c o r t e x 9 1 ( 2 0 1 7 ) 2 2 1 e2 3 3

Available online at www.sciencedirect.com

ScienceDirect
Journal homepage: www.elsevier.com/locate/cortex

Special issue: Research report

Brainebehavior relationships in source memory:

Effects of age and memory ability

Liesel-Ann Meusel a, Cheryl L. Grady a,b, Patricia E. Ebert a,1 and

Nicole D. Anderson a,b,*
a
Rotman Research Institute, Baycrest, Toronto, Ontario, Canada
b
Departments of Psychiatry & Psychology, University of Toronto, Canada

article info abstract

Article history: There is considerable evidence for age-related decrements in source memory retrieval, but
Received 27 June 2016 the literature on the neural correlates of these impairments is mixed. In this study, we used
Reviewed 26 August 2016 functional magnetic resonance imaging to examine source memory retrieval-related brain
Revised 2 November 2016 activity, and the monotonic relationship between retrieval-related brain activity and
Accepted 30 December 2016 source memory accuracy, as a function of both healthy aging (younger vs older) and
Published online 12 January 2017 memory ability within the older adult group (Hi-Old vs Lo-Old). Participants studied lists of
word pairs, half visually, half aurally; these were re-presented visually in a scanned test
Keywords: phase and participants indicated if the pair was ‘seen’ or ‘heard’ in the study phase. The
Aging Lo-Old, but not the Hi-Old, showed source memory performance decrements compared to
Source memory the Young. During retrieval of source memories, younger and older adults engaged lateral
Functional magnetic resonance and medial prefrontal cortex (PFC) and medial posterior parietal (and occipital) cortices.
imaging The groups differed in how brain activity related to source memory accuracy in dorsal
Brain-behavior anterior cingulate cortex, precuneus/cuneus, and the inferior parietal cortex; in each of
these areas, greater activity was associated with poorer accuracy in the Young, but with
higher accuracy in the Hi-Old (anterior cingulate and precuneus/cuneus) and Lo-Old
(inferior parietal lobe). Follow-up pairwise group interaction analyses revealed that
greater activity in right parahippocampal gyrus was associated with better source memory
in the Hi-Old, but not in the Lo-Old. We conclude that older adults recruit additional brain
regions to compensate for age-related decline in source memory, but the specific regions
involved differ depending on their episodic memory ability.
© 2017 Elsevier Ltd. All rights reserved.

* Corresponding author. Rotman Research Institute, Baycrest, 3560 Bathurst Street, Toronto, Ontario M6A 2E1, Canada.
E-mail addresses: lmeusel@research.baycrest.org (L.-A. Meusel), cgrady@research.baycrest.org (C.L. Grady), Patricia.Ebert@alberta-
healthservices.ca (P.E. Ebert), nanderson@research.baycrest.org (N.D. Anderson).
1
Present address: Rockyview General Hospital, 7007 14 St SW, Calgary, Alberta T2V 1P9, Canada.
http://dx.doi.org/10.1016/j.cortex.2016.12.023
0010-9452/© 2017 Elsevier Ltd. All rights reserved.
222 c o r t e x 9 1 ( 2 0 1 7 ) 2 2 1 e2 3 3
1. Introduction
In daily life, it is often important for us to remember the
context in which we learned information. For example, as
academics, it is important for us to remember where we read
something so that the correct source can be properly cited.
Healthy aging is associated with source memory impairments
(e.g., Anderson et al., 2008; Cansino et al., 2013; Hashtroudi,
Johnson, & Chrosniak, 1989; McIntyre & Craik, 1987;
Wegesin, Jacobs, Zubin, Ventura, & Stern, 2000; see Mitchell
& Johnson, 2009, for review), particularly when strategies to
unitize the memoranda are not encouraged (Bastin et al., 2013;
Dulas & Duarte, 2014), and when item or fact memory per-
formance is not equated between younger and older adult
groups (Schacter, Kaszniak, Kihlstrom, & Valdiserri, 1991). The
objective of this study was to examine age- and memory
ability-related differences in the relationship between brain
activity during source memory retrieval and task accuracy.
1.1. Brain activation during source memory retrieval
Functional neuroimaging studies of young adults have begun
to identify the neural bases of source memory retrieval. Three
brain regions have been most consistently found to be acti-
vated preferentially during retrieval of source compared to
item information: the left prefrontal cortex (PFC), posterior
parietal cortices including the precuneus, and medial tem-
poral lobe regions including the hippocampus and para-
hippocampus (Cansino, Maquet, Dolan, & Rugg, 2002;
Dobbins, Foley, Schacter, & Wagner, 2002; Lundstrom,
Ingvar, & Petersson, 2005; Lundstrom et al., 2003; Rugg,
Fletcher, Chua, & Dolan, 1999). The role of the left PFC in
source memory is supported by the finding of source memory
impairments in patients with left, but not right, unilateral PFC
lesions (Duarte, Ranganath, & Knight, 2005). Posterior parietal
regions are known to be involved in memory retrieval
€ rfel, Werner, Shaefer, von Kummer, &
(Daselaar et al., 2009; Do
Karl, 2009; Huijbers et al., 2013; Hutchinson, Uncapher, &
Wagner, 2009). One study found that activation of posterior
parietal regions was associated with source memory retrieval
success, whereas activation of lateral prefrontal regions was
associated with source memory errors, suggesting that the
latter may be more involved in source memory retrieval
monitoring than retrieval per se (Donaldson, Wheeler, &
Petersen, 2009). The notion that the precuneus plays a cen-
tral role in source memory retrieval is supported by the fact
that continuous theta burst stimulation over the precuneus
prior to retrieval reduced source memory errors in a healthy
sample (Bonnı̀ et al., 2015). Finally, in terms of the role of
medial temporal lobe regions in source memory, the Binding
Items and Context model (Diana, Yonelinas, & Ranganath,
2007; Eichenbaum, Yonelinas, & Ranganath, 2007) proposes
that retrieval of items is supported by the perirhinal cortex,
retrieval of context details is supported by parahippocampal
cortex, and retrieval of item-context bindings is supported by
the hippocampus. Consistent with this model, Diana,
Yonelinas, and Ranganath (2009) reported activation of the
parahippocampal cortex and hippocampus during source
memory retrieval.
1.2. Age-related changes in source memory brain
activation
Given the role of left PFC, posterior parietal cortex, and hip-
pocampus/parahippocampus in source memory retrieval in
young adults, important questions are how activation of these
regions differs with age and how this activation relates to age-
related decrements in source memory. Indirect support for the
involvement of the PFC in older adults' source memory im-
pairments is the finding that source memory impairments
among older compared to younger adults have also been
associated with poorer performance on tests of executive
functioning that are sensitive to frontal lobe damage (e.g.,
Craik, Morris, Morris, & Loewen, 1990; Glisky, Rubin, &
Davidson, 2001). More direct evidence for age-related differ-
ences in left PFC activation during source retrieval comes from
functional neuroimaging studies, but the results are mixed.
Mitchell, Raye, Johnson, and Greene (2006) reported greater
left dorsolateral PFC activity in younger than older adults
during retrieval of one type of source memory (whether items
were presented as pictures or words), but not another
(whether items had appeared on the left or right), relative to
item recognition trials. Cansino et al. (2015) reported age-
related reductions in bilateral PFC activity during correct
versus incorrect source memory retrieval. More recently,
Ankudowich, Pasvanis, and Rajah (2016) reported a multivar-
iate analysis that identified age-related decreases in left PFC
activation during retrieval (when contrasted with encoding).
However, in an electroencephalography study, Swick,
Senkfor, and Van Petten (2006) examined event-related po-
tentials (ERPs) during source memory retrieval and found a
left frontal negative potential 600e1200 msec post stimulus in
older adults that was not present in the younger adults. This
late potential was interpreted as the older adults' use of
alternate strategies to compensate for source memory de-
clines (see Rajah, Languay, & Valiquette, 2010, for a similar
interpretation of older adults' enhanced engagement of this
region in an fMRI study).
In terms of posterior parietal and medial temporal lobe
regions, Kukolja, Thiel, Wilms, Mirzazade, and Fink (2009)
found that retrieval-related activity involving left lateral and
medial parietal cortex, including the precuneus, was greater
in older than younger adults, and associated with more cor-
rect than incorrect spatial source judgments in both groups;
left hippocampal activity associated with correct source
memory only in the younger group. Further evidence for age-
related reductions in bilateral hippocampal activation during
source retrieval is provided by Cansino et al. (2015), and one
study found a relationship between age-related source
memory decrements and hippocampal volume in seniors
(Rajah, Kromas, Han, & Pruessner, 2010). One discrepant
finding is that of Ankudowich et al. (2016), who report age-
related increases in bilateral parahippocampal activation
during source retrieval, relative to source encoding, but that
may be due to the encoding versus retrieval contrast derived
from the analysis. In each of the studies mentioned so far,
younger adults' source memory performance exceeded that of
older adults; collectively the results point to altered (some-
times reduced, sometimes enhanced) activation of left PFC,
 c o r t e x 9 1 ( 2 0 1 7 ) 2 2 1 e2 3 3
intact parietal activation, and reduced hippocampal involve-
ment during source memory retrieval in older compared to
younger adults.
Other studies that included manipulations to ensure equal
performance in younger and older adults have reported rela-
tively more age-invariant patterns of brain activity during
retrieval of source information (Duverne, Habibi, & Rugg, 2008;
Leshikar & Duarte, 2014), or greater activation in older than
younger adults, including in posterior hippocampus (Duverne
et al., 2008), and in bilateral PFC and precuneus (Morcom, Li, &
Rugg, 2007), which was interpreted as less efficient brain
processing in support of source memory in older adults.
Leshikar and Duarte (2014) reported greater left PFC activation
in older than younger adults when the younger cohort per-
formed better (see also Spaniol & Grady, 2012), but not when
performance was equated. A related finding is that of de
Chastelaine, Mattson, Wang, Donley, and Rugg (2016), who
examined associative memory rather than source memory,
although both draw on memory for an item's context: acti-
vation of medial PFC, bilateral hippocampus, and posterior
cingulate during correct identification of intact word pairs
compared with false endorsements of rearranged pairs
declined with age, but this age-related effect was eliminated
when memory performance was statistically controlled.
Clearly, more research is needed to understand age-related
changes in brain activation patterns during source memory,
and how these are influenced by task performance.
All of the above studies contrasted correct and incorrect
trials on a source memory task, included all trials regardless
of accuracy, or included only correct trials, methods which
do not allow for a very fine-grained understanding of the
extent to which activation across these regions varies with
behavior. Two alternate approaches examining individual
differences in the influence of source memory performance
on age-related differences in brain activation are: 1) to recruit
two groups of healthy seniors, one with relatively higher
memory functioning than the other, and, 2) to identify brain
regions in which activation is monotonically associated with
source memory performance. In an early positron emission
tomography study, Cabeza, Anderson, Locantore, and
McIntosh (2002) examined brain activity during a source
memory task compared to a cued recall task in younger
adults and two groups of seniors defined by how well they
performed on neuropsychological tests of memory (c.f.,
Glisky, Polster, & Routhieaux, 1995). Age-related decrements
on both memory tasks were evident only in the lower func-
tioning older group. They found similar activation of right
PFC in all three groups during source compared to associative
memory retrieval, but the higher-functioning older adult
group additionally recruited left PFC, which was interpreted
as evidence of compensatory brain activity. Duarte, Henson,
and Graham (2008) examined source memory, as well as
recollection and familiarity using remember/know judg-
ments. The older adult group was split into Hi-Old and Lo-
Old groups based on their task recognition performance
(remember or know, corrected for false alarms). Compared to
the younger adult group, the Hi-Old group had intact sub-
jective recollection (remember) but impaired objective
recollection (correct source judgments that were accompa-
nied by a remember response), whereas the Lo-Old was
223
impaired on both measures of source recollection. Subjective
recollection was associated with increased activation of left
lateral and bilateral medial parietal cortices comparably in
the Young and Hi-Old, but not the Lo-Old. The objective
recollection results revealed brain areas in which the Hi-Old
showed hyperactivation in comparison to the Young and Lo-
Old; these regions included bilateral PFC and temporal lobe
regions.
We are aware of only two studies that examined the
monotonic relationship between source memory perfor-
mance and retrieval-related brain activity in younger and
older adults. Rajah, Languay, et al (2010) reported significant
Pearson correlations between activity in bilateral dorsolateral
(BA 46) and anterior (BA 10) PFC regions of interest (ROI) and
both spatial and temporal source memory that were numeri-
cally greater in older than young adults. McDonough, Wong,
and Gallo (2013) reported that the association between
source memory performance and activity in a right dorsolat-
eral PFC ROI was positive in young adults, but negative in older
adults. However, limitations of these approaches are that re-
lationships between performance and activation of other
brain regions outside the ROI might have been missed, and
Rajah et al.’s analyses did not control for increased Type I
error associated with multiple correlations in a way that a
thresholded voxel-by-voxel approach would.
1.3. The present study
On balance, prior research on the neural bases of source
memory decrements in aging provides a mixed picture of age-
related differences in left PFC, posterior parietal and pre-
cuneal cortices, and medial temporal lobe regions, and no
study to date has provided a whole-brain analysis of age-
related differences in the monotonic relationship between
source memory performance and brain activity. The primary
goal of the present study was to examine both age-related and
episodic memory performance-related effects on the rela-
tionship between brain activity during a source retrieval task
and task performance. Prior to addressing these
brainebehavior relationships, however, we wanted to ensure
that our task was activating the brain regions expected during
a source memory task, and examine whether there were age-
and/or episodic memory performance-related differences in
source memory retrieval brain activation (regardless of task
accuracy). Healthy younger and older adults, the latter of
whom were split into a Hi-Old and Lo-Old group based on
neuropsychological tests of memory (c.f., Cabeza et al., 2002),
performed a source memory task in the scanner. Word pairs
were presented visually or aurally during an unscanned study
phase. In the scanned test phase, the word pairs were pre-
sented again visually, and participants indicated whether
they had seen or heard that word pair in the study phase. We
predicted intact source memory performance in the Hi-Old
but not the Lo-Old, compared to the young adults (c.f.,
Cabeza et al., 2002). Based on the literature review above, we
predicted source retrieval-related activity in left PFC, posterior
parietal cortices, and medial temporal lobe regions, and that
the positive association between brain activity and source
memory performance in these regions would be stronger in
the Young and Hi-Old than in the Lo-Old.
224 c o r t e x 9 1 ( 2 0 1 7 ) 2 2 1 e2 3 3

LawtoneBrody scale (Lawton & Brody, 1969) to assess activities

2. Material and methods of daily living; Shipley vocabulary test (Zachary, 1986) and
Wechsler Abbreviated Scale of Intelligence (WASI; Wechsler,
2.1. Participants
1999) Matrix Reasoning to estimate intelligence; WASI Block
Design test to measure visuospatial construction; Wechsler
Sixteen younger adults (10 female, Mage ¼ 23.0, age
Adult Intelligence Scale-III (WAIS-III; Wechsler, 1991) Digit Span
range ¼ 19e29) and 32 older adults (24 female, Mage ¼ 70.3, age
and Digit Symbol tests to measure attention; semantic fluency
range ¼ 64e78) were recruited via the participant database at
(Animals; Barr & Brandt, 1996) and a 15-item version of the
the Rotman Research Institute, flyers, newspaper advertise-
Boston Naming Test (Fastenau, Denburg, & Mauer, 1998) to
ments, and community talks. The older adult group for the
assess semantic memory; phonemic fluency (FAS; Benton &
study presented here was selected from another study in
Hamsher, 1976), Wechsler Memory Scale-III (WMS-III;
order to meet particular criteria related to episodic memory
Wechsler, 1997) Mental Control, Wechsler Adult Intelligence
functioning, age, gender, and education, as described in the
Scale-Revised Arithmetic (Wechsler, 1981), and the modified
next section. Participants were included if they were right-
Wisconsin Card Sorting Task (Hart, Kwentus, Wade, & Taylor,
handed (defined as  .85 on the Edinburgh Handedness In-
1988) to measure executive functioning; California Verbal
ventory, Oldfield, 1971), native English speakers, and high
Learning Test-II (Delis, Kramer, Kaplan, & Ober, 2000), WMS-III
school graduates. The older adults also had to pass the
Logical Memory, Wechsler Memory Scale-R (WMS-R;
Modified Telephone Interview for Cognitive Status (Welsh,
Wechsler, 1987) Visual Paired Associates, the WMS-R Verbal
Breitner, & Magruder-Habib, 1993). Exclusion criteria
Paired Associates, and WMS-III Faces I to assess episodic
included a diagnosis of or belief that they had a learning
memory; and, Hospital Anxiety and Depression Scale (Zigmond
disability, hearing aid user, failed criteria for magnetic reso-
& Snaith, 1983) to measure mood.
nance (MR) imaging scanning (under 200 lbs, not claustro-
Following Cabeza et al. (2002), who found altered source
phobic, and no implanted ferromagnetic metals), and the
memory brain activation in seniors with high episodic mem-
presence of a medical or psychiatric condition affecting
ory functioning, we split the older adults into two groups.
cognition or brain imaging, including: dementia; lupus; mul-
Based on the neuropsychological assessment, and relative to a
tiple sclerosis; brain injury; loss of consciousness
larger sample of 217 cognitively intact older adults (aged
lasting > 5 min; encephalitis; meningitis; stroke; transient
64e88, M age ¼ 72.4, SD ¼ 5.8) who had received this same
ischemic attack; epilepsy; brain tumour; migraines; diabetes
battery of assessments, the older adult group was selected
or high blood sugar; uncontrolled thyroid or cholesterol con-
such that there were 16 “Hi-Old” and 16 “Lo-Old”, representing
ditions; hypertension; heart disease; prior myocardial infarct
better (z > .1) or worse (z < .1) than average memory, respec-
or cardiac bypass surgery; arteriosclerosis; peripheral vascular
tively, on an age regressed standardized composite of the
disease; congestive heart failure; heart arrhythmia; heart
CVLT-II long delay cued recall, the WMS-III Logical Memory I
murmur; exposure to toxic chemicals or fumes; chemo-
(first recalls only), the WMS-R Verbal Paired Associates I, the
therapy; radiation to head or neck; chronic obstructive sleep
WMS-R Visual Paired Associates II, and the WMS-III Faces I
apnea; major depression, anxiety, or other psychiatric disor-
(c.f., Glisky et al., 1995), who were additionally matched for
der requiring hospitalization or current depression or anxiety;
gender, and, as closely as possible, for age and education. Note
excessive alcohol use exceeding the Canadian recommenda-
that all participants' memory functioning was within normal
tions (7 or 10 drinks/week for women and men, respectively,
limits. Age regressed standardized composites of executive
Canadian Centre on Substance Abuse, 2011); and, regular drug
functioning were also computed for the older adults, based on
(e.g., marijuana) use. Participants also had to have a better-ear
phonemic fluency, WMS-III Mental Control, WAIS-R Arith-
pure-tone average over 500, 1000, 2000, and 4000 Hz of 40 dB or
metic, and Wisconsin Card Sorting Task categories, although
less, or correctly identify 18 or more words out of 20 in a novel
these were not used to select the participants.
“speech-in-MR-noise” word identification task in which single
multi-syllabic words were presented aurally at 80 dB against a
2.3. fMRI source memory task
recording of 80 dB scanner noise. Participants also had to have
a normal neuropsychological profile, as determined in the first
In the second session, participants practiced the source mem-
session of the study (see 2.2).
ory task in an MR simulator, to familiarize them with the task
In the telephone screening, the Morningness-Eveningness
and ensure that they were not experiencing claustrophobia. The
Questionnaire (Horne & Ostberg, 1976) was also adminis-
third session took place in the scanner, a Siemens 3T Tim Trio
tered. For both sessions, ‘morning’ people were tested in the
using a 12-channel head coil. All images were acquired in an
morning and ‘neutral’ and ‘evening’ people were tested in the
oblique axial plane. Structural imaging included a T1-weighted
afternoon. Participants were paid $110 for three sessions (see
MP-RAGE 3D SPGR structural image (TR ¼ 2000 msec, TE ¼ 2.63,
2.2 and 2.3) and parking or transit costs were reimbursed. This
FOV ¼ 256 mm, 160 slices, 1  1  1 mm voxels), and a T2-
study was approved by the Baycrest Research Ethics Board.
weighted fluid-attenuated inversion recovery sequence
(TR ¼ 9000, TE ¼ 96, 165 flip angle, T1 ¼ 2000, FOV ¼ 220, 30
2.2. Neuropsychological assessment slices, 1.1  .9  5 mm voxels). The functional scans for the
source memory task were T2*-weighted EPI sequences
In the first session, participants were administered a 3- (TR ¼ 2000, TE ¼ 20, 70 flip angle, FOV ¼ 200, 30 slices,
h neuropsychological assessment. This consisted of the 3MS 3.125  3.125  5 mm voxels, yielding 66 TRs per run). Partici-
(Teng & Chui, 1987) to assess general cognitive status; pants were fitted with MR-compatible corrective eyeglasses if
 c o r t e x 9 1 ( 2 0 1 7 ) 2 2 1 e2 3 3
needed. Visual stimuli were projected onto a screen viewed by
participants via a mirror positioned in the scanner bore, and
auditory stimuli were presented at 80 dB using headphones
(Avotec, Inc.). Responses were collected using four-button fibre-
optic response pads (Current Designs, Inc.). Stimulus delivery
and response collection were performed using an E-Prime
(Version 1.2) program (Psychology Software Tools, Inc.).
For the source memory task, a list of 36 weakly related word
pairs (e.g., minister-satin) was created and split into three sets
such that word frequency, word length, and number of sylla-
bles did not differ across pair member or sets. The order of
the sets and pairs within sets were randomized separately for
each person. There were three runs of the source memory task.
Each run began with an unscanned encoding phase, in which 12
pairs were presented. Half of the noun pairs were presented
visually in black 36-point Tahoma font against a white back-
ground, and half were presented aurally, in a random order
with the restriction that no more than four trials occurred
sequentially in the same modality. Each trial lasted 2 sec, with
either the visual noun pair presented for that duration, or the
first auditory pair member presented within the first
1000 msec, and the second auditory pair member presented
within the second 1000 msec. Participants were instructed to
try to remember whether they saw or heard each word pair, as
they would be asked for this information in the later test phase.
A black fixation cross was presented for 1500 msec after
each trial. On one trial, chosen at random, the fixation cross
was red. Participants were instructed to press a button when
they saw the red fixation cross. This was done to prevent
participants from closing their eyes during the encoding
phase (a strategy one could use to improve source memory by
encoding only the auditory pairs).
After the last pair was presented, there was a 5 sec break,
after which the list was repeated in exactly the same way. The
list was repeated based on pilot testing to avoid behavioral
floor effects. In order to eliminate recency effects, after the
second presentation of the list, participants were shown eight
numbers (1e8) sequentially in a random order, each for
1500 msec, and had to indicate by button press if the number
was odd or even.
In the scanned test phase of each run, each studied word
pair and six control trials (i.e., XXXXXeXXXXX) were pre-
sented visually for 2 sec each, in a random order with the
restriction that no more than four trials of each type (seen at
study, heard at study, control) occurred in a row, each fol-
lowed by a jittered 2, 4, or 6 sec fixation cross interstimulus
interval (ISI), with equal numbers of each fixation duration.
Participants pressed a button with the index finger of one
hand to indicate that they had “seen” that pair in the study
phase, with the index finger of their other hand to indicate
that they had “heard” that pair in the study phase, and with
either index finger in response to a control trial. Assignment of
hand to response (seen, heard) was counterbalanced across
participants. Participants had 4 sec to respond (i.e., up to 2 sec
after the word pair or control stimuli disappeared).
2.4. Demographic and neuropsychological data analysis
Univariate ANOVAs were used to examine group differences
in demographics (age and education), neuropsychological
225
variables, and source memory performance. Source memory
reaction time data were analyzed only for correct responses.
Partial h2 is presented as a measure of effect size. Sidak post-
hoc tests were used to identify pairwise differences for sig-
nificant group main effects.
2.5. Image preprocessing and analysis
2.5.1. Preprocessing
Preprocessing was carried out in AFNI (ver 16.1.15; May 11,
2016) according to the following steps: physiological motion
correction, removal of the first 6 TRs of each functional run to
allow for stabilization of the magnetic field, interleaved slice
time correction, alignment of the anatomical and functional
datasets, motion correction, spatial normalization to Talair-
ach space and resampling to an isotropic voxel size of
3  3  3 mm, spatial smoothing with a Gaussian kernel of
4 mm FWHM, and voxelwise scaling.
2.5.2. Individual-level analysis
A general linear model was constructed for each participant,
which included two event-related task regressors of interest
(representing the presentation of the word pair and control
stimuli), each modeled by seven basis functions from 0 to
12 sec after stimulus onset. Note that all word pair trials,
correct and incorrect, were included in the analysis as we
were interested in the monotonic relationship between task-
related brain activity and task performance. Six motion re-
gressors were included as covariates of no interest, and the
data were highpass filtered prior to analysis. Based on
censoring of TRs related to motion and signal outliers, one
participant in the Young group was excluded from the anal-
ysis because of excessive motion (almost one-third of TRs
required censoring because of motion > .3 mm). Another
participant in the Young group was excluded as a significant
signal outlier in the group analysis, affecting all imaging
comparisons.
2.5.3. Group analysis
To measure task-related brain activation while controlling for
the visual and motor aspects of the task, word pair stimuli,
representing brain activation during source retrieval (SRC),
were contrasted with control stimuli (CON). First, we exam-
ined group differences in source memory retrieval-related
brain activity using an ANOVA (3dMVM, Chen, Adleman,
Saad, Leibenluft, & Cox, 2014), with Group (Young, Hi-Old,
Lo-Old) as the between-subjects factor and Task (SRC, CON)
as the within-subjects factor. Next, to explore associations
between task-related brain activation and task accuracy
across groups, we constructed an ANCOVA (3dMVM, Chen
et al., 2014) on the SCR-CON contrast, with Group as the
between-subjects factor (Young, Hi-Old, Lo-Old), and source
memory accuracy (proportion correct) entered as a covariate
of interest. For all analyses, the resultant voxelwise statistical
maps were thresholded at p < .01 and cluster-wise correction
was applied using 3dClustSim (AFNI, ver 16.1.15), using two-
sided thresholding and incorporating an estimated non-
Gaussian autocorrelation function. Based on this, voxel clus-
ters greater than 30 were considered significant at a family-
wise alpha of p < .05.
226 c o r t e x 9 1 ( 2 0 1 7 ) 2 2 1 e2 3 3

outperformed the Lo-Old, or that memory composite scores

3. Results were higher in the Hi-Old than Lo-Old.

3.1. Group demographics and neuropsychological

3.2. Source memory performance
performance

The three groups differed in the likelihood of correctly

Table 1 displays demographic and neuropsychological test
detecting the red fixation crosses that were presented during
performance data for the three groups. The two older adult
the encoding phases, F(2, 43) ¼ 4.20, p ¼ .022, h2 ¼ .163. The
groups did not differ in age, but despite our best attempts to
Young responded correctly to more red fixations (M ¼ 5.34,
match them on education, the Hi-Old had acquired signifi-
SD ¼ 1.60, out of a maximum of six e three lists each pre-
cantly more years of education than the Lo-Old (with the
sented twice) than did the Hi-Old group (M ¼ 2.87, SD ¼ 2.73),
Young gaining an intermediate level of education that did not
with the Lo-Old group not differing from either of the other
differ from either older adult group). Vocabulary scores were
two groups (M ¼ 3.63, SD ¼ 2.58). Many participants reported
higher in the two older adult groups than in the Young. Overall
having forgotten about the instructions to press a button in
cognition (3MS) was higher in the Young and Hi-Old than Lo-
response to the red fixation cross. One would expect that
Old. All three groups had similar levels of language skills
greater misses of the red fixation cross would associate with
(naming, semantic fluency), and, importantly, executive
better task performance if participants were adopting the
functioning, and mood. Indeed, executive composite scores
strategy to close their eyes during encoding. Thus, we would
did not differ between the Hi-Old and Lo-Old. There was an
expect the Young to have the lowest accuracy, the Hi-Old the
age-related difference favoring the young in fluid intelligence,
highest, with the Lo-Old in between; this was not the case at
visuospatial construction, working memory (Lo-Old differed
all, as we show next, and so taken together with the self-
from the young, but the two older adult groups did not differ
report data, we are confident that participants were not clos-
on this measure), and psychomotor speed, consistent with
ing their eyes during encoding. During source recall, the
much previous research. The two older groups were selected
groups did not differ in the number of response omissions
based on their memory functioning, and thus it is not sur-
(M < 1), F(2, 43) < 1, p ¼ .616, h2 ¼ .022.
prising that on most memory measures, the Hi-Old
Source memory performance is displayed in Table 2. A
significant effect of Group on source memory accuracy, F(2,
43) ¼ 5.01, p ¼ .011, h2 ¼ .189, was driven by significantly higher
Table 1 e Demographic and neuropsychological test
performance for the Young, Hi-Old, and Lo-Old. accuracy in the Young (M ¼ .90, SD ¼ .06) than Lo-Old group
(M ¼ .79, SD ¼ .13). The Hi-Old group's source memory accu-
Young Hi-Old Lo-Old p- racy (M ¼ .85, SD ¼ .11) was intermediate and not significantly
value
different from the other two groups. Accuracy to ‘seen’ versus
N 14 16 16 ‘heard’ words did not differ overall, F(1, 43) < 1, p ¼ .499,
Age (Yrs)a 23.4 (3.0) 70.1 (3.6) 70.4 (4.0) h2 ¼ .011, but this did interact with Group, F(2, 43) ¼ 4.61,
Education (Yrs)b 15.8 (1.5) 17.3 (2.4) 15.4 (2.3) .035
p ¼ .015, h2 ¼ .176. Paired t-tests within Group revealed that
3MSe 97.4 (2.0) 97.1 (2.4) 94.8 (4.0) .035
the Young adults were more accurate in making ‘heard’ than
Shipley Vocabularya 31.7 (3.1) 36.3 (3.1) 35.4 (2.8) <.001
Matrix Reasoningc 29.2 (2.5) 24.9 (4.5) 23.6 (5.6) .003 ‘seen’ source judgments, t(13) ¼ 3.21, p ¼ .007, h2 ¼ .442, but no
Block Designc 59.5 (8.1) 41.7 (13.3) 35.0 (13.9) <.001 differences in source memory accuracy by modality were
Digit Span Totald 21.8 (3.6) 19.6 (3.9) 16.8 (4.1) .004 evident in the Hi-Old or Lo-Old (p > .147). When the interaction
Digit Symbol Totalc 90.8 (13.5) 67.5 (10.3) 68.8 (13.0) <.001 was decomposed within modality, memory for visual items
Boston Naming 14.1 (1.0) 14.3 (.8) 14.1 (1.5) .790 did not differ between groups (F < 1), but memory for heard
Semantic Fluency 22.7 (5.4) 18.8 (5.6) 18.4 (6.8) .116
items did, F(2, 43) ¼ 9.13, p < .001, h2 ¼ .298, with the Young
Phonemic Fluency 53.4 (10.8) 44.88 (9.2) 46.6 (10.2) .064
Mental Control 31.5 (4.9) 28.1 (6.0) 28.9 (5.7) .230
outperforming only the Lo-Old. Thus, age differences were
Arithmetic 11.1 (3.3) 13.1 (3.8) 12.6 (2.8) .278 eliminated when the study context was reinstated at retrieval,
WCST Categories 6.0 (.0) 5.3 (1.4) 5.2 (1.3) .120 and while neural activity in the Young group may have been
Executive Composite (z) n/a .03 (.57) .05 (.50) .883 influenced by this modality effect, this seems unlikely to be
CVLT-II Long Delay 13.5 (2.3) 13.2 (1.8) 10.2 (2.3) <.001 the case in the two older adult groups.
Cued Recalle
Time to make a source memory decision also differed
Logical Memory Ic 33.8 (5.6) 28.9 (4.3) 24.5 (5.8) <.001
among groups, F(2, 43) ¼ 10.22, p < .001, h2 ¼ .322, but this
Verbal Paired If 22.5 (1.2) 20.3 (2.0) 17.1 (2.8) <.001
Visual Paired IIe 6.0 (.0) 5.9 (.3) 5.0 (1.1) <.001
Faces Id 40.2 (3.5) 37.1 (3.9) 34.2 (3.9) <.001
Table 2 e Source memory performance for the Young, Hi-
Memory Composite (z) n/a .35 (.24) .54 (.25) <.001
Old, and Lo-Old.
HADS Anxiety 5.7 (3.0) 4.8 (3.1) 5.3 (3.6) .719
HADS Depression 3.6 (3.2) 1.7 (1.5) 2.3 (2.6) .108 Young Hi-Old Lo-Old
Note: WCST ¼ Wisconsin Card Sorting Test; CVLT-II ¼ California Accuracy
Verbal Learning Test-II; HADS ¼ Hospital Anxiety and Depression Seen .85 (.12) .83 (.13) .81 (.15)
Scale. Heard .95 (.05) .85 (.13) .74 (.18)
a
Young < Hi-Old ¼ Lo-Old; b Hi-Old > Lo-Old; c Young > Hi-Old ¼ Lo- Reaction Time (ms)
Old; d Young > Lo-Old; e Young ¼ Hi-Old > Lo-Old; f Young > Hi- Seen 1150 (120) 1526 (339) 1526 (213)
Old > Lo-Old. Heard 1266 (212) 1558 (281) 1541 (257)
 c o r t e x 9 1 ( 2 0 1 7 ) 2 2 1 e2 3 3 227

reflected a simple age effect, with the Young responding interactions in several brain regions, including dorsal anterior
significantly faster (M ¼ 1185 msec, SD ¼ 151) than either the cingulate cortex, midline precuneus and cuneus, and the right
Hi-Old (M ¼ 1517, SD ¼ 271) or Lo-Old (M ¼ 1472, SD ¼ 251) who inferior parietal lobule (Table 4). To better understand these
responded at comparable speeds. Source memory decision interactions, we extracted mean task-related activation
time did not differ as a function of studied item modality, F(1, across each region (SRC e CON) and correlated this with task
43) ¼ 3.28, p ¼ .077, h2 ¼ .071, which also did not interact with accuracy, within each group. In each region, we found that the
Group, F(2, 43) ¼ 1.03, p ¼ .365, h2 ¼ .046. relation between mean task-related activation and source
memory accuracy differed between groups. Specifically, sig-
nificant negative correlations between accuracy and task-
3.3. Source memory brain activation
related activation in each region were evident in Young
adults. Significant positive associations between accuracy and
The ANOVA on task-related brain activation revealed a main
activation in dorsal anterior cingulate and posterior parietal
effect of Group in medial parietooccipital regions, primarily,
regions (precuneus and cuneus) were found for the Hi-Old,
but also in medial and lateral prefrontal regions, and the infe-
while a significant positive association between accuracy
rior parietal lobule, bilaterally. We do not interpret these Group
and activation in the right inferior parietal lobule appeared for
main effects further, as they could be artefacts of age-related
the Lo-Old (Fig. 1).
effects on cerebrovasculature (Samanez-Larkin & D'Esposito,
To gain a better understanding of the Group by Accuracy
2008). Main effects of Task (SRC e CON) were present across
interaction, we examined pairwise interaction comparisons.
medial and lateral parietal regions, medial and lateral frontal
The Young versus Hi-Old comparison and the Young versus
cortex, and the bilateral insula (Table 3). There were no regions
Lo-Old comparison both showed Group by Accuracy in-
that showed significant Group  Task interactions.
teractions within dorsal anterior cingulate cortex, the right
inferior parietal lobule, and midline posterior regions
3.4. Relationship between source memory accuracy and including precuneus and cuneus, with the Young versus Hi-
brain activation Old comparison also identifying Group by Accuracy in-
teractions in the bilateral insula, right angular gyrus, and left
The ANCOVA on task-related brain activation (SRC e CON) by middle frontal gyrus (Table 5). Again, to better understand
Group (Young, Hi-Old, Lo-Old) found Group by Accuracy these interactions, we extracted mean activation across each
region and correlated this with task accuracy, for each group.
Each of these regions showed the same age-related patterns
Table 3 e Effects of group (Young, Hi-Old, Lo-Old) and task
(SRC, CON) on task-related activation. that were evident in the three-group ANCOVA e positive as-
sociations between source memory accuracy and brain ac-
Cluster Side x y z tivity in the Hi-Old and Lo-Old, and negative associations in
Size the Young. Comparison of the Hi-Old and Lo-Old group
Effect of Group revealed that in the right parahippocampal gyrus, the Hi-Old
Precuneus/posterior cingulate 262a L 0 42 43 group showed a significant positive association between
Posterior cingulate 58a R 12 51 8
activation and task accuracy, whereas the Lo-Old group
Posterior cingulate 41b L 11 55 11
showed a negative, albeit non-significant association (Table 5
Cuneus 65a L 6 88 18
Middle occipital gyrus 59a L 52 59 3 and Fig. 2).
Precentral/postcentral gyrus 204a L 35 25 51
Middle temporal gyrus 70a L 41 66 18
Medial frontal gyrus 85a R 1 48 13 4. Discussion
Middle frontal gyrus 50a L 34 14 38
Inferior parietal lobule 40b L 54 30 37
In the present study, we compared brain activation during
Inferior parietal lobule 41b R 38 34 42
Effect of Task (SRC e CON)
source memory retrieval and associations between brain
Medial/posterior parietooccipital 3563a L 11 69 21 activation and task accuracy in younger and cognitively-intact
48a R 3 49 31 older adults, the latter of whom were split into Lo-Old and Hi-
Lateral frontal/temporal/parietal 1193a R 53 32 20
Lateral frontal 1179a L 39 18 28
Medial prefrontal/anterior 639a R 3 45 20 Table 4 e Group (Young, Hi-Old, Lo-Old) by task accuracy
cingulate interactions on task-related (SCR-CON) activation.
Lateral frontal/parietal 633a L 46 21 25
Cluster Size Side x y z
147a R 36 24 56
Lateral frontal 543a R 39 31 16 Group £ Accuracy Interaction
103a R 29 2 51 Dorsal anterior cingulate 61a R 2 22 37
Dorsal anterior cingulate/medial 515a R 1 16 46 Precuneus/cuneus 53a L 3 76 30
frontal 43a L 0 58 41
Paracentral lobule 212a R 2 20 45 Inferior parietal lobule 39b R 38 53 41

Note: xyz coordinates (in Talairach space) are reported for each Note: xyz coordinates (in Talairach space) are reported for each
cluster's center of mass. cluster's center of mass.
a a
p < .01. p < .01.
b b
p < .02. p < .02.
228 c o r t e x 9 1 ( 2 0 1 7 ) 2 2 1 e2 3 3

Fig. 1 e Group by accuracy interactions on task-related activation. For each brain region identified by the Group by Accuracy
interaction, mean task-related activation (SRC e CON) was plotted against task accuracy, for each group. The two posterior
midline regions identified in the interaction (cuneus and precuneus) were averaged together for presentation because they
showed a similar pattern of findings. Young adults showed significant negative correlations between task accuracy and
task-related activation across regions. The Hi-Old group showed significant positive associations between accuracy and
activation in dorsal anterior cingulate and posterior parietal regions (precuneus and cuneus), while the Lo-Old group
showed a significant positive association between accuracy and activation in the right inferior parietal lobule.

Old groups based upon their performance on neuropsycho-
logical tests of episodic memory. We predicted that source
memory performance would be impaired in the Lo-Old, but
not in the Hi-Old, compared to younger adults' performance.
This behavioral pattern was confirmed, with the Young dis-
playing better source memory than the Lo-Old; the Hi-Old's
performance was intermediate and not significantly different
from either of the other two groups.
4.1. Group differences in source memory retrieval-
related brain activity
Although our primary goal was to investigate group differ-
ences in the relationship between brain activation and source
memory performance, we also wanted to confirm that our
task recruited the expected brain regions based on the litera-
ture. Specifically, we predicted to find source memory-related
activations in left PFC, posterior parietal cortices including the
precuneus, and medial temporal lobe regions including hip-
pocampus and parahippocampal cortex (Cansino et al., 2002;
Dobbins et al., 2002; Lundstrom et al., 2005, 2003; Rugg et al.,
1999). Indeed, in an initial analysis identifying regions that
were more activated following a source memory cue than our
control stimuli (regardless of task accuracy), we did find acti-
vation across prefrontal regions, bilaterally, and across pos-
terior parietal cortices extending into the occipital lobe.
However, this was not true for the medial temporal lobe, and
there were no regions of source retrieval activation that
differed across groups. By including older adults with high
episodic memory abilities, we may have minimized our ability
to detect group effects on task-related brain activity, particu-
larly considering that task accuracy in this Hi-Old group did
not differ with either the Young group or the Lo-Old group.
Thus, this failure to find group differences in source memory
retrieval-related brain activity may reflect the relative age-
invariance in brain activation seen in other studies when
 c o r t e x 9 1 ( 2 0 1 7 ) 2 2 1 e2 3 3 229

Table 5 e Pairwise Interaction Comparisons of Group and Task Accuracy on task-related activation (r-values are the
correlations between activity and task accuracy for each group).

Cluster Size Side x y z r-value

Young/Hi-Old by Accuracy Y Hi
Precuneus/cuneus 171a R 1 70 38 .713 .565
Dorsal anterior cingulate 148a R 1 15 38 .632 .730
Inferior parietal lobule 55a R 39 53 42 .745 .355
Inferior frontal gyrus/anterior insula 53a R 47 13 0 .712 .628
Insula 50a L 36 0 1 .636 .775
Angular gyrus 34b R 47 68 35 .653 .669
Middle frontal gyrus 31b L 34 46 22 .834 .417
Young/Lo-Old by Accuracy Y Lo
Dorsal anterior cingulate 83a R 1 15 41 .705 .518
Precuneus/cuneus 74a L 1 75 32 .724 .256
68a L 0 59 41 .655 .432
Inferior parietal lobule 61a R 38 54 42 .738 .569
32b L 52 32 43 .735 .258
Hi-Old/Lo-Old by Accuracy Hi Lo
Parahippocampal gyrus 30b R 25 3 16 .760 .409

Note: xyz coordinates (in Talairach space) are reported for each cluster's center of mass. Y ¼ Young; Hi ¼ Hi-Old; Lo ¼ Lo-Old.
a
p < .01.
b
p < .05.

Fig. 2 e Pairwise interaction comparison between the Hi-Old and Lo-Old group. There was a significant Group (Hi-Old vs. Lo-
Old) by Accuracy interaction in the right parahippocampal gyrus. The graph reflecting this interaction plots mean activation
across this region as a function of task performance within each older adult group. As evidenced by the graph, greater
activation supported better performance among the Hi-Old adults, but was not significantly related to performance among
the Lo-Old adults.

behavioral performance is fairly comparable across groups
(Duverne et al., 2008; Leshikar & Duarte, 2014).
4.2. Group differences in the relationship between source
memory accuracy and brain activation
The primary focus of this study was to examine age-related
differences in the association between whole brain activation
during a source memory task and task accuracy. In addition,
within the older adult group, we examined how these asso-
ciations were modulated by episodic memory ability. We
predicted stronger associations between source memory-
related brain activity and task accuracy in the Young and Hi-
Old groups, compared to the Lo-Old group, in PFC, posterior
parietal regions, and medial temporal lobe regions. We found
that greater activation of posterior parietal regions, including
the precuneus, and dorsal anterior cingulate cortex was
associated with better source memory accuracy in the Hi-Old,
while greater activation of the right inferior parietal lobule
was associated with better source memory in the Lo-Old.
However, pairwise comparisons between each older adult
group and the young adult group showed positive activation-
accuracy associations in all three of these regions in both the
Hi-Old and Lo-Old groups. Thus, there appears to be a general
age-related increase in the association of engagement of re-
gions associated with retrieval success (posterior parietal and
anterior cingulate cortex; Donaldson et al., 2009) and source
memory performance, suggesting compensatory activity to
offset declines in source retrieval accuracy.
In the Young/Hi-Old comparison, the Hi-Old group showed
significant associations between task accuracy and task-
related activation in additional brain regions, including left
230 c o r t e x 9 1 ( 2 0 1 7 ) 2 2 1 e2 3 3
PFC and bilateral insula. In general, recruitment of left PFC
regions to support source memory performance is consistent
with a number of previous studies (e.g., Cabeza et al., 2002;
Duarte et al., 2008; Morcom et al., 2007; Rajah, Languay, et al.,
2010; Spaniol & Grady, 2012), and the positive relationship be-
tween activation and task accuracy in the Hi-Old group could
reflect either compensatory engagement (c.f., Cabeza et al.,
2002; Rajah, Languay, et al., 2010), or the increased need for
retrieval monitoring to support performance (Donaldson et al.,
2009). Insular cortices, along with dorsal anterior cingulate
regions, form part of a salience network (Seeley et al., 2007)
involved in directing awareness towards behaviorally-relevant
stimuli, whether it be external sensory information or internal
physiological signals (e.g., emotional and bodily state infor-
mation; see Uddin, 2015 for a review). Functional subdivision of
the insula implicates dorsal anterior portions in cognitive
processing (Chang, Yarkoni, Khaw, & Sanfey, 2013), and the
right anterior insula in particular, in the context of the salience
network, plays an important role in “switching” between
engagement of the central executive network and the default-
mode network, making it a key player in cognitive control
processes (Menon & Uddin, 2010; Sridharan, Levitin, & Menon,
2008). Not surprisingly then, in relation to memory perfor-
mance, a number of studies have linked insula and anterior
cingulate activation to error awareness (Cosentino et al., 2015;
Grady, St.-Laurent, & Burianova  , 2015; Hester, Fassbender, &
Garavan, 2004), and to monitoring and decision-making more
generally (e.g., Harle, Chang, van't Wout, & Sanfey, 2012; Bastin
et al., 2016). In the present study, the positive relation between
activation and task accuracy in the Hi-Old suggests that
engagement of these regions is a compensatory process in
support of task performance, perhaps reflecting greater effort
to stay on task and monitor performance.
The pairwise comparison of Hi-Old and Lo-Old revealed an
association between activation of the right parahippocampal
gyrus and source memory accuracy in the Hi-Old; this finding
echoes that of Ankudowich et al. (2016), who found age-related
increases in parahippocampal activation during source mem-
ory retrieval, but our results suggest that these increases may be
most evident in older adults with better episodic memory skills.
Previous work supports a role for the parahippocampal cortex in
source memory retrieval (Diana et al., 2009) and memory
for contextual information (Diana et al., 2007). The differential
association between parahippocampal engagement and task
accuracy as a function of episodic memory performance
may indicate that these high memory performers, via reserve
mechanisms, have been able to maintain integrity of this region
in light of known age-related reductions in parahippocampal
volume (e.g., Fjell et al., 2009; Persson et al., 2016).
In contrast to their older counterparts, a uniform finding in
the Young group was that source memory retrieval-related
brain activation was negatively associated with task accu-
racy. We believe the consistently negative activationeaccur-
acy relationship was a function of the relatively high accuracy
rate amongst the young adults, which ranged from 78% to
97%. Our data suggest that among the younger adults, the
better their source memory, the less they needed to engage
lateral and medial PFC, posterior parietal cortices including
precuneus, and bilateral insula. This should not be construed
as meaning that younger adults do not rely on these regions
for source memory retrieval, however, as our examination of
overall task effects (i.e., regardless of accuracy) identified
these regions as being more active during source retrieval
trials than during control trials across all groups.
It is also possible that younger adults were engaging these
regions differently than the older adults during task perfor-
mance. For example, different functional subregions of the
precuneus appear to be differentially associated with memory
task performance. The ventral precuneus and posterior
cingulate cortex, as well as other default-mode regions, have
been implicated in the encoding/retrieval “flip”, which de-
scribes the finding of performance-related deactivations dur-
ing memory encoding, and performance-related activations
during retrieval (Huijbers et al., 2012; Huijbers et al., 2013). In
contrast, the dorsal precuneus shows performance-related
deactivations across both encoding and retrieval (Huijbers
et al., 2013). Further, these subregions show differential con-
nectivity with a number of brain networks, including the
default mode network, the dorsal attention network, and the
frontoparietal control network (Huijbers et al., 2013; Utevsky,
Smith, & Huettel, 2014; Zhang & Li, 2012). Thus, in the pre-
sent study, it is possible that the poorer performing younger
adults were engaging regions of the default-mode network
more during the source memory task, and not attending to the
task as much as their older counterparts (c.f. Rypma &
D'Esposito, 2000; Rypma, Berger, Genova, Rebbechi, &
D'Esposito, 2005 for similar negative accuracy/activation as-
sociations in younger adults). The fact that young adults
identified more of the red fixation crosses than did older
adults may also play a role, as it is possible that some younger
adults engaged in less source encoding (and instead focused
on detecting the red fixation crosses) which may have resulted
in poorer performance and less source memory activity.
5. Limitations
Despite providing new insights into differences in neural
mechanisms supporting source memory retrieval as a func-
tion of aging and episodic memory ability, this study did have
some limitations. The group sizes were relatively small, which
likely limited power to detect differences in source memory
performance between the Hi-Old and either the Young or Lo-
Old groups. Furthermore, although presenting the study list
twice prevented floor effects in the older groups, it may have
rendered the task too simple for the young adults. We also
remind readers that the Young group only had better source
memory for word pairs that were encoded aurally than visu-
ally; we did not analyze brain activity as a function of modality
due to power issues, but we doubt that this modality effect
had a large influence on our results, as temporal and occipital
regions involved in audition and vision, as well as in source
memory reinstatement effects (Thakral, Wang, & Rugg, 2015)
were not evident in any of the results.
6. Conclusions
The present study found that during retrieval of source
memories, compared to a control condition, young and older
 c o r t e x 9 1 ( 2 0 1 7 ) 2 2 1 e2 3 3
adults engaged lateral and medial PFC and medial posterior
parietooccipital cortices. Analysis of group differences in how
this brain activation related to source memory accuracy
revealed positive associations in dorsal anterior cingulate,
medial posterior parietal, and right inferior parietal cortices in
older adults, suggestive of an age-related increase in the
recruitment of these regions to compensate for age-related
source memory declines. Those with better episodic mem-
ory showed additional, positive activity-accuracy associations
in left PFC, bilateral insula, and the right angular gyrus
compared to Young, and in the right parahippocampus
compared to those older adults with lower episodic memory
performance. We argue that the consistent finding of negative
activity-accuracy source memory associations in the Young is
a by-product of their relatively good performance on the task.
Taken together, these findings suggest an age- and ability-
related shift in engagement of brain networks mediating
retrieval processes.
Funding
This study was supported by the Canadian Institutes of Health
Research (MOP 67015).
Acknowledgments
We thank Sandra Priselac, Andrea Maione, Brian Mainland,
Magdelena Lysenko, Sarah Waters-Schulte, and Lu Jin for their
assistance in recruiting participants and running the study.
references
Anderson, N. D., Ebert, P. L., Jennings, J. M., Grady, C. L.,
Cabeza, R., & Graham, S. J. (2008). Recollection- and
familiarity-based memory in healthy aging and amnestic mild
cognitive impairment. Neuropsychology, 22, 177e187. http://
dx.doi.org/10.1037/0894-4105.22.2.177.
Ankudowich, E., Pasvanis, S., & Rajah, M. N. (2016). Changes in the
modulation of brain activity during context encoding vs.
retrieval across the adult lifespan. NeuroImage, 139, 103e113.
http://dx.doi.org/10.1016/j.neuroimage.2016.06.022.
Barr, A., & Brandt, J. (1996). Word-list generation deficits in
dementia. Journal of Clinical and Experimental Neuropsychology,
18, 810e822.
Bastin, J., Deman, P., David, O., Gueguen, M., Benis, D., Minotti, L.,
et al. (2016). Direct recordings from human anterior insula
reveal its leading role within the error-monitoring network.
Cerebral Cortex. http://dx.doi.org/10.1093/cercor/bhv352 [Epub
ahead of print].
Bastin, C., Diana, R. A., Simon, J., Collette, F., Yonelinas, A. P., &
Salmon, E. (2013). Associative memory and aging: The effect of
unitization on source memory. Psychology and Aging, 28,
275e283. http://dx.doi.org/10.1037/a0031566.
Benton, A. L., & Hamsher, K. de S. (1976). Multilingual aphasia
examination manual. Iowa City: University of Iowa.
Bonnı̀, S., Veneiro, D., Mastropasqua, C., Ponzo, V.,
Caltagirone, C., Bozzali, M., et al. (2015). TMS evidence for a
selective role of the precuneus in source memory retrieval.
231
Behavioral Brain Research, 282, 70e75. http://dx.doi.org/10.1016/
j.bbr.2014.12.032.
Cabeza, R., Anderson, N. D., Locantore, J. K., & McIntosh, A. R.
(2002). Aging gracefully: Compensatory brain activity in high-
performing older adults. NeuroImage, 17, 1394e1402. http://
dx.doi.org/10.1016/nimg.2002.1280.
Canadian Centre on Substance Abuse. (2011). Alcohol and Health in
Canada: A summary of evidence and guidelines for low-risk
drinking. Retrieved from http://www.ccsa.ca/Eng/topics/
alcohol/drinking-guidelines/Pages/default.aspx.
Cansino, S., Estrada-Manilla, C., Herna  ndez-Ramos, E., Martinez-
Galindo, J. G., Torres-Trejo, F., Go  mez-Ferna  ndez, T., et al.
(2013). The rate of source memory decline across the adult life
span. Developmental Psychology, 49, 973e985. http://dx.doi.org/
10.1037/a0028894.
Cansino, S., Maquet, P., Dolan, R. J., & Rugg, M. D. (2002). Brain
activity underlying encoding and retrieval of source memory.
Cerebral Cortex, 12, 1048e1056. http://dx.doi.org/10.1093/
cercor/12.10.1048.
Cansino, S., Trejo-Morales, P., Estrada-Manilla, C., Pasaye-
~ eda, E., Salgado-Lujambio, P.,
Alcaraz, E. H., Aguilar-Castan
et al. (2015). Brain activity during source memory retrieval in
young, middle-aged, and old adults. Brain Research, 1618,
168e180. http://dx.doi.org/10.1016/j.brainres.2015.05.032.
Chang, L. J., Yarkoni, T., Khaw, M. W., & Sanfey, A. G. (2013).
Decoding the role of the insula in human cognition: Functional
parcellation and large-scale reverse inference. Cerebral Cortex,
23, 739e749. http://dx.doi.org/10.1093/cercor/bhs065.
de Chastelaine, M., Mattson, J. T., Wang, T. H., Donley, B. E., &
Rugg, M. D. (2016). The neural correlates of recollection and
retrieval monitoring: Relationships with age and recollection
performance. NeuroImage, 138, 164e175. http://dx.doi.org/
10.1016/j.neuroimage.2016.04.07.
Chen, G., Adleman, N. E., Saad, Z. S., Leibenluft, E., & Cox, R. W.
(2014). Applications of multivariate modeling to neuroimaging
group analysis: A comprehensive alternative to univariate
general model. NeuroImage, 99, 571e588. http://dx.doi.org/
10.1016/j.neuroimage.2014.06.027.
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, 163e169. http://dx.doi.org/10.1016/
j.neuropsychologia.2015.05.032.
Craik, F. I., Morris, L. W., Morris, R. G., & Loewen, E. R. (1990).
Relations between source amnesia and frontal lobe
functioning in older adults. Psychology and Aging, 5, 148e151.
http://dx.doi.org/10.1037/0882-7974.5.1.148.
Daselaar, S. M., Prince, S. E., Dennis, N. A., Hayes, S. M., Kim, H., &
Cabeza, R. (2009). Posterior midline and ventral parietal
activity is associated with retrieval success and encoding
failure. Frontiers in Human Neuroscience, 3, 13. http://dx.doi.org/
10.3389/neuro.09.013.2009.
Delis, D. C., Kramer, J. H., Kaplan, E., & Ober, B. A. (2000). California
verbal learning test (2nd Ed.). San Antonio, TX: The
Psychological Corporation.
Diana, R. A., Yonelinas, A. P., & Ranganath, C. (2007). Imaging
recollection and familiarity in the medial temporal lobe: A
three-component model. Trends in Cognitive Science, 11,
379e386. http://dx.doi.org/10.1016/j.tics.2007.0.001.
Diana, R. A., Yonelinas, A. P., & Ranganath, C. (2009). Medial
temporal lobe activity during source retrieval reflects
information type, not memory strength. Journal of Cognitive
Neuroscience, 22, 1808e1818. http://dx.doi.org/10.1162/
jocn.2009.21335.
Dobbins, I. G., Foley, H., Schacter, D. L., & Wagner, A. D. (2002).
Executive control during episodic retrieval: Multiple prefrontal
processes subserve source memory. Neuron, 35, 989e996.
http://dx.doi.org/10.1016/s0896-6273(02)00858-9.
232 c o r t e x 9 1 ( 2 0 1 7 ) 2 2 1 e2 3 3
Donaldson, D. I., Wheeler, M. E., & Petersen, S. E. (2009).
Remember the source: Dissociating frontal and parietal
contributions to episodic memory. Journal of Cognitive
Neuroscience, 22, 377e391. http://dx.doi.org/10.1162/
jocn.2009.21242.
Do€ rfel, D., Werner, A., Schaefer, M., von Kummer, R., & Karl, A.
(2009). Distinct brain networks in recognition memory share a
defined region in the precuneus. European Journal of
Neuroscience, 30, 1947e1959. http://dx.doi.org/10.1111/j.1460-
9568.2009.06973.x.
Duarte, A., Henson, R. N., & Graham, K. S. (2008). The effects of
aging on the neural correlates of subjective and objective
recollection. Cerebral Cortex, 18, 2169e2180. http://dx.doi.org/
10.1093/cercor/bhm243.
Duarte, A., Ranganath, C., & Knight, R. T. (2005). Effects of
unilateral prefrontal lesions on familiarity, recollection,
and source memory. The Journal of Neuroscience, 25,
8333e8337. http://dx.doi.org/10.1523/JNEUROSCIE.1392-
05.2005.
Dulas, M. R., & Duarte, A. (2014). Aging affects the interaction
between attentional control and source memory: An fMRI
study. Journal of Cognitive Neuroscience, 26, 2653e2669. http://
dx.doi.org/10.1162/jocn_a_00663.
Duverne, S., Habibi, A., & Rugg, M. D. (2008). Regional specificity of
age effects on the neural correlates of episodic retrieval.
Neurobiology of Aging, 29, 1902e1916. http://dx.doi.org/10.1016/
j.neurobiolaging.2007.04.022.
Eichenbaum, H., Yonelinas, A. P., & Ranganath, C. (2007). The
medial temporal lobe and recognition memory. Annual Review
of Neuroscience, 30, 123e152. http://dx.doi.org/10.1146/
annurev.neuro.30.051606.094328.
Fastenau, P. S., Denburg, N. L., & Mauer, B. A. (1998). Parallel short
forms for the Boston Naming Test: Psychometric properties
and norms for older adults. Journal of Clinical and Experimental
Neuropsychology, 20, 828e834. http://dx.doi.org/10.1076/
jcen.20.6.828.1105.
Fjell, A. M., Walhovd, K. B., Fennema-Notestine, C., McEvoy, L. K.,
Hagler, D. J., et al. (2009). One year brain atrophy evident in
healthy aging. Journal of Neuroscience, 29, 15223e15231. http://
dx.doi.org/10.1523/JNEUROSCI.3252-09.2009.
Glisky, E. L., Polster, M. R., & Routhieaux, B. C. (1995). Double
dissociation between item and source memory.
Neuropsychology, 9, 229e235. http://dx.doi.org/10.1037/0894-
4105.9.2.229.
Glisky, E. L., Rubin, S. R., & Davidson, P. S. (2001). Source memory
in older adults: An encoding or retrieval problem? Journal of
Experimental Psychology: Learning, Memory, & Cognition, 27,
1131e1146. http://dx.doi.org/10.1037/0278-7393.27.5.1131.
Grady, C. L., St.-Laurent, M., & Burianova  , H. (2015). Age
differences in brain activity related to unsuccessful
declarative memory retrieval. Brain Research, 1612, 30e47.
http://dx.doi.org/10.1016/j.brainres.2014.12.036.
Harle , K. M., Chang, L. J., van't Wout, M., & Sanfey, A. G. (2012).
The neural mechanisms of affect infusion in social economic
decision-making: A mediating role of the anterior insula.
NeuroImage, 61, 32e40. http://dx.doi.org/10.1016/
j.neuroimage.2012.02.027.
Hart, R. P., Kwentus, J. A., Wade, J. B., & Taylor, J. R. (1988).
Modified Wisconsin Card Sorting Test in elderly normal,
depressed and demented patients. Clinical Neuropsychologist, 2,
49e56. http://dx.doi.org/10.1080/13854048808520085.
Hashtroudi, S., Johnson, M. K., & Chrosniak, L. D. (1989). Aging
and source memory. Psychology and Aging, 4, 106e112. http://
dx.doi.org/10.1037/0882-7974.4.1.106.
Hester, R., Fassbender, C., & Garavan, H. (2004). Individual
differences in error processing: A review and reanalysis of
three event-based studies using the GO/NOGO task. Cerebral
Cortex, 14, 986e994. http://dx.doi.org/10.1093/cercor/bhh059.
Horne, J., & Ostberg, O. (1976). A self-assessment questionnaire to
determine morningness-eveningness in human circadian
rhythms. International Journal of Chronobiology, 4, 97e110.
Huijbers, W., Schultz, A. P., Vannini, P., McLaren, D. G.,
Wigman, S. E., Ward, A. M., et al. (2013). The encoding/retrieval
flip: Interactions between memory performance and stage and
relationship to intrinsic cortical networks. Journal of Cognitive
Neuroscience, 25, 1163e1179. http://dx.doi.org/10.1162/
jocn_a_00366.
Huijbers, W., Vannini, P., Sperling, R. A., Pennartz, C. M.,
Cabeza, R., & Daselaar, S. M. (2012). Explaining the encoding/
retrieval flip: Memory-related deactivations and activations in
the posteromedial cortex. Neuropsychologia, 50(14), 3764e3774.
Hutchinson, J. B., Uncapher, M. R., & Wagner, A. D. (2009).
Posterior parietal cortex and episodic retrieval: Convergent
and divergent effects of attention and memory. Learning &
Memory, 16, 343e356. http://dx.doi.org/10.1101/lm.919109.
Kukolja, J., Thiel, C. M., Wilms, M., Mirzazade, S., & Fink, G. R.
(2009). Ageing-related changes of neural activity associated
with spatial contextual memory. Neurobiology of Aging, 30,
630e645. http://dx.doi.org/10.1016/
j.neurobiolaging.2007.08.015.
Lawton, M. P., & Brody, E. M. (1969). Assessment of older people:
Self-maintaining and instrumental activities of daily living.
Gerontologist, 9, 179e186. http://dx.doi.org/10.1093/geront/
9.3_part_1.179.
Leshikar, E. D., & Duarte, A. (2014). Medial prefrontal cortex
supports source memory for self-referenced materials in
young and older adults. Cognitive, Affective, & Behavioral
Neuroscience, 14, 236e252. http://dx.doi.org/10.3758/s13415-
013-0198-y.
Lundstrom, B. N., Ingvar, M., & Petersson, K. M. (2005). The role of
the precuneus and left inferior frontal cortex during source
memory episodic retrieval. NeuroImage, 27, 824e834. http://
dx.doi.org/10.1016/j.neuroimage.2005.05.008.
Lundstrom, B. N., Petersson, K. M., Andersson, J., Johansson, M.,
Fransson, P., & Ingvar, M. (2003). Isolating the retrieval of
imagined pictures during episodic memory: Activation of the
left precuneus and left prefrontal cortex. NeuroImage, 20,
1934e1943. http://dx.doi.org/10.1016/
j.neuroimage.2003.07.017.
McDonough, I. M., Wong, J. T., & Gallo, D. A. (2013). Age-related
differences in prefrontal cortex activity during retrieval
monitoring: Testing the compensation and dysfunction
accounts. Cerebral Cortex, 23, 1049e1060. http://dx.doi.org/
10.1093/cercor/bhs064.
McIntyre, J. S., & Craik, F. I. (1987). Age differences in memory for
item and source information. Canadian Journal of Psychology, 41,
175e192. http://dx.doi.org/10.1037/h0084154.
Menon, V., & Uddin, L. Q. (2010). Salience, switching, attention
and control: A network model of insula function. Brain
Structure & Function, 214, 655e667. http://dx.doi.org/10.1007/
s00429-010-0262-0.
Mitchell, K. J., & Johnson, M. K. (2009). Source monitoring 15 years
later: What have we learned from fMRI about the neural
mechanisms of source memory? Psychological Bulletin, 135,
638e677. http://dx.doi.org/10.1037/a0015849.
Mitchell, K. J., Raye, C. L., Johnson, M. K., & Greene, E. J. (2006). An
fMRI investigation of short-term source memory in young and
older adults. NeuroImage, 30, 627e633. http://dx.doi.org/
10.1016/j.neuroimage.2005.09.039.
Morcom, A. M., Li, J., & Rugg, M. D. (2007). Age effects on the
neural correlates of episodic retrieval: Increased cortical
recruitment with matched performance. Cerebral Cortex, 17,
2491e2506. http://dx.doi.org/10.1093/cercor/bhl155.
Oldfield, R. C. (1971). The assessment and analysis of handedness:
The Edinburgh inventory. Neuropsychologia, 9, 97e113. http://
dx.doi.org/10.1016/0028-3932(71)90067-4.
 c o r t e x 9 1 ( 2 0 1 7 ) 2 2 1 e2 3 3
Persson, N., Ghisletta, P., Dahle, C. L., Bender, A. R., Yang, Y.,
Yuan, P., et al. (2016). Regional brain shrinkage and change in
cognitive performance over two years: The bidirectional
influences of the brain and cognitive reserve factors.
NeuroImage, 126, 15e26. http://dx.doi.org/10.1016/
j.neuroimage.2015.11.028.
Rajah, M. N., Kromas, M., Han, J. E., & Pruessner, J. C. (2010a).
Group differences in anterior hippocampal volume and in the
retrieval of spatial and temporal context memory in healthy
young versus older adults. Neuropsychologia, 48, 4020e4030.
http://dx.doi.org/10.1016/j.neuropsychologia.2010.10.010.
Rajah, M. N., Languay, R., & Valiquette, L. (2010b). Age-related
changes in prefrontal cortex activity are associated with
behavioral deficits in both temporal and spatial context
memory retrieval in older adults. Cortex, 46, 535e549. http://
dx.doi.org/10.1016/j.cortex.2009.07.006.
Rugg, M. D., Fletcher, P. C., Chua, P. M.-L., & Dolan, R. J. (1999). The
role of the prefrontal cortex in recognition memory and
memory for source: An fMRI study. NeuroImage, 10, 520e529.
http://dx.doi.org/10.1006/nimg.1999.0488.
Rypma, B., Berger, J. S., Genova, H. M., Rebbechi, D., &
D'Esposito, M. (2005). Dissociating age-related changes in
cognitive strategy and neural efficiency using event-related
fMRI. Cortex, 41, 582e594. http://dx.doi.org/10.1016/s0010-
9452(08)70198-9.
Rypma, B., & D'Esposito, M. (2000). Isolating the neural
mechanisms of age-related changes in human working
memory. Nature Neuroscience, 3, 509e515. http://dx.doi.org/
10.1038/74889.
Samanez-Larkin, G. R., & D'Esposito, M. (2008). Group
comparisons: Imaging the aging brain. Social Cognitive and
Affective Neuroscience, 3, 290e297. http://dx.doi.org/10.1093/
scan/nsn029.
Schacter, D. L., Kaszniak, A. W., Kihlstrom, J. F., & Valdiserri, M.
(1991). The relation between source memory and aging.
Psychology and Aging, 6, 559e568. http://dx.doi.org/10.1037/
0882-7974.6.4.559.
Seeley, W. W., Menon, V., Schatzberg, A. F., Keller, J., Glover, G. H.,
Kenna, H., et al. (2007). Dissociable intrinsic connectivity
networks for salience processing and executive control.
Journal of Neuroscience, 28, 2349e2356. http://dx.doi.org/
10.1523/jneurosci.5587-06.2007.
Spaniol, J., & Grady, C. (2012). Aging and the neural correlates of
source memory: Over-recruitment and functional
reorganization. Neurobiology of Aging, 33, e3ee18. http://
dx.doi.org/10.1016/j.neurobiolaging.2010.10.005.
Sridharan, D., Levitin, D. J., & Menon, V. (2008). A critical role for
the right fronto-insular cortex in switching between central-
233
executive and default-mode networks. Proceedings of the
National Academy of Sciences, 105, 12569e12574. http://
dx.doi.org/10.1073/pnas.0800005105.
Swick, D., Senkfor, A. J., & Van Petten, C. (2006). Source memory
retrieval is affected by aging and prefrontal lesions: Behavioral
and ERP evidence. Brain Research, 1107, 161e176. http://
dx.doi.org/10.1016/j.brainres.2006.06.013.
Teng, E. L., & Chui, H. C. (1987). The modified mini-mental state
(3MS) examination. Journal of Clinical Psychiatry, 48, 314e318.
Thakral, P. P., Wang, T. H., & Rugg, M. D. (2015). Cortical
reinstatement and the confidence and accuracy of source
memory. NeuroImage, 109, 118e129. http://dx.doi.org/10.1016/
j.neuroimage.2015.01.003.
Uddin, L. Q. (2015). Salience processing and insular cortical
function and dysfunction. Nature Reviews Neuroscience, 16,
55e61. http://dx.doi.org/10.1038/nrn3857.
Utevsky, A. V., Smith, D. V., & Huettel, S. A. (2014). Precuneus is a
functional core of the default-mode network. The Journal of
Neuroscience, 34(3), 932e940.
Wechsler, D. (1981). The wechsler adult intelligence scale-revised
manual. San Antonio, TX: The Psychological Corporation.
Wechsler, D. (1987). The wechsler memory scale e revised manual.
San Antonio, TX: The Psychological Corporation.
Wechsler, D. (1991). The wechsler adult intelligence scale-III manual.
San Antonio, TX: The Psychological Corporation.
Wechsler, D. (1997). The wechsler memory scale e III manual. San
Antonio, TX: The Psychological Corporation.
Wechsler, D. (1999). The wechsler abbreviated scale of intelligence
manual. San Antonio, TX: The Psychological Corporation.
Wegesin, D. J., Jacobs, D. M., Zubin, N. R., Ventura, P. R., &
Stern, Y. (2000). Source memory and encoding strategy in
normal aging. Journal of Clinical and Experimental
Neuropsychology, 22, 455e464. http://dx.doi.org/10.1076/1380-
3395(200008)22:4;1e0;FT455.
Welsh, K., Breitner, J., & Magruder-Habib, K. (1993). Detection of
dementia in the elderly using telephone screening of cognitive
status. Neuropsychiatry, Neuropsychology and Behavioral
Neurology, 6, 103e110.
Zachary, R. A. (1986). Shipley institute of living scale: Revised manual.
Los Angeles, CA: Western Psychological Services.
Zhang, S., & Li, C. R. (2012). Functional connectivity mapping of
the human precuneus by resting state fMRI. NeuroImage, 59(4),
3548e3562.
Zigmond, A. S., & Snaith, R. P. (1983). The hospital anxiety and
depression scale. Acta Psychiatrica Scandinavia, 67, 361e370.
http://dx.doi.org/10.1111/j.1600-0447.1983.tb09716.x.