (set size: small, medium, large) mixed ANOVA (See Table 1). Older adults had higher false recognition rates (mean = 0.32, SD = 0.10) than younger adults (mean = 0.24, SD = 0.10), F(1,30) = 4.70, p = 0.04, partial η2 = 0.14. A main effect of set size emerged, F(2, 60) = 42.66, p < 0.001, partial η2 = 0.59. Across age, false alarm rate was greater for the large set size (mean = 0.34, SD =

Discussion

The present study investigated the neural underpinnings of how increased use of gist-based processing affects false memory for visual stimuli in aging. We had predicted involvement of the hippocampus, as previous work has suggested that the more closely items are related (e.g., items that share the same verbal label), the more processes involved in false recognition are activated (Koutstaal and Schacter, 1997). This yields a greater need for hippocampal-based reconstruction as the degree of

Conclusions

This research extends prior gist and false memory research by investigating neural age differences in false memories for graded gist information. Our results suggest that hippocampal involvement extends to visual stimuli in regards to false memory and gist. That younger adults exhibited enhanced connectivity with inferior and superior temporal regions relative to older adults at the large versus small set size suggests that younger versus older adults use additional resources (e.g., semantic

Disclosure statement

The authors have no conflicts of interest to disclose.

Acknowledgements

The authors gratefully acknowledge support from the National Institutes of Health, grants NIA R01 AG008441 and NIMH R01 MH60941 (to Daniel L. Schacter). The Athinoula A. Martinos Center for Biomedical Imaging is supported by the National Center for Research Resources (grant P41 RR14075) and by the MIND Institute. The authors also thank Becky Sokal and Rachel Garoff-Eaton for thoughtful contributions and John Ksander for technical assistance.

References (49)

  • M.A. Yassa et al.

    Pattern separation in the hippocampus

    Trends Neurosci.

    (2011)
  • A. Yonelinas

    The nature of recollection and familiarity: a review of 30 years of research

    J. Mem. Lang.

    (2002)
  • D.A. Balota et al.

    Veridical and false memories in healthy older adults and in dementia of the Alzheimer's type

    Cogn. Neuropsychol.

    (1999)
  • C. Bastin et al.

    The contribution of recollection and familiarity to recognition memory: a study of the effects of test format and aging

    Neuropsychology

    (2003)
  • C.J. Brainerd et al.

    Fuzzy-trace theory and false memory

    Curr. Dir. Psychol. Sci.

    (2002)
  • R. Cabeza et al.

    Can medial temporal lobe regions distinguish true from false? an event-related functional MRI study of veridical and illusory recognition memory

    Proc. Natl. Acad. Sci.

    (2001)
  • C.S. Carter et al.

    Anterior cingulate cortex, error detection, and the online monitoring of performance

    Science

    (1998)
  • A.M. Dale

    Optimal experimental design for event-related fMRI

    Hum. Brain Mapp.

    (1999)
  • S.M. Daselaar et al.

    Effects of healthy aging on hippocampal and rhinal memory functions: an event-related fMRI study

    Cereb. Cortex

    (2006)
  • P.S. Davidson et al.

    Neuropsychological correlates of recollection and familiarity in normal aging

    Cogn. Affect. Behav. Neurosci.

    (2002)
  • S.W. Davis et al.

    Que PASA? the posterior-anterior shift in aging

    Cereb. Cortex

    (2008)
  • N.A. Dennis et al.

    Effects of aging on the neural correlates of successful item and source memory encoding

    J. Exp. Psychol. Learn. Mem. Cogn.

    (2008)
  • N.A. Dennis et al.

    Age-related differences in brain activity during true and false memory retrieval

    J. Cogn. Neurosci.

    (2008)
  • A. Duarte et al.

    The effects of aging on the neural correlates of subjective and objective recollection

    Cereb. Cortex

    (2008)
  • Cited by (17)

    • Age-related differences in encoding-retrieval similarity and their relationship to false memory

      2022, Neurobiology of Aging
      Citation Excerpt :

      Results of these studies have identified several critical findings related to perceptual false memories. Specifically, the studies find that older adults tend to underrecruit occipital regions supporting true memories and overrecruit both prefrontal and lateral temporal regions during the execution of false memories (Dennis, Johnson, et al., 2014; Dennis & Turney, 2018; Paige et al., 2016; Webb, Turney, & Dennis, 2016; Webb & Dennis, 2018). Activity in occipital regions, including early visual cortex and lingual gyrus, are posited to reflect sensory signals that are sensitive to the original presentation of previously viewed items (e.g., targets; Dennis et al., 2012; Slotnick & Schacter, 2004; Wheeler & Buckner, 2004).

    • Fornix white matter microstructure differentially predicts false recollection rates in older and younger adults

      2021, Neuropsychologia
      Citation Excerpt :

      Older adults, compared to younger adults, display reduced hippocampal activity associated with true recollection (Daselaar et al., 2006) as well as impaired monitoring during recollection-based retrieval (Buschert et al., 2010). Age-related reductions in prefrontal and hippocampal activation associated with monitoring of item-specific details have also been observed during retrieval of false memories (Dennis et al., 2008, 2014; Fandakova et al., 2014; Paige et al., 2016) and is posited to contribute to increased false memories in older adults. Given the coupling between the PFC and MTL as part of the retrieval process, it stands that the white matter connections among such regions are also a critical component of their functional relationship.

    • The production of false recognition and the associated state of consciousness following encoding in a naturalistic context in aging

      2021, Consciousness and Cognition
      Citation Excerpt :

      Therefore, it is conceivable that in our study, for both age groups, the production of different types of false recognition (critical, perceptually and semantically related) was due to reliance on the shared meaning among study items (gist): items that are consistent with the gist of the lists are considered as familiar and can in turn be associated with false remember judgments, also referred to as “phantom recollections”. The rate of false recognitions in both age groups and the increase in false remember judgments associated with critical items in the older adults are in line with gist-based processing and its preservation in aging (Koutstaal & Schacter, 1997; Paige, Cassidy, Schacter, & Gutchess, 2016; Tun et al., 1998). FTT also assumes a process called “recollection rejection” that reduces the incidence of false alarms at retrieval when the recollection of a studied item is used correctly to reject an associated distractor despite its enhanced familiarity (Brainerd, Reyna, Wright, & Mojardin, 2003; Carneiro et al., 2012).

    • Episodic reconstruction contributes to high-confidence false recognition memories in older adults: Evidence from event-related potentials

      2019, Brain and Cognition
      Citation Excerpt :

      First, we observed higher raw false alarm rates of lures but similar corrected false memories for older adults compared to young adults. It has been suggested that the strength of gist traces increased with the number of exemplars presented at encoding, and accordingly the age differences in false memories were magnified under the larger category size condition (Koutstaal & Schacter, 1997; Koutstaal, Schacter, Galluccio, & Stofer, 1999; Paige, Cassidy, Schacter, & Gutchess, 2016). Therefore, one possible explanation for the nonsignificant age differences in the corrected false memories is that a relatively small category size (four category exemplars) was used in the present study.

    • The influence of perceptual similarity and individual differences on false memories in aging

      2018, Neurobiology of Aging
      Citation Excerpt :

      While behavioral results supported this preposition (i.e., showing an increased rate of false memories as a function of target-lure similarity), the absence of neural modulation in older adults suggests that the 2 factors are not related in the same way they are in young (Turney and Dennis, 2017).1 The absence of neural modulation as a function of gist mirrors recent findings from Paige et al. (2016). Taken together, results may suggest that despite higher false alarms rates in older compared with younger adults,2 there is a threshold above which false memories are not differentially influenced by factors that modulate the relationship between targets and lures (e.g., perceptual relatedness, gist).

    View all citing articles on Scopus
    View full text