Download Recall and recognition memory in amnesia

Neuropsychologia 45 (2007) 1232–1246
Recall and recognition memory in amnesia: Patients with hippocampal,
medial temporal, temporal lobe or frontal pathology
Michael D. Kopelman a,∗ , Peter Bright a,1 , Joseph Buckman a , Alex Fradera a ,
Haruo Yoshimasu a,2 , Clare Jacobson a , Alan C.F. Colchester b
a
King’s College London, Institute of Psychiatry, London, United Kingdom
b Kent Institute of Medicine and Health Sciences, University of Kent,
Canterbury, Kent, United Kingdom
Received 28 July 2005; received in revised form 10 October 2006; accepted 22 October 2006
Available online 30 November 2006
Abstract
The relationship between recall and recognition memory impairments was examined in memory-disordered patients with either hippocampal,
medial temporal, more widespread temporal lobe or frontal pathology. The Hirst [Hirst, W., Johnson, M. K., Phelps, E. A., & Volpe, B. T. (1988).
More on recognition and recall in amnesics. Journal of Experimental Psychology: Learning, Memory, & Cognition, 14, 758–762] technique for
titrating exposure times was used to match recognition memory performance as closely as possible before comparing recall memory scores.
Data were available from two different control groups given differing exposure times. Each of the patient groups showed poorer recall memory
performance than recognition scores, proportionate to the difference seen in healthy participants. When patients’ scores were converted to Zscores, there was no significant difference between mean Z-recall and Z-recognition scores. When plotted on a scatterplot, the majority of the
data-points indicating disproportionately low recall memory scores came from healthy controls or patients with pathology extending into the
lateral temporal lobes, rather than from patients with pathology confined to the medial temporal lobes. Patients with atrophy extending into the
parahippocampal gyrus (H+) performed worse than patients with atrophy confined to the hippocampi (H−); but, when H− patients were given
a shorter exposure time (5 s) and compared with H+ at a longer exposure (10 s), their performance was virtually identical and did not indicate
any disproportionate recall memory impairment in the H− group. Parahippocampal volumes on MRI correlated significantly with both recall and
recognition memory. The possibility that findings were confounded by inter-stimulus artefacts was examined and rejected. These findings argue
against the view that hippocampal amnesia or memory disorders in general are typically characterised by a disproportionate impairment in recall
memory. Disproportionate recall memory impairment has been observed in a number of published cases, and the reason for the varying pattern
obtained across hippocampal patients requires further examination.
© 2006 Elsevier Ltd. All rights reserved.
Keywords: Amnesia; Memory disorders; Hippocampus; Medial temporal lobes; Recognition; Recall
1. Introduction
It is widely accepted that recognition memory reflects
a combination of a familiarity judgement and a degree of
conscious recollection, whereas recall memory depends upon
∗
Corresponding author at: 3rd Floor, Block 8, South Wing, St. Thomas’s
Hospital, London SE1 7EH, United Kingdom. Tel.: +44 207 188 5396;
fax: +44 207 633 0061.
E-mail address: [email protected] (M.D. Kopelman).
1 Now at Anglia Ruskin University, Cambridge, United Kingdom.
2 Now at Dept. of Neuropsychiatry, Showa University, Northern Yokohama
Hospital, Japan.
0028-3932/$ – see front matter © 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.neuropsychologia.2006.10.005
recollective processes (Giovanello & Verfaellie, 2001; Jacoby,
Toth, & Yonelinas, 1993; Mayes, Holdstock, Isaac, Hunkin,
& Roberts, 2002). However, there is considerable controversy
concerning the effects of amnesia upon recall and recognition
memory, respectively. One view is that hippocampal amnesia,
including cases of developmental amnesia, is specifically characterised by a disproportionate impairment in recall memory,
whereas recognition memory is preserved. A second view is
that amnesic or memory-disordered patients in general manifest
disproportionate recall memory impairment. A third view is
that amnesia, including that which follows focal hippocampal
pathology, produces a proportionate impairment in both recall
and recognition memory. This controversy relates to views of
M.D. Kopelman et al. / Neuropsychologia 45 (2007) 1232–1246
hippocampal function—whether the hippocampi are involved
in encoding/retrieval processes in general, or whether they
contribute specifically to the contextual/associative/relational
memory processes which characterise recollection. This, in turn,
relates to whether recollection (recall) and familiarity (recognition) should be viewed as ‘redundant’ processes (recollection
incorporates whatever happens in familiarity plus further operations), ‘independent’ (different but overlapping operations) or
‘exclusive’ processes (different and non-overlapping).
On the basis of a meta-analysis of single case and small
group studies of memory-disordered patients, Aggleton and
Shaw (1996) (see also Aggleton & Brown, 1999) argued that
patients with pathology within the hippocampi, fornices, mamillary bodies, mamillo-thalamic tract or anterior thalami showed
impairments on verbal and visual recall but not recognition
memory. In such patients with damage to what they called the
‘extended hippocampal circuit’, memory based on familiarity
judgements (recognition) was intact, whereas recall memory,
involving recollection of contextual features, such as time and
spatial location, was impaired. They argued that combined
hippocampal and parahippocampal (including entorhinal and
perirhinal) lesions were required to produce an impairment in
familiarity-based or recognition memory. However, there was
a ‘floor’ effect in the recall scores of the subjects with larger
lesions in their meta-analysis, making interpretation difficult.
There are other cases, which provide support for this hypothesis. Vargha-Khadem, Gadian, Watkins, and Connelly (1997)
described three patients with a developmental amnesia for
everyday events, resulting from brain injuries in infancy or
early childhood. These patients showed a pronounced loss of
hippocampal volume bilaterally, and their neuropsychological
test performance revealed impairments on verbal and visual
recall but not recognition memory, the latter being tested
with material that included lists of words, non-words, familiar faces and unfamiliar faces. These findings suggested that,
whilst recall of episodic memories was impaired as a result of
these patients’ hippocampal pathology, recognition memory and
semantic memory were spared. More detailed evidence in support of this in one of these cases was published by Baddeley,
Vargha-Khadem, and Mishkin (2001), using the Doors and People Test battery (Baddeley, Emslie, & Nimmo-Smith, 1994).
Moreover, Mayes et al. (2002); (Holdstock et al., 2002; Mayes
et al., 2004) have described in detail an adult-onset patient,
YR, who suffered selective bilateral lesions to the hippocampi.
Across 43 recognition memory tests, YR showed significant
impairment relative to controls, but the impairment was very
minor (mean Z = −0.5) and clinically significant (>2S.D.) in
only 10% of tests. By contrast, YR showed a severe and disproportionate impairment on recall tests (mean Z = −3.6), which
was clinically significant in 95% of tests (Mayes et al., 2002).
Further investigations showed that YR was unimpaired on a
forced-choice object recognition memory test, but was clearly
impaired at an equivalently difficult yes/no object recognition
test (Holdstock et al., 2002), and she was also impaired at recognition of associations between different kinds of information,
even when tested by forced-choice tasks (Mayes et al., 2004).
Bastin et al. (2004) and Aggleton et al. (2005) have reported a
1233
similar cases, and Holdstock, Mayes, Gong, Roberts, and Kapur
(2005) a patient in whom non-verbal (but not verbal) recognition
memory was relatively spared, possibly related to asymmetrical
SPECT findings. Interestingly, Henke et al. (1999) described a
patient with hypoxic bilateral hippocampal damage whose initial recall and recognition memory impairment evolved through
time to a more selective recall deficit. Moreover, Yonelinas et
al. (2002) reported that 56 cardiac arrest patients with presumed
hypoxic brain damage involving the hippocampi showed disproportionate impairment on a word-list recall test, relative to
recognition memory performance, standardised according to Zscores. In addition, some functional imaging investigations have
produced evidence of differential medial temporal activations
during tasks involving recollection or familiarity processes, consistent with this hypothesis (Davachi, Mitchell, & Wagner, 2003;
Eldridge, Knowlton, Furmanski, Bookheimer, & Engel, 2000;
Ranganath et al., 2004).
There is, however, an older tradition, which argues that
disproportionate impairment in recall memory or recollective
processes is characteristic of amnesic patients in general, and
that memories based on familiarity alone are relatively preserved in amnesia (Giovanello & Verfaellie, 2001; Hirst et al.,
1986; Hirst, Johnson, Phelps, & Volpe, 1988; Huppert & Piercy,
1976, 1978; Warrington & Weiskrantz, 1982; Yonelinas, Kroll,
Dobbins, Lazzara, & Knight, 1998). For example, Huppert and
Piercy (1976, 1978) found that amnesic patients made memory
judgements purely on the basis of ‘trace strength’ or familiarity,
even when they had been asked to make more specific evaluations about item recency or frequency. Hirst et al. (1986, 1988)
showed that, after matching amnesic patients’ performance to
that of healthy subjects in two different ways on a recognition
memory test, the amnesic group’s recall scores were disproportionately impaired, relative to the controls. Giovanello and
Verfaellie (2001) employed a very similar design to that of Hirst
et al. (1986, 1988), finding that they replicated Hirst et al.’s result
in one task, but not the other. These authors argued that amnesic
patients and healthy participants performed the two tasks in
different ways, and that this was consistent with a differential
impairment of recollective memory in the amnesic patients.
The third view – namely, that (verbal and visual) recall and
recognition memory are proportionately impaired in amnesia –
has been advocated by Squire and colleagues in a series of publications (Haist, Shimamura, & Squire, 1992; Manns, Hopkins,
Reed, Kitchener, & Squire, 2003; Manns & Squire, 1999; Reed
& Squire, 1997; Stark, Bayley, & Squire, 2002; Stark & Squire,
2003). These authors have argued that patients with damage
thought to be limited to the hippocampal region consistently
show impairments on tests, such as the Recognition Memory
Test, especially if a delay is introduced (Reed & Squire, 1997),
the Doors and People Test (Manns & Squire, 1999), the recognition component of the Rey Auditory Verbal Learning Test
(Manns et al., 2003), as well as on a wide variety of other recognition memory tests, whether tested by forced-choice or yes/no
recognition procedures (Reed & Squire, 1997; Stark & Squire,
2003). Consistent with these findings, Kopelman and Stanhope
(1998) used a variant of Hirst et al. (1988) technique, matching performance on recognition memory testing and avoiding
1234
M.D. Kopelman et al. / Neuropsychologia 45 (2007) 1232–1246
‘ceiling’ and ‘floor’ effects: these authors found, unlike Hirst
et al., that there was no evidence of disproportionate recall
memory impairment in memory-disordered patients with diencephalic, temporal lobe or frontal lesions. Moreover, Coleshill
et al. (2004) have shown that unilateral electrical stimulation to
the left or right hippocampus produced material-specific disruption of yes/no recognition memory performance, and Kramer
et al. (2005) have recently employed MRI volume measurements to show that hippocampal volumes are the best predictor
of both delayed recall and recognition memory discriminability.
Kopelman et al. (2001) also found that MRI measures of hippocampal volume were the most consistent volumetric correlates
of both recall and recognition memory.
Hence, considerable uncertainty remains concerning which
patients will show a disproportionate recall memory impairment,
and under what circumstances. The most persuasive descriptions
of disproportionate recall memory impairment have occurred
either in developmental cases (Vargha-Khadem et al., 1997) or
in single case-reports of adult acquired lesions (e.g. Henke et al.,
1999; Mayes et al., 2002). Other investigations have included
larger numbers, but have lacked either detailed neuro-imaging
(Yonelinas et al., 2002) or appropriate ‘matching’ procedures on
which recall/recognition comparisons rely. Moreover, there are
suggestions that the delay between stimulus presentation and
recall/recognition testing (Kopelman & Stanhope, 1997, 1998)
or the interpolation of an inter-item distractor task (Giovanello
& Verfaellie, 2001) might be critical in determining the pattern
of findings obtained.
In the present investigation, we have examined this issue in
19 memory-disordered patients, whose amnesia resulted from
either lesions confined to the medial temporal lobes, or from
lesions extending more laterally from the medial temporal lobe
to the lateral temporal cortex, or from focal frontal pathology.
Quantitative structural MRI brain measurements from these participants were available, and, in particular, it was possible to
make a comparison between a subgroup from the medial temporal group of three patients (H−) in whom the hippocampi
alone were atrophied, relative to healthy control values, and
a subgroup of two patients (H+) in whom the parahippocampal structures (including perirhinal and entorhinal cortex) were
also atrophied (see Bright et al., 2006, for findings in these two
subgroups at retrograde amnesia tasks). We again employed
a modification of Hirst et al. (1988) technique, which allows
for appropriate ‘matching’ of recognition memory performance
between amnesic patients and healthy participants, but, unlike
our earlier investigation (Kopelman & Stanhope, 1998), comparison of the findings was made across three different delay
conditions (30 s, 2 and 10 min)—a period over which our earlier studies (Green & Kopelman, 2002; Kopelman & Stanhope,
1997) as well as those of others (Isaac & Mayes, 1999a, 1999b)
have suggested that critical differences in forgetting rates on
recall memory might occur. Secondly, we have examined the
scatter of our participants’ recall scores plotted against their
recognition scores in a manner analogous to Yonelinas et al.
(2002). Thirdly, we sought differential patterns of correlation
between hippocampal and parahippocampal volumes with our
measures of recall and recognition memory performance across
the total patient group, and we compared the pattern of performance of our H+ and H− subgroups on our recall/recognition
memory measures. Finally, we examined whether the interpolation of an inter-item distractor task influences the pattern of
results by examining this issue in healthy participants.
2. Method
2.1. Participants—clinical and MRI description
2.1.1. Medial temporal lesion group
Five patients were selected on the basis of significant anterograde memory
loss and MRI evidence that regional brain atrophy was restricted to the medial
temporal lobe structures: the former was defined as clinical evidence of significant memory impairment and a NART-R minus WMS-R Delayed Recall
index discrepancy of at least 15 points (range 17–64 points). Table 1 shows
the mean for WMS-R general and delayed recall indexes. The atrophy was
attributable to acute hypoxic episodes in three of these patients. A fourth patient
had experienced an acute encephalopathy of uncertain origin at 13, associated
with presumed hypoxia and subsequent left-sided mesial temporal sclerosis and
partial seizures. The fifth patient had suffered complex partial seizures over a
period of many years. In all cases, the atrophy was bilateral. Fig. 1 shows coronal
sections from the brains of these patients revealing medial temporal lobe atrophy, confined to the hippocampi (top row) and also involving parahippocampal
structures (bottom row, left).
Table 1
Background cognitive test scores
Controls
Frontal
Medial temporal
Temporal
ANOVA
Mean
S.D.
Mean
S.D.
Mean
S.D.
Mean
S.D.
N
Mean age
NART-R IQ
9
49.4
117.6
18.1
13.0
7
53.3
107.4
11.2
17.3
5
41.8
114.2
7.1
6.3
7
41.6
104.1
12.5
13.8
N.S.
N.S.
Memory
WMS-R GM index
WMS-R DM index
119.0
120.8
11.9
14.5
94.1
88.0
21.7
21.0
80.4
67.4
7.2
16.8
65.9
70.0
14.3
15.7
p < 0.0001
p < 0.0001
Frontal/executive
FAS verbal fluency
Card-sort categories
Card-sort persev’ns
53.4
6.0
0.8
13.3
0.8
1.4
29.0
3.4
6.3
16.3
2.3
5.4
40.6
5.8
1.4
13.8
0.8
1.5
36.1
4.3
2.3
8.9
2.7
2.4
p < 0.02
p < 0.01
p < 0.05
M.D. Kopelman et al. / Neuropsychologia 45 (2007) 1232–1246
1235
Fig. 1. Top row: Coronal sections showing hippocampal atrophy only (H−) in DL, JB and DH (cerebral hypoxia). Bottom row, left: Parahippocampal and hippocampal
atrophy (H+) in JM (cerebral hypoxia). Centre: Axial sections showing bilateral medial temporal temporal lobe pathology and extensive right antero-lateral pathogy
in SM (herpes encephalitis) and (right) bilateral frontal pathology in JW (contusion and haemorrhage).
2.1.2. Temporal lobe lesion group
These patients were chosen on the basis of their all having significant anterograde memory impairments (based again on clinical evidence and minimum
NART-R minus WMS Delayed index difference of at least 15 points (range
15–63 points, see Table 1 for mean values)), in association with MRI evidence of extensive medial and antero-lateral temporal lobe damage. In this
paper, these patients will sometimes be referred to as the ‘lateral’ temporal
lobe group to distinguish them from those with pathology confined to the
medial temporal lobes, but it should be understood that this group’s pathology
also involved medial temporal lobe structures. Of the seven patients selected
for this group, five had been diagnosed with (antibody confirmed) herpes
encephalitis. In four of these patients, there was evidence of temporal lobe
damage in both hemispheres, although the extent of damage was predominantly left lateralised in three patients and predominantly right lateralised in
one patient; the remaining patient (DJ) showed unilateral left temporal lobe
damage, as previously described by Stanhope and Kopelman (2000). In all
except DJ, the signal alteration on MRI implicated the medial temporal lobes
bilaterally (in DJ unilaterally), involving the hippocampi and parahippocampal
structures including the entorhinal, perirhinal and parahippocampal cortices.
In the more affected hemisphere, the signal alteration involved the anterolateral temporal lobe cortex (see Fig. 1, bottom row, middle). Two further
patients were included in this group. One patient had suffered an encephalitic
illness at 20, resulting in residual temporal lobe epilepsy. The other patient
had had a temporal lobe abscess at 17 resulting in (predominantly) verbal
memory impairment. An MRI carried out when she was 34 showed a large
left temporal CSF-filled lesion, involving medial and lateral temporal lobe
structures.
2.1.3. Frontal lesion patients
Seven patients with focal frontal lesions and deficits on measures of executive function were recruited (e.g. Fig. 1, bottom row, right). All showed some
‘frontal’ behavioural symptoms, such as apathy, irritability, emotional lability or
disinhibition. In two patients, the pathology resulted from acute head injury and
associated contusions and haematomas, worse on the right than the left. Another
two patients had undergone surgery for removal of tumours: one a left frontal
meningioma arising from the planum sphenoidale, which had been only partially
resected, the other a transfrontal craniotomy for removal of a pituitary tumour
resulting in right anteromedial frontal damage. There were a further two cases
with frontal infarcts. In one of these patients, the damage was restricted to the
left hemisphere, but the other patient showed bilateral frontal signal alteration:
both showed pronounced ‘frontal’ behavioural changes. Finally, one patient had
suffered a large right frontal cerebral abscess, following a tooth infection, and
he showed extensive residual signal alteration in the right prefrontal cortex on
MRI.
None of the patients in any of the groups had any known psychiatric problems, evidence of substance abuse or other conditions, which might have affected
ability to understand instructions or to complete the tasks.
2.1.4. Controls
Two sets of healthy control participants were recruited for this study. The
first set (Controls A) (N = 9) were recruited from a local further education college as well as non-clinical staff in the hospital, matched as closely as possible
to the patients for age, sex, NART-R and years of education. The second set
(Controls B) (N = 12) were recruited from non-clinical hospital staff, and were
again matched as closely as possible in terms of the same variables.
2.2. Quantitative structural MRI
MRI scans were axially acquired on a 1.5T Philips scanner, using a protocol
of T1 and T2 weighted gradient and PD echo 3D volume datasets. Slice thickness
was 1.5 mm and the matrix size 256 × 256, giving a voxel size of 1.3 mm3 . A
HP735 graphics workstation was used to segment (delineate) brain structures of
1236
M.D. Kopelman et al. / Neuropsychologia 45 (2007) 1232–1246
interest across sequential MR slices. The data were analysed using a hierarchical
segmentation program, allowing detailed volumetric assessment. The program
incorporates visualisation, manipulation and storage/retrieval functions in its
interface, and segmentation tools include a multi-slice 2D hierarchical segmentation program, a 2D polyline tool for drawing a sequence of connected straight
lines and a 3D plane cutting tool. Quantitative structural MRI measurements of
the left and right temporal lobes, antero-lateral temporal lobes, medial temporal
lobes and hippocampi were taken from planimetric measurements determined
according anatomical definitions and segmentation criteria described elsewhere
(Colchester et al., 2001; Kopelman et al., 2003).
Fig. 2 shows total, lateral and medial temporal lobe mean volumes in the
patient groups, relative to 10 volunteers in a reference control sample (Colchester
et al., 2001; Kopelman et al., 2003) who did not differ significantly from either of
our control groups in terms of mean age, sex ratio or NART-R IQ. It shows that
the temporal lobe lesion group showed significant atrophy across total temporal
lobe, total lateral temporal and total medial temporal volumes. The frontal lesion
and medial temporal lesion groups did not differ significantly from controls in
terms of total temporal lobe or lateral temporal volumes. The medial temporal
lesion group showed a mean medial temporal lobe volume approximately half
way between the controls and the temporal lobe lesion group: they differed
significantly from controls in terms of mean medial temporal volume on a t-test
(t = 2.86, p < 0.025), but not on a Bonferroni post hoc test following one-way
ANOVA across all four groups. Fig. 3 shows that the medial temporal lesion
group and the temporal lesion group both showed highly significant atrophy in
terms of left and right hippocampal volumes. These quantitative MRI data show
that, despite the variability in underlying aetiology, the allocation of patients to
these groups is valid in terms of regional brain volumes.
2.3. Background neuropsychological findings
Background cognitive test scores were collected, and are summarised in
Table 1. Statistical comparisons are given between the patient groups and the
total control group (N = 21). On a measure of estimated premorbid IQ (NARTR) (Nelson & Willison, 1991), there were no differences among the groups
(F(3,36) = 1.22, N.S.). However, there were significant differences across the
groups for the general memory index (F(3,36) = 14.74, p < 0.0001), delayed
memory index (F(3,36) = 17.87, p < 0.0001), as well as the individual visual
and verbal memory indexes (p < 0.0001). In terms of Bonferroni post hoc tests,
all patient groups performed significantly more poorly than controls on the
delayed memory index (p < 0.02), and both temporal lobe lesion groups (but
not the frontal lobe group) performed significantly worse than controls on general memory (p < 0.01). For general memory, the ‘lateral’ temporal lobe lesion
patients performed more poorly than the frontal lesion patients (p < 0.02) but the
two temporal lesion groups did not differ significantly from each other. None
of the patient groups differed significantly from one another on the delayed
memory index.
Table 1 also shows significant differences in card-sorting categories
(F(3,36) = 4.49, p < 0.01) and perseverations (F(3,36) = 3.27, p < 0.05), and on
verbal fluency (F(3,36) = 4.17, p < 0.02) with the control group performing best
and the frontal lesion group performing worst in each case. On Bonferroni post
hoc tests the frontal lesion group performed significantly worse than controls
in each case (p < 0.05). Neither the ‘lateral’ temporal nor the medial temporal
lesion group differed significantly from the controls.
3. Experiment 1: Analysis 1
This experiment examined performance on recall memory,
relative to recognition, at delays of 30 s, 2 and 10 min to examine
whether a differential pattern of performance across the patient
groups (frontal, medial temporal and ‘lateral’ temporal) might
emerge at the longer delays.
Fig. 2. Volumetric measures of temporal lobe structures for controls and each
patient group. Figures in square brackets show mean percentage deviation from
control group volumes.
M.D. Kopelman et al. / Neuropsychologia 45 (2007) 1232–1246
1237
were typed side by side on 15 cm × 10 cm cards. Lists A and B
were employed equally as often as targets and distractors both
within and across the participants.
3.2. Procedure
Fig. 3. Left and right hippocampal volumes for controls and the two temporal
lobe groups (temporal and medial temporal). Figures in square brackets show
mean percentage deviation from control group volumes.
3.1. Materials
The materials used in this study were adapted from Kopelman
and Stanhope (1998). Three sets of 32 words of 1 or 2 syllables were selected, within each of which half of the words were
related, and belonged to one of two categories, and the other
half were unrelated (stimulus details are provided in Appendix
A). The words were selected from Battig and Montague (1969)
norms. The mean word frequency of both the related and unrelated words was 128 occurrences per million (Francis & Kucera,
1982) and the three sets of words were matched for word
frequency and syllable length. A different set of words was
presented to subjects at each time interval, with the order of
exposure to each set counterbalanced within and across subject
groups.
Each set of 32 words was divided into 2 (matched) lists of
16 words (A and B), 1 of which would be used for the recall
task. Within each list there were eight unrelated words and four
related words from two categories, and the words in each were
matched for frequency and syllable length. Words were typed
in large font on to 15 cm × 10 cm cards. The 16 words in each
list to be presented were divided into 4 blocks, 2 related and
2 unrelated, and were arranged so as the two types of block
alternated. Each block was preceded by a card on which was
typed either “unrelated” or the name of the category from which
the related words were taken, e.g. “colours”. Within each block,
word order was randomised, and the order of block presentation
was counterbalanced across participants.
For the forced-choice recognition test, the words from the
matched list, which had not been used in the recall task were
employed as distractors. Each “target” word, i.e. a word that
had been presented in the recall task, was paired with a word
from the same category from the distractor list. The two words
The procedure follows that employed originally by Hirst et
al. (1988) and subsequently by Kopelman and Stanhope (1998).
Participants were presented with one of the word-lists consisting of 16 items. The experimenter told the participants that they
would be presented with two blocks of related words, and two
blocks of unrelated words, and that each block would be introduced with a card on which was written either “unrelated” or
the category from which the words were drawn. The participants
were told to read each word aloud and remember it.
In order to match recognition memory scores, exposure times
were titrated across groups. The exposure time to each stimulus
was 7 s in the frontal lobe group, 10 s in the medial temporal
lobe group and 12 s in the temporal lobe group. Our initial
control group (Controls A) was given an exposure time of
3 s per stimulus. These exposure times were determined on
the basis of extensive pre-study piloting, which, in turn, was
informed by the findings of an earlier study (Kopelman &
Stanhope, 1998). Subsequently, we tested a second control group
(Controls B), who were given an exposure time of 0.5 s per
slide.
It has been pointed out, however, that in such designs the
total duration of the presentation phase of the task, and the mean
item-to-test delay, is longer for the memory-impaired patients
than the healthy controls (Mayes, 1986). This may lead to an
underestimation of the patients’ performance. Consequently, as
in the Kopelman and Stanhope (1998) investigation, an interstimulus task was employed to match the inter-item delay across
the groups. The inter-stimulus task consisted of counting backwards in multiples of three until the next stimulus was presented,
starting from a number between 100 and 1000 indicated on a
flash card (see also Experiment 2, below). This procedure gave
an inter-item delay of 12 s in each group from initial presentation
of each item until presentation of the next.
Following presentation of the last word, a 30 s, 2 or 10 min
delay (filled with normal conversation) preceded free recall
memory testing, in which participants were asked to recall as
many words as they could in any order. When they could recall
no further words, a forced-choice recognition memory task commenced, in which the experimenter held up the forced-choice
cards described earlier, and asked which of the two words they
had seen before.
This procedure was conducted at each of the three delay
conditions (30 s, 2 and 10 min) using a different set of words
for each condition. The word-sets for use at each delay were
counterbalanced across subjects.
3.3. Results
In order to assess any disproportionate impairment in recall
relative to recognition in the patient groups, it was first necessary
to check whether the groups were matched in terms of recog-
1238
M.D. Kopelman et al. / Neuropsychologia 45 (2007) 1232–1246
nition scores at each time interval. First, we compared the two
control groups (A and B) finding that, following their more prolonged exposure time, Controls A performed significantly better
than Controls B at 30 s (t = 4.98, p < 0.0001), 2 min (t = 3.27,
p < 0.005) and 10 min (t = 3.88, p = 0.001). However, when the
two control groups were compared with the three patient groups,
there were no significant differences on one-way ANOVAs in
terms of recognition memory score at either 30 s (Controls A:
F = 0.76, Controls B: F = 1.39), 2 min (A: F = 0.53, B: F = 1.13)
or 10 min (A: F = 2.43, B: F = 2.17). However, there were a number of subjects scoring at ceiling (16/16) on the recognition task
at each time-delay, although none of the controls at a 0.5 s exposure (Controls B) did so. As in Kopelman and Stanhope (1998),
we then excluded participants with perfect scores on recognition
memory, finding that there were still no significant differences
between the groups at 30 s (A: F = 1.53, B: F = 0.24), 2 min (A:
F = 1.16, B: F = 0.15) or 10 min (A: F = 2.68, B: F = 0.99). This
confirmed that, following our manipulation of the patient versus
control exposure times, our two control groups did not differ significantly from the three patient groups in terms of recognition
memory. Having ‘matched’ the groups on recognition memory,
we then examined their performance at recall memory for the
words.
Fig. 4 shows the recognition and recall performance of the
three patient groups and Control Group B at each of the three
time-periods after excluding subjects at ceiling. Controls A are
also shown as a dotted line. Comparing the patient groups
with Controls B, there was a highly significant main effect
of (recognition/recall) condition (F(1,18) = 462.50; p < 0.001),
but the main effect of group failed to reach statistical significance (F(3,18) = 1.42, N.S.), and of particular importance the
group by condition interaction (F(3,18) = 1.25) was not statistically significant. However, there was a significant main
effect of delay (F(2,36) = 3.63, p < 0.05) and significant group
by delay (F(2,36) = 6.02, p < 0.01) and group by condition by
delay (F(6,36) = 3.16, p < 0.05) interactions. This reflected the
relatively superior performance of the frontal group at recall at
the 2 min delay (confirmed by individual ANOVAs at each delay,
showing a significant effect at this delay only).
When all patients were analysed (including those at ceiling),
and compared with Control Group B, neither the main effect
of group (F(3,26) = 2.70), nor the group by condition interaction (F(3,26) = 2.87), nor the group by condition by delay
interaction (F(6,52) = 2.10) were statistically significant. There
was a significant main effect of (recognition/recall) condition
(F(1,26) = 379.2; p < 0.0001).
Statistical comparisons between the three patient groups and
(the less well-matched) Control Group A also failed to find
any evidence of a recall–recognition discrepancy across the
groups. When subjects at ceiling were excluded, neither the
group by condition (F(3,8) = 1.24) nor the group by condition by
delay (F(6,16) = 1.23) interactions was statistically significant.
When all subjects were analysed (including those at ceiling),
the group by condition interaction was statistically significant
(F(3,23 = 3.42, p < 0.05), but not the group by condition by delay
interaction (F(6,46 = 2.67, N.S.). However, when all subjects
were analysed in the 10 min condition taken in isolation (in
Fig. 4. Recall and recognition performance for each patient group at each test
delay compared with Control B (0.5 s exposure, solid line and + symbol) and
Control A (3 s exposure, dotted line), after eliminating subjects performing at
ceiling on recognition testing.
which fewer subjects were at ceiling than in the 30 s and 2 min
conditions), the group by condition interaction was not statistically significant (F(3,24) = 1.45, N.S.). In other words, when
the subjects were away from ceiling in the recognition condition (either by examining scores at the longest delay only or after
deliberately excluding those with ‘perfect’ recognition memory
scores), there was no evidence of a differential effect of group
on recall relative to recognition memory performance.
In a further analysis, we examined whether these findings
might relate: (i) to the fact that overall medial temporal volume
was lower in the lateral temporal group (7578 mm3 ) than the
medial temporal group (10,872 mm3 ) or (ii) to the fact that the
lateral temporal group had greater atrophy on the left (Fig. 3).
We chose three patients from each group with approximately
equal medial temporal volumes (mean volume for lateral
temporal subset: 8520 mm3 ; for medial temporal subset:
M.D. Kopelman et al. / Neuropsychologia 45 (2007) 1232–1246
9095 mm3 ). The recognition/recall discrepancies showed only
a trend difference at 30 s (F(1,4) = 3.78, p > 0.10), and were
closely similar, not approaching significance, at subsequent
delays (2 min: (F(1,4) = 0.07, N.S.; 10 min: (F(1,4) = 0.02,
N.S.). We then compared three lateral temporal patients with
more evenly matched left and right medial temporal mean
volumes (left: 3719 mm3 ; right: 4094 mm3 ) with three medial
temporal patients (left: 4470 mm3 ; right: 4625 mm3 ). Again,
the recognition–recall difference did not approach significance
(30 s: F(1,4) = 0.10, N.S.; 2 min: F(1,4) = 1.66, N.S.; 10 min:
F(1,4) = 0.84, N.S.). Given the small sample sizes, we also
computed the same six comparisons with the non-parametric
Mann–Whitney test for independent samples. The results were
entirely consistent with the parametric analyses (p > 0.10).
In comparing performance across related and unrelated
words, there was a non-significant trend for the frontal group
to show a greater related-unrelated difference in the recall condition (i.e. better performance in recalling related words) than
the two temporal lobe groups (compare Kopelman & Stanhope,
1998). However, neither the group by condition interaction (all
participants: F = 0.32; excluding ceiling: F = 0.71) nor the group
by condition by delay interaction (all participants: F = 0.49;
excluding ceiling: F = 0.72) were statistically significant.
3.4. Summary
These results indicate that none of the patient groups (frontal,
medial temporal or ‘lateral’ temporal) showed a disproportionate
impairment on recall relative to recognition memory compared
with healthy controls, after matching for recognition memory
performance as closely as possible by manipulating exposure
times and after excluding subjects at ceiling in some of the analyses. Nor did this finding appear to result from differences in mean
medial temporal lobe volumes between the groups, laterality
effects or differences between related and unrelated words.
4. Experiment 1: Analysis 2—scatterplots of recall and
recognition scores
In order to investigate the relationship between recall and
recognition memory more thoroughly, we examined the scatter
of individuals’ recall scores plotted against recognition scores
at each time-point in a manner similar to that of Yonelinas et al.
(2002).
4.1. Method
To standardise the scales along each axis, each subject’s
recall and recognition scores were converted to Z-scores at each
time-point. For these analyses, the Z-scores were based on the
means and standard deviations (expressed as percentages) in the
combined control group (A plus B) in order that the standard
deviation in the controls at each delay should approximately
match that of the patients’ (Recognition, C versus P, 30 s: 10.9%
versus 13.4%; 2 min: 10.3% versus 10.7%; 10 min: 12.6% versus
12.0%; Recall, 30 s: 28.6% versus 21.0%; 2 min: 29.6% versus
27.3%; 10 min: 28.9% versus 29.8%). After inspecting the scat-
1239
terplots at each delay separately, we decided it was legitimate to
merge the plots into a scatter of 56 points (3 points for each of
the 19 patients, one at each delay, 1 missing datum) to compare
with Yonelinas et al.’s (2002) scatter of 56 points.
4.2. Results
As a first step, we compared control-referenced mean
Z-scores for the whole patient group on recall and recognition memory (56 ‘paired’ observations), analogous to Fig. 1a
in Yonelinas et al.’s (2002). Unlike their investigation of
hypoxic patients, we obtained highly similar mean Z-recall and
Z-recognition scores (recognition = 0.18, recall = 0.05; pairedsamples t = 1.15, N.S.).
Where the medial temporal group were examined in isolation (15 paired observations), analogous to Yonelinas et
al.’s (2002) study of hypoxic patients, the mean Z-scores
were still closely similar (mean Z-recognition = 0.04, mean Zrecall = −0.04; paired −t = 0.46, N.S.).
Kopelman and Stanhope (1998) included a group of five
(different) hypoxic patients tested on this same task, but at
the 30 s delay only. These participants were closely matched
to the present medial temporal group in terms of mean age,
NART-R and memory indexes. In order to enlarge the size
of the present hypoxic group to N = 10, we converted these
patients’ recall and recognition scores to Z-scores, using the
present controls’ values at this delay as the reference means and
standard deviations. When this was done, the mean Z-values in
this group of 10 hypoxic patients were again not significantly
different (mean Z-recognition = 0.14, mean Z-recall = −0.25;
paired −t = 0.64, N.S.).
Fig. 5a shows the scatter of Z-recall scores against Zrecognition scores for all 56 paired observations in the patients.
The line plots where the points would fall if Z-recall was equal
to Z-recognition for each obtained value. The data-points were
distributed approximately equally either side of our ‘idealised’
line, consistent with neither recall nor recognition being disproportionately impaired in this patient group.
Fig. 5b shows the scatter for the entire group (controls + patients combined), and it also indicates lines through the
zero intercepts. The distribution of these data-points across the
quadrants has been influenced by the fact that the exposure times
have been manipulated to match patients’ performance to controls’; where Z-scores for patients were referenced to controls A
in isolation, more patients fell into the bottom left-hand quadrant
(indicating impairments in both recall and recognition memory) and many fewer in the top right-hand column (the quadrant
indicating relatively spared recall and recognition memory at
these exposure times). The values in the scattergram were put
into a multiple regression examining the predictive value of Zrecognition and group membership on Z-recall. As we were
interested in differences in performance between the healthy
control and patient samples, group membership was entered as
0 for control, 1 for patient. The product of this binary value and
the recognition score was then computed to produce a dummy
variable. This dummy variable was entered as a further regressor
enabling a test of parallelism to be carried out, as outlined by
1240
M.D. Kopelman et al. / Neuropsychologia 45 (2007) 1232–1246
combined control groups (A and B) for recall and recognition,
there was no overall evidence of significantly disproportionate recall (or recognition) memory impairment in either: (i) the
whole patient group, (ii) the medial temporal group, taken in
isolation or (iii) the medial temporal group ‘pooled’ with that
from Kopelman and Stanhope (1998). Only one data-point from
a medial temporal lobe patient fell within the quadrant on a scattergram indicating disproportionate recall memory impairment.
5. Experiment 1: Analysis 3—recall and recognition
memory scores and medial temporal MRI measures
In order to examine the relative contribution of hippocampal and parahippocampal (including entorhinal and perirhinal)
structures to recall and recognition memory, we employed our
quantitative structural MRI measurements of the hippocampi
and the parahippocampal gyri to examine for correlations with
recall and recognition memory performance within the total
patient group. Secondly, we examined the findings from the five
patients in the medial temporal group in more detail in order to
compare performance in subjects whose atrophy was confined to
the hippocampi (H− subgroup) with those whose pathology also
involved the parahippocampal gyri (H+ subgroup). This analysis was initially conducted using the exposure times and data
described above, but, in addition, the H− subgroup were subsequently re-tested at a shorter (5 s) exposure time and a further
analysis conducted.
5.1. Method
Fig. 5. Control-referenced Z-recall and recognition scores: (a) pattern of performance for each patient group shown with a line of idealised fit; (b) combined
scatterplot of both controls and patients, showing dotted lines through zero intercepts. These scatters show one point for each subject at every test delay, resulting
in three points per subject.
Kleinbaum, Kupper, Muller, and Nizam (1988). This test for parallelism revealed a highly significant effect of the regression of
Z-recall against Z-recognition (t = 8.15, p < 0.0001), but no significant effect of group (patient/control) membership (t = −0.37,
N.S.) or group by condition interaction (t = −1.34, N.S.). In other
words, this analysis confirms that, as recognition memory scores
diminished, recall memory scores showed a proportionate fall,
regardless of group.
The important quadrant in Fig. 5b is the bottom right-hand
one, which was the quadrant indicating disproportionate recall
memory impairment with preserved recognition memory at
these exposure times. There were 19 data-points which fell
within this quadrant: 10 were from control participants, 6 from
the temporal lobe lesion group, 2 from frontal patients and only
1 from the medial temporal group.
4.3. Summary
In summary, when the patients’ recall and recognition scores
were converted to Z-scores, based on the means and S.D.s of the
Separate segmentations were carried out on coronal sections to measure hippocampal volumes and medial temporal
(combined hippocampal and parahippocampal) volumes. Our
boundary definitions for the hippocampi were closely similar to
those described by Mori et al. (1997), except that we included the
subiculum as part of the hippocampus. Anteriorly, the alveolar
covering of the hippocampus provided a border with the amygdala. The posterior limit of the hippocampus was the coronal
slice in which the fornix clearly emerged from the fimbria of the
hippocampus, just anterior to the splenium of the corpus callosum. These margins were checked in sagittal and axial sections.
The medial temporal measurements employed the same anterior
and posterior margins but, in the coronal plane, segmentations
were taken from the subiculum across the cortical surface of the
parahippocampal gyrus, and then deep into the collateral (rhinal) sulcus until it met the inferolateral point of the hippocampus
(see Colchester et al., 2001; Kopelman et al., 2003).
Inspection of the quantified MRI measurements revealed that
the medial temporal group could be subdivided into a subgroup
of three patients (H−) in whom the hippocampi alone appeared
to be atrophied, relative to healthy control values, and a subgroup
of two patients (H+) in whom the parahippocampal structures
were also atrophied. Table 2 shows the mean hippocampal and
parahippocampal volumes in these two subgroups, relative to
healthy controls. F-values from one-way analyses of variance
and their significance values are shown, together with the results
of (C versus H− and H− versus H+) Bonferroni post hoc tests.
M.D. Kopelman et al. / Neuropsychologia 45 (2007) 1232–1246
1241
Table 2
Mean hippocampal and parahippocampal volumes in H+ and H− subgroups and controls (figures in square brackets indicate % difference from controls’ mean)
(mm3 )
Parahippocampal volume
Hippocampal volume (mm3 )
Controls
H− group
H+ group
F
p
C vs. H−
H− vs. H+
6553 (± 960)
7772 (± 1008)
8368 (± 250) [+28%]
4864 (± 293) [−37%]
3741 (± 1983) [−43%]
3590 (± 1482) [−54%]
12.48
21.30
<0.001
<0.001
N.S.
<0.002
<0.001
N.S.
Note: H−: patients with focal hippocampal atrophy; H+: patients in whom the parahippocampal structures were also atrophied.
The table shows that the H− subgroup did not differ significantly from the controls in terms of total (left and right)
parahippocampal volume, but that the H+ subgroup showed
significantly smaller parahippocampal volumes than the H−
subgroup (p < 0.001); the H+ group also differed significantly from controls in terms of total parahippocampal volume
(p < 0.02). By contrast, both H+ and H− subgroups showed
significantly smaller total hippocampal volumes than controls
(p < 0.001, p < 0.002, respectively), but they did not differ significantly from each other with respect to total hippocampal
volumes.
5.2. Results
Table 3 shows correlations between recall and recognition
scores and ‘total’ (left and right) hippocampal, parahippocampal and medial temporal volumes across the total patient group.
Included in the table are the findings for recall and recognition
at all delays, and also for the mean Z-scores for recall and recognition averaged across delays, using the data from Analyses 1
and 2 above.
Contrary to the prediction that recall memory impairment
might be specifically associated with hippocampal atrophy, and
recognition memory impairment with parahippocampal atrophy,
both recall and recognition memory scores were significantly
correlated with parahippocampal volumes but not with hippocampal volumes at the 30 s delay and when Z-scores were
averaged across all delays. Although there were no significant correlations at 2 and 10 min, the trends were in the same
direction, i.e. stronger relationships with parahippocampal than
hippocampal volumes.
Fig. 6 shows the findings from the H+ and H− subgroups
taken in isolation. The findings for H+ and H− at 10 s exposure per stimulus is shown, but, in addition, there is a curve
for the H− subgroup when re-tested over 4 years later at 5 s
exposure. Controls A (3 s exposure) and B (0.5 s) are shown
for comparison. If the H− group were specifically impaired
in recall, but not recognition, that group would show a steeper
Fig. 6. Recall and recognition memory performance at each time interval for
the H+ and H− subgroups of the medial temporal group at 10 s exposure per
stimulus. Performance is shown at each of the three delays in comparison with
Controls A (3 s/stimulus) and B (0.5 s). Also shown are the findings at each delay
on subsequent re-testing of the H− subgroup at 5 s. per stimulus.
Table 3
Correlations between recall/recognition scores and MRI volumes of medial temporal structures
Total hippocampal volumes
Total parahippocampal volumes
Total medial temporal volumes
***
**
*
p < 0.001 (one-tailed).
p = 0.01 (one-tailed).
p < 0.05 (one-tailed).
Recall
(30 s)
Recog.
(30 s)
Recall
(2 min)
Recog.
(2 min)
Recall
(10 min)
Recog.
(10 min)
Mean Z-recall:
all delays
Mean Z-recog.
all delays
0.23
0.75***
0.57**
0.40
0.56**
0.53*
0.10
0.43
0.31
0.07
0.23
0.18
−0.03
0.32
0.19
0.14
0.28
0.24
0.10
0.53*
0.37
0.26
0.44*
0.39
1242
M.D. Kopelman et al. / Neuropsychologia 45 (2007) 1232–1246
recognition/recall gradient than the H+ group. In fact, the figure shows that, at each time-delay, the H+ subgroup (with the
larger lesions) performed somewhat worse than the H− subgroup (when each subgroup received a 10 s exposure time per
stimulus), but that there was no difference in the slopes of the
two subgroups’ curves. The curves were approximately parallel
and, if anything, the H+ group showed a (minimally) steeper
gradient, contrary to prediction.
Statistical analysis was initially conducted comparing the H+
and H− subgroups, each at 10 s exposure per stimulus. There
was an overall significant effect of condition, i.e. recognition
versus recall (Wilcoxon test, Z = −2.0, p < 0.05), but no significant effects of (H+ versus H−) group on either recall (Z = 1.8,
N.S.), recognition (Z = −1.7, N.S.) or recognition–recall discrepancy (Z = −1.2, N.S.). There were also no significant effects
of delay on recall (Friedman test, χ2 = 5.4, N.S.) or on recognition (χ2 = 2.0, N.S.), although there was on recognition–recall
discrepancy (χ2 = 6.6, p < 0.05). Subtracting 10 min scores from
30 s scores, the groups did not differ in terms of either recall,
recognition or recognition–recall discrepancy scores as a function of delay (Wilcoxon, Z > −1.6, p > 0.1 in all cases).
Because such statistical comparisons lack power in small
subgroups, we also compared individual patients with controls.
Fig. 6 shows that the H− group’s scores at 10 s exposure were
closely matched with Controls A, especially at 30 s and 10 min,
and that the H+ group’s were very similar to Controls B at all
delays. Comparing the individual patients in the H− and H+ subgroups with ‘their’ respective control group, using the Crawford
and Garthwaite (2005) Revised Standardised Difference Test,
one H− patient differed significantly from Controls A on a onetailed test at 30 s only (t = 2.01, p < 0.05), but this patient did not
differ from Controls A at 2 or 10 min. The other H− patients and
also the H+ patients did not differ significantly from controls at
any delay.
Because of the possibility that a ceiling effect on recognition might have obscured the presence of a steeper gradient in
the H− subgroup, the same three participants were re-tested
on the same material using a shorter (5 s) exposure test more
than 4 years after their original testing. Fig. 6 shows that the
performance of the H− subgroup at this exposure time closely
matched that of Controls B (0.5 s/stimulus) and also that of the
H+ subgroup given a 10 s exposure. Using the Crawford and
Garthwaite (2005) Revised Standardised Difference Test, none
of the individual H− participants at this exposure time differed
significantly from Controls B at any delay. In short, H− patients
at 10 s/stimulus did not differ significantly from Controls A (3 s),
and H− at 5 s did not differ from Controls B (0.5 s).
but not hippocampal, measurements within the total patient
group. Although the H+ subgroup at 10 s exposure performed
worse than the H− subgroup at 10 s per stimulus, there were
no significant group differences between them in terms of
recognition–recall discrepancy (or in recognition–recall discrepancy as a function of delay). Moreover, when the individuals
within the H+ subgroup at 10 s and H− subgroup at 5 s exposure were compared with Controls B (0.5 s/stimulus), none of
these participants differed from the controls in terms of the
recognition–recall difference at any delay.
5.3. Summary
Fig. 7 shows that the two groups were matched in terms
of mean recognition memory scores, and that there was
only minimal difference in terms of mean recall scores.
Whilst the recall/recognition difference was highly significant (F(1,20) = 48.67, p < 0.001), neither the main effect of
group (F(1,20) = 0.16) nor the group by condition interaction
(F(1,20) = 0.31) approached significance. Excluding subjects at
ceiling, the group by condition (recall/recognition) interaction
remained non-significant (F = 0.06).
Use of the quantitative structural MRI hippocampal and
parahippocampal volumes allowed us to examine the correlation of these measures with recall and recognition scores,
and to differentiate a ‘hippocampal only’ (H−) subgroup
from a ‘hippocampal plus parahippocampal’ (H+) subgroup
among the medial temporal patients. Overall recall and recognition measures correlated significantly with parahippocampal,
6. Experiment 2
In the above investigation, we employed an inter-item distractor task to match mean delays-to-testing of individual
items across the groups. While the present data were being
collected, Giovanello and Verfaellie (2001) criticised this technique arguing that it might eliminate recollection/familiarity or
recall/recognition differences across groups, because it “likely
interfere(s) with the establishment of inter-item associations
known to benefit recollection . . . [leading to] suboptimal recollection [recall] in control participants.” On the other hand, it
can be argued that the inter-item distractor task prevents active
rehearsal and gives a better matching of exposure times across
groups (Mayes, 1986), and that there is no empirical evidence
that it affects the relationship between recall and recognition
memory.
In order to test whether or not the distractor task would indeed
affect the relationship between recall and recognition memory,
we compared a group of healthy subjects using the same exposure time as Controls A, but with an unfilled gap between stimuli,
i.e. no distractor task.
6.1. Method
We compared a group of 13 healthy participants (mean age,
48.2 ± 15.0; mean NART-R IQ = 108.5 ± 8.2) with our 9 participants in Controls A (mean age, 49.4 ± 18.1; mean NART-R
IQ = 117.5 ± 13.0) on the recall/recognition task.
The exposure time and presentation conditions of the stimuli were identical across the two groups, except that the ‘new’
group of healthy subjects had an unfilled gap between the presentation of individual items, whereas the ‘old’ group of controls
performed the distractor task between items as described above.
6.2. Results
M.D. Kopelman et al. / Neuropsychologia 45 (2007) 1232–1246
Fig. 7. Effect of filled and unfilled intervals upon recognition and recall scores
in healthy controls.
6.3. Summary
The findings of this experiment do not support the view that
the use of an inter-item distractor task interferes with recollection
in such a way as to affect the relationship between recall and
recognition memory. Therefore, our use of a distractor task in
Experiment 1 above (and in Kopelman & Stanhope, 1998) is
most unlikely to have influenced the relationship between recall
and recognition memory.
7. General discussion
In this paper, we have examined the relationship of recall
and recognition memory impairments in memory-disordered
patients in a number of different ways. We employed Hirst et al.
(1988) technique of titrating exposure times in order to ‘match’
recognition memory performance across participant groups as
closely as possible, and then making a comparison of recall
memory performance. We found that, when ceiling effects were
avoided, each of the patient groups showed a fall in recall memory performance (relative to recognition scores), which was
proportionate to that seen in healthy participants. Moreover,
there was no significant interaction with the delay until testing:
in other words, contrary to the speculation in our earlier paper
(Kopelman & Stanhope, 1998), a recall/recognition dissociation
did not emerge as the delay until testing increased (a difference
from findings in forgetting rate studies presumably relating to
the differing procedures and material employed). These findings
held good both when the patients were compared with a control
group given a 3.0 s exposure per word (Controls A), and also
when comparison was made with controls given a 0.5 s exposure
per word (Controls B).
When we converted the patients’ scores to Z-scores, based on
the mean and standard deviation of the combined control groups’
1243
scores, there was no significant difference between their mean
Z-recall and Z-recognition scores. The distribution of points on
a scatterplot of the patients’ Z-scores (Fig. 5a) was plotted in
a similar manner to that of Yonelinas et al. (2002), except that
we used an ‘idealised’ line to indicate where the points would
fall if Z-recall were equal to Z-recognition for each obtained
value, whereas Yonelinas et al. employed a regression line based
on their controls’ data. The data-points were distributed fairly
evenly either side of this line. When the data-points from the
combined control group (A and B) and the patients were plotted, and lines through the zero intercepts superimposed (Fig. 5b),
relatively few data-points fell in the bottom right-hand quadrant,
which was the quadrant indicating disproportionate recall memory impairment. More particularly, 10 of these latter data-points
were from control participants, 6 from ‘lateral’ temporal lobe
patients, 2 from frontal patients, and only 1 from the medial
temporal group.
Examination of MRI correlates indicated that it was total
parahippocampal, rather than hippocampal, volumes which correlated significantly with both recognition and recall memory
in this study. Moreover, although the subgroup with combined
hippocampal and parahippocampal (H+) atrophy consistently
performed worse than the subgroup whose atrophy was confined
to the hippocampi (H−), when both were given a 10 s/stimulus
exposure time, there was no evidence of disproportionate recall
memory impairment in the latter group. This was analysed in a
number of ways, including direct comparison of the H− and H+
subgroups, as well as comparison of the individuals within each
subgroup (using the Crawford & Garthwaite, 2005, test) with a
closely matched control group (A and B, respectively). Moreover, when the H− subgroup was re-tested at a shorter exposure
time (5 s/stimulus), their performance closely matched both
Controls B (0.5 s/stimulus) and the H+ subgroup (10 s/stimulus)
(Fig. 6). Taken together, these findings argue strongly against
the view that hippocampal amnesia is always characterised by
a disproportionate impairment in recall memory, at least on this
particular test.
Moreover, these findings were obtained using a forced-choice
recognition memory test, a design which at least some studies
have indicated is more likely to show preserved verbal and visual
recognition memory performance (and disproportionate recall
memory impairment) in hippocampal patients (Holdstock et al.,
2002). The purpose of the ‘titration’ experimental design was
to match recognition memory performance across the groups
as closely as possible, as has previously been employed in
several of the better designed studies (Giovanello & Verfaellie,
2001; Huppert & Piercy, 1976, 1978; Hirst et al., 1986, 1988;
Kopelman & Stanhope, 1998): in the absence of such an
experimental design, comparison of recall and recognition
memory performance, even after statistical manipulations such
as the use of Z-scores, is fraught with difficulty. The use of
an interpolated distractor task allows the delay until testing
to be matched across the participant groups whilst preventing
rehearsal between stimuli. An empirical test (Experiment 2)
indicated that the interpolation of an inter-stimulus distractor
task does not influence the relationship between recall and
recognition memory performance, as had previously been sug-
1244
M.D. Kopelman et al. / Neuropsychologia 45 (2007) 1232–1246
gested (Giovanello & Verfaellie, 2001). A potential criticism
is that we tested recall and recognition memory in succession
on the same word-lists. However, this design was based on that
employed by Hirst et al. (1988), and was also used by Janowsky,
Shimamura, Kritchevsky, and Squire (1989), Giovanello and
Verfaellie (2001) and Yonelinas et al. (2002), who obtained
contrasting results. As was also true of those investigations, only
one experimental task was employed, and it is certainly arguable
that the present findings need to be replicated using different
types and modalities of stimuli, and that a design using counterbalanced lists for recall and recognition could also usefully be
employed.
Although a series of recent publications have pointed to
differential patterns of activation in medial temporal lobe structures during recollection or familiarity processes (Davachi et al.,
2003; Eldridge et al., 2000; Ranganath et al., 2004), and case
reports of individual patients or small groups clearly indicate
that some patients with pathology confined to the hippocampi
show disproportionate verbal and visual recall memory impairment (Aggleton et al., 2005; Bastin et al., 2004; Henke et
al., 1999; Mayes et al., 2002), the only larger study purporting to show this (known to the present authors) is that by
Yonelinas et al. (2002). In that study, the volumes of brain
structures were not actually measured, which places severe limitations on the interpretation of the findings, and imposition
on their Fig. 1b of zero intercept lines, comparable with our
Fig. 5b, would indicate that the majority of their subjects in
fact showed combined recall and recognition memory impairments. By contrast, there are other investigations which have
shown proportionate effects upon verbal and visual recall and
recognition memory in groups of patients with pathology either
confined to the hippocampi or more extensively distributed in
medial temporal and thalamic structures (Haist et al., 1992;
Kopelman & Stanhope, 1998; Manns et al., 2003; Reed &
Squire, 1997). Kopelman et al. (2001) and Kramer et al. (2005)
found that it was MRI hippocampal volumes which showed
the most consistent correlations with both recall and recognition memory. This latter finding is at odds with the present
observation that parahippocampal volumes best correlated with
both recall and recognition memory performance, but both sets
of findings argue against a differential contribution of hippocampal/parahippocampal (perirhinal) pathology to recall and
recognition memory impairments, respectively. Coleshill et al.
(2004) have shown that unilateral electrical stimulation to the left
or right hippocampus produced material-specific disruption of
recognition memory. Moreover, Adlam (2003) has found that 11
out of 12 adolescents or young adults with developmental amnesia failed to show disproportionate recall memory impairment
on the Doors and People Test.
This brings us to the important question posed by Holdstock
et al. (2002)—namely, under what conditions is recognition
memory spared relative to recall after selective hippocampal
damage in humans? One possibility is that disproportionate
recall memory impairment simply reflects a milder memory
disorder, as clinical observation and comparison of amnesia
severity across some published studies suggest (Kopelman,
2002). However, this cannot account for all the published
observations (Kopelman, 2002). Similarly, the duration of time
elapsed (or the degree of functional or strategic cognitive reorganisation) since acute hypoxic brain damage may also be as
important, as Henke et al. (1999) indicated. (The present first
author has unpublished data on at least two similar cases). Again,
however, this is unlikely to account for all the discrepancies in
the literature. As already alluded to, Holdstock et al. (2002)
indicated that forced-choice object recognition memory was
unimpaired in their hippocampal amnesic patient, whereas her
performance on an equivalently difficult yes/no object recognition memory task was impaired when the targets and foils were
very similar. Their patient was also impaired on recognition of
object-location associations, whether tested by forced-choice or
yes/no memory tasks (Holdstock et al., 2002) and at recognition of other types of association between differing kinds of
information (Mayes et al., 2004). Nevertheless, these latter dissociations occurred within a single patient, and they do not
account for the discrepancies in the findings on simpler verbal and visual recognition memory tasks across other patients
with apparently similar pathology. Other possible explanations
for these discrepancies in the literature could include variation in the type, location or extent of hippocampal pathology
and/or variation in the type, location and extent of pathology
or dysfunction beyond the hippocampi and medial temporal
lobes (Mayes et al., 2004). For example, patients with known
hippocampal atrophy and/or sclerosis (in epilepsy) have been
shown to have concurrent thalamic hypometabolism (Kapur,
Thompson, Kartsounis, & Abbott, 1999) or combined thalamic
and retrosplenial hypometabolism (Reed et al., 1999). However,
there is no clear evidence, either in the literature or from the
present investigation, that aetiological differences underlie these
discrepancies.
At present, we do not have the relevant information available
to choose between these alternative explanations for the conflicts in the literature, which may not necessarily be mutually
exclusive or equally applicable to all comparisons. However, the
present findings serve as an antidote to the argument, based on
very few case reports or functional imaging studies alone, that
hippocampal amnesia or amnesia in general is typically characterised by disproportionate recall memory impairment, and
that contiguous and closely inter-connected structures subserve
very different memory functions. This may be another instance
(compare the literature on retrograde amnesia), where the findings from functional imaging studies in healthy participants and
those from neuropsychological investigations in patients with
focal lesions point in different directions. The present findings suggest that many patients with medial temporal or more
widespread temporal lobe damage show proportionate recall and
recognition memory impairments, and that, where discrepancies
between recall and recognition memory exist, it will require further fine-grained cognitive and neuro-imaging investigations to
disentangle their basis.
Acknowledgements
The research was funded by the Special Trustees of Guy’s &
St Thomas’s Hospital.
M.D. Kopelman et al. / Neuropsychologia 45 (2007) 1232–1246
Appendix A. Stimulus lists
Related words
Properties of buildings
Wall
Basement
Hall
Window
1245
References
Door
Stairs
Roof
Floor
Colours
Green
Pink
Brown
Yellow
Orange
Grey
Blue
White
Vehicles
Car
Taxi
Airplane
Ferry
Lorry
Train
Coach
Van
Clothing
Scarf
Jacket
Sock
Skirt
Collar
Tie
Trousers
Slipper
Birds
Falcon
Robin
Pigeon
Seagull
Thrush
Swallow
Magpie
Raven
Vegetables
Corn
Onion
Cabbage
Beetroot
Parsnip
Bean
Lettuce
Carrot
Unrelated words
Unrelated 1
Spring
Wine
Street
Town
Tail
Pretty
Game
Picture
Unrelated 2
Breakfast
Letter
Horse
Oil
Limit
Machine
Money
Club
Unrelated 3
Speaker
Careful
Hammer
Marker
Partner
Marsh
Rubber
Athlete
Unrelated 4
Clear
Plastic
Fuse
Planet
Pole
Suitcase
Column
Wine
Unrelated 5
Iron
Carry
Sunshine
Record
Cushion
Wave
Drawing
Scissors
Unrelated 6
Key
Cloud
Heavy
Paper
Shoulder
Marble
Table
Watch
Adlam, A. R. (2003). Associations in Memory: A Study of Developmental Amnesia. Ph.D. thesis. University of London.
Aggleton, J. P., & Brown, M. W. (1999). Episodic memory, amnesia, and
the hippocampal-anterior thalamic axis. Behavioral & Brain Sciences, 22,
425–489.
Aggleton, J. P., & Shaw, C. (1996). Amnesia and recognition memory: A reanalysis of psychometric data. Neuropsychologia, 34, 51–62.
Aggleton, J. P., Vann, S. D., Denby, C., Dix, S., Mayes, A. R., Roberts, N.,
et al. (2005). Sparing of the familiarity component of recognition memory
in a patient with hippocampal pathology. Neuropsychologia, 43(12), 1810–
1823.
Baddeley, A. D., Emslie, H., & Nimmo-Smith, I. (1994). The Doors and People
Test. Bury St. Edmunds, UK: Thames Valley Test Company.
Baddeley, A., Vargha-Khadem, F., & Mishkin, M. (2001). Preserved recognition
in a case of developmental amnesia: Implications for acquisition of semantic
memory? Journal of Cognitive Neuroscience, 13, 357–369.
Bastin, C., Van der Linden, M., Charnallet, A., Denby, C., Montaldi, N., Roberts,
N., et al. (2004). Dissociation between recall and recognition memory performance in an amnesic patient with hippocampal damage following carbon
monoxide poisoning. Neurocase, 10, 330–344.
Battig, W. F., & Montague, W. E. (1969). Category norms of verbal items in 56
categories. A replication and extension of the Connecticut category norms.
Journal of Experimental Psychology, 80, 1–46.
Bright, P., Buckman, J., Fradera, A., Yoshimasu, H., Colchester, A. C. F., &
Kopelman, M. D. (2006). Retrograde amnesia in patients with hippocampal,
medial temporal, temporal lobe or frontal pathology. Learning and Memory,
13, 545–557.
Colchester, A., Kingsley, D., Lasserson, D., Kendall, B., Bello, F., Rush, C., et al.
(2001). Structural MRI volumetric analysis in patients with organic amnesia,
1: Methods and comparative findings across diagnostic groups. Journal of
Neurology Neurosurgery & Psychiatry, 71, 13–22.
Coleshill, S. G., Binnie, C. D., Morris, R. G., Alarcon, G., Van Emde Boas,
W., Demetrios, V. N., et al. (2004). Material-specific recognition memory
deficits elicited by unilateral hippocampal electrical stimulation. Journal of
Neuroscience, 24, 1612–1616.
Crawford, J. R., & Garthwaite, P. H. (2005). Testing for suspected impairments and dissociations in neuropsychology. Evaluations of alternatives
using Monte Carlo simulations and revised tests for dissociations. Neuropsychology, 19, 318–331.
Davachi, L., Mitchell, J. P., & Wagner, A. D. (2003). Multiple routes to memory:
Distinct medial temporal lobe processes build item and source memories.
Proceedings of the National Academy of Sciences, United States of America,
100, 2157–2162.
Eldridge, L. L., Knowlton, B. J., Furmanski, C. S., Bookheimer, S. Y., & Engel,
S. A. (2000). Remembering episodes: A selective role for the hippocampus
during retrieval. Nature Neuroscience, 3, 1149–1152.
Francis, W. N., & Kucera, H. (1982). Frequency analysis of english usage:
Lexicon and grammar. Boston: Houghton Mifflin Co.
Giovanello, K. S., & Verfaellie, M. (2001). The relationship between recall and
recognition in amnesia: Effects of matching recognition between patients
with amnesia and controls. Neuropsychology, 15, 444–451.
Green, R. E. A., & Kopelman, M. D. (2002). Contribution of recollection and
familiarity judgments to rate of forgetting in organic amnesia. Cortex, 38,
161–178.
Haist, F., Shimamura, A. P., & Squire, L. R. (1992). On the relationship between
recall and recognition memory. Journal of Experimental Psychology, 18,
691–702.
Henke, K., Kroll, N. E. A., Behniea, H., Amaral, D. G., Miller, M. B., Rafal,
R., et al. (1999). Memory lost and regained following bilateral hippocampal
damage. Journal of Cognitive Neuroscience, 11, 682–697.
Hirst, W., Johnson, M. K., Kim, J. K., Phelps, E. A., Risse, G., & Volpe, B.
T. (1986). Recognition and recall in amnesics. Journal of Experimental
Psychology: Learning, Memory, & Cognition, 12, 445–451.
Hirst, W., Johnson, M. K., Phelps, E. A., & Volpe, B. T. (1988). More on recognition and recall in amnesics. Journal of Experimental Psychology: Learning,
Memory, & Cognition, 14, 758–762.
1246
M.D. Kopelman et al. / Neuropsychologia 45 (2007) 1232–1246
Holdstock, J. S., Mayes, A. R., Gong, Q. Y., Roberts, N., & Kapur, N. (2005).
Item recognition is less impaired than recall and associative recognition in
a patient with selective hippocampal damage. Hippocampus, 15, 203–215.
Holdstock, J. S., Mayes, A. R., Roberts, N., Cezayirli, E., Isaac, C. L., O’Reilly,
R. C., et al. (2002). Under what conditions is recognition spared relative to
recall after selective hippocampal damage in humans? Hippocampus, 12,
341–351.
Huppert, F. A., & Piercy, M. (1976). Recognition memory in amnesic patients:
Effects of temporal context and familiarity of material. Cortex, 12, 3–20.
Huppert, F. A., & Piercy, M. (1978). The role of trace strength in recency and
frequency judgements by amnesic and control subjects. Quarterly Journal
of Experimental Psychology, 30, 347–354.
Isaac, C. L., & Mayes, A. R. (1999a). Rate of forgetting in amnesia: I. Recall
and recognition of prose. Journal of Experimental Psychology: Learning,
Memory, & Cognition, 25, 942–962.
Isaac, C. L., & Mayes, A. R. (1999b). Rate of forgetting in amnesia: II. Recall
and recognition of word lists at different levels of organization. Journal of
Experimental Psychology: Learning, Memory, & Cognition, 25, 963–977.
Jacoby, L. L., Toth, J. P., & Yonelinas, A. P. (1993). Separating conscious
and unconscious influences of memory: Measuring recollection. Journal
of Experimental Psychology: General, 122, 139–154.
Janowsky, J. S., Shimamura, A. P., Kritchevsky, M., & Squire, L. R. (1989).
Cognitive impairment following frontal lobe damage and its relevance to
human amnesia. Behavioral Neuroscience, 103, 548–560.
Kapur, N., Thompson, P., Kartsounis, L. D., & Abbott, P. (1999). Retrograde
amnesia: Clinical and methodological caveats. Neuropsychologia, 3, 27–30.
Kleinbaum, D. G., Kupper, L. L., Muller, K. E., & Nizam, A. (1998). Applied
regression analysis and other multivariable methods (3rd Ed.). Belmont,
CA, US: Brooks, Cole Publishing Co.
Kopelman, M. D. (2002). Disorders of memory. Brain, 125, 2152–2190.
Kopelman, M. D., & Stanhope, N. (1997). Rates of forgetting in organic amnesia
following temporal lobe, diencephalic, or frontal lobe lesions. Neuropsychology, 11, 343–356.
Kopelman, M. D., & Stanhope, N. (1998). Recall and recognition memory in
patients with focal frontal, temporal lobe and diencephalic lesions. Neuropsychologia, 36, 785–796.
Kopelman, M. D., Lasserson, D., Kingsley, D., Bello, F., Rush, C., Stanhope,
N., et al. (2001). Structural MRI volumetric analysis in patients with organic
amnesia, 2: Correlations with anterograde memory and executive tests in 40
patients. Journal of Neurology, Neurosurgery & Psychiatry, 17, 23–28.
Kopelman, M. D., Lasserson, D., Kingsley, D. R., Bello, F., Rush, C., Stanhope, N., et al. (2003). Retrograde amnesia and the volume of critical brain
structures. Hippocampus, 13, 879–891.
Kramer, J. H., Howard, J. R., An-Tao, D., Schuff, N., Odell, C., Delis, D. C., et al.
(2005). Dissociations in hippocampal and frontal contributions to episodic
memory performance. Neuropsychology, 19(6), 799–805.
Manns, J. R., Hopkins, R. O., Reed, J. M., Kitchener, E. G., & Squire, L. R.
(2003). Recognition memory and the human hippocampus. Neuron, 37,
171–180.
Manns, J. R., & Squire, L. R. (1999). Impaired recognition memory on the
Doors and People Test after damage limited to the hippocampal region.
Hippocampus, 9, 495–499.
Mayes, A. R. (1986). Learning and memory disorders and their assessment.
Neuropsychologia, 24, 25–39.
Mayes, A. R., Holdstock, J. S., Isaac, C. L., Hunkin, N. M., & Roberts,
N. (2002). Relative sparing of item recognition memory in a patient
with adult-onset damage limited to the hippocampus. Hippocampus, 12,
325–340.
Mayes, A. R., Holdstock, J. S., Isaac, C. L., Montaldi, D., Grigor, J., Gummer,
A., et al. (2004). Associative recognition in a patient with selective hippocampal lesions and relatively normal item recognition. Hippocampus, 14,
763–784.
Mori, E., Yoneda, Y., Yamashita, H., Hirono, N., Ikeda, M., & Yamadori,
A. (1997). Medial temporal structures relate to memory impairment in
Alzheimer’s Disease: An MRI volumetric study. Journal of Neurology, Neurosurgery & Psychiatry, 63, 214–221.
Nelson, H. E., & Willison, J. R. (1991). The National Adult Reading Test (2nd
ed.). Windsor: NFER-Nelson.
Ranganath, C., Yonelinas, A. P., Cohen, M. X., Dy, C. J., Tom, S. M., &
D’Esposito, M. (2004). Dissociable correlates of recollection and familiarity
within the medial temporal lobes. Neuropsychologia, 42, 2–13.
Reed, L. J., Marsden, P., Lasserson, D., Sheldon, N., Lewis, P., Stanhope, N.,
et al. (1999). FDG-PET analysis and findings in amnesia resulting from
hypoxia. Memory, 7, 599–612.
Reed, J. M., & Squire, L. R. (1997). Impaired recognition memory in patients
with lesions limited to the hippocampal formation. Behavioral Neuroscience,
11, 667–675.
Stanhope, N., & Kopelman, M. D. (2000). Art and memory: A 7 year follow-up
of herpes encephalitis in a professional artist. Neurocase, 6, 99–110.
Stark, C. E. L., Bayley, P. J., & Squire, L. R. (2002). Recognition memory for
single items and for associations is similarly impaired following damage to
the hippocampal region. Learning & Memory, 9, 238–242.
Stark, C. E. L., & Squire, L. R. (2003). Hippocampal damage equally impairs
memory for single items and memory conjunctions. Hippocampus, 13,
281–292.
Vargha-Khadem, F., Gadian, D. G., Watkins, K. E., & Connelly, A. (1997).
Differential effects of early hippocampal pathology on episodic and semantic
memory. Science, 277, 376–380.
Warrington, E. K., & Weiskrantz, L. (1982). Amnesia: A disconnection syndrome? Neuropsychologia, 20, 233–248.
Yonelinas, A. P., Kroll, N. E. A., Dobbins, I., Lazzara, M. M., & Knight, T.
G. (1998). Recollection and familiarity deficits in amnesia: Convergence of
remember-know, process dissociation, and receiver operating characteristic
data. Neuropsychology, 12, 323–339.
Yonelinas, A. P., Kroll, N. E. A., Quamme, J. R., Lazzara, M. M., Sauve, M.,
Widaman, K. F., et al. (2002). Effects of extensive temporal lobe damage
or mild hypoxia on recollections and familiarity. Nature Neuroscience, 5,
1236–1241.