Neuropsychologia 48 (2010) 921–932
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/neuropsychologia
Memory, metamemory and their dissociation in temporal lobe epilepsy
Charlotte E. Howarda, Pilar Andrésb,∗,1, Paul Broksa,2, Rupert Noadc,2, Martin Sadlerd,2,
Debbie Cokerd,2, Giuliana Mazzonie,3
aSchool of Psychology, University of Plymouth, United Kingdom
bDepartment of Psychology, University of the Balearic Islands, Cra de Valldemossa, Km 7,5, 07122 Palma de Mallorca, Spain
cDepartment of Clinical Neuropsychology, Derriford Hospital, Plymouth Hospitals NHS Trust, United Kingdom
dDepartment of Neurology, Derriford Hospital, Plymouth Hospitals NHS Trust, United Kingdom
eDepartment of Psychology, University of Hull, United Kingdom
a r t i c l ei n f o
Received 15 July 2009
Received in revised form
12 November 2009
Accepted 14 November 2009
Available online 22 November 2009
a b s t r a c t
Patients with temporal-lobe epilepsy (TLE) present with memory difficulties. The aim of the current
study was to determine to what extent these difficulties could be related to a metamemory impairment.
Fifteen patients with TLE and 15 matched healthy controls carried out a paired-associates learning task.
Memory recall was measured at intervals of 30min and 4 weeks. We employed a combined Judgement-
of-Learning (JOL) and Feeling-of-Knowing (FOK) task to investigate whether participants could monitor
deficit of episodic memory in patients with epilepsy compared with controls, but metamemory in TLE
patients was intact. Patients were able to monitor their memory successfully at the item-by-item level,
and tended to be even more accurate than controls when making global judgements.
© 2009 Elsevier Ltd. All rights reserved.
Epilepsy is a chronic and common neurological disorder char-
acterised by recurrent seizures. Temporal-lobe epilepsy (TLE), the
most common form of epilepsy, is associated with cell loss in
the hippocampus and the surrounding areas resulting in mem-
ory difficulties such as episodic memory impairments, as well as
impairments in long-term consolidation and remote memory (see
Bell & Giovagnoli, 2007; Leritz, Grande, & Bauer, 2006 for reviews).
Several studies have shown discrepancies between subjective
reports of memory problems and objective measures from neu-
ropsychological tasks in TLE patients (Gleissner, Helmstaedter,
Quiske, & Elger, 1998; Thompson & Corcoran, 1992; Vermeulen,
Aldenkamp, & Alpherts, 1993). For example, some studies have
shown that TLE patients present with memory complaints, but
perform adequately when assessed objectively with standardised
memory tasks (Gallassi, Morreale, Lorusso, Pazzaglia, & Lugaresi,
1988; Hermann, Wyler, Steenman, & Richey, 1988; O’Shea, Saling,
E-mail address: email@example.com (P. Andrés).
1Pilar Andrés was the Principal Supervisor during the realisation of this project
at the University of Plymouth.
2These authors facilitated the access to the patients at Derriford Hospital.
during the writing up of this paper.
Bladin, & Berkovic, 1996; Thompson & Corcoran, 1992). Three main
factors have been suggested to explain these underestimations
of memory in TLE patients. The first is the existence of acceler-
ated forgetting (AF) which has been attributed to the presence
of seizures during the retention period. Since seizures affect the
consolidation process, immediate recall should not be affected by
TLE, and therefore, only delayed recall tasks would show differ-
ences in memory performance between controls and TLE patients.
Blake, Wroe, Breen, and McCarthy (2000; also see Mameniskiene,
Jatuzis, Kaubrys, & Budrys, 2006) showed for example significant
differences between TLE patients and controls at delayed recall
for complex verbal material for which the initial level of encod-
ing was equated between TLE patients and controls. AF, however,
is far from being a constant feature of TLE, and numerous studies
have shown equivalent differences between TLE patients and con-
Hermann, 2005; Bell, 2006; Giovagnoli, Casazza, & Avanzini, 1995;
Helmstaedter, Hauff, & Elger, 1998).
The second factor refers to the presence of mood disturbances
(anxiety and depression), which interfere with the subjective
perception of memory performance, leading to underestimations
(Ba˜ nos et al., 2004; Elixhauser, Leidy, Meador, Means, & Willian,
1999; Giovagnoli, Mascheroni, & Avanzini, 1997; Vermeulen et al.,
The third factor is a specific deficit in metamemory. Metamem-
ory plays a central role in human learning through development
0028-3932/$ – see front matter © 2009 Elsevier Ltd. All rights reserved.
C.E. Howard et al. / Neuropsychologia 48 (2010) 921–932
(Flavell & Wellman, 1977), and a deficit in this set of processes has
tion in several populations (e.g., Light, 1991; Shimamura & Squire,
Metamemory is one component of metacognition, which can
be broadly defined as the knowledge about one’s own cognitive
abilities. According to Nelson and Narens (1990), metamemory,
in common with other metacognitive processes, is comprised of
two key processes: ‘monitoring’ and ‘control’. Monitoring refers to
the collection of information and awareness about one’s memory
processes, including encoding, level of knowledge, retrieval, and
performance outcome, whereas control acts as a self-regulation
process, activating and directing these same cognitive processes.
One example of control processes is allocating sufficient time to
studying material for a successful later recall. Depending on the
difficulty of the material, different amounts of study time should
are closer to the recall threshold (see Son & Metcalfe, 2000 for
review). Memory monitoring is usually measured with tasks such
as Judgements-of-Learning (JOL) and Feeling-of-Knowing (FOK),
As applied to clinical populations, the hypothesis is that poor
memory can result from inadequate metamemory monitoring,
inadequate metamemory control, or both. Deficits in metamem-
ory have been observed in some types of neurological patients, but
not in others (see review by Pannu & Kaszniak, 2005). In the case of
Alzheimerˇ ıs Disease, it has been proposed that the loss in episodic
memory experienced by some patients can be explained by the
observed impairment in metamemory functions, and in particu-
lar by the deficit in metacognitive monitoring (e.g., Correa, Graves,
& Costa, 1996; McGlynn & Kaszniak, 1991; Souchay, 2007). How-
ever, it has also been shown that a metamemory deficit is not an
obligatory trait of AD, and that some patients can also show unim-
paired metamemory ability (Moulin, Perfect, & Jones, 2000). In that
context, Cosentino, Metcalfe, Butterfield, and Stern (2007) have
ory loss show poor monitoring processes whereas patients who
are aware of their memory loss demonstrate metamemory that is
comparable to healthy older adults. It has also been shown that
patients with severe anterograde amnesia can produce accurate
metamemory (Feeling-of-Knowing) judgements. Thus, impaired
metamemory accuracy is not an obligatory feature of amnesia
(Shimamura & Squire, 1986).
A relationship has been suggested between metacognition and
executive processes (Fernandez-Duque, Baird, & Posner, 2000;
Shimamura, 2000; Souchay, Isingrini, & Espagnet, 2000). Neu-
Schwartz, & Schacter, 2003) and neuropsychological (Janowsky,
Shimamura, & Squire, 1989; Modirrousta & Fellows, 2008; Schnyer
et al., 2004; Vilkki, Servo, & Surma-aho, 1998; Vilkki, Surma-aho,
& Servo, 1999) studies have confirmed a primary role of the pre-
frontal cortex in metamemory processing. Although deficits in the
prefrontal areas seem to be more likely to be linked to metamem-
ory problems, there are reasons to predict that patients with TLE
tions in general are sustained by a diffuse neural network rather
than by only prefrontal areas (Andrés, 2003; Collette & Van der
Linden, 2002). Second, in a study in which the neural correlates
of FOK judgements were assessed in a face-name association task,
Kikyo and Miyashita (2004; see also Schnyer, Nicholls, & Verfaellie,
ing FOKs on higher-order information processing of face images or
Additionally, Modirrousta and Fellows (2008) showed an interest-
ing dissociation between impaired FOK judgements and intact JOLs
in patients with prefrontal damage. Pannu, Kaszniak, and Rapcsak
(2005) and Schnyer et al. (2004) also showed important disso-
ciations in frontal patients, with some metamemory tasks (for
example, FOKs) impaired and others (for example, JOLs) within
normal range. These findings suggest that JOL accuracy is likely to
be dependent on other areas than the prefrontal cortex, for exam-
ple the temporal cortex. Thirdly, it has been shown that patients
with early Alzheimer’s disease, who, like TLE patients, suffer from
hippocampal and temporal atrophy, present with metamemory
deficits (see Souchay, 2007 for a review). Finally, several studies
have documented that cognitive dysfunction in TLE affects func-
tions supported by the frontal cortex such as mental flexibility and
et al., 2000). More specifically, Hermann, Seidenberg, Haltiner and
Wyler (1991; also see Keller, Baker, Downes, & Roberts, 2009) pos-
the “spread of temporal lobe hypometabolism to the thalamus sec-
ondarily affecting the frontal lobe” or possibly the direct “spread
of temporal lobe hypometabolism to the frontal lobe” (p. 1214).
This has lead to the ‘nociferous cortex hypothesis’, postulating that
there are electrophysiological abnormalities in distal extratempo-
likely that metamemory processes, intimately related to execu-
tive functions (Fernandez-Duque et al., 2000; Shimamura, 2000;
Souchay et al., 2000) are also disrupted in TLE patients.
Although scarce, some neuropsychological studies have looked
at metacognitive deficits in TLEs. In two studies, Prevey, Delaney
and Mattson (1988) and Prevey, Delaney, Mattson and Tice (1991)
concluded that TLE patients present a deficit in metacognitive
monitoring. Prevey et al. (1988) conducted two experiments in
which metamemory functioning was explored at encoding and
retrieval in TLE patients and controls. In Experiment 1, partici-
pants were presented with two memory span tasks consisting of
lists of single syllable nouns (verbal task) and non-meaningful geo-
metrical shapes (visual task). Lists were of increasing length, from
1 to 10 items per list, and after learning each list, participants
were instructed to provide a yes/no judgement as to whether they
thought they could remember the words/non-meaningful geomet-
ric shapes in the list in the order presented. The results showed
that TLE patients anticipated that they would perform just as well
as the controls, but in fact, they performed less well than the con-
trols on the recall tasks. It was also noted that the site of the lesion
(left–right) mediated prediction accuracy depending on the exper-
imental materials used (verbal/non-verbal).
In Experiment 2 participants were asked to make FOK judge-
ments on general information questions they had previously
answered incorrectly, by providing a ‘yes’ or ‘no’ response as to
whether they would be able to recognise the correct answer from
a range of six alternatives. Although also in this case the authors
conclude that monitoring is impaired in TLE patients, the results
are actually not clear, and depend on which measure of FOK accu-
racy is used. Gamma correlations, which are the most commonly
criminate between which items will or will not be recalled and
no significant differences between controls and patients, either left
or right. Only when proportion of positive FOK recognitions was
used to assess accuracy, TLE patients resulted to be less accurate
This measure reflects the proportion of correctly recognised items
over the total number of items for which positive (yes) recogni-
tion was predicted. In the assessment of relative metamnemonic
accuracy, this conditional probability measure has long been aban-
by Nelson (1984) in which Gamma was demonstrated to be supe-
rior to a number of other measures of association, including scores
C.E. Howard et al. / Neuropsychologia 48 (2010) 921–932
based on conditional probabilities (see Benjamin & Díaz, 2008, for
a more thorough exam of accuracy measures in metacognition).
In two subsequent experiments examining Feeling-of-Knowing
in TLE patients and using the same procedure as Experiment 2 in
pattern of differences between TLE patients and controls when
making FOK judgements. In both Experiment 1 and Experiment 2,
Gamma correlations were statistically not different between con-
trols and TLE patients, although numerically they were higher in
controls. Statistically significant differences were found in Experi-
ment 1 in the proportion of correct FOK, calculated as proportion
of positive FOK recognition. Also for these two experiments, the
authors concluded that FOK accuracy is lower in TLE patients than
controls, reflecting poor memory monitoring in this population.
However, given the procedure and the data analysis, this con-
clusion does not seem warranted. First, the memory task was a
span task, which assesses serial short-term and working memory,
and not typical episodic long-term memory. Second, the memory
task used to assess FOK was a fact retrieval task, commonly used
in those years. As such, however, it tests semantic memory, not
episodic memory, and thus these data have little to say about pos-
sible monitoring deficits in episodic memory in TLE patients. Third,
even when using this semantic memory task, Gamma correlations
showed no significant difference in FOK accuracy between controls
and patients, only proportion of correctly predicted recognition
did. Fourth, no difference between groups was found in span recall
prediction, only in actual span recall, and the conclusion about
comparison not supported by any data analysis.
In contrast to the claims by Prevey et al. (1988, 1991), in
their review of metamemory experiments in various types of
neurological patients, Pannu and Kaszniak (2005, p. 116) report
almost perfect metacognitive accuracy in a face recognition task
(gamma=.90) in FOK judgements in a patient with prosopag-
nosia due most likely to a right temporal epilepsy focus (Rapcsak,
Pannu, & Kaszniak, 2005, as in Pannu & Kaszniak, 2005, p.116).
Furthermore, this result was not simply due to this patient giving
constantly very low ratings. Rather, it was due to her giving higher
ratings to faces she was then able to recognise, and lower ratings
to faces she was not able to recognise.
Given these previous unclear and mixed results, the objective
of the present study was to investigate whether metamemory
monitoring processes are disrupted in TLE patients when mon-
itoring is tested by assessing accuracy of online monitoring for
both recall (item-by-item Judgement-of-Learning predictions) and
recognition (item-by-item Feeling-of-Knowing predictions) in an
episodic memory task. In addition, we also examined global pre-
dictions of episodic memory (Global JOLs), which have been found
to be impaired in Alzheimer’s patients (e.g. Correa et al., 1996;
McGlynn & Kaszniak, 1991; however see Cosentino et al., 2007 and
Moulin et al., 2000 for clinical variability in metamemory in AD). In
TLE patients the observed discrepancy between severe complaints
about memory loss and their relatively adequate performance in
objective memory tests (e.g. Gallassi et al., 1988; Hermann et al.,
1988; O’Shea et al., 1996) suggests that a metamemory deficit
ory performance. At the same time, the findings by Prevey et al.
(1988) would suggest the opposite, i.e. a clear overestimation of
memory performance in TLE patients. Given this disparity, in the
present study we cannot predict the specific direction of the dis-
crepancy between memory evaluations and memory performance.
Finally, we also tested whether impaired control processes
might be responsible for the memory problems commonly affect-
ing TLE patients. Control processes were examined by assessing
how study-time is allocated (Mazzoni & Cornoldi, 1993; Mazzoni,
encoding by devoting more time to items that are either more dif-
ficult to learn (Mazzoni & Cornoldi, 1993), or are closer to the recall
in the use of this strategy, which would be revealed if more time is
devoted to items that are easy to recall, or to items that are too dif-
ficult to be learned (Metcalfe & Kornell, 2003), could be responsible
for observed deficits in episodic memory in TLE patients.
2. The present study
In the present study a paired-associates learning task was pre-
sented to 15 patients with TLE and 15 matched healthy controls.
ory tasks, we employed the combined JOL and FOK task. A memory
memory performance. Furthermore, anxiety and depression were
assessed to control for the possible effect of these variables on
Based on the results of previous studies (Bell & Giovagnoli,
2007; Leritz et al., 2006 for reviews), our predictions were that
TLE patients would present with a deficit in episodic memory,
which would be greater at delayed recall (i.e., 4 weeks after encod-
ing). Similarities between metacognition and executive control
processes suggested in previous literature (Fernandez-Duque et
al., 2000; Shimamura, 2000; Souchay et al., 2000) supported the
prediction that there would also be the potential for a degree of
executive dysfunction in TLE patients. Finally, and crucially, based
on the methodological problems in previous studies (Prevey et al.,
1988, 1991) and the mixed results obtained in the literature, we
aimed at exploring further metamemory abilities in TLE patients.
Fifteen temporal-lobe patients with epilepsy (M=38.33 years; SD=12.41; range
18–63) and 15 controls (M=33.67 years; SD=10.90; range 18–52) participated in
this study. TLE patients were recruited from Derriford Hospital’s (Plymouth Hos-
pitals NHS Trust) neurology out-patients clinic, whereas control participants were
recruited from the University of Plymouth’s School of Psychology undergraduate
ers Group received a small remuneration to cover any travel or parking expenses.
Undergraduate participants received participation points as part of their course
TLE patients were considered suitable for investigation based on the follow-
ing screening criteria: (1) TLE out-patients; (2) aged between 18 and 65 years; (3)
English as their native language; (4) normal hearing and normal/corrected vision;
(5) a minimum of 8 years education; (6) evidence of an abnormal EEG recording
and/or MRI/CT scan to confirm condition and epileptic focus; (7) dosage and type
of anti-epileptic drugs stable for a minimum of 1 month; (8) no presence of any
current or past psychiatric disorders (including alcohol, substance abuse or clinical
depression); (9) no other degenerative or cognitive disease that may prevent them
from participating (i.e. learning disability, aphasia); (10) not undergone corrective
surgery for their epilepsy; (11) not experienced a seizure in the past 24h prior to
testing (determined on day of testing).
3.2. Demographic characteristics
Demographic characteristics of both groups and epilepsy features of the TLE
patients can be found in Table 1. Control participants and TLE patients did not
significantly differ in terms of age [F (1, 28)=1.20, MSE=136.45, p=.28, ?2p=.04],
years of formal education [F (1, 28)=.25, MSE=5.66, p=.62, ?2p=.01], gender4[X2
?2p=.03]. Twelve (80%) of the TLE patients were diagnosed as having complex par-
tial seizures, one (7%) patient experienced complex partial seizures with secondary
4A Chi Square test was performed on the frequency of males and females in each
C.E. Howard et al. / Neuropsychologia 48 (2010) 921–932
Demographic characteristics and epilepsy features for TLE and control groups (standard deviations are in parentheses).
TLE n=15Controls n=15
Age of onset
Seizure frequency (# per month)
Evidence provided by only an abnormal EEGa, MRIbor combination of bothc
Summary of the neuropsychological test battery and MFQ results (standard deviations are in parentheses).
Test TLE n=15Controls n=15
Harris test of lateral dominance(Handedness)
D-KEFS design fluency
D-KEFS color-word interference
Hayling sentence completion test
Logical Memory I
Logical Memory II
Predictive Verbal IQ
Predictive Performance IQ
General frequency of forgetting
Seriousness of forgetting
1.07 (.26) 1.07 (.26).00 1.00
D-KEFS: Delis–Kaplan Executive Function System; WAIS-III: Wechsler adult intelligence scale 3rd Edition; WMS-III: Wechsler memory scale 3rd Edition; NART: National
Adult Reading Test; MFQ: memory functioning questionnaire.
generalisation, another (7%) had simple partial seizures and one (7%) other patient
was classified as having both complex partial and simple partial seizures. Two (13%)
patients were seizure free5at the time of testing. Eight (53%) were on monotherapy
and seven (47%) were on polytherapy (maximum combination of 3 anti-epileptic
drugs (AEDs)). Eleven (73%) TLE patients had seizures during the 4-week interval
between Session 1 and 2. The number of seizures experienced during the 4-week
interval did not significantly correlate with recall performance at Time 2 [r=−.33,
erPoint 2003 and run on a Toshiba Tablet laptop computer. One-hundred and
twenty word items (see Appendix A) were selected from the MRC Psycholinguistics
Database (Retrieved http://www.psy.uwa.edu.au/mrcdatabase/uwa mrc.htm, 19th
September 2006) to form the sixty paired-associates for this task. Words chosen
were of similar length, frequency of occurrence and level of concreteness in the
English language. Words differed in their level of relatedness. Thirty of the word
pairs were semantically related (i.e. hammer–saw), and the remaining thirty were
5These two seizure free patients reported not having experienced a seizure for
at least 6 months at the moment of testing (one for over a year and the other for
6 months). Patients were advised by their medical team to keep their own seizure
diary, which enabled us to consult the frequency of the seizures, although it should
be noted that we cannot completely rule out the possibility that patients experi-
enced seizures that they did not record.
not related (i.e. duck–cloth). Word pairs were presented to participants one at a
time in the centre of the screen in Arial font size 44 in black on a white background.
Presentation time (study-time) of all word pairs was self-paced.
All participants were individually tested in a quiet room at either the University
of Plymouth, School of Psychology, or in one of the neurology clinic rooms at Derri-
ford Hospital. All participants gave written consent prior to taking part in the study.
The protocol was approved by the South West Devon Research Ethics Committee
(NHS REC) and also by the University of Plymouth, Faculty of Science Human Ethics
Committee. Participants were made aware that the study would be completed over
two sessions. Session 2 (Time 2) followed on 4 weeks from Session 1 (Time 1).
3.5. JOL task
Participants were informed that they were going to be shown sixty-word pairs
was presented one at a time and participants used the spacebar to declare recall
readiness and proceed onto the next item. A practise block consisting of four-word
and the words could be clearly read. Practise word pairs were not included in the
Immediately after studying each word pair, participants were asked to rate
how certain they felt they would recall the second part of that particular word
pair, if presented with only the first word as a cue later on in the session. The
actual time of when participants would be asked to recall the words was not
C.E. Howard et al. / Neuropsychologia 48 (2010) 921–932
Number of errors (mean and SD) in the four tasks of the D-KEFS Color-Word Interference Test for TLE and control groups.
Group Color namingWord reading InhibitionInhibition/switching
TLE latency times
Controls latency times
Cor: self-corrected; Noncor: uncorrected. Mean (SD) latencies to complete each of the four tasks are also included.
mentioned. Item-by-item JOLs were requested on a 6-point scale set at 20% inter-
vals (0%=definitely will not recall, 20%=20% sure, 40%=40% sure, 60%=60% sure,
verbally responded to give their rating on a particular word pair and the researcher
recorded their responses on a record sheet. The time taken studying each word pair
was recorded by the laptop in order to measure study-time allocation. At the time
of making a JOL the word pair was no longer visible to the participant. After JOLs
had been recorded for all sixty-word pairs, participants were asked to make a global
JOL as to how many of the sixty items they thought they would recall later on in
the session. Responses were given as a figure out of sixty. A 30-min delay was then
introduced in which non-verbal neuropsychological tests were administered (see
of the sixty-word pairs (one by one) and asked to recall aloud the target word. At
test, the presentation order of the cue words was different from the order presented
during the study phase, to prevent possible recency and primacy effects at recall.
Participants were given 5s to respond to each uncompleted word pair before the
screen refreshed and moved onto the next word pair. Responses were recorded by
the researcher on a record sheet.
3.6. FOK task
For every non-recalled or incorrectly recalled word pairs, participants were
then asked to make FOK judgements, which were made on the same 6-point scale
described for JOLs (from 0% to 100% at 20% intervals) as to whether they would
be able to recognise the second part of the word pair when the first part was pre-
sented along with four possible alternatives, one of which was the target word. The
recognition task was presented after the FOK judgements had been completed for
all non-recalled pairs.
with three distracters (semantic, phonological and neutral) (i.e. shovel-? radio-
spade-space-rake, queen-? pine-oak-pint-bone). It was emphasised to participants
not to guess at a particular word but to only respond if they thought it was the
correct word. Participants were given 8s in which to read the four alternatives and
choose the answer. Responses were recorded by the researcher on a record sheet.
At Session 2 (4 weeks later) participants were asked again to make a global JOL
as to how many of the word pairs they thought they could remember from four
weeks ago (as a figure out of 60). Participants were then tested as previously at
Session 1, by presenting the cue word for 5seconds and asking to recall the target
word. Participants then followed the same procedure for the FOK task for all the
word pairs they either failed to recall or incorrectly recalled at the time of test. The
final neuropsychological tests followed to complete the battery (see Table 2).
3.7. Memory Functioning Questionnaire
The Memory Functioning Questionnaire (MFQ, Gilewski et al., 1990) was
tioning. The questionnaire consists of four factors; General Frequency of Forgetting,
Seriousness of Forgetting, Retrospective Functioning and Mnemonic Usage.
3.8. Neuropsychological evaluation
A standard neuropsychological test battery (see Table 2 for a summary of the
individual tests) was completed by all participants. The battery was split between
the two sessions. The following tests were administered to form the neuropsycho-
logical test battery;
dominance in all participants.
(2) The Hospital Anxiety and Depression Scale (HADS, Zigmond & Snaith, 1983)
was selected to provide a severity score of anxiety and depression for each
(3) The Delis–Kaplan Executive Function System (D-KEFS, Delis, Kaplan, & Kramer,
2001) Design Fluency Test, D-KEFS Color-Word Interference Test (Delis et al.,
administered to measure executive functions.
(4) Similarities, Arithmetic and Comprehension subtests were selected from the
Wechsler Adult Intelligence Scale 3rd Ed. (WAIS-III, Wechsler, 1997a).
ory II were chosen from the Wechsler Memory Scale 3rd Edition. (WMS-III,
provided a test of pre-morbid intelligence. Predicted full scale IQ, verbal IQ
and performance IQ scores were obtained in both control participants and TLE
4.1. Neuropsychological test battery
The results from the neuropsychological test battery are pre-
sented in Table 2. The neuropsychological tests which yielded a
significant difference between TLE patients and controls included
the depression scores of the HADS [F (1, 28)=7.83, MSE=10.65,
p<.01, ?2p=.22], the composite scaled scores of the D-KEFS
Color-Word Interference Test [F (1, 28)=6.19, MSE=5.52, p<.05,
?2p=.18], the subtests similarities [F (1, 28)=4.22, MSE=3.82,
p<.05, ?2p=.13] and comprehension [F (1, 28)=7.84, MSE=6.81,
p<.01, ?2p=.22] from the WAIS-III, and the subtests Logical Mem-
ory I [F (1, 28)=6.49, MSE=6.65, p<.05, ?2p=.19], Logical Memory
II [F (1, 28)=17.98, MSE=6.45, p<.001, ?2p=.39] and Faces I [F
(1, 28)=5.37, MSE=6.76, p<.05, ?2p=.16] from the WMS-III. The
percentage retention scores from the story recall subtests (Logical
between groups [F (1, 28)=13.92, MSE=152.80, p<.001, ?2p=.33].
The direction of these differences indicated that the TLE patients
performed more poorly than the controls. The findings from the
subtests of the WMS-III provide the first indication of a memory
deficit in the TLE patients for both immediate and delayed recall.
Sentence Completion Test showed a tendency in TLE patients
to have some level of executive dysfunction [F (1, 28)=3.21,
MSE=1.50, p=.08, ?2p=.10].
No significant differences were obtained on the NART predicted
FSIQ scores [F (1, 28)=.87, MSE=99.22, p=.36, ?2p=.03], pre-
dicted verbal IQ scores [F (1, 28)=.95, MSE=84.60, p=.34, ?2p=.03]
and predicted performance IQ scores [F (1, 28)=1.05, MSE=79.16,
p=.31, ?2p=.04] or number of years of education [F (1, 28)=.25,
MSE=5.66, p=.62, ?2p=.01].
Given the significant result in the D-KEFS Color-Word Inter-
ference Test, further analysis of the components within this test
was carried out. The number of uncorrected and self-corrected
errors produced in each of the four conditions of the D-KEFS
Color-Word Interference Test were rare (see Table 3), and were
consequently not analysed. Latency times for the four conditions
(see Table 3) were analysed using a 2 (group)×4 (condition)
repeated measures ANOVA. The results showed a main effect of
group [F (1, 28)=10.50, MSE=259.58, p<.01, ?2p=.27], condition
[F (3, 84)=128.70, MSE=107.77, p<.001, ?2p=.82] and an interac-
tion between condition and group [F (3, 84)=3.89, MSE=107.77,
p<.01, ?2p=.12]. The analysis of the interaction showed that the
C.E. Howard et al. / Neuropsychologia 48 (2010) 921–932
at Time 1 and Time 2 for controls and TLE patients (standard deviations are in
Time of recallTLE Controls
Time 1 total recall performance
Time 1 related word pairs recalled
Time 1 unrelated word pairs recalled
Time 2 total recall performance
Time 2 related word pairs recalled
Time 2 unrelated word pairs recalled
Note: Time 2 (Session 2) followed on 4 weeks from Time 1 (Session 1).
in the inhibition/switching condition [t (28)=−2.97, p<.01]. More-
over, the interference (inhibition – color naming) and switching
and showed equivalent interference effects in the two groups [t
(28)=−.76, p=.46] but a greater switch cost [t (28)=−2.26, p<.05]
in the TLE patients than in the control participants.
Individual items from the subjective memory questionnaire
(MFQ) were rated on a Likert scale ranging from 1 to 7, whereby
lower values signify more of a perceived memory problem. The
cumulative mean scores for all four factors in both groups were
calculated and analysed. The factor ‘General Frequency of Forget-
ting’, which measures memory self-efficacy, was the only factor
to yield a significant result. This factor indicated that TLE patients
rated the occurrence of forgetting more frequently (represented
by a lower cumulative mean score, M=4.00, SD=.82) than con-
trol participants (M=4.79, SD=1.02), F (1, 28)=5.49, MSE=.86,
p<.05, ?2p=.16. The factor ‘Mnemonic Usage’, which measures
whether participants frequently implement daily strategies to
support memory or the effort made to avoid failures of mem-
ory, showed an almost significant difference [F (1, 28)=3.84,
MSE=1.36, p=.06, ?2p=.12] between TLE patients and controls,
suggesting that TLE patients tended to state that they use more
mnemonic strategies than controls. The remaining two factors
showed ratings which did not significantly differ between con-
trol participants and TLE patients (Seriousness of Forgetting [F (1,
28)=.00, MSE=1.06, p=.97, ?2p=.00], Retrospective Functioning [F
(1, 28)=1.35, MSE=1.47, p=.26, ?2p=.05]).
4.2. Recall performance
Recall performance at sessions 1 and 2 is illustrated in Table 4
(percentages). A 2 (group)×2 (list type)×2 (time of recall)
repeated measures ANOVA was carried out on the items recalled
at Time 1 and Time 2. There was a main effect of group [F (1,
28)=13.82, MSE=40.13, p<.001, ?2p=.33] indicating that total
recall was lower in TLE patients than in control participants, a
main effect of time of recall [F (1, 28)=149.33, MSE=14.06, p<.001,
?2p=.84], showing that controls and TLE patients recalled fewer
unrelated) [F (1, 28)=196.28, MSE=19.40, p<.001, ?2p=.88] indi-
cating that participants recalled more items from the related list.
interactions between group and time of recall [F (1, 28)=16.73,
MSE=14.06, p<.001, ?2p=.37], and between time of recall and list
type [F (1, 28)=12.59, MSE=10.51, p<.001, ?2p=.31]. None of the
other interactions reached significance.
The interaction between group and time of recall is illustrated
in Fig. 1 (percentages), which shows a steep decline in recall over
time among controls but not among patients. However, the steeper
decline could be due to differences in initial baseline recall scores.
differences at Time 1, and differences in change between Time 1
and Time 2 in two separate analyses of variance. The latter anal-
Fig. 1. Recall performance at Time 1 and Time 2 for TLE and control groups. Error
bars relate to standard error.
ysis was conducted with and without using Time 1 recall as a
covariate. At Time 1 controls recalled a greater number of words
(M=35.07, SD=8.40) than TLE patients (M=20.87, SD=9.01), F (1,
28)=19.95, MSE=75.81, p<.001, ?2p=.42. Between Time 1 and
Time 2 their recall decreased more than that of patients, F (1,
28)=16.73, MSE=56.25, p<.001, ?2p=.37. However, when recall at
Time 1 was added as a covariate, the effect of group was no longer
significant, F (1, 27)=.63, MSE=24.48, p=.43, ?2p=.02, indicating
that the apparent difference between controls and patients on rate
of decline in recall was due to baseline differences.
list type, we analysed differences in recall for related and unrelated
words at Time 1, and differences in change scores between Time
1 and Time 2. The latter analysis was conducted with and with-
out using Time 1 recall scores for related and unrelated words as
covariates. At Time 1 recall for related items was greater than for
unrelated items F (1, 29)=209.22, MSE=12.81, p<.001, ?2p=.88.
Recall decreased between Time 1 and Time 2 more for related than
for unrelated items, F (1, 29)=12.76, MSE=20.74, p<.001, ?2p=.31.
However, when recall at Time 1 was added as a covariate, the
effect of type of items was no longer significant, F (1, 27)=0.07,
MSE=15.13, p<.79, ?2p=.00, indicating that the apparent differ-
ences between related and unrelated items on rate of decline in
recall was due to baseline differences.
Given significant differences between groups in depression
levels (p<.01) and the subtests Similarities (p<.05) and Com-
prehension (p<.01), these measures were entered separately as
covariates into the main initial 2×2×2 ANOVA to control for
possible effects on recall performance. The analysis revealed
that depression [F (1, 25)=1.91, MSE=37.80, p=.18, ?2p=.07],
prehension [F (1, 25)=.01, MSE=37.80, p=.94, ?2p=.00] failed to
reach significance and had no influence on recall performance.
4.3. Metamemory accuracy: Judgement-of-Learning paradigm
Item-by-item JOLs were collected only at Time 1.6
Goodman–Kruskal’s Gamma correlation (which ranges from +1 to
−1) was used to calculate the relationship between item-by-item
JOL predictions and actual recall performance for all sixty-word
pairs (30 semantically related and 30 unrelated) at Time 1 (see
sample) had to be excluded from the sample since they recalled 0 unrelated words,
and therefore Gammas could not be computed for them, as this would create an
extreme variability in the Gamma values of the remaining 10 TLE patients. For this
reason, the effect of list type was not calculated on JOLs.
C.E. Howard et al. / Neuropsychologia 48 (2010) 921–932
JOL and FOK Gamma correlations for TLE and control groups (standard deviations
are in parentheses).
TLE n=15Controls n=15
JOL Gamma (Time 1)
FOK Gamma (Time 1)
FOK Gamma (Time 2)
Note: For FOK Gamma Time 1 TLE n=14, controls n=13.
Table 5 for Gamma correlations). A score nearer +1 indicates a high
relationship between the item-by-item JOLs and recall, whereas a
One-sample t-tests revealed that control participants’ [t
(14)=12.24, p<.001] and TLE patients’ [t (14)=12.55, p<.001] JOL
ing that both groups demonstrated a level of metacognitive ability
and that their item-by-item JOLs were not made by chance. This
a degree of metacognitive competence when making their item-
by-item JOLs. Moreover, independent-samples t-tests revealed no
significant differences in JOL Gammas between controls and TLE
patients [t (28)=1.09, p=.29] when considering all sixty-word
In order to test whether the two groups used the ratings for
ANOVA was also carried out on the number of times (proportions
of use) each JOL rating was used (see Fig. 2). There was no main
ing that overall use of ratings did not significantly differ between
groups, a main effect of rating type [F (5, 140)=2.67, MSE=190.68,
p<.05, ?2p=.09], showing that some ratings were more frequently
used than others. Finally, the interaction between group and rating
of JOL ratings across the entire list.
4.4. Recall readiness/study-time allocation
Metamemory control was measured as the overall study-time
allocated to studying the sixty-word pairs between groups. The
overall mean study-time (in seconds) for the sixty-word pairs was
calculated for both groups, as well as the mean time spent studying
(see Table 6). To determine whether there was a difference in the
group’s ability to adjust the time spent studying the words depen-
dent on their level of difficulty, a 2 (group)×2 (list type) ANOVA
was carried out on study-time allocation. The results showed
no main effect of group [F (1, 28)=1.96, MSE=17435.22, p=.17,
?2p=.07], indicating that the amount of time spent by control par-
ticipants and TLE patients studying the words was not significantly
different. Moreover, the significant main effect of list type [F (1,
Fig. 2. Judgement-of-Learning ratings’ proportions of use in TLE patients and con-
trols. Error bars refer to standard errors.
TLE n=15 M
Controls n=15 M
Related word pair study-time
Unrelated word pair study-time
28)=21.62, MSE=3962.81, p<.001, ?2p=.44] indicated that both
groups spent more time studying the unrelated than the semanti-
cally related word pairs. The interaction between group and list
type did not reach significance [F (1, 28)=2.74, MSE=3962.81,
p=.11, ?2p=.09], suggesting that metamemory control processes
at the item-by-item level were not different in patients with TLE
4.5. Global JOLs
A 2 (group)×2 (time of global JOL) repeated measures ANOVA
was carried out on the global JOL predictions made at Time 1
and Time 2. There was no main effect of group [F (1, 28)=.00,
MSE=134.16, p=.99 ?2p=.00] indicating that global JOLs were
not significantly different between groups. As expected, the main
effect of time of global JOLs [F (1, 28)=27.81, MSE=72.16, p<.001,
?2p=.50] demonstrated that global JOLs were higher at Time 1
(nearer to the time of encoding) than at Time 2. There was no evi-
dence of an interaction between group and time of global JOL [F (1,
28)=.60, MSE=72.16, p=.45, ?2p=.02], revealing that groups were
not significantly different at either Time 1 or Time 2 and suggesting
that both groups were able to adjust the global prediction after the
Time 1 recall.
Finally, following a reviewer’s suggestion, we also calculated
what could be called a ‘total JOL’ by adding the number of items
that received a positive JOL (60%, 80%, and 100%) and compared it
with the global JOL made by each participant with the group as the
metacognitive awareness’ did not significantly differ between TLE
patients and controls either [F (1, 28)=1.10, MSE=41.49, p=.30,
4.6. Metamemory accuracy: non-directional discrepancy scores
In line with the argument already proposed by Moulin et al.
(2000) for Alzheimer Disease patients, here too non-directional
discrepancy scores (Hertzog, Saylor, Fleece, & Dixon, 1994) were
used as a direct measure of the participantsˇ ı accuracy in predicting
their memory performance (instead of simply inferring under- or
accompanied by similar predictions between groups, Prevey et al.,
1988). Non-directional discrepancy scores were calculated as the
modulus of difference between global JOLs (predictions) and actual
recall both at Time 1 and Time 2, for both controls and TLE patients
(see Fig. 3). The rationale for using a non-directional method can
be understood by considering that the use of a directional score
can lead to erroneous levels of over- and underestimation when
analysing group performance. For instance, if two participants in
the same group were to over- and under-estimate by two items
this would lead to a group mean of zero, as both participants scores
would cancel each other out.
The unsigned absolute difference for each group at both Time
1 and Time 2 was analysed in a 2 (group)×2 (time of non-
directional discrepancy score) repeated measures ANOVA. The
results revealed no main effect of group [F (1, 28)=.69, MSE=83.55,
p=.41, ?2p=.02]. There was a main effect of time [F (1, 28)=10.36,
MSE=35.72, p<.01, ?2p=.27], with Time 1 having higher levels
than Time 2. No evidence of an interaction was revealed between
C.E. Howard et al. / Neuropsychologia 48 (2010) 921–932
Fig. 3. Non-directional discrepancy scores at Time 1 and Time 2 for TLE and control
groups. The modulus difference between global JOL scores (predictions) and actual
recall performance was used to calculate non-directional discrepancy scores. Error
bars relate to standard error.
group×time of the discrepancy score [F (1, 28)=2.22, MSE=35.72,
4.7. Metamemory accuracy: recognition and Feeling-of-Knowing
Non-recalled or incorrectly recalled items were used in a recog-
nition task. This was done separately for Time 1 and Time 2. A 2
(group)×2 (time) ANOVA was carried out on the proportion of
ipants (Time 1: M=76.74, SD=14.16; Time 2: M=60.87, SD=13.60)
recognised a significantly greater percentage of target words than
the TLE patients (Time 1: M=51.26, SD=23.50; Time 2: M=40.71,
SD=20.56) [F (1, 28)=13.32; MSE=586.53; p=.001; ?2p=.32] and
that both groups recognised more items at Time 1 than at Time
2 [F (1, 28)=27.95; MSE=93.64; p=.001; ?2p=.50]. The interac-
tion between group and time of recall was not significant [F (1,
28)=1.13; MSE=93.64; p=.30; ?2p=.04].
Rate of forgetting in recognition, which was obtained by cal-
culating the difference in recognition scores between Time 1 and
Time 2 and dividing it by recognition at Time 1 was almost identi-
cal in controls and patients (controls M=.19, SD=.19; TLE patients
M=.21, SD=.22), t (28)=−.19; p=.86.
Goodman–Kruskal Gamma correlations between the FOK
answering correctly to all word pairs tested and one TLE patient
responding with the same rating (20%) for all word pairs in this
part of the task at Time 1 were excluded from the analyses.
tions were significantly different from zero for control participants
[t (12)=3.95, p<.01] and marginally different from zero for TLE
patients [t (13)=1.98, p=.07] indicating that both groups were
metacognitively competent when making their FOK Gamma cor-
7We did not analyze the effect of list type on FOKs because during recall a sig-
nificantly greater proportion of related word pairs were recalled than unrelated in
both groups (see Table 4). As the FOK task only involves testing the word pairs that
were not recalled, the majority of word pairs tested would therefore be unrelated in
the recognition test, creating also an unbalanced set of data. All together, although
it would be interesting to look at the effects of list type (or difficulty, in general)
on metamemory abilities in future research, we considered our current data set not
suitable for this purpose.
relations at Time 1. FOK Gamma correlations made at Time 2 were
not significantly different from zero (p>.05) for either group.
Independent-samples t-tests revealed that FOK Gamma corre-
lations were not significantly different between the control group
1, t (25)=.96, p=.34, nor at Time 2, t (28)=−.27, p=.79 (controls:
M=.09, SD=.26; and TLE patients: M=.13, SD=.42).
4.9. Correlation analysis of the MFQ factors with recall
In order to determine whether there was any relationship
between subjective views of memory measured by the MFQ and
actual recall performance, Pearson’s correlation coefficients (r)
formance at Time 1 and Time 2 for controls and TLE patients. The
General Frequency of Forgetting score (high scores indicate less
perceived forgetting) correlated (r=.57, p<.05) with recall perfor-
mance at Time 2 in the TLE patients (n=15), indicating that recall
was higher in patients who reported lower forgetting rates. No
other significant correlations between the MFQ factors and recall
performance were found for either the controls or TLE patients.
4.10. Correlation analysis of epilepsy variables and recall
In order to determine whether there were any specific epilepsy
variables (laterality, seizure type, age of onset, duration, frequency
of seizures, number of AEDs, number of seizures within the 4-week
interval) which had an influence on recall performance at Time 1
and Time 2, Pearson’s correlation coefficients (r) were computed.
None of the epilepsy variables significantly correlated with recall
performance (all p>.05).
Pearson’s correlation coefficients (r) were also computed
between all epilepsy variables and standardised subtests of the
WMS-III (n=15). Significant correlations were revealed between
the subtest Digit Span of the WMS-III and laterality (r=−.73,
p<.001), showing that laterality appeared to be negatively related
to scores on the Digit Span task accounting for 53% of the variation
in scores; age of onset (r=.59, p<.05) appeared to be positively
related to scores on the Digit Span task accounting for 35% of the
variance; and duration (r=−.61, p<.05) appeared to be negatively
relations were also detected between the number of AEDs with
Logical Memory I (r=−.53, p<.05) indicating that AEDs appeared
to be negatively related to Logical Memory I accounting for 28% of
the variation in scores.
To our knowledge, this is the first study to investigate the exis-
tence of a metamemory deficit in TLE patients, and to examine
whether the episodic memory impairment typically observed in
TLE patients (Bell & Giovagnoli, 2007; Leritz et al., 2006) might be
due to it. In that context, it is worth mentioning that other clini-
cal populations with episodic memory declines have shown poor
2007) and hence a link between metamemory and memory might
help to understand the memory deficits observed in TLE patients.
In previous research examining metamemory abilities in TLE
patients, Prevey et al. (1988, 1991) concluded that metamemory,
and specifically monitoring processes, are impaired in TLE patients
after observing that, differently from controls, patients tended to
overestimate their memory performance. However, in those stud-
ies this conclusion was based solely on the fact that whereas
memory predictions were equivalent for TLE patients and con-
C.E. Howard et al. / Neuropsychologia 48 (2010) 921–932
trols, the recall performance was lower for TLE patients than for
controls. No adequate measure of accuracy was reported to sup-
port such claim (e.g. discrepancy scores). Moreover, these studies
did not examine monitoring for episodic memory, but measured
metamemory for short-term memory span (serial memory), and
for factual (semantic) information.
In the present study we examined verbal episodic memory
and metamemory abilities related to it. Accuracy in mon-
itoring processes for verbal episodic memory was assessed
by measuring item-by-item Judgements-of-learning (JOL) and
itoring. We also assessed accuracy of a more global form of recall
prediction (Global JOL). Finally, a measure of metacognitive control
processes (study time allocation, Mazzoni & Cornoldi, 1993; Son &
in TLE patients.
In the present study, TLE patients showed a clear and signifi-
cant impairment in episodic memory. However, no impairment in
metamemory was observed in either monitoring or control pro-
cesses. We discuss these results separately.
Consistent with previous results (see Bell & Giovagnoli, 2007;
Leritz et al., 2006 for reviews), in this study TLE patients performed
significantly worse than controls in recall when memory was mea-
sured 30min after acquisition. TLE patients also showed poorer
performance at this time when tested for recognition for the non-
Four weeks after acquisition, and contrary to the hypothesis of
accelerated forgetting (AF), the results showed in both groups no
significant difference in forgetting rates for recall.
After 4 weeks, recognition of non-recalled items was still sig-
nificantly poorer for TLE participants than for controls, although
the forgetting rate was not significantly either different between
the two groups. This persistent difficulty of TLE patients in recog-
nising the words they could not recall is consistent with the idea
that the memory deficits observed in TLE patients are mainly due
to a deficit at the encoding stage of the memory process, which is
typically observed in patients with damage in their temporal lobes
(Shimamura, Janowsky, & Squire, 1991). Indeed, if the information
had been encoded to a similar level as in the control participants,
presentation of the non-recalled items should have facilitated their
retrieval, as is typically observed in patients with intact temporal
lobes but damaged frontal cortex (Aggleton & Brown, 1999).
ated forgetting. Although testing directly AF was not the main aim
of our study, and thus initial encoding was not equated between
tend to forget more quickly than controls during the first 4 weeks.
Accordingly, at least in the present study, AF is not a viable expla-
nation of the initial poorer recall in TLE patients. These results,
however, are in line with previous research showing that AF is
not always present in TLE patients (Bell, 2006; Bell et al., 2005;
Giovagnoli et al., 1995; Helmstaedter et al., 1998). In this context,
it is worth noting that Bell and Giovagnoli (2007) noticed that the
majority of the single cases showing AF had relatively atypical TLE,
in that they had a history of either severe traumatic brain injury
with extensive bilateral cortical atrophy and frequent seizures
(Holdstock, Mayes, Isaac, Gong, & Roberts, 2002), probable para-
neoplastic encephalitis, bilateral EEG and MRI abnormalities, and
an extremely high seizure frequency (O’Connor, Sieggreen, Ahern,
Schomer, & Mesulam, 1997). AF may also be more prominent in
certain subtypes of TLE, such as transient epileptic amnesia (TEA),
in which 44% patients report symptoms of AF (Butler et al., 2007).
Other possible factors determining when AF is more likely to be
present in TLE have been discussed by Bell and Giovagnoli (2007)
and Butler and Zeman (2008) in their recent reviews.
Finally, it is worth mentioning that the differences between TLE
in the current study could not be explained by greater levels of
on the similarities and comprehension subtests of the WAIS-III)
as this was ruled out when their role was tested in the ANCO-
VAs. However, age of onset, duration of the epilepsy, and number
of anti-epileptic drugs had an effect on subtests of the WMS-III
5.2. Metacognitive monitoring and control (metamemory)
Although some have suggested and found that deficits in
memory in TLE patients can be attributed to poor metacogni-
tive monitoring and control processes (Prevey et al., 1988, 1991),
this hypothesis is not supported by the present data. TLE patients
were not significantly different from controls in any of the online
metacognitive measures used in this study.
Both controls and patients were able to predict with relative
accuracy which items they would have been able to recall and
which they would have not, as Judgements-of-Learning accuracy
was above chance in both groups, and not significantly different
in TLE and controls. This suggests that TLE patients, similarly to
controls, monitor effectively online learning of verbal material. In
ease with which each single item is learned. In this case accuracy
depends on the extent to which perceived ease of learning corre-
sponds to later probability of retrieval.
Accuracy of FOK judgements was also not significantly differ-
ent between the two groups. FOK ratings are supposedly based
on partial accessibility of unrecallable items, and on the sense of
is also determined by accessibility and familiarity.
Overall, the results on accuracy of these online monitoring tasks
fail to show any clear metacognitive deficit in TLE patients, sug-
gesting that online monitoring might be adequate and memory
problems in this group should be attributed to a deficit in other
supported by the observation that differences in memory between
TLE patients and controls obtained after 4 weeks disappear when
the difference in initial performance is taken into account.
Global judgements are another type of metamemory evalua-
tion, which reflects beliefs about oneself and about the difficulty
of the task. Thus, individuals who believe to be poor at memory
tasks should give lower global estimates than those who believe
to be good. Also in this case no difference was found between
groups, in either in magnitude or accuracy (discrepancy scores)
of global JOLs. If anything, the numbers go in the opposite direc-
(Global JOL=22.47, recall=35.07) and patients slightly overesti-
mating it (Global JOL=24.13; recall=20.87). Although this slight
overestimation in patients might be indicative of a mild metacog-
nitive deficit, we do not think this is the case in our sample for two
reasons. First, the absolute amount of overestimation is quite small
and non-significant (t (14)=.98, p=.35). Patients seem to be rather
‘on target’, and there is no theoretical reason to think that ‘on tar-
Second, the analysis of discrepancy scores (i.e. accuracy of global
JOLs) did not reveal any significant effect of group, nor a significant
interaction, suggesting that the discrepancy between judgement
and recall is not different in the two groups. In addition, one should
consider that accuracy was not significantly different between TLE
C.E. Howard et al. / Neuropsychologia 48 (2010) 921–932
patients and controls also when metacognitive judgements were
assessed item-by-item. Although global and item-by-item judge-
ments are probably based on partly different information, this
additional lack of difference in accuracy of judgements provides
further support to the claim that metacognitive evaluations are not
impaired in TLE patients.
Had the difference (between underestimation in controls and
on-target estimation in patients) been clearly significant, it might
have suggested that patients are more aware of their memory abil-
ities than controls. Although this conclusion cannot be drawn from
the present data, self-report measures (MFQ) from this study pro-
vide some support to it. Patients report more frequent forgetting
(p=.03), and a clear trend in using more strategies than controls
(p=.06). Both elements indicate that patients might be aware of
their memory problems. Besides the clear indication that they for-
get more, reporting a greater use of strategies might represent an
attempt to compensate for the perceived memory impairment.
gies (control processes) are employed. The way study-time was
allocated was also not significantly different in the two groups,
although numerically patients tended to devote more study time
to studying more difficult (unrelated) than easier (related) pairs,
and no interaction involved groups. In spite of this similarity, how-
ever, patients initially recalled fewer items than controls, an effect
that again suggests encoding problems rather than monitoring and
nitive deficit in TLE patients, which provides additional support
to previous dissociations between memory and metamemory in
neuropsychological patients with temporal damage (Shimamura &
aware of their memory problems, and may be better at predicting
global memory performance (global JOLs). Since this is potentially
a counterintuitive result with crucial implications we are currently
focusing our research on this specific issue.
5.3. Executive functions
Delis–Kaplan Executive Function Test, with a significantly greater
switch cost in the Inhibition/switching condition in the Stroop
ficulty to inhibit the automatic response in the Inhibition section of
the Hayling test. Both these tests measure inhibition abilities, and
confirm that a focal frontal lesion is not necessary to observe this
type of deficit in clinical populations (Andrés, 2003; Andrés & Van
der Linden, 2001). The presence of this relative executive dysfunc-
tion in a sample of TLE patients with intact metamemory abilities
indicates that metamemory is likely to run independently at least
from inhibitory mechanisms. It would therefore be interesting to
investigate to what extent metamemory is dependent from other
al., 2000). Finally, it is worth mentioning that the significant group
effects on executive measures (or the memory measures described
earlier) speak against the possibility that the absence of deficits
observed in our TLE sample is due to a lack of statistical power.
to what has been suggested in previous studies, this deficit is not
explained by metamemory problems (Prevey et al., 1988, 1991),
accelerated forgetting (Blake et al., 2000) or mood disturbances
(e.g. Ba˜ nos et al., 2004). It is more likely that the poor memory
performance shown by the TLE patients in the current sample is
due to impairments occurring at encoding resulting from temporal
damage (see Aggleton & Brown, 1999; Squire, 1992 for reviews),
and similar to the deficits observed in stronger forms of amnesia
(Mayes et al., 2003; O’Connor et al., 1997).
This research was funded by an Economic and Social Research
Council (ESRC) Collaborative Awards in Science and Engineering
(CASE) Studentship (PTA-033-2006-00006) obtained by Giuliana
Mazzoni, in collaboration between the University of Plymouth and
to thank the out-patients from Derriford Hospital, Plymouth Hos-
pitals NHS Trust who participated in this study, the psychology
undergraduates from the School of Psychology at the University
of Plymouth and members from the Paid Supporters Group.
Word pairs presented in the JOL task.
Related word pairsUnrelated word pairs
NB: The second words in the pairs are the target words.
Aggleton, J., & Brown, M. (1999). Episodic memory, amnesia, and the hippocampal-
anterior thalamic axis. Behavioural and Brain Sciences, 22, 425–489.
Andrés, P. (2003). Frontal cortex as the central executive of working memory: Time
to revise our view. Cortex, 39, 871–895.
Andrés, P., & Van der Linden, M. (2001). Supervisory attentional system in patients
with focal frontal lesions. Journal of Clinical and Experimental Neuropsychology,
report of cognitive abilities in temporal lobe epilepsy: Cognitive, psychosocial,
and emotional factors. Epilepsy & Behavior, 5, 575–579.
Bell, B. D. (2006). WMS-III Logical memory performance after a two-week delay in
temporal lobe epilepsy and control groups. Journal of Clinical and Experimental
Neuropsychology, 28, 1435–1443.
Bell, B. D., & Giovagnoli, A. R. (2007). Recent innovative studies of memory in tem-
poral lobe epilepsy. Neuropsychological Review, 17, 455–476.
and the selective reminding test. The conventional 30-minute delay suffices.
Psychological Assessment, 17, 103–109.
C.E. Howard et al. / Neuropsychologia 48 (2010) 921–932
In J. Dunlosky, & R. A. Bjork (Eds.), Handbook of metamemory and memory (pp.
Blake, R. V., Wroe, S. J., Breen, E. K., & McCarthy, R. A. (2000). Accelerated forgetting
in patients with epilepsy. Brain, 123, 472–483.
Burgess, P. W., & Shallice, T. (1997). Hayling and Brixton tests. London: The Psycho-
The syndrome of transient epileptic amnesia. Annals of Neurology, 61, 587–598.
Butler, C., & Zeman, A. (2008). Recent insights into the impairment of memory
in epilepsy: Transient epileptic amnesia, accelerated long-term forgetting and
remote memory impairment. Brain, 131, 2243–2263.
Collette, F., & Van der Linden, M. (2002). Brain imaging of the central executive
component of working memory. Neuroscience and Biobehavioral Reviews, 26,
Corcoran, R., & Upton, D. (1993). A role for the hippocampus in card sorting? Cortex,
Correa, D. D., Graves, R. E., & Costa, L. (1996). Awareness of memory deficit in
Alzheimer’s disease patients and memory impaired older adults. Aging, Neu-
ropsychology and Cognition, 3, 215–228.
Cosentino, S., Metcalfe, J., Butterfield, B., & Stern, Y. (2007). Objective metamem-
ory testing captures awareness of deficit in Alzheimerˇ ıs disease. Cortex, 43,
Delis, D. C., Kaplan, E., & Kramer, J. H. (2001). Delis–Kaplan executive function system:
Examiner’s manual. San Antonio, TX: The Psychological Corporation.
Elixhauser, A., Leidy, N. K., Meador, K., Means, E., & Willian, M. K. (1999). The rela-
tionship between memory performance, perceived cognitive functioning, and
mood in patients with epilepsy. Epilepsy Research, 37, 13–24.
Fernandez-Duque, D., Baird, J. A., & Posner, M. (2000). Executive atten-
tion and metacognitive regulation. Consciousness and Cognition, 9, 288–
Perspectives on the development of memory and cognition. Hillsdale, NJ: Erlbaum.
Gallassi, R., Morreale, A., Lorusso, S., Pazzaglia, P., & Lugaresi, E. (1988). Epilepsy
presenting as memory disturbance. Epilepsia, 29, 624–629.
Gilewski, M. J., Zelinski, E. M., & Warner Schaie, K. (1990). The memory functioning
questionnaire for assessment of memory complaints in adulthood and old age.
Psychology and Aging, 5, 482–490.
Giovagnoli, A. R., Casazza, M., & Avanzini, G. (1995). Visual learning on a selective
Epilepsia, 36, 704–711.
Giovagnoli, A. R., Mascheroni, S., & Avanzini, G. (1997). Self-reporting of every-
day memory in patients with epilepsy: Relation to neuropsychological, clinical,
pathological and treatment factors. Epilepsy Research, 28, 119–128.
performance–complaint relationship in patients with epilepsy: A matter
of daily demands? Epilepsy Research, 32, 401–409.
Harris, A. J. (1974). Harris test of lateral dominance. New York: The Psychological
Helmstaedter, C., Hauff, M., & Elger, C. E. (1998). Ecological validity of list-learning
tests and self-reported memory in healthy individuals and those with tem-
poral lobe epilepsy. Journal of Clinical and Experimental Neuropsychology, 20,
Hermann, B. P., Wyler, A. R., Steenman, H., & Richey, C. T. (1988). The interrelation-
ship between language function and verbal learning/memory performance in
patients with complex partial seizures. Cortex, 24, 345–353.
Hermann, B. P., Seidenberg, M., Haltiner, A., & Wyler, A. R. (1991). Mood state in
unilateral temporal lobe epilepsy. Biological Psychiatry, 30, 1205–1218.
Hermann, B. P., Seidenberg, M., Schoenfeld, J., Peterson, J., Leveroni, C., & Wyler,
A. R. (1996). Empirical techniques for determining the reliability, magnitude,
and pattern of neuropsychological change after epilepsy surgery. Epilepsia, 37,
Hertzog, C., Saylor, L. L., Fleece, A. M., & Dixon, R. A. (1994). Metamemory and aging:
Relations between predicted, actual and perceived memory, task performance.
Aging and Cognition, 1, 203–237.
Holdstock, J. S., Mayes, A. R., Isaac, C. L., Gong, Q., & Roberts, N. (2002). Differential
involvement of the hippocampus and temporal lobe cortices in rapid and slow
learning of new semantic information. Neuropsychologia, 40, 748–768.
Janowsky, J. S., Shimamura, A. P., & Squire, L. R. (1989). Memory and metamemory:
Comparisons between frontal lobe lesions and amnesic patients. Psychobiology,
Keleman, W. L., & Weaver, C. A. (1997). Enhanced metamemory at delays: Why do
judgements of learning improve over time? Journal of Experimental Psychology:
Learning Memory and Cognition, 23, 1394–1409.
Keller, S., Baker, G., Downes, J., & Roberts, N. (2009). Quantitative MRI of the pre-
frontal cortex and executive function in patients with temporal lobe epilepsy.
Epilepsy and Behavior, 15, 186–195.
Kikyo, H., & Miyashita, Y. (2004). Temporal lobe activations of “feeling-of-knowing”
induced by face-name associations. Neuroimage, 23, 1348–1357.
Kikyo, H., Ohki, K., & Miyashita, Y. (2002). Neural correlates for feeling-of-knowing:
An fMRI parametric analysis. Neuron, 36, 177–186.
Leritz, E. C., Grande, L. J., & Bauer, R. M. (2006). Temporal lobe epilepsy as a model
to understand human memory: The distinction between explicit and implicit
memory. Epilepsy & Behavior, 9, 1–13.
Light, L. L. (1991). Memory and aging: Four hypotheses in search of data. Annual
Review of Psychology, 42, 333–376.
A.,& Elger,C. E.(1998). The
Mameniskiene, R., Jatuzis, D., Kaubrys, G., & Budrys, V. (2006). The decay of memory
between delayed and long-term recall in patients with temporal lobe epilepsy.
Epilepsy & Behavior, 8, 278–288.
Maril, A., Simons, J. S., Mitchell, J. P., Schwartz, B. L., & Schacter, D. L. (2003). Feeling-
of-knowing in episodic memory: An event-related fMRI study. Neuroimage, 18,
Martin, R. C., Sawrie, S. M., Edwards, R., Roth, D. L., Faught, E., Kuzniecky, R. J., et
al. (2000). Investigation of executive function change following anterior tempo-
ral lobectomy: Selective normalization of verbal fluency. Neuropsychology, 14,
Mayes, A. R., Isaac, C. L., Holdstock, J. S., Cariga, P., Gummer, A., & Roberts, N. (2003).
Long-term amnesia: A review and detailed illustrative case study. Cortex, 39,
Mazzoni, G., & Cornoldi, C. (1993). Strategies in study time allocation: Why is study
time sometimes not effective? Journal of Experimental Psychology: General, 122,
Mazzoni, G., Cornoldi, C., & Marchitelli, G. (1990). Do memorability ratings affect
study-time allocation? Memory & Cognition, 18, 196–204.
McGlynn, S. M., & Kaszniak, A. W. (1991). When metacognition fails: Impaired
awareness of deficit in Alzheimer’s disease. Journal of Cognitive Neuroscience,
Metcalfe, J., & Kornell, N. (2003). The dynamics of learning and allocation of study
Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., & Wager, T. D.
(2000). The unity and diversity of executive functions and their contributions to
complex frontal lobe tasks: A latent variable analysis. Cognitive Psychology, 41,
Modirrousta, M., & Fellows, L. K. (2008). Medial prefrontal cortex plays a critical
and selective role in ‘feeling-of-knowing’ meta-memory judgements. Neuropsy-
chologia, 46, 2958–2965.
Moulin, C. J. A., Perfect, T. J., & Jones, R. W. (2000). Global predictions of memory
in Alzheimer’s disease: Evidence for preserved metamemory monitoring. Aging,
Neuropsychology, and Cognition, 7, 230–244.
Nelson, T. O. (1984). A comparison of current measures of the accuracy of feeling-
of-knowing predictions. Psychological Bulletin, 95, 109–133.
Nelson, T. O., & Narens, L. (1990). Metamemory: A theoretical framework and new
findings. The Psychology of Learning and Motivation, 26, 125–173.
Nelson, H. E., & Willison, J. (1991). Restandardisation of the NART against the WAIS-
R. In H. E. Nelson (Ed.), National adult reading test (NART). Test manual (2nd ed.,
pp. 13–23). Windsor: Nelson.
O’Connor, M., Sieggreen, M. A., Ahern, G., Schomer, D., & Mesulam, M. (1997). Accel-
erated forgetting in association with temporal lobe epilepsy and paraneoplastic
encephalitis. Brain and Cognition, 35, 71–84.
O’Shea, M. F., Saling, M. M., Bladin, P. F., & Berkovic, S. F. (1996). Does naming con-
tribute to memory self-report in temporal lobe epilepsy? Journal of Clinical and
Experimental Neuropsychology, 18, 98–109.
Pannu, J. K., & Kaszniak, A. W. (2005). Metamemory experiments in neurological
populations: A review. Neuropsychology Review, 15, 105–129.
frontal lobe damage. Journal of the International Neuropsychological Society, 11,
Prevey, M. L., Delaney, R. C., & Mattson, R. H. (1988). Metamemory in temporal lobe
epilepsy: Self-monitoring of memory functions. Brain and Cognition, 7, 298–
temporal lobe epilepsy: Monitoring knowledge inaccessible to conscious recall.
Cortex, 27, 81–92.
Rapcsak, S. Z., Pannu, J. K., & Kaszniak, A. W. (2005). Monitoring and control process-
ing in prosopagnosia. Manuscript in preparation.
Schnyer, D. M., Verfaellie, M., Alexander, M., LaFleche, G., Nicholls, L., & Kaszniak,
A. W. (2004). A role for right medial prefrontal cortex in accurate feeling-of-
knowing judgements: Evidence from patients with lesions to frontal cortex.
Neuropsychologia, 42, 957–966.
Schnyer, D. M., Nicholls, L., & Verfaellie, M. (2005). The role of VMPC in metamemo-
rial judgements of content retrievability. Journal of Cognitive Neuroscience, 17,
Shimamura, A. (2000). Toward a cognitive neuroscience of metacognition. Con-
sciousness and Cognition, 9, 313–323.
Shimamura, A. P., & Squire, L. R. (1986). Memory and metamemory: A study of
feeling-of-knowing phenomenon in amnesic patients. Journal of Experimental
Psychology: Learning, Memory and Cognition, 12, 452–460.
Shimamura, A., Janowsky, J., & Squire, L. (1991). What is the role of frontal lobe dam-
age in memory disorders? In H. D. Levin, H. M. Eisenberg, & A. L. Benton (Eds.),
Frontal lobe functioning and dysfunction. New York: Oxford University Press.
Son, L. K., & Metcalfe, J. (2000). Metacognitive and control strategies in study-time
allocation. Journal of Experimental Psychology: Learning, Memory, and Cognition,
Souchay, C. (2007). Metamemory in Alzheimer’s disease. Cortex, 43, 987–1003.
Souchay, C., Isingrini, M., & Espagnet, L. (2000). Relations between feeling-of-
knowing and frontal lobe functioning in older adults. Neuropsychology, 14,
Squire, L. R. (1992). Memory and the hippocampus: A synthesis from findings with
rats, monkeys, and humans. Psychological Review, 99, 195–231.
Thompson, P. J., & Corcoran, R. (1992). Everyday memory failures in people with
epilepsy. Epilepsia, 33, 18–20.
C.E. Howard et al. / Neuropsychologia 48 (2010) 921–932
Vermeulen, J., Aldenkamp, A. P., & Alpherts, W. C. (1993). Memory complaints in
epilepsy: Correlations with cognitive performance and neuroticism. Epilepsy
Research, 15, 157–170.
after frontal lobe lesions. Neuropsychology, 12, 268–277.
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,
Wechsler, D. (1997a). Manual for the Wechsler adult intelligence scale (3rd ed.). San
Antonio, TX: The Psychological Corporation.
Wechsler, D. (1997b). Manual for the Wechsler memory scale (3rd ed.). San Antonio,
TX: The Psychological Corporation.
Zigmond, A. S., & Snaith, R. P. (1983). The hospital anxiety depression scale. Acta
Psychiatrica Scandinavica, 67, 361–370.