Archives of Clinical Neuropsychology 36 (2021) 1404–1425

Prospective Metamemory Monitoring of Episodic Visual Memory

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

in Community-Dwelling Older Adults with Subjective
Cognitive Decline and Mild Cognitive Impairment
Susan Y. Chi1,2,3,4, *, Elizabeth F. Chua2,5 , Dustin W. Kieschnick3 , Laura A. Rabin1,2,5,6
1
Queens College, City University of New York, Queens, NY, USA
2
Graduate Center, City University of New York, New York, NY, USA
3
University of California at San Francisco, Weill Institute for Neurosciences, San Francisco, CA, USA
4
Framework Associates, Santa Monica, CA, USA
5
Brooklyn College, City University of New York, Brooklyn, NY, USA
6
Albert Einstein College of Medicine, Einstein Aging Study, Bronx, NY, USA
*Corresponding author at: Psychology Department, Queens College, City University of New York, 65-30 Kissena Boulevard Flushing, NY 11367,
USA; Framework Associates, 7720 Painter Avenue Unit C, Whittier, CA 90602 USA. Tel.: (323) 999-1395; Fax: (855) 568-6438.
E-mail address: [email protected], [email protected] (Susan Y. Chi).
Editorial Decision 30 January 2021; Accepted 24 January 2021

Abstract
Objective: Metamemory tasks have been utilized to investigate anosognosia in older adults with dementia, though previous
research has not systematically compared memory self-awareness in prodromal dementia groups. This represents an important
oversight given that remedial and interventional efforts may be most beneficial before individuals’ transition to clinical dementia.
We examine differences in memory self-awareness and memory self-monitoring between cognitively healthy elderly controls
and prodromal dementia groups.
Methods: Participants with subjective cognitive decline despite intact objective neuropsychological functioning (SCD; n = 82),
amnestic mild cognitive impairment (aMCI; n = 18), nonamnestic mild cognitive impairment (naMCI; n = 38), and normal cog-
nitive functioning (HC; n = 120) were recruited from the Einstein Aging Study for a cross-sectional study. Participants completed
an experimental visual memory-based global metamemory prediction task and subjective assessments of memory/cognition and
self-awareness.
Results: While, relative to HC, memory self-awareness and memory self-monitoring were preserved for delayed memory
performance in SCD and aMCI, these processes were impaired in naMCI. Furthermore, results suggest that poor metamemory
accuracy captured by our experimental task can be generalized to everyday memory problems.
Conclusions: Within the framework of the Cognitive Awareness Model, our findings provide preliminary evidence that poor
memory self-awareness/self-monitoring in naMCI may reflect an executive or primary anosognosia, with implications for
tailored rehabilitative interventions.
Keywords: Mild cognitive impairment; Meta cognition; Alzheimer’s disease; Dementia; Rehabilitation; Executive functions

Introduction

The ability to accurately self-assess one’s own memory functioning has been shown to be vulnerable to the neuropathological
changes associated with Alzheimer’s disease (AD) (Brandt, Carvalho, Belfort, & Dourado, 2018; Morris & Mograbi, 2013). Poor
self-awareness of memory ability can be conceived as a problem with metamemory—broadly defined as knowledge about one’s
own memory functioning (Nelson & Narens, 1990). Metamemory is supported by monitoring and control mechanisms that,

© The Author(s) 2021. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: [email protected]
https://doi.org/10.1093/arclin/acab008 Advance Access publication 23 April 2021
 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425 1405

respectively, assess the status of ongoing memory performance and direct behavior to optimize memory functioning (Nelson
& Narens, 1990). Impairment of metamemory functioning, including poor awareness of memory/cognitive deficits, has been
linked to objective functional difficulties (Steward, Bull, Kennedy, Crowe, & Wadley, 2019), as well as poorer utilization of
compensatory strategies (Schmitter-Edgecombe & Seelye, 2011) and cognitive rehabilitation outcomes (Clare, Wilson, Carter,
Roth, & Hodges, 2004). Taken together, the inclusion of metamemory assessments in diagnostic assessments of AD and other
neurocognitive disease conditions has been recommended (Brandt et al., 2018; Chapman, Colvin, & Cosentino, 2020; Morris &
Mograbi, 2013; Ryals, O’Neil, Mesulam, Weintraub, & Voss, 2018). Additionally, recent research has focused on investigating
the translational utility of experimental metamemory tasks in clinical settings, especially with regard to clarifying diagnoses and

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

guiding therapeutic strategies for older adults with various levels of disordered awareness (Chapman et al., 2020).
Individuals in the earliest stages of a neurocognitive disease process (i.e., before progression to AD/dementia) can potentially
benefit the most from diagnostic evaluation and cognitive interventions (Smart et al., 2017; Sohlberg & Mateer, 2001). Mild
cognitive impairment (MCI) is generally recognized as a transition stage between normal cognitive functioning and clinical
dementia (Osuna et al., 2019; Winblad et al., 2004). While the amnestic MCI (aMCI) subtype is characterized by a primary
impairment in memory and most likely represents prodromal AD, the nonamnestic MCI (naMCI) subtype is characterized by
impairments in nonmemory cognitive domain(s) and is associated with progression to non-AD dementias (Ferman et al., 2013;
Petersen, 2004, 2011; Petersen et al., 2001). Furthermore, subjective cognitive decline (SCD) in older adults—characterized
by a subjective sense that one’s cognition has declined despite intact performance on objective cognitive testing and preserved
daily functioning (Jessen et al., 2014; Molinuevo et al., 2017)—is a relatively new diagnostic entity increasingly recognized as
representing one of the earliest preclinical stages of AD and other dementias (Rabin, Smart, & Amariglio, 2017). Individuals
with SCD are thought to be detecting changes in their own cognition before the detection threshold of standard cognitive testing
(Jessen et al., 2014; Rabin et al., 2017). In fact, SCD may precede the MCI stage by 10–15 years (Reisberg et al., 2008) and
has been linked to biomarkers and neurological markers associated with AD (Amariglio et al., 2015; Jessen et al., 2006; Osuna
et al., 2019; Perrotin et al., 2015; Saykin et al., 2006; Striepens et al., 2010) and other dementias (Beckett et al., 2015; Jessen
et al., 2014). Therefore, SCD is thought to reflect prodromal forms of different MCI subtypes (Beckett et al., 2015; Jessen, 2014;
Osuna et al., 2019; Reisberg et al., 2008; Reisberg & Gauthier, 2008).
In the MCI literature, several studies utilizing a judgment of learning (JOL) paradigm, which assesses online prospective
metamemory judgments made during encoding, have reported intact metamemory functioning in aMCI (Akhtar, Moulin, &
Bowie, 2006; Ryals et al., 2018; Seelye, Schmitter-Edgecombe, & Flores, 2010). Although only one of these studies (e.g., Seelye
et al., 2010) have included both aMCI and naMCI groups, the findings offered some preliminary support that JOL tasks may
be useful in differentiating MCI subtypes. The researchers used a global prediction paradigm (which included a JOL measure),
where participants made predictions about their total recall performance on each trial of a verbal list-learning task both before
and after learning. Metamemory accuracy is typically calculated using measures of absolute accuracy (i.e., calibration scores)
and/or relative accuracy (i.e., correlational or sensitivity scores), which, respectively, assess the difference/precision and/or the
relationship between subjective metamemory judgments and objective performance on a criterion task (Schraw, 2009). Seelye
et al. (2010) used calibration scores for “pre-experience” predictions (i.e., predictions made prior to learning) to measure memory
self-awareness and calibration scores for “post-experience” predictions (i.e., predictions made after learning, representing global
JOLs) to measure memory self-monitoring. All groups demonstrated significantly better calibration at post- compared with pre-
experience, suggesting that, similar to HC, aMCI and naMCI participants were able to utilize experience from the ongoing
task to accurately update memory self-knowledge (i.e., showing intact “prediction upgrading”; Devolder, Brigham, & Pressley,
1990). However, results also showed that naMCI participants were significantly more underconfident (i.e., demonstrating poorer
calibration) in their delayed recall performance compared with aMCI and HC both at pre- and post-experience, suggesting that
memory self-awareness and self-monitoring for delayed memory could be differentially impaired in naMCI.
Given that past research has established a critical role of the frontal cortex but not the temporal lobes for JOLs (see Chua,
Pergolizzi, & Weintraub, 2014), it is possible that executive functions/frontal systems impairment in naMCI individuals (e.g.,
who typically present with a dysexecutive profile compared with those with aMCI) could have negatively impacted their self-
awareness and self-monitoring abilities. In fact, a recent study from our laboratory (Chi, Chua, Kieschnick & Rabin, 2020),
utilizing a retrospective confidence judgment paradigm, also showed that metamemory monitoring accuracy was impaired
in naMCI (but not aMCI) relative to HC and that this selective metamemory deficit was potentially attributable to poor
performance monitoring. Although Seelye et al. (2010) concluded that metamemory accuracy was preserved in naMCI because
these participants also demonstrated intact prediction upgrading and relative metamemory accuracy, we contend that significant
underconfidence (or poor calibration in general) in those with naMCI may represent a behavioral marker of early decline in
memory/cognitive awareness, as well as reflect a possible target for therapeutic intervention.
In the SCD literature, metamemory is also highly relevant for both diagnosis and treatment. The prevailing conceptualization
of SCD as a marker of preclinical MCI/dementia is that those with SCD are likely able to self-perceive and compensate for
1406 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425

subtle cognitive changes that are not yet detectable on standardized testing during this early disease stage and it is only after a
critical point in the disease process, when the level of cognitive decline associated with the underlying disease burden surpasses
these individuals’ ability to compensate, that objective cognitive impairment and the onset of MCI becomes apparent (e.g., see
Jessen et al., 2014 and Rabin et al., 2017). Therefore, a key diagnostic assumption is that individuals with SCD are capable of
making accurate, global self-assessments about their cognitive and memory functioning. Although this assumption is supported
by a recent study conducted by our laboratory that showed comparable performance between SCD and HC participants on
online measures of absolute and relative metamemory accuracy, SCD has also been attributed to other factors, including poor
metacognition and depression (Buckley, Laming, Chen, Crole, & Hester, 2016; Chapman et al., 2020; Metternich, Schmidtke,

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

& Hüll, 2009; Reid & MacLullich, 2006).
The Cognitive Awareness Model (CAM; Agnew & Morris, 1998; Hannesdottir & Morris, 2007; Morris & Mograbi, 2013),
originally developed to understand anosognosia in AD, may also be instructive in understanding memory/cognitive awareness in
MCI and potentially SCD. The CAM theorizes that accurate metacognitive output relies on attentional, executive, and memory
systems (Morris & Mograbi, 2013; Piras, Piras, Orfei, Caltagirone, & Spalletta, 2016; Steward et al., 2019). The CAM also
differentiates between unawareness that is secondary to memory (i.e., mnemonic anosognosia) or executive dysfunction (i.e.,
executive anosognosia), and primary unawareness (i.e., primary anosognosia), which may arise due to disruptions in connectivity
between multiple brain areas (Hannesdottir & Morris, 2007; Morris & Mograbi, 2013; Piras et al., 2016). In the latter scenario,
impaired awareness is likely when measured explicitly, but awareness of memory/cognitive failures may still emerge through
intact implicit systems and manifest as emotional reactions (see Morris & Mograbi, 2013). Central to the CAM is the idea that
self-knowledge is updated when a “mismatch” is detected between incoming and existing information about self-ability. Failure
to consolidate information about changes in current functioning (e.g., due to an mnemonic anosognosia), impairment in error
detection and/or performance monitoring (e.g., due to an executive anosognosia), or a lack of conscious awareness of memory
failures (e.g., due to a primary anosognosia) could result in possessing inaccurate self-knowledge about one’s cognitive abilities
over time (Hannesdottir & Morris, 2007; Morris & Mograbi, 2013). Moreover, according to the CAM, unawareness may differ
based on MCI subtype (especially, amnestic vs. frontal/dysexecutive forms; Morris & Mograbi, 2013; Piras et al., 2016) and
increase with greater cognitive impairment (Wolfsgruber et al., 2014).
In the current study, we used a global metamemory prediction paradigm (which included a JOL measure) previously
applied in MCI and AD research (Schmitter-Edgecombe & Seelye, 2011; Seelye et al., 2010). We examined differences in
memory self-awareness, as measured by “pre-experience” prediction accuracy, and memory self-monitoring, as measured
by “post-experience” prediction accuracy (i.e., JOL accuracy), in three predementia groups—aMCI, naMCI, and SCD—and
healthy elderly controls. Specifically, we embedded metamemory judgment queries into the Brief Visuospatial Memory Test-
Revised (BVMT-R) (Benedict, Schretlen, Groninger, Dobraski, & Shpritz, 1996) to evaluate group differences in metamemory
monitoring of visual memory. Given that past research has established a critical role of the frontal cortex but not the temporal
lobes for JOLs (see Chua et al., 2014), a JOL paradigm is particularly suitable for differentiating metamemory functioning in
aMCI and naMCI. Moreover, there is some support that a JOL task based on visual (as opposed to verbal) memory, which would
target frontal and right hemispheric regions, could be more sensitive in detecting specific monitoring and awareness deficits in
naMCI. For example, lesion studies have shown impaired global JOL accuracy in patients with right frontal lesions relative
to controls and those with left frontal and/or right posterior lesions (Vilkki, Servo, & Surma-aho, 1998; Vilkki, Surma-aho, &
Servo, 1999). Studies in early AD have also reported a link between disordered awareness and the right hemisphere (Cosentino
et al., 2007), particularly impairment of the right insula (Cosentino et al., 2015). Finally, a recent metamemory study in aMCI
(Ryals et al., 2018) reported that both HC and aMCI participants demonstrated overconfidence on a visual (but not verbal)
memory global JOL task, suggesting that an overconfidence bias could be associated with right hemispheric mechanisms. To
our knowledge, this is the first study to use a visual-memory-based JOL task to evaluate metamemory functioning in both MCI
subgroups and SCD.
We hypothesized that metamemory deficits in aMCI and naMCI can arise as a function of their primary neuropsychological
impairments, which could interfere with different mechanisms in the brain’s cognitive awareness system. Given that temporal
lobe integrity is not critical for JOL accuracy (Chua et al., 2014), we predicted that, relative to HC, aMCI participants would
demonstrate comparable global JOL accuracy on a visual memory task, indicating intact memory self-monitoring. In contrast,
given the known sensitivity of JOL to frontal systems, we predicted that, relative to HC, naMCI participants would demonstrate
significantly worse global JOL accuracy on a visual memory task, indicating impaired memory self-monitoring. Also, together
with the presumption that cognitive decline is likely to be very subtle early in the disease process and preliminary evidence of
intact metamemory accuracy in SCD from a previous study (e.g., Chi et al., 2020), we predicted that global JOL accuracy should
be intact in SCD participants, indicating intact memory self-monitoring. Finally, a secondary goal was to explore the relationship
between global JOL measures from our experimental task and self-report measures related to everyday memory/cognitive
problems, which could shed light on the generalizability of laboratory-based measures to everyday experience.
 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425 1407

Method

Participants

Participants were recruited from the Einstein Aging Study (EAS), which enrolls ethnically and social-economically diverse
community-dwelling individuals who reside in the Bronx, NY. EAS participants are recruited through systematic sampling from
voter registration and Medicare lists (Katz et al., 2012; Lipton et al., 2003), with the following exclusion criteria: age < 70 years,
active psychiatric symptomatology and/or visual/auditory impairments that would interfere with neuropsychological testing,

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

institutionalized, nonambulatory, and non-English-speaking. We did not invite EAS participants with clinical depression or
dementia. The protocol was approved by the institutional review boards of the City University of New York and Albert Einstein
College of Medicine, and all participants provided written informed consent. We provided transportation, lunch, and $25 for
participation.
Participant classification was carried out using a previously established psychometric approach (see Chi et al., 2014; Chi
et al., 2020; Rabin, Wang, Katz, & Lipton, 2014; and refer to Appendix A for specific information about classification
procedures). Briefly, we first established norms for 13 neuropsychological tests for 411 EAS participants who were dementia
free for 3 years (i.e. the “robust sample”; these individuals were not participants in the current study). Next, consistent with
a previous study of cognitively normal older adults using a similar set of neuropsychological tests (Holtzer, Verghese, Xue,
& Lipton, 2006), a principal component analysis yielded three cognitive factors, memory, global/verbal, executive/processing
speed. For participants in the current study, cognitive domain scores were calculated as the average of the Z scores of the
neuropsychological tests within each cognitive factor using means and SDs from the robust sample, stratified by age group
(70–79 and 80 and above).
MCI was classified in participants whose cognitive domain scores were more than 1 SD below the mean of the robust sample
on one or more cognitive factors and who endorsed at least one cognitive complaint on EAS self-report measures. MCI was
further subdivided into aMCI, for participants whose cognitive factor Z scores were below 1 SD on the memory or memory plus
executive/processing speed and/or global domains, and into naMCI, for participants whose cognitive factor Z scores were below
1 SD on the executive and/or global domains.
SCD was classified in cognitively intact participants (i.e., cognitive factor Z scores for all three domains did not fall >1 SD
below the mean of the robust sample) who also exceeded an optimal cut point for self- and/or informant concerns. Subjective
cognitive items that have demonstrated reliability and predictive validity for dementia from previous research (Rabin et al., 2012)
were used to derive a self- and/or informant-concern score that was based on the proportion of positive responses. Subsequently,
we derived optimal cut points for self- and informant-concern scores utilizing a receiver operating characteristic (ROC) analysis,
stratified by young-old (age 70–79) and old-old (age 80 and above) groups, which used the robust sample and was based on
the cross-sectional association between the self or informant concern and MCI (Rabin, Wang, et al., 2014). Thus, we chose to
set cut points for “significant” subjective cognitive concerns that related to the predictive value for subsequent MCI (as SCD is
considered to be a pre-MCI condition). Results showed that the optimal cut for the self-concern scores was 12.5% and 22.2% of
subjective cognitive items endorsed for the younger and older groups, respectively. The optimal cut for the informant-concern
score was 21.0% and 10.0% of subjective cognitive items endorsed for the younger and older groups, respectively. Participants
who met one or both optimal cuts for their age group on subjective cognition, but who were neuropsychologically intact, received
a classification of SCD.
HC were classified in participants whose cognitive factor Z scores for all three domains did not fall considerably lower (>1
SD) than the mean of the robust sample and who did not exceed the optimal cut point for self- and/or informant complaints.

Procedure

As noted above, this study was part of a larger longitudinal study of cognitive aging. Participants were first assessed during
their annual EAS visit, which included neuropsychological and neurological examinations (see Katz et al., 2012 for details);
∼2 weeks later, they completed a second assessment session that included our visual memory global prediction task, as well as
other objective and subjective assessments.
For the metamemory task, we embedded queries to elicit pre- and post-experience predictions about prospective memory
performance into the standard administration of the BVMT-R (Benedict et al., 1996) to capture memory monitoring and episodic
visual memory performance within a single paradigm. As per the standardized task instructions, participants were informed that
they would have 10 s to study “six geometric figures” presented on a stimulus sheet (BVMT-R, Form 1), after which they
would “draw each figure exactly as it appeared and in its correct location on the page.” Prior to learning the stimuli, participants
made predictions about their future performance for immediate and delayed memory (see Appendix B for exact instructions).
1408 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425

Their responses to these queries were recorded and represented their “pre-experience predictions” for immediate recall (Learning
Trials 1, 2, 3) and delayed recall (Delayed Recall Trial). Then, participants continued with the standardized administration of the
BVMT-R to complete the three learning trials. Specifically, the stimulus sheet included six simple geometric designs arranged in
a 2 × 3 matrix. On each learning trial, participants studied the stimulus sheet for 10 s and then freely recalled as many figures as
they could by drawing them from memory. Following completion of all three learning trials, participants were queried to make a
global JOL about their future performance on the delayed recall trial: “If I ask you about the figures later, how many do you think
you will remember?” Their response was recorded and represented a global JOL rating, an assessment of how much information
one feels has been acquired subsequent to a period of learning. Participants then completed questionnaires for ∼25 min with

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

items related to self-perceptions about cognition and mood, including the Comprehensive Assessment of Prospective Memory,
Section B (Chau, Lee, Fleming, Roche, & Shum, 2007), Cognitive Change Index (Rattanabannakit et al., 2016), Profile of
Mood States (Terry, Lane, & Fogarty, 2003), Social Support Questionnaire (Sarason, Levine, Basham, & Sarason, 1983), and
Satisfaction With Life Scale (Diener, Emmons, Larsen, & Griffin, 1985).
Following the delay period, participants continued with the standardized administration of the BVMT-R to complete the
delayed recall trial. Specifically, they were asked to reproduce all the figures that they could remember in their correct locations
on the page. Then, the experimenter positioned the BVMT-R Recognition Stimulus Booklet in front of participants and explained
that more figures would be shown, one at a time (e.g., figures included 6 targets and 6 distractors). Participants were instructed
to respond yes/no to indicate whether each figure was included on the original stimulus sheet. Before exposure to the recognition
stimulus, participants were queried to make an estimate about their recognition performance. Their response was recorded and
represented a global recognition estimate, as it required participants to estimate their overall recognition performance as opposed
to making a judgment about whether they will recognize each specific target. Finally, participants were exposed to the recognition
stimulus and their yes/no responses were recorded.

Measures

Experimental Task Measures. Objective memory performance: episodic memory scores—BVMT-R Standardized measures of
visual memory performance based on the BVMT-R protocol were calculated for immediate free recall/learning trials (i.e., Trials
1, 2, and 3) and delayed memory trials, including free delayed recall (Delayed Recall Trial) and recognition (for which we
used two scores, Recognition Hits and Recognition Discrimination Index). For all recall trials, participants received 1 point for a
correctly recalled figure and 1 point for a figure that was drawn in the correct location. Therefore, the score for each figure ranged
from 0 to 2 (where 0 indicated neither correct recall nor correct location, 1 indicated either correct recall or correct location, 2
indicated both correct recall and location), and the total score for each trial (sum of all figure scores) ranged from 0 to 12 (where
0 indicated no points were obtained for any of the 6 figures and 12 indicated that the maximum points were obtained for all 6
figures). The raw score for BVMT-R-Delayed Recall equaled the total score for the Delayed Recall Trial, ranging from 0 to 12.
The raw score for BVMT-R-Recognition Hits equaled the sum of correctly recognized target figures, ranging from 0 to 6. The
raw score for the BVMT-R-Recognition Discrimination Index was calculated by subtracting the number of Recognition False
Alarms from the Recognition Hits score, where Recognition False Alarms equaled the number of distractor figures incorrectly
recognized as target figures. Thus, the range for the Recognition Discrimination Index score was −6 to 6, with 6 representing
normal discrimination and −6 representing very poor discrimination (i.e., endorsement of 6 false positive and no target figures).
Objective memory performance: episodic memory scores—global metamemory prediction task. Objective memory perfor-
mance scores for all immediate (i.e., Trial 1, 2, 3) and delayed recall trials were calculated using the standardized BVMT-R
scoring criteria based on recall accuracy for figure but not for location because metamemory queries pertained only to number
of figures that would be remembered in the future. This was done to simplify the comparison between subjective and objective
memory performance, as well as to reduce participant confusion when describing the task prior to learning, which could have
led to unwanted influences on judgments and evaluations. Therefore, participants received 1 point for each figure that was
accurately drawn for each recall trial regardless of whether the figures were drawn in their correct locations. The range for all
recall measures was also 0 to 6, where 0 indicated no figures and 6 indicated all figures were correctly recalled. For delayed
recognition, we utilized the standardized scoring criteria for BVMT-R, Recognition Hits and Recognition Discrimination Index
(see above).
 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425 1409

Subjective memory performance: pre-experience predictions. Our global metamemory prediction task yielded three trial-
specific pre-experience predictions pertaining to immediate memory performance (i.e., for Trials 1, 2, and 3) and one trial-
specific pre-experience prediction for delayed memory performance (i.e., for the Delayed Recall trial). Scores ranged from 0 to
6, where 0 indicated no figures and 6 indicated all figures would be correctly recalled on the respective recall trial.
Subjective memory performance: post-experience predictions. Our procedure also yielded a global JOL rating about delayed
memory performance (on the Delayed Recall trial; range = 0 to 6, where 0 indicated no figures and 6 indicated all figures would
be correctly recalled on delayed recall), which was recorded after the completion of all three learning trials. Lastly, there was
one global recognition estimate (range = 0 to 6, where 0 indicated no figures and 6 indicated all figures would be correctly

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

identified on the recognition task), which was recorded prior to beginning the delayed recognition trial.
Subjective memory performance: metamemory accuracy. Calibration is the degree to which the level of predicted performance
(i.e., a metacognitive judgment) corresponds to the actual level of performance, and it is a common measure of metamemory
accuracy (Dunlosky & Metcalfe, 2008). We calculated calibration by comparing participants’ predicted and actual memory
performance for each specific trial (i.e., number of figures participant predicted he/she would remember − number of figures
he/she recalled or recognized). For example, if a participant’s global JOL rating for delayed recall was “6” (i.e., she predicted she
would recall 6 figures when asked later) and her objective delayed recall score was also 6 (i.e., she recalled all 6 figures following
the delay period), her global JOL accuracy score would equal 0, indicating perfect calibration. For pre-experience predictions,
we calculated three calibration scores to measure prediction accuracy related to immediate memory (i.e., Trials 1, 2, and 3) and
one calibration score to measure prediction accuracy related to delayed memory (i.e., Delayed Recall Trial). For post-experience
predictions, we calculated three calibration scores to measure global JOL accuracy (i.e., Delayed Recall Trial) and the accuracy
of global recognition estimates (based on Recognition Hits and Recognition Discrimination Index scores). Calibration scores
can range between −6 and 6, where 0 indicates perfect calibration; a positive score represents overconfidence with predicted
memory being greater than actual memory, and a negative score indicates underconfidence with predicted memory being weaker
than actual memory.

Subjective report measures

The Cognitive Change Index (CCI) (Rattanabannakit et al., 2016) is a 20-item measure of participants’ ability level on certain
tasks and cognitive skills (e.g., recalling information, making decisions) compared with 5 years ago. The CCI includes both self
and informant forms and uses a 5-point Likert scale (1 = normal ability/no change to 5 = severe problem/much worse). Of the 20
CCI items, 12 items focus on memory, 5 on executive functioning, and 3 on language. The CCI self (CCI-S) and CCI informant
(CCI-I) scores are the sum of all items on the self-reported and informant reported versions of the assessment, respectively
(range = 20–100), with greater scores representing increased cognitive problems/cognitive change. In addition, the CCI includes
a difference score between self and informant reports (CCI-D, range = −80 to 80), which is calculated by using, CCI-S − CCI-I,
and represents the discrepancy between self- and informant-reports (Rattanabannakit et al., 2016). A positive score indicates
that the participant reported greater cognitive impairment relative to the informant, while a negative score indicates the reverse.
Compared with self-report measures of cognitive decline, informant-report measures have been more strongly correlated to
participants’ objective neuropsychological test scores (Gavett, Dunn, Stoddard, Harty, & Weintraub, 2011; Rami et al., 2014);
therefore, greater distances between self-and informant-ratings have been conceptualized to reflect poorer awareness of overall
cognitive change/functioning.
The Comprehensive Assessment of Prospective Memory, Section B (CAPM B; Chau et al., 2007) is a 39-item questionnaire
that assesses how problematic everyday prospective memory failures are to an individual, thus measuring his/her level of concern
(Chau et al., 2007). Specifically, prospective memory refers to memory for intended actions that are to be carried out at a specific
time in the future, such as remembering to pass on a phone message, take medication, or turn off the stove after a set period
of time. The CAPM B includes self- and informant-rated versions. Each item describes a memory failure and participants and
informants indicate “how much of a problem” each listed failure has been in the past month (scale 1–5) with “1” representing
“no problem at all” and “5” representing “a very serious problem.” A “not applicable” (N/A) option is also available. We used
the CAPM B total score, the average rating of all items answered, excluding N/A responses, with scores ranging from 0 to 5.
We also calculated a difference score between self and informant reports (CAPM B-D; range = −5 to 5) using, CAPM B-self
score − CAPM B-informant score, with positive values reflecting greater concern reported by participants relative to informants
and negative values representing the reverse. We considered the magnitude of the CAPM B-D score to reflect the distance
between informant and participant ratings, with larger magnitudes likely reflecting poorer awareness of everyday prospective
memory failures.
1410 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425

The short form of the Geriatric Depression Scale (GDS; Sheikh & Yesavage, 1986) is a self-reported measure of depressive
symptoms, using a yes/no rating scale. Scores range from 0 to 15, and scores of 5 or higher suggest clinical depression (Almeida
& Almeida, 1999; Marc, Raue, & Bruce, 2008).

Statistical Analyses

We used SPSS Version 26 for analyses, and all p-values were two-tailed with an alpha level of .05. Effect sizes for analysis
of variance (ANOVA) and analysis of covariance (ANCOVA) were assessed using partial eta squared. With regard to the

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

demographic characteristics of our participants, we calculated the descriptive statistics of mean and standard deviation for
continuous variables and frequency and percentage for categorical variables. We used ANOVA to compare the continuous
variables and the Pearson chi-square test to compare the categorical variables, followed by post-hoc analyses utilizing multiple
comparisons with Bonferroni correction or the Mann–Whitney U test, respectively.
For performance-based measures, we utilized separate mixed ANCOVAs, adjusting for significantly different variables
from the demographic and clinical characteristic comparisons, to examine between-group and within-subject differences on
measures of episodic memory, memory self-awareness, and memory self-monitoring related to participants’ actual and predicted
memory performance on Trials 1, 2, 3, Delayed Recall, and Delayed Recognition. Mauchly’s test of sphericity was used to
determine whether there was a violation in the assumption of sphericity. If the assumption was not met, the Huynh–Feldt
correction was applied and we reported the epsilon (ε). Significant interactions were further investigated utilizing separate
independent one-way ANCOVAs, accounting for appropriate covariates or contrasts. Specifically, for our analyses of self-
awareness of memory general knowledge (i.e., predicted memory performance at pre-experience), contrasts were performed
comparing each trial to its previous trial to examine how participants’ memory predictions may change over time, as research
has shown that older adults generally expect (immediate) recall to improve with repeated learning and to decline after a
long delay (see Seelye et al., 2010). Contrasts were also used to compare all groups to HC (baseline). Moreover, Group
X Trial contrasts were used to detect group differences at specific intervals over the time course of the learning/memory
trials. Simple effects plots and/or separate paired t-tests were used to further examine group differences between trials of
interest (i.e., where slopes significantly differed between groups). For our analysis of episodic memory and self-awareness of
memory ability (i.e., calibration scores at pre-experience), simple effect tests using separate independent one-way ANCOVAs,
accounting for appropriate covariates, were conducted to examine group differences at each level of trial. For our analysis
of memory self-monitoring, we examined group differences in metamemory accuracy, as measured by calibration scores
based on participants’ post-experience memory predictions for Delayed Recall (i.e., global JOL accuracy). Also, following
the convention of Seelye et al. (2010), memory self-monitoring was also evaluated by comparing the accuracy of pre-
experience memory predictions for Delayed Recall to the accuracy of post-experience memory predictions for Delayed Recall
in order to detect signs of prediction upgrading. Notably, for our analyses of Delayed Recognition, because we only elicited
post-experience memory predictions, we examined between-group differences for recognition predictions and the accuracy
of these predictions (based on Recognition Discrimination Index and Recognition Hits) using separate independent one-
way ANCOVAs, controlling for appropriate covariates. All post-hoc analyses used multiple comparisons with Bonferroni
correction.
Finally, we carried out two Pearson correlation analyses to explore the relationship between performance-based metamemory
(i.e., JOL ratings and JOL accuracy) and subjective reports of everyday cognitive problems, as measured by self- and informant-
reports and subjective rating discrepancy (SRD) scores, using the entire sample.

Results

Participant Characteristics

Table 1 shows the demographic and clinical characteristic comparisons for the HC, SCD1 , aMCI, and naMCI groups.
Education significantly differed and post-hoc tests showed the mean years of education was significantly lower for naMCI
compared to HC and SCD (ps<.001). Ethnicity also significantly differed, with post-hoc tests showing a greater proportion

1 Only 8 SCD participants were classified based on informant-report of cognitive concerns. Box-plot analyses showed that these 8 participants were not outliers in any of the relevant demographic
or experimental measures.
 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425 1411

Table 1. Demographic and clinical characteristics of the healthy control, subjective cognitive decline, amnestic mild cognitive impairment, and nonamnestic
mild cognitive impairment groups (n = 260)
HC SCD aMCI naMCI
M (SD) or # (%) M (SD) or # (%) M (SD) or # (%) M (SD) or # (%)
Variable n = 120 n = 84 n = 18 n = 38 p
Age (years) 80.28 (5.53) 81.56 (5.18) 81.56 (7.00) 80.42 (5.72) ns
Sex (women) 80.00 (66.67) 54.00 (64.30) 11.00 (61.10) 31.00 (81.60) ns
Education (years) 14.99 (3.13) 15.10 (3.20) 13.56 (4.44) 11.84 (2.91)∗∗∗++ <.001
Ethnicity (non-white) 45.00 (37.50) 24.00 (28.60) 8.00 (44.40) 27.00 (71.10)∗∗∗++ <.001

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

GDS 1.24 (1.60) 2.13 (2.71)∗ 2.00 (2.33) 1.42 (1.46) <.05
Notes. M = mean; SD = standard deviations; HC = healthy control; SCD = subjective cognitive decline; aMCI = amnestic mild cognitive impairment;
naMCI = nonamnestic mild cognitive impairment; GDS = Geriatric Depression Scale. For continuous variables: p values are based on univariate analysis of
variance. For non-continuous variables: p values are based on the Pearson chi-square test.
Significantly different from HC (p < .05)∗ , (p < .01)∗∗ , (p < .001)∗∗∗
Significantly different compared to SCD (p < .05)+ , (p < .001)++

of non-white participants in the naMCI group compared to HC (p<.001) and SCD (p<.001). GDS2 was also significant and
post-hoc tests showed that the mean GDS score was significantly higher for SCD compared to HC (p<.05). However, all group
means for GDS were below the cut-off score associated with clinical depression (i.e., GDS > 4). There were no significant
between-group differences of age or sex.

Episodic Memory Performance

Tables 2 and 3 summarize the means, standard deviations, and group and within-subjects comparisons for HC, SCD, aMCI,
and naMCI on measures of immediate recall, delayed recall, and delayed recognition obtained using the BVMT-R and global
prediction task, respectively. Figure 1 visually depicts group and within-subjects comparisons across trials for objective and
subjective memory/metamemory measures. As expected, using both sets of scores, the MCI groups demonstrated poor overall
immediate recall and delayed recall compared with HC; however, only aMCI demonstrated impaired recognition (see also
Figure 1a). This pattern shows that, although both MCI groups struggled with recall, memory storage was intact for naMCI but
not aMCI, highlighting the latter group’s primary deficit in episodic memory.

Memory Self-awareness

Table 4 summarizes the group means and standard deviations for HC, SCD, aMCI, and naMCI, as well as note significant
between-group and within-subject differences, on measures of memory self-awareness, as indexed by pre-experience memory
predictions and predictive accuracy.
Memory predictions. The 4 Group (HC, SCD, aMCI, and naMCI) X 4 Trial (1, 2, 3, Delayed Recall) mixed ANCOVA,
using pre-experience memory prediction (with education, ethnicity, and GDS as covariates), showed no significant main effect
of group (F < 1). There was a significant within-subjects effect of Trial (F(2.23, 570.95) = 10.71, MSE = .90, ε = .76,
p < .001, ηp 2 = .04) that was qualified by a significant Group X Trial interaction (F(6.88, 570.95) = 3.45, MSE = .897, ε = .76,
p <, ηp 2 = .04). Group X Trial contrasts revealed significant Group X Trial interactions between Trial 2 and Trial 3 (F(3,
249) = 2.96, MSE = .64, p < .05, ηp 2 = .03), as well as between Trial 3 and Delayed Recall (F(3, 249) = 7.04, MSE = .64,
p < .001, ηp 2 = .08). The remaining contrast (Trial 2 vs. Trial 1) did not reveal a significant interaction term (F < 1). Post-hoc
tests using separate paired t-tests for each group, together with analysis of a simple effects plot (Figure 1b), showed that in the
pre-experience phase all groups expected their recall performance to significantly improve between Trial 1 and Trial 2 (HC:
t(119) = −9.69, p < .001; SCD: t(81) = −10.01, p < .001; aMCI: t(17) = −5.58, p < .001; naMCI: t(37) = −4.91, p < .001).
In addition, all groups except aMCI also expected their recall performance to continue to significantly improve between
Trial 2 and Trial 3 (HC: t(119) = −6.95, p < .001; SCD: t(81) = −6.24, p < .001; aMCI: t(17) = −1.49, p = ns; naMCI:
t(37) = −2.14, p < .05). Furthermore, all groups except naMCI expected their recall performance to markedly decline (see
Figure 1b to compare slopes) between Trial 3 and Delayed Recall (HC: t(117) = 14.59, p < .001; SCD: t(81) = 12.84, p < .001;

2 17 participants demonstrated GDS scores above the cut-off associated with depression. Results from a Pearson chi-square test showed that the frequency of these depressed participants did
not significantly differ between groups. Results from all statistical tests that included these participants with elevated clinical depression scores did not differ from those that excluded them;
therefore, we reported findings from the analyses that included all participants.
 1412

Table 2. Analyses for group and within-subjects differences for the healthy control, subjective cognitive decline, amnestic mild cognitive impairment, and nonamnestic mild cognitive impairment groups
on standardized measurements of immediate recall, delayed recall, and recognition obtained using the Brief Visual Memory Test-Revised (N = 260)
Group Trial

HC SCD aMCI naMCI Trial 1 Trial 2 Trial 3 DR RDI ANCOVA

Variable M (SD/SE) M (SD/SE) M (SD/SE) M (SD/SE) M (SE) M (SE) M (SE) M (SE) (SE) F-value (p) ηp 2
n = 120 n = 84 n = 18 n = 38
BVMT-R
Main Effects
Group 4.67 (.17) 5.09 (.19) 2.96 (.40) 3.22 (.29) — — — — — 13.94(<.001) .14
Trial — — — — 2.15 (.12) 3.68 (.18) 4.79 (.21) 4.59 (.22) 4.71 (.10) 10.09 (<.001) .04
Group X Trial — — — — — — — — — 6.19 (<.001) .07
Simple Effects
Trial 1 2.56 (1.70) 2.67 (1.61) 1.44 (.62)+ 1.61 (1.18) — — — — — 7.12 (<.05) .04
Trial 2 4.61 (2.32)a 4.90 (2.56)a 2.33 (1.28)∗∗ 2.45 (1.48)∗∗ — — — — — 9.02 (<.001) .10
Trial 3 5.92 (2.43)a 6.57 (2.72)a 3.33 (1.78)∗∗∗ 2.90 (1.94)∗∗∗ — — — — — 14.90 (<.001) .15
DR 5.70 (2.89) 6.35 (2.93) 3.00 (1.88)∗∗ 2.84 (1.73)∗∗∗ — — — — — 12.36 (<.001) .13
RDI 4.97 (.14) 5.35 (.93) 4.00 (1.41)∗ 4.40 (1.37) — — — — — 6.34 (<.001) .07
Notes. M = mean; SD = standard deviation; SE = standard error; HC = healthy control; SCD = subjective cognitive decline; aMCI = amnestic mild cognitive impairment; naMCI = non-amnestic
mild cognitive impairment; BVMT-R = Brief Visual Memory Test-Revised; DR = Delayed Recall; RDI = Recognition Discrimination Index; ANCOVA = analysis of covariance. All scores are raw
scores. Sample sizes slightly vary due to omission of scores by certain participants. Mixed factorial and independent one-way ANCOVA were used to compare group and within-subjects differences of
all variables (for main and simple effects analyses, respectively), adjusting for ethnicity, education, and GDS. All effect sizes are partial eta square. ns = not significant.
Significantly different from HC (p < .05)∗ , (p < .01)∗∗ , (p < .001)∗∗
Significant trend compared to SCD (p = .06)+
S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425

Recall increased from subsequent trial (ps < .001)a

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425 1413

Table 3. Analyses for group and within-subjects differences for the healthy control, subjective cognitive decline, amnestic mild cognitive impairment, and
nonamnestic mild cognitive impairment groups on measurements of immediate recall, delayed recall, and recognition obtained using the global metamemory
prediction task (N = 260)
Group Trial

HC SCD aMCI naMCI Trial 1 Trial 2 Trial 3 DR RDI ANCOVA ηp 2

Variable M (SD/SE) M (SD) M (SD) M (SD) M (SE) M (SE) M (SE) M (SE) M (SE) F-value (p)
n = 120 n = 84 n = 18 n = 38
GPT

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

Main Effects
Group 2.36 (.09) 2.57 (.11) 1.54 (.23) 1.62 (.17) — — — — — 10.48 (p < .001) .12
Trial — — — — .56 (.07) 1.2 (.10) 1.82 (.12) 1.73 (.12) 4.72 (.10) 35.57 (p < .001) .13
Group X Trial — — — — — — — — — 2.05 (p < .05) .03
Simple Effects
Trial 1 0.79 (0.88) 0.91 (0.82) 0.29 (0.47)+ 0.31 (0.68) — — — — — 3.66 (<.05) .04
Trial 2 1.66 (1.27)a 1.73 (1.31)a 0.59 (0.80)∗ 0.77 (0.81)a — — — — — 5.41 (<.01) .06
Trial 3 2.26 (1.50)a 2.54 (1.48)a 1.41 (1.18)a 0.89(0.87)∗∗ — — — — — 8.46 (<.001) .10
DR 2.16 (1.52) 2.54 (1.48) 1.24 (1.03)∗ 0.89 (0.90)∗∗ — — — — — 7.01 (<.001) .08
RDI 4.97 (.14)a 2.54 (1.48)a 4.00 (1.41)∗ a 4.40 (1.37)a — — — — — 6.34 (<.001) .07
Notes. M = mean; SD = standard deviation; SE = standard error; HC = healthy control; SCD = subjective cognitive decline; aMCI = amnestic mild cognitive
impairment; naMCI = non-amnestic mild cognitive impairment; GPT = Global Prediction Task; DR = Delayed Recall; RDI = Recognition Discrimination
Index; ANCOVA = analysis of covariance. All scores are raw scores. Sample sizes slightly vary due to omission of scores by certain participants. Mixed
factorial and independent one-way ANCOVA were used to compare group and within-subjects differences of all variables (for main and simple effects analyses,
respectively), adjusting for ethnicity, education, and GDS. All effect sizes are partial eta square. ns = not significant.
Significantly different from HC (p < .05)∗ , (p < .01)∗
Significant trend compared to SCD (p = .06)+
Recall or recognition significantly increased from subsequent trial (ps < .001)a

aMCI: t(17) = 3.74, p < .01; naMCI: t(37) = 3.14, p < .01). Taken together, findings suggest that, prior to exposure to task
stimuli/learning, all groups demonstrated knowledge that memory performance typically improves with repetition and may
decline after a delay.

Accuracy of memory predictions. The 4 Group (HC, SCD, aMCI, and naMCI) X 4 Trial (1, 2, 3, Delayed Recall) mixed
ANCOVA conducted using calibration scores (e.g., predictive/metamemory accuracy), with education, ethnicity, and GDS as
covariates, showed no significant main effect of group (F = 1.43, p > .05). There was, however, a significant main effect of Trial
(F(2.61, 615.26) = 3.42, MSE = 1.44, ε = .87, p < .05, ηp 2 = .02) that was qualified by a significant Group X Trial interaction
(F(7.82, 615.26) = 3.31, MSE = 1.44, ε = .87, p < .01, ηp 2 = .04). Separate one-way independent ANCOVAs, controlling for
appropriate covariates, revealed no significant group differences in calibration for Trial 1 (F < 1), Trial 2 (F = 1.31), or Trail
3 (F < 1). However, groups did significantly differ on calibration on Delayed Recall (F(3, 236) = 4.67, p < .01, ηp 2 = .06).
Post-hoc tests showed that HC participants were significantly more accurate (demonstrated by a smaller mean calibration score
because perfect calibration is 0) in their pre-experience memory predictions compared with naMCI. Analysis of simple effects
plot (see Figure 1c) also showed that all groups were significantly more accurate in their pre-experience memory predictions
for delayed compared with immediate recall trials. Taken together, findings indicate that, at pre-experience, the predementia
groups were just as accurate as HC in predicting their immediate memory performance; however, the naMCI group (but no
other predementia group) was less accurate in predicting their delayed memory performance relative to HC.

Memory Self-monitoring. Table 5 summarizes the group means and standard deviations for HC, SCD, aMCI, and naMCI, as
well as note significant between-group and within-subject differences, on measures of memory self-monitoring, as indexed by
post-experience memory predictions and their accuracy.

Memory predictions: Delayed recall. The 4 Group (HC, SCD, aMCI, and naMCI) by 2 Experience (pre-experience vs. post-
experience) mixed ANCOVA using memory predictions for Delayed Recall (with education, ethnicity, and GDS as covariates),
showed no significant main effect of Group (F < 1). There was, however, a significant effect of Experience, (F(1, 242) = 7.90,
MSE = 1.34, ε = 1, p < .01, ηp 2 = .03), where the predicted performance for Delayed Recall was higher at pre-experience
(M = 3.06, SE = .12) compared with post-experience (M = 2.71, SE = .12). The Group X Experience interaction effect was not
1414 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

Fig. 1. (A) Episodic memory performance for each group across immediate memory and delayed memory trials of the experimental global metacognitive
prediction task. Mean number of figures recalled or recognized for each trial depicted separately for each group. Error bars represent 95% confidence intervals.
Trials showing significant group differences are represented with “∗ ”. (B) Predicted memory performance for each group across immediate and delayed recall
trials at pre-experience. Mean number of figures predicted at recall for each trial depicted separately for each group. Error bars represent 95% confidence
intervals. Significant interactions detected between trials are represented with “∗ ”. (C) Metamemory accuracy (calibration) of predicted memory performance
for each group across immediate and delayed recall trials at pre-experience. Mean calibration score for each recall trial depicted separately for each group. Error
bars represent 95% confidence intervals. Significant interactions detected between trials are represented with “∗ ”. ∗ p < .05, ∗∗ p < .01, ∗∗∗ p < .001.

significant (F < 1). Taken together, findings suggest that all participants, irrespective of group classification, predicted higher
delayed recall performance prior to learning compared with after learning, indicating that the experience of learning impacted
(i.e., lowered) their performance expectations.

Accuracy of memory predictions: Delayed recall. The 4 Group (HC, SCD, aMCI, and naMCI) by 2 Experience (pre-
experience vs. post-experience) mixed ANCOVA using calibration scores for Delayed Recall (with education, ethnicity, and
GDS as covariates), showed a significant main effect of Group (F(3, 230) = 6.54, MSE = 2.71, p < .001, ηp 2 = .08). Post-
hoc tests showed the naMCI group (M = 1.84, SE = .26) but not aMCI (M = 1.39, SE = .37) demonstrated significantly
worse calibration with higher calibration scores (perfect calibration = 0) compared with the HC (M = 0.67, SE = .15) and
SCD (M = 0.48, SE = .18) groups (ps < .01 and .001, respectively). There was also a significant main effect of Trial (F(1,
230) = 7.69, MSE = 1.36, ε = 1, p < .01, ηp 2 = .03), which showed that all participants were more accurate in their memory
predictions at post-experience (M = 1.27, SE = .17) compared with pre-experience (M = 0.92, SE = .12). Lastly, the Group
 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425 1415

Table 4. Means and standard deviations reported for the healthy control, subjective cognitive decline, amnestic mild cognitive impairment, and nonamnestic mild
cognitive impairment groups on measurements of memory self-awareness, indexed by pre-experience memory predictions and the accuracy of pre-experience
memory predictions (calibration scores), obtained from the global metamemory prediction task (N = 260)
HC SCD aMCI naMCI
Variable M (SD) M (SD) M (SD) M (SD)
n = 120 n = 84 n = 18 n = 38
Memory Self-Awareness
Predictions
Trial 1 3.50 (1.25) 3.42 (1.38) 3.11 (1.18) 2.92 (1.08)

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

Trial 2 4.26 (1.34)aa 4.26 (1.29)aa 4.00 (1.14)aa 3.75 (1.37)aa
Trial 3 4.76 (1.32)aa 4.83 (1.30)aa 4.33 (1.46) 3.97 (1.40)a
DR 2.92 (1.45)bb 2.94 (1.36)bb 2.94 (1.30)b 3.34 (1.60)b
Prediction Accuracy
Trial 1 2.7 (1.4) 2.51 (1.54) 2.88 (1.18) 2.49 (1.17)
Trial 2 2.58 (1.76) 2.51 (1.73) 3.41(1.42) 2.80 (1.69)
Trial 3 2.47 (1.83) 2.27 (2.11) 2.82 (1.91) 2.97 (1.67)
DR 0.75 (1.87) 0.48 (2.22) 1.65 (1.58) 2.34 (2.10)∗+
Notes. M = mean; SD = standard deviations; HC = healthy control; SCD = subjective cognitive decline; aMCI = amnestic mild cognitive impairment;
naMCI = non-amnestic mild cognitive impairment; n = sample size
Predictions significantly increased from subsequent trial (p < .05)a , (p < .001)aa
Predictions significantly decreased from subsequent trial (p < .01)b , (p < .001)bb
Significantly different from HC (p < .05)∗
Significantly different from SCD (p < .01)+

Table 5. Means and standard deviations for the healthy control, subjective cognitive decline, amnestic mild cognitive impairment, and nonamnestic mild
cognitive impairment groups on measurements of memory self-monitoring, indexed by post-experience memory predictions and the accuracy of post-experience
predictions (calibration scores), obtained from the global metamemory prediction task (N = 260)
HC SCD aMCI naMCI ANCOVA ηp 2
Variable M (SD) M (SD) M (SD) M (SD) F-value (p)
n = 120 n = 84 n = 18 n = 38
Memory Self-Monitoring
Predictions
Pre-Exp, DR 2.92 (1.45) 2.94 (1.36) 2.94 (1.30) 3.34 (1.60) — —
Post-Exp, DR (Global JOL) 2.67 (1.32)a 2.87 (1.36)a 2.44 (1.15)a 2.68 (1.34)a — —
Prediction Accuracy
Delayed Recall
Pre-Exp, DR 0.75 (1.87) 0.48 (2.22) 1.65 (1.58) 2.34 (2.10)∗++ — —
Post-Exp, DR (Global JOL) 0.51 (1.43)b 0.30 (1.28)b 1.12 (1.87)b 1.74 (1.50)b ∗∗++ — —
Delayed Recognition
Post-Exp, GRE (RDI) −0.92 (1.82) −1.01 (1.46) −0.73 (2.15) −0.78 (2.11) 0.10 (ns) .00
Post-Exp, Rec 4.14 (1.50) 4.37 (1.44) 3.31 (1.20) 3.51 (1.40)∗∗∗+ 3.76 (<.05) .05
Notes. M = mean; SD = standard deviations; HC = healthy control; SCD = subjective cognitive decline; aMCI = amnestic mild cognitive impairment;
naMCI = nonamnestic mild cognitive impairment; ANCOVA = analysis of covariance. Sample sizes slightly vary due to omission of scores by certain
participants. Pre-Exp = Pre-experience; Post-Exp = Post-Experience; DR = delayed recall; Rec = delayed recognition; GRE = Global Recognition Estimate;
Global JOL = global judgment of learning; Rec Hits = recognition hits; RDI = Recognition Discrimination Index; ns = not significant. ANCOVA was used to
compare group differences of all variables, adjusting for education, ethnicity, and GDS. All effect sizes are partial eta square.
Significantly different from HC (p < .05)∗ , (p < .01)∗∗
Significant trend compared to HC (p = .06)∗∗∗
Significantly different from SCD (p < .05)+ , (p < .01)++
Predicted recall performance declined from subsequent trial (ps < .001)a
Prediction accuracy better at post-experience compared with pre-experience (ps < .001)b

X Experience interaction effect was not significant (F < 1). Taken together, findings suggest that, although all participants
demonstrated an increase in prediction accuracy from pre- to post-experience, naMCI participants were less accurate (in the
direction of overconfidence) in their delayed memory performance both prior and after learning compared with HC and SCD
participants.
Accuracy of memory predictions: Delayed recognition. We evaluated group differences in post-experience prediction
accuracy for delayed recognition performance (i.e., global recognition estimate accuracy) for both RDI and Recognition Hits
1416 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425

Table 6. Summary of results for Pearson’s correlations between global prediction measures and subjective report scores related to everyday prospective memory
failures and overall cognitive change
Subjective Report Scores JOL rating JOL Accuracy
CAPM-S .048 -.067
CAPM-I .133 .221∗∗
CAPM-D .083 -.015
CCI-S .073 .031
CCI-I -.110 -.036
CCI-D .240∗∗ .153

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

Note. JOL = Judgment of learning; CAMP = Comprehensive Assessment of Prospective Memory, Section B, which measures levels of concern/problem with
everyday prospective memory failures; S = self-reported; I = informant-reported; D = difference score (subjective—informant report score); CCI = Cognitive
Change Index, which measures level of change/problem with overall cognitive functioning.
∗ p < .05
∗∗ p < .01

using separate one-way ANCOVAs, where education, ethnicity, and GDS were entered as covariates. There was a significant
group effect on global recognition estimate accuracy when the calibration score was based on recognition hits. Post-hoc tests
showed that naMCI demonstrated worse calibration with a larger mean global recognition estimate score compared with HC
(p = .06) and SCD (p < .05). However, there was no significant group effect on global recognition estimate accuracy when
the calibration score was based on the Recognition Discrimination Index score. Notably, group means for global recognition
estimate accuracy (i.e., using both recognition hits and the Recognition Discrimination Index score) were negative overall,
suggesting that all groups were underconfident in prospectively judging their recognition performance. In addition, group means
for global recognition estimate accuracy reflected greater underconfidence when the calibration score was based on recognition
hits compared with the Recognition Discrimination Index score, which was due to the fact that participants’ actual recognition
scores were higher when false positives were not accounted for. Participants likely did not take penalties for endorsing false
alarms into consideration when making their global recognition estimates. Therefore, the fact that naMCI only differed from
HC and SCD participants on global recognition estimates when the recall hits score (but not Recognition Discrimination Index)
was used for calibration, suggests that naMCI participants were prone to false recognition errors.

Correlation Between Global Prediction Measures and Self- and Informant-reported Inventories of Everyday Cognitive
Functioning. We explored the relationship between global JOL ratings, as well as global JOL accuracy, and the six subjective
measures associated with the CAPM-B and CCI (i.e., self-report, informant-report, and self-informant difference scores for each
inventory) utilizing the entire sample in two different Pearson correlation analyses. See Table 6 for a summary of results.
There was a significant small positive correlation between global JOL ratings and the CCI-D score, suggesting that,
subsequent to learning, participants who gave higher global JOL ratings also showed poorer self-awareness (i.e., a larger positive
discrepancy between self- and informant-reported cognitive decline). Unexpectedly, however, poor self-awareness was in the
direction of underconfidence.
There was also a significant small positive correlation between global JOL calibration scores and CAPM-I, suggesting that
participants who demonstrated poorer global JOL accuracy (i.e., larger calibration scores) also had informants who were more
concerned about their everyday prospective memory failures. No other significant correlations emerged between experimental
measures and self-report assessments.

Discussion

Our primary goal was to investigate differences in metamemory accuracy between cognitively healthy elderly controls (HC)
and those with SCD, aMCI, and naMCI using a global metamemory prediction task based on visual memory. Compared with HC,
all prodromal dementia groups demonstrated comparable metamemory accuracy with regard to immediate recall performance at
pre-experience, indicating intact self-awareness about immediate memory processes. In addition, all participant groups were able
to appropriately modify their memory predictions after task exposure/learning, demonstrating significantly better metamemory
accuracy at post-experience compared with pre-experience. However, compared with HC, only the naMCI group demonstrated
significantly worse metamemory accuracy at pre-experience and significantly worse JOL accuracy at post-experience when
predicting their delayed recall performance, indicating that, despite demonstrating prediction upgrading, naMCI (but not aMCI)
participants still struggled with poor self-awareness and self-monitoring of delayed memory processes. Altogether, results
 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425 1417

suggest neuropsychological deficits specific to MCI subtypes may interfere with different mechanisms in the brain’s cognitive
awareness system resulting in differential levels of awareness and metamemory accuracy. Moreover, our findings provide
confirmatory evidence that metamemory abilities remain intact in SCD. A second goal was to explore the relationship between
global JOL measures and self-report measures related to everyday memory/cognitive problems. Results offer some support that
metamemory difficulties captured using our experimental metamemory task may generalize to everyday memory difficulties.

Episodic Memory Performance

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

First, objective memory performance measured using the standardized BVMT-R protocol and our global metamemory
prediction task (which included modifications to the BVMT-R protocol/scoring) showed the same general between-group
patterns. As expected, relative to HC, the MCI groups demonstrated poor overall immediate recall and delayed recall; however,
only aMCI additionally demonstrated impaired delayed recognition. This pattern suggests that although both MCI groups
struggled with recall, memory storage was intact for naMCI but not for aMCI, highlighting the latter group’s primary deficit in
episodic memory (Petersen, 2004, 2011). Recall difficulties for naMCI participants were presumably for nonmemory-related
reasons, such as inefficient retrieval secondary to executive function deficits (Anderson, 2009; Brooks, Weaver, & Scialfa, 2006;
Redondo, Beltrán-Brotóns, Reales, & Ballesteros, 2016). Additionally, consistent with their diagnostic profile, SCD participants
performed comparably to HC on both standardized and experimental episodic memory measures. Altogether, results show that
the modified BVMT-R scoring procedures retained sensitivity to diagnostic differences.

Subjective Memory Performance

Memory self-awareness. Relative to HC, participants with SCD, aMCI, and naMCI demonstrated intact general knowledge
and self-awareness related to immediate memory processes. These findings are consistent with previous research showing that
memory self-knowledge and self-awareness of working memory and immediate memory abilities are preserved in MCI and
AD (Bertrand et al., 2019; Seelye et al., 2010; Silva, Pinho, Macedo, Souchay, & Moulin, 2017; Thomas, Lee, & Balota,
2013). Notably, relative to HC, the naMCI participants (and no other preclinical dementia group) demonstrated evidence of
poor memory self-awareness of delayed recall, which could potentially suggest that those with naMCI base their metamemory
judgments on inaccurate representations of self-ability. However, given that memory self-awareness, especially as measured by
predictive accuracy at pre-experience, has also been linked to other factors, such as familiarly with task procedures (Connor,
Dunlosky, & Hertzog, 1997), future research is needed to investigate the determinates of impaired memory self-awareness
observed in those with naMCI. Lastly, given that SCD is conceived as a pre-MCI condition (Rabin et al., 2017), findings of
intact general knowledge and self-awareness of memory processes were expected.

Memory self-monitoring: Mild Cognitive Impairment. Consistent with our predictions, those with naMCI (but not aMCI)
demonstrated deficits in memory self-monitoring (i.e., JOL accuracy) compared with HC. Given that past research has shown
that the frontal cortex but not the temporal lobes is critical in supporting JOL accuracy (Andrés, Mazzoni, & Howard, 2010;
Howard et al., 2010; Howard, Andrés, & Mazzoni, 2013; Vilkki et al., 1998; Vilkki et al., 1999), our results suggest that poor
JOL calibration observed in naMCI participants was likely attributable to primary deficits in executive functions/frontal systems,
while intact JOL calibration observed in aMCI participants was likely because JOL accuracy is not dependent on the integrity
of episodic memory/Medial Temporal Lobe systems. In addition, because the naMCI group included participants characterized
by low executive and/or global/verbal functioning factor scores, deficits in global functioning, which may also interfere with
the efficiency of executive functioning processes, could have also contributed to poor JOL calibration in some of our naMCI
participants. Importantly, all groups demonstrated the ability to significantly improve the accuracy of their memory predictions
for Delayed Recall after learning/task exposure (i.e., prediction upgrading) essentially by lowering their memory predictions.
This suggests that even the naMCI group was able to utilize experience with the task to update memory self-knowledge, albeit
less efficiently compared with HC.
Altogether, our findings offer support for the hypothesis that differences in metamemory accuracy may arise in MCI subtypes
as a function of their primary neuropsychological impairments that may interfere with specific mechanisms in the brain’s
cognitive awareness system. Findings for aMCI participants are consistent with past research showing preserved JOL accuracy
in aMCI (Akhtar et al., 2006; Ryals et al., 2018; Seelye et al., 2010). Results for naMCI are also generally consistent with those
reported by Seelye et al. (2010)—who, using a verbal memory-based metamemory prediction paradigm, showed evidence that
naMCI (but not aMCI) participants were significantly more poorly calibrated in their delayed recall predictions both at pre- and
post-experience compared with HC. Finally, findings for naMCI in the current study are consistent with those from our previous
1418 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425

study (Chi et al., 2020) using a retrospective metamemory task, which showed that metamemory monitoring processes were
impaired in naMCI (but not aMCI).
Memory self-monitoring: Subjective Cognitive Decline. Consistent with our prediction, SCD participants were able to
accurately predict their memory performance relative to HC, demonstrating intact memory self-monitoring. These findings
are also consistent with our previous study (Chi et al., 2020), which showed comparable performance between SCD and HC
on a retrospective metamemory task. Notably, although SCD participants demonstrated a significantly higher mean GDS score
compared with HC in the current study (e.g., depression symptoms were in the nonclinical range for all groups), performance
between SCD and HC participants did not significantly differ on any of our metamemory accuracy measures after controlling for

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

depression and other significant demographic variables. Therefore, although our results support the idea that even nonclinical
depression symptomology is an important factor to consider in understanding and treating SCD (Buckley et al., 2016; Chapman
et al., 2020), the current findings, together with those from our previous study, show that elevated levels of nonclinical depression
symptomology in those with SCD likely do not interfere with their ability to accurately evaluate their memory functioning either
prospectively or retrospectively.

Global recognition estimates. We found that participants underestimated their delayed recognition performance. Notably,
participants were queried about their prospective recognition performance immediately after the delayed recall trial. It is possible
that the perception of poor performance on the delayed recall trial or perception that the task was difficult lowered participants’
confidence in recognition performance. Confidence in one’s response on a given task/trial has been reported to influence
confidence on the following task/trial (Rahnev, Koizumi, McCurdy, D’Esposito, & Lau, 2015). Unfortunately, we did not query
participants about their predicted delayed recognition performance before learning, which precludes further analysis of this
issue. However, future research should explore the effect of task order.

Relationship Between JOL Accuracy and Subjective Reports of Everyday Cognitive Functioning

JOL predictions were positively correlated with an SRD measure that was based on overall cognitive change/problem,
suggesting that participants who made higher predictions for delayed recall performance at post-experience also had poorer
self-awareness scores (i.e., a larger discrepancy between self- and informant-report), however, unexpectedly in the direction
of underconfidence. One explanation is that participants may have been able to accurately assess their memory/cognitive
functioning when completing offline metamemory measures (e.g., questionnaires) because these are based on metacognitive
knowledge (i.e., general knowledge and beliefs about their memory; Flavell, 1979). In spite of this, it is possible that those
with cognitive difficulties struggled to spontaneously use their metacognitive knowledge to support their predictions while
engaged in online performance monitoring (Perrotin, Belleville, & Isingrini, 2007), given the high cognitive load of experimental
metamemory tasks, resulting in their higher JOL ratings. Lastly, although the size of both MCI groups is comparable to those
reported in other metacognition in MCI studies (e.g., Ryals et al., 2018; Seelye et al., 2010; see also Piras et al., 2016, Table 2), it
limited our ability to explore correlational differences between “online” JOL measures and “offline” measures of both cognitive
change and prospective memory difficulties for each group individually due to insufficient power.
In addition, JOL calibration scores (e.g., higher scores indicate poorer JOL accuracy) were positively correlated with
informant but not self-reported concern about prospective memory failures, suggesting that participants who demonstrated the
worst JOL accuracy also had informants who were the most concerned about their everyday memory difficulties. Given that
prospective memory failures are associated with safety implications for activities of daily life, such as remembering to take
medication or turn off the stove (Chau et al., 2007), it is not unexpected that everyday PM failures related to poor JOL accuracy
would be linked to greater informant concern as these failures may be more salient. In addition, the lack of a relationship
between self-reported concern and poorer JOL accuracy could provide more evidence of poorer self-awareness of memory
functioning in participants with poor visual memory JOL accuracy. Importantly, given that anosognosia is also associated
with safety risks (Starkstein, Jorge, Mizrahi, Adrian, & Robinson, 2007), we offer support that outcome measures from our
experimental metamemory task can generalize to memory difficulties in everyday life.

The CAM and Clinical Implications for Guiding Treatment

Although our global metamemory prediction task is not suitable for detection of a mnemonic anosognosia, it holds potential
for detecting the executive and primary forms of anosognosia. In the context of CAM, failure to detect errors and/or impaired
evaluative processes during performance monitoring (e.g., due to an executive anosognosia) or lack of conscious awareness
 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425 1419

of memory failures (e.g., due to a primary anosognosia) could result in one possessing inaccurate self-knowledge about
his/her cognitive abilities over time (Hannesdottir & Morris, 2007; Morris & Mograbi, 2013). Interestingly, although naMCI
participants struggled with self-monitoring of delayed memory processes (i.e., demonstrating poor JOL accuracy relative to
HC) in our study, they were still able to improve their prediction accuracy after learning, demonstrating the ability to use their
experience with the task to update memory self-knowledge. Overall, this finding may provide preliminary support to the idea
that poor JOL accuracy in naMCI could be more attributable to problems with performance monitoring rather than to a lack of
conscious awareness of memory failures. However, modifications to our global metamemory prediction task will be necessary
to definitively distinguish between executive and primary forms of anosognosia. For example, incorporating a separate feedback

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

condition to our metamemory task may allow us to directly differentiate executive and primary anosognosia (as was recently
demonstrated in a study in AD; see Cosentino et al., 2016). Specifically, given that an executive anosognosia is thought to arise
due to deficits in performance monitoring and not due to disruptions in conscious awareness of memory failures (as in the case
of a primary anosognosia) (Cosentino et al., 2016; Morris & Mograbi, 2013), a significant improvement in predictive accuracy
following feedback in naMCI may be indicative that poor memory awareness was due to an executive anosognosia and vice
versa.
Overall, because individuals who are aware of their cognitive strengths and weaknesses may be more likely to benefit
from interventions (Clare et al., 2004; Moulin, James, Perfect, & Jones, 2003), our findings suggest that those with SCD and
aMCI are likely to benefit from cognitive rehabilitation or remediation strategies (Barnes et al., 2013; Hoogenhout, de Groot,
Van der Elst, & Jolles, 2012; Jessen et al., 2014). Importantly, for those with SCD, providing cognitive treatment from the
standpoint of preventative care could empower these individuals to proactively support their own cognitive and emotional
well-being rather than waiting until significant clinical symptoms are present (Smart et al., 2017). Furthermore, for those
with aMCI, our findings suggest that intact prospective metamemory processes represent an important area of strength that
is important when recommending compensatory strategies. Moreover, for those with naMCI, interventions that directly target
error detection and reasoning errors, such as Metacognitive Training Therapy (Moritz et al., 2014), may be most useful for those
who demonstrate signs of an executive anosognosia, while recommendations for environmental adaption or use of mnemonic
assistive devises, which do not require high levels of awareness, could be more appropriate for those who demonstrate signs
of a primary anosognosia. Finally, given that our task was able to differentiate between MCI subtypes, as well as demonstrate
generalizability between outcome measures and everyday memory difficulties, the results of this study support the potential
utility of experimental metamemory tasks for informing diagnostic clarity and treatment recommendations.

Funding

This work was supported by the National Institute on Aging (NIA) and National Institute of General Medical Sciences
(SC2AG039235), NIA (P01 AG03949, R01AG039409-0), The Czap Foundation, and The Leonard and Sylvia Marx Foundation.

Acknowledgements

The authors are appreciative of Milushka Elbulok-Charcape, Moisey Abramov, Valdiva Da Silva, Dr. Avner Aronov, Dr. Ashu
Kapoor, Dr. Erica Meltzer, Charlotte Magnotta, Dr. Wendy Ramratan, Dr. Molly Zimmerman, Dr. Richard Lipton, and Mindy
Katz for their contributions.

Conflict of Interest

None declared.

References

Agnew, S. K., & Morris, R. (1998). The heterogeneity of anosognosia for memory impairment in Alzheimer’s disease: A review of the literature and a proposed
model. Aging & Mental Health, 2(1), 7–19.
Akhtar, S., Moulin, C. J., & Bowie, P. C. (2006). Are people with mild cognitive impairment aware of the benefits of errorless learning? Neuropsychological
Rehabilitation, 16(3), 329–346.
Almeida, O. P., & Almeida, S. A. (1999). Confiabilidade da versão Brasileira da Escala de Depressão em Geriatria (GDS) versão reduzida. Arq Neuropsiquiatr,
57(2B), 421–426.
Amariglio, R. E., Mormino, E. C., Pietras, A. C., Marshall, G. A., Vannini, P., Johnson, K. A. et al. (2015). Subjective cognitive concerns, amyloid-β, and
neurodegeneration in clinically normal elderly. Neurology, 85(1), 56–62.
1420 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425

Anderson, S.-E. (2009). Predictions of episodic memory following moderate to severe traumatic brain injury during inpatient rehabilitation. Journal of Clinical
and Experimental Neuropsychology, 31(4), 425–438. doi: 10.1080/13803390802232667.
Andrés, P., Mazzoni, G., & Howard, C. E. (2010). Preserved monitoring and control processes in temporal lobe epilepsy. Neuropsychology, 24(6), 775.
Barnes, D. E., Santos-Modesitt, W., Poelke, G., Kramer, A. F., Castro, C., Middleton, L. E. et al. (2013). The mental activity and eXercise (MAX) trial: A
randomized controlled trial to enhance cognitive function in older adults. Journal of the American Medical Society Internal Medicine, 173(9), 797–804.
doi: 10.1001/jamainternmed.2013.189.
Beckett, L. A., Donohue, M. C., Wang, C., Aisen, P., Harvey, D. J., Saito, N. et al. (2015). The Alzheimer’s disease neuroimaging initiative phase 2: Increasing
the length, breadth, and depth of our understanding. Alzheimer’s and Dementia, 11(7), 823–831. doi: 10.1016/j.jalz.2015.05.004.
Benedict, R. H., Schretlen, D., Groninger, L., Dobraski, M., & Shpritz, B. (1996). Revision of the brief visuospatial memory test: Studies of normal performance,
reliability, and validity. Psychological Assessment, 8(2), 145.

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

Bertrand, J. M., Mazancieux, A., Moulin, C. J., Béjot, Y., Rouaud, O., & Souchay, C. (2019). In the here and now: Short term memory predictions are preserved
in Alzheimer’s disease. Cortex, 119, 158–164.
Brandt, M., Carvalho, R. L. S. d., Belfort, T., & Dourado, M. C. N. (2018). Metamemory monitoring in Alzheimer’s disease a systematic review. Dementia &
Neuropsychologia, 12(4), 337–352.
Brooks, B. L., Weaver, L. E., & Scialfa, C. T. (2006). Does impaired executive functioning differentially impact verbal memory measures in older adults with
suspected dementia? Clinical Neuropsychology, 20(2), 230–242. doi: 10.1080/13854040590947461.
Buckley, R. F., Laming, G., Chen, L. P., Crole, A., & Hester, R. (2016). Assessing error awareness as a mediator of the relationship between subjective concerns
and cognitive performance in older adults. PLoS One, 11(11), e0166315. doi: 10.1371/journal.pone.0166315.
Chapman, S., Colvin, L. E., & Cosentino, S. (2020). Translational aspects of the multidisciplinary study of metacognition. Translational Issues in Psychological
Science, 6(1), 26–31. doi: 10.1037/tps0000224.
Chau, L. T., Lee, J. B., Fleming, J., Roche, N., & Shum, D. (2007). Reliability and normative data for the comprehensive assessment of prospective memory
(CAPM). Neuropsychological Rehabilitation, 17(6), 707–722.
Chi, S. Y., Chua, E. F., Kieschnick, D. W., & Rabin, L. A. (2020). Retrospective metamemory monitoring of semantic memory in community-dwelling older
adults with subjective cognitive decline and mild cognitive impairment. Neuropsychological rehabilitation, 1–35. doi: 10.1080/09602011.2020.1831552.
Chi, S. Y., Rabin, L. A., Aronov, A., Fogel, J., Kapoor, A., & Wang, C. (2014). Differential focal and nonfocal prospective memory accuracy in a demographically
diverse group of nondemented community-dwelling older adults. Journal of the International Neuropsychological Society, 20(10), 1015–1027. doi:
10.1017/S1355617714000964.
Chua, E. F., Pergolizzi, D., & Weintraub, R. R. (2014). The Cognitive Neuroscience of Metamemory Monitoring: Understanding Metamemory Processes,
Subjective Levels Expressed, and Metacognitive Accuracy. In The Cognitive Neuroscience of Metacognition (pp. 267–291). London, UK: Springer.
Clare, L., Wilson, B. A., Carter, G., Roth, I., & Hodges, J. R. (2004). Awareness in early-stage Alzheimer’s disease: Relationship to outcome of cognitive
rehabilitation. Journal of Clinical and Experimental Neuropsychology, 26(2), 215–226.
Connor, L. T., Dunlosky, J., & Hertzog, C. (1997). Age-related differences in absolute but not relative metamemory accuracy. Psychology and Aging, 12(1), 50.
Cosentino, S., Metcalfe, J., Butterfield, B., & Stern, Y. (2007). Objective metamemory testing captures awareness of deficit in Alzheimer’s disease. Cortex; a
journal devoted to the study of the nervous system and behavior, 43(7), 1004–1019. doi: 10.1016/s0010-9452(08)70697-x.
Cosentino, S., Brickman, A. M., Griffith, E., Habeck, C., Cines, S., Farrell, M. et al. (2015). The right insula contributes to memory awareness in cognitively
diverse older adults. Neuropsychologia, 75, 163–169. doi: 10.1016/j.neuropsychologia.2015.05.032.
Cosentino, S., Zhu, C., Bertrand, E., Metcalfe, J., Janicki, S., & Cines, S. (2016). Examination of the metacognitive errors that contribute to anosognosia in
Alzheimer’s disease. Cortex, 84, 101–110.
Devolder, P. A., Brigham, M. C., & Pressley, M. (1990). Memory performance awareness in younger and older adults. Psychology and Aging, 5(2), 291.
Diener, E., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). The satisfaction with life scale. Journal of Personality Assessment, 49(1), 71–75.
Dunlosky, J., & Metcalfe, J. (2008). Metacognition. Thousand Oaks, CA: Sage Publications.
Ferman, T. J., Smith, G. E., Kantarci, K., Boeve, B. F., Pankratz, V. S., Dickson, D. W. et al. (2013). Nonamnestic mild cognitive impairment progresses to
dementia with Lewy bodies. Neurology, 81(23), 2032–2038.
Flavell, J. H. (1979). Metacognition and cognitive monitoring: A new area of cognitive–developmental inquiry. American Psychologist, 34(10), 906.
Gavett, R., Dunn, J. E., Stoddard, A., Harty, B., & Weintraub, S. (2011). The cognitive change in women study (CCW): Informant ratings of cognitive change
but not self-ratings are associated with neuropsychological performance over 3 years. Alzheimer Disease and Associated Disorders, 25(4), 305–311. doi:
10.1097/WAD.0b013e31820d8652.
Goodglass, H., Kaplan, E., & Weintraub, S. (1983). Boston naming test. Philadelphia, PA: Lea & Febiger.
Grober, E., & Buschke, H. (1987). Genuine memory deficits in dementia. Developmental neuropsychology, 3(1), 13–36.
Hannesdottir, K., & Morris, R. G. (2007). Primary and secondary anosognosia for memory impairment in patients with Alzheimer’s disease. Cortex, 43(7),
1020–1030.
Holtzer, R., Verghese, J., Xue, X., & Lipton, R. B. (2006). Cognitive processes related to gait velocity: Results from the Einstein aging study. Neuropsychology,
20(2), 215.
Hoogenhout, E. M., de Groot, R. H., Van der Elst, W., & Jolles, J. (2012). Effects of a comprehensive educational group intervention in older women with
cognitive complaints: A randomized controlled trial. Aging & Mental Health, 16(2), 135–144.
Howard, C. E., Andrés, P., Broks, P., Noad, R., Sadler, M., Coker, D. et al. (2010). Memory, metamemory and their dissociation in temporal lobe epilepsy.
Neuropsychologia, 48(4), 921–932.
Howard, C. E., Andrés, P., & Mazzoni, G. (2013). Metamemory in temporal lobe epilepsy: A study of sensitivity to repetition at encoding. Journal of the
International Neuropsychological Society, 19(4), 453–462.
Jessen, F. (2014). Subjective and objective cognitive decline at the pre-dementia stage of Alzheimer’s disease. European Archives of Psychiatry and Clinical
Neuroscience, 264(Suppl 1), S3–S7. doi: 10.1007/s00406-014-0539-z.
Jessen, F., Amariglio, R. E., van Boxtel, M., Breteler, M., Ceccaldi, M., Chetelat, G. et al. (2014). A conceptual framework for research on subjective cognitive
decline in preclinical Alzheimer’s disease. Alzheimer’s and Dementia, 10(6), 844–852. doi: 10.1016/j.jalz.2014.01.001.
 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425 1421

Jessen, F., Feyen, L., Freymann, K., Tepest, R., Maier, W., Heun, R. et al. (2006). Volume reduction of the entorhinal cortex in subjective memory impairment.
Neurobiological Aging, 27(12), 1751–1756. doi: 10.1016/j.neurobiolaging.2005.10.010.
Katz, M. J., Lipton, R. B., Hall, C. B., Zimmerman, M. E., Sanders, A. E., Verghese, J., Dickson, D. W., & Derby, C. A. (2012). Age-specific and sex-specific
prevalence and incidence of mild cognitive impairment, dementia, and Alzheimer dementia in blacks and whites: a report from the Einstein Aging Study.
Alzheimer disease and associated disorders, 26(4), 335–343. doi: 10.1097/WAD.0b013e31823dbcfc.
Lipton, R. B., Katz, M. J., Kuslansky, G., Sliwinski, M. J., Stewart, W. F., Verghese, J., Crystal, H. A., & Buschke, H. (2003). Screening for dementia by
telephone using the memory impairment screen. Journal of the American Geriatrics Society, 51(10), 1382–1390. doi: 10.1046/j.1532-5415.2003.51455.x.
Marc, L. G., Raue, P. J., & Bruce, M. L. (2008). Screening performance of the 15-item geriatric depression scale in a diverse elderly home care population.
American Journal of Geriatric Psychiatry, 16(11), 914–921. doi: 10.1097/JGP.0b013e318186bd67.
Metternich, B., Schmidtke, K., & Hüll, M. (2009). How are memory complaints in functional memory disorder related to measures of affect, metamemory and

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

cognition? Journal of Psychosomatic Research, 66(5), 435–444.
Molinuevo, J. L., Rabin, L. A., Amariglio, R., Buckley, R., Dubois, B., Ellis, K. A. et al. (2017). Implementation of subjective cognitive decline criteria in
research studies. Alzheimer’s & Dementia, 13(3), 296–311.
Moritz, S., Andreou, C., Schneider, B. C., Wittekind, C. E., Menon, M., Balzan, R. P. et al. (2014). Sowing the seeds of doubt: A narrative review on metacognitive
training in schizophrenia. Clinical Psychology Review, 34(4), 358–366.
Morris, J. C., Edland, S., Clark, C., Galasko, D., Koss, E., Mohs, R. et al. (1993). The consortium to establish a registry for Alzheimer’s disease (CERAD) part
IV. Rates of cognitive change in the longitudinal assessment of probable Alzheimer’s disease. Neurology, 43(12), 2457–2457.
Morris, R. G., & Mograbi, D. C. (2013). Anosognosia, autobiographical memory and self knowledge in Alzheimer’s disease. Cortex, 49(6), 1553–1565. doi:
10.1016/j.cortex.2012.09.006.
Moulin, C. J., James, N., Perfect, T. J., & Jones, R. W. (2003). Knowing what you cannot recognise: Further evidence for intact metacognition in Alzheimer’s
disease. Aging, Neuropsychology, and Cognition, 10(1), 74–82.
Nelson, T. O. & Narens, T. (1990). Metamemory: A theoretical framework and new findings. In >Psychology of learning and motivation (Vol. 26, pp. 125–173).
London, UK: Elsevier Academic Press.
Osuna, J., Thomas, K., Edmonds, E., Bangen, K., Weigand, A., Wong, C. et al. (2019). Subtle cognitive decline predicts progression to mild cognitive impairment
above and beyond Alzheimer’s disease risk factors. Archives of Clinical Neuropsychology, 34(6), 846–846. doi: 10.1093/arclin/acz035.14.
Perrotin, A., Belleville, S., & Isingrini, M. (2007). Metamemory monitoring in mild cognitive impairment: Evidence of a less accurate episodic feeling-of-
knowing. Neuropsychologia, 45(12), 2811–2826. doi: 10.1016/j.neuropsychologia.2007.05.003.
Perrotin, A., de Flores, R., Lamberton, F., Poisnel, G., La Joie, R., de la Sayette, V. et al. (2015). Hippocampal subfield volumetry and 3D surface mapping in
subjective cognitive decline. Journal of Alzheimer’s Disease, 48(s1), S141–S150.
Petersen, R. C. (2004). Mild cognitive impairment as a diagnostic entity. Journal of Internal Medicine, 256(3), 183–194.
Petersen, R. C. (2011). Mild cognitive impairment. New England Journal of Medicine, 364(23), 2227–2234.
Petersen, R. C., Doody, R., Kurz, A., Mohs, R. C., Morris, J. C., Rabins, P. V. et al. (2001). Current concepts in mild cognitive impairment. Archives of Neurology,
58(12), 1985–1992.
Piras, F., Piras, F., Orfei, M. D., Caltagirone, C., & Spalletta, G. (2016). Self-awareness in mild cognitive impairment: Quantitative evidence from systematic
review and meta-analysis. Neuroscience & Biobehavioral Reviews, 61, 90–107.
Rabin, L., Wang, C., Katz, M., & Lipton, R. (2014). A psychometric approach to the classification of subjective cognitive decline and mild cognitive impairment.
Paper presented at the Annual Meeting of the International Neuropsychological Society, Seattle, WA.
Rabin, L., Wang, C., Katz, M. J., Derby, C. A., Buschke, H., & Lipton, R. B. (2012). Predicting Alzheimer’s disease: neuropsychological tests, self-reports, and
informant reports of cognitive difficulties. Journal of the American Geriatric Society, 60(6), 1128–1134. doi: 10.1111/j.1532-5415.2012.03956.x.
Rabin, L. A., Smart, C. M., & Amariglio, R. E. (2017). Subjective cognitive decline in preclinical Alzheimer’s disease. Annual review of clinical psychology,
13, 369–396. doi: 10.1146/annurev-clinpsy-032816-045136.
Rahnev, D., Koizumi, A., McCurdy, L. Y., D’Esposito, M., & Lau, H. (2015). Confidence leak in perceptual decision making. Psychological Science, 26(11),
1664–1680.
Rami, L., Mollica, M. A., García-Sanchez, C., Saldaña, J., Sanchez, B., Sala, I. et al. (2014). The subjective cognitive decline questionnaire (SCD-Q): A
validation study. Journal of Alzheimer’s Disease, 41(2), 453–466.
Rattanabannakit, C., Risacher, S. L., Gao, S., Lane, K. A., Brown, S. A., McDonald, B. C. et al. (2016). The cognitive change index as a measure of
self and informant perception of cognitive decline: Relation to neuropsychological tests. Journal of Alzheimer’s Disease, 51(4), 1145–1155. doi:
10.3233/JAD-150729.
Redondo, M. T., Beltrán-Brotóns, J. L., Reales, J. M., & Ballesteros, S. (2016). Executive functions in patients with Alzheimer’s disease, type 2 diabetes mellitus
patients and cognitively healthy older adults. Experimental Gerontology, 83, 47–55.
Reid, L. M., & MacLullich, A. M. (2006). Subjective memory complaints and cognitive impairment in older people. Dementia and Geriatric Cognitive Disorders,
22(5–6), 471–485.
Reisberg, B., & Gauthier, S. (2008). Current evidence for subjective cognitive impairment (SCI) as the pre-mild cognitive impairment (MCI) stage of
subsequently manifest Alzheimer’s disease. International Psychogeriatrics, 20(1), 1–16.
Reisberg, B., Prichep, L., Mosconi, L., John, E. R., Glodzik-Sobanska, L., Boksay, I. et al. (2008). The pre-mild cognitive impairment, subjective
cognitive impairment stage of Alzheimer’s disease. Alzheimer’s & dementia: the journal of the Alzheimer’s Association, 4(1 Suppl 1), S98–S108. doi:
10.1016/j.jalz.2007.11.017.
Reitan, R. M. (1958). Validity of the trail making test as an indicator of organic brain damage. Perceptual and Motor Skills, 8(3), 271–276.
Rosen, W. G. (1980). Verbal fluency in aging and dementia. Journal of clinical and experimental neuropsychology, 2(2), 135–146.
Ryals, A. J., O’Neil, J. T., Mesulam, M.-M., Weintraub, S., & Voss, J. L. (2018). Memory awareness disruptions in amnestic mild cognitive impairment:
Comparison of multiple awareness types for verbal and visuospatial material. Aging, Neuropsychology, and Cognition, 26(4), 577–598.
Sarason, I. G., Levine, H. M., Basham, R. B., & Sarason, B. R. (1983). Assessing social support: The social support questionnaire. Journal of Personality and
Social Psychology, 44(1), 127–139. doi: 10.1037/0022-3514.44.1.127.
1422 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425

Saykin, A., Wishart, H., Rabin, L., Santulli, R., Flashman, L., West, J. et al. (2006). Older adults with cognitive complaints show brain atrophy similar to that
of amnestic MCI. Neurology, 67(5), 834–842.
Schmitter-Edgecombe, M., & Seelye, A. M. (2011). Predictions of verbal episodic memory in persons with Alzheimer’s disease. Journal of Clinical and
Experimental Neuropsychology, 33(2), 218–225. doi: 10.1080/13803395.2010.507184.
Schraw, G. (2009). 21 measuring metacognitive judgments. Handbook of metacognition in education, 415.
Seelye, A. M., Schmitter-Edgecombe, M., & Flores, J. (2010). Episodic memory predictions in persons with amnestic and nonamnestic mild cognitive
impairment. Journal of Clinical and Experimental Neuropsychology, 32(4), 433–441.
Sheikh, J. I., & Yesavage, J. A. (1986). Geriatric Depression Scale (GDS): recent evidence and development of a shorter version. Clinical Gerontologist: The
Journal of Aging and Mental Health.
Silva, A. R., Pinho, M. S., Macedo, L., Souchay, C., & Moulin, C. (2017). Mnemonic anosognosia in Alzheimer’s disease is caused by a failure to transfer

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

online evaluations of performance: Evidence from memory training programs. Journal of Clinical Experimental Neuropsychology, 39(5), 419–433. doi:
10.1080/13803395.2016.1231799.
Smart, C. M., Karr, J. E., Areshenkoff, C. N., Rabin, L. A., Hudon, C., Gates, N. et al. (2017). Non-pharmacologic interventions for older adults with
subjective cognitive decline: Systematic review, meta-analysis, and preliminary recommendations. Neuropsychological Review, 27(3), 245–257. doi:
10.1007/s11065-017-9342-8.
Sohlberg, M. M., & Mateer, C. A. (2001). Improving attention and managing attentional problems. Annals of the New York Academy of Sciences, 931(1),
359–375.
Spreen, O. & Strauss, E. (1998). A compendium of neuropsychological tests: Administration, norms, and commentary. New York, NY: Oxford University Press.
Starkstein, S. E., Jorge, R., Mizrahi, R., Adrian, J., & Robinson, R. (2007). Insight and danger in Alzheimer’s disease. European Journal of Neurology, 14(4),
455–460.
Steward, K. A., Bull, T. P., Kennedy, R., Crowe, M., & Wadley, V. G. (2019). Neuropsychological correlates of anosognosia for objective functional difficulties
in older adults on the mild cognitive impairment Spectrum. Archives of Clinical Neuropsychology: the official journal of the National Academy of
Neuropsychologists, 35(4), 365–376. doi: 10.1093/arclin/acz065.
Striepens, N., Scheef, L., Wind, A., Popp, J., Spottke, A., Cooper-Mahkorn, D. et al. (2010). Volume loss of the medial temporal lobe structures in subjective
memory impairment. Dementia and Geriatric Cognitive Disorders, 29(1), 75–81.
Terry, P. C., Lane, A. M., & Fogarty, G. J. (2003). Construct validity of the profile of mood states—Adolescents for use with adults. Psychology of Sport and
Exercise, 4(2), 125–139.
Thomas, A. K., Lee, M., & Balota, D. A. (2013). Metacognitive monitoring and dementia: How intrinsic and extrinsic cues influence judgments of learning in
people with early-stage Alzheimer’s disease. Neuropsychology, 27(4), 452.
Vilkki, J., Servo, A., & Surma-aho, O. (1998). Word list learning and prediction of recall after frontal lobe lesions. Neuropsychology, 12(2), 268.
Vilkki, J., Surma-aho, O., & Servo, A. (1999). Inaccurate prediction of retrieval in a face matrix learning task after right frontal lobe lesions. Neuropsychology,
13(2), 298.
Wechsler, D. (1987). Manual for the Wechsler memory scale-revised. San Antonio, TX: Psychological Corporation.
Wechsler, D. (1987). Wechsler memory scale-revised. San Antonio, TX: The Psychological Corporation.
Winblad, B., Palmer, K., Kivipelto, M., Jelic, V., Fratiglioni, L., Wahlund, L. O. et al. (2004). Mild cognitive impairment–beyond controversies, towards a
consensus: Report of the international working group on mild cognitive impairment. Journal of Internal Medicine, 256(3), 240–246.
Wolfsgruber, S., Wagner, M., Schmidtke, K., Frolich, L., Kurz, A., Schulz, S. et al. (2014). Memory concerns, memory performance and risk of dementia in
patients with mild cognitive impairment. PLoS One, 9(7), e100812. doi: 10.1371/journal.pone.0100812.

APPENDIX A

First, we established robust norms for the 13 neuropsychological tests utilizing 411 independent EAS participants who were
dementia-free for 3 years, who were not participants in the current study, and whom we refer to as the “robust sample.” Second,
three underlying cognitive factors were identified using a principal component analysis: global/verbal; executive/processing
speed; and memory. Third, for participants in the current study, global/verbal, executive/processing speed, and memory
cognitive domain scores were calculated as the average Z score of each neuropsychological test associated within a given
factor, derived using means and standard deviations (SD) of the robust sample stratified by age group (70–79 and 80 and
above).
MCI was classified in participants whose cognitive domain scores were considerably lower (>1 SD) than the mean
of the robust sample on one or more cognitive factors and who endorsed at least one cognitive complaint on EAS self-
report measures—i.e., items that assess participants’ self-perceptions of their cognitive abilities taken from the Consortium
to Establish a Registry for Alzheimer’s Disease (Morris et al., 1993), a yes-no rating scale of current functioning of
several cognitive domains; or the “cognitive item” from the GDS (Sheikh & Yesavage, 1986), a dichotomous item that
asks participants whether they feel they have “more memory problems than most.” MCI was further subdivided into aMCI
and naMCI. Participants whose cognitive factor Z scores were below 1 SD on memory or memory plus global, and/or
 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425 1423

executive/processing speed domains of the robust sample were classified as aMCI. Participants whose cognitive factor Z
scores were below 1 SD on the executive/processing speed and/or global domains of the robust sample were classified as
naMCI.
SCD was classified in cognitively intact participants (i.e., cognitive factor Z scores for all three domains did not fall
considerably lower [>1 SD] than the mean of the robust sample) who exceeded an optimal cut point for self- and/or informant
complaints. We used cognitive complaints items from previous research (Rabin et al., 2012) to derive scores that were the

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

proportion of “positive” responses (i.e., responses in the direction of having a cognitive problem/change). Subsequently, we
derived an optimal cut point from an ROC analysis, stratified by young-old (age 70–79) and old-old (age 80 and above) groups,
which used the robust sample and was based on the cross-sectional association between the self- or informant complaint and
MCI (see Rabin, 2014). The optimal cut for self-complaint score was 12.5% for the younger group and 22.2% for the older
group. The optimal cut for the informant-complaint score was 21.0% for the younger group and 10.0% for the older group.
HC was classified in 120 cognitively intact participants whose cognitive factor Z scores for all three domains did not fall
considerably lower (>1 SD) than the mean of the robust sample and who did not exceed the optimal cut point for self- and/or
informant-reported cognitive complaints.
Notably, the 13 neuropsychological tests that were used to establish robust norms were: verbal episodic memory/word
learning—free recall from the Free and Cued Selective Reminding Test (Grober & Buschke, 1987); verbal episodic memory/story
recall—Logical Memory I subtest of the Wechsler Memory Scale-Revised (Wechsler, 1987); verbal fluency/word generation
according to an initial letter—Letter Fluency (Spreen & Strauss, 1998); verbal fluency/naming exemplars from a category
–Category Fluency (Rosen, 1980); confrontation naming—short form of the Boston Naming Test (Goodglass, Kaplan, &
Weintraub, 1983); visuomotor tracking, divided attention, and cognitive flexibility—Trail Making Test Parts A and B (Reitan,
1958); and select subtests of the Wechsler Adult Intelligence Scale—Third Edition (Wechsler, 1987), including visuospatial
organization—Block Design, psychomotor processing speed—Digit Symbol-Coding, auditory attention and working memory—
Digit Span, general fund of knowledge—Information, vocabulary level—Vocabulary, and verbal abstraction of categories—
Similarities.

APPENDIX B

Before presentation of each learning trial, the respondent’s attention should be fixed at the point where the Recall Stimulus
Booklet will be positioned. Then say:
I will show you a sheet that has six geometric figures on it. I want you to study the figures so that you can remember as
many of them as possible. You will have just 10 seconds to study the entire display. I will present the figures right here (place
hand at eye level approximately 16 inches in front of respondent). After I take the display away, try to draw each figure exactly
as it appeared and in its correct location on the page.
Repeat instructions and clarify as often as necessary. Then say:
Before we begin the task, I have a few questions for you. How many of the six geometric figures do you think you will
recall after they are displayed for a total of 10 seconds?
∗
Record response in the upper right-hand corner of the response sheet for (Trial 1). ∗ .
Then say: How many of the six geometric figures do you think you will recall after they are displayed a second time for a
total of 10 seconds? (Clarify if necessary).
∗
Record response in the upper right hand corner of the response sheet for (Trail 2). ∗ .
1424 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425

Then say: How many of the six geometric figures do you think you will recall after they are displayed a third time for a
total of 10 seconds? (Clarify if necessary).
∗
Record response in the upper right hand corner of the response sheet for (Trail 3). ∗ .
Then say: How many of the six geometric figures do you think you will recall after a 25 minute delay period where you
are performing other tasks? (Clarify if necessary).
∗
Record response in the upper right hand corner of the response sheet for (DR Trial). ∗ .

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

Open the Recall Stimulus Booklet to the appropriate form and place it down in front of the respondent. Remind participants:
Remember, you will have just 10 seconds to study the entire display. After I take the display away, try to draw each figure
exactly as it appeared and in its correct location on the page.
When ready, expose the stimulus from a distance of ∼16 inches. The booklet may be held at eye level or rested on the table
top (held in an upright position).
It is imperative that the display is exposed for 10 full seconds. Afterward, remove the booklet and say:
Then say: Now draw as many of the figures as you can in their correct location on the page.
Respondent is permitted as much time as necessary and encouraged to draw the designs as precisely as possible
(use of an eraser is permitted). Examiner may encourage guessing. After the respondent indicates being done, ask
the respondent to put pencil down. Turn to (T2) in booklet (and make sure response sheet for T1 is out of view).
Then say:
That was fine. Now I would like to see whether you can remember more of the figures if you have another chance. I will
present the display again for 10 seconds. Try to remember as many of the figures as you can this time, including the ones you
remembered on your last attempt. Try to draw each figure precisely and in its correct location.
Pause to answer any questions, make any behavioral observations, and again expose the stimulus for exactly 10 s. Then,
remove the booklet and have respondent draw responses on (T2). When respondent indicates being done, immediately turn to
(T3) and remove (T2) from view. Then say:
That was fine. Now I would like to see whether you can remember more of the figures if you have another chance. I will
present the display again for 10 seconds. Try to remember as many of the figures as you can this time, including the ones you
remembered on your last attempt. Try to draw each figure precisely and in its correct location.
After the respondent indicates that he/she is finished, remove the response form. Then say:
Try not to forget the display because I may ask you to remember the figures later.
If I ask you about the figures later, how many do you think you will remember?
∗
Record response in the upper right-hand corner of the response sheet for (T3). ∗ .
Turn to the front page of the Response Form and record the time in the space provided.

Delayed Recall

Delay should consist of predominantly questionnaires and verbal tasks. After 25 min, position the response sheet for the
Delayed Recall Trial and say:
Remember the figures I showed you before? I want to see how many you can remember now. I know it sounds difficult, but
try to draw as many of the figures as you can in their correct location on the page. Remember, try to draw them accurately.
Just do the best you can.
After the respondent indicates being finished drawing, remove the Response Form. Record the time, determine the delay
interval in minutes, and also record this number in the appropriate location.
 S. Y. Chi et al. / Archives of Clinical Neuropsychology 36 (2021); 1404–1425 1425

Recognition Trial

Immediately after the Delayed Recall Trial, position the Recognition Booklet in front of the respondent with the card indicating
the form of the test and instructions visible to the administrator. Then say:
Now I will show you some more figures, one at a time. Some were on the display I showed you before and others are new
figures you have not seen before. Say “yes” for those figures that were on the display and say “no” if I show you a figure that
was not on the display. Do you understand?

Downloaded from https://academic.oup.com/acn/article/36/8/1404/6238274 by Columbia University Libraries user on 09 May 2025

How many of the original six figures do you think you will be able to correctly identify?
Now, yes or no, was this one of the figures I showed you before?
∗
Record response in the upper right-hand corner of the response sheet for (Recognition). ∗ .
Respondent is permitted as much time as needed for each item. If respondent says “I don’t know” or something similar,
encourage a response even if it means guessing. Circle “yes” or “no” in the spaces provided for each item on the back page of
the Response Form. Be sure to record the response to each item.