The Status of Semantic and Episodic Memory in Amnesia
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Chapter
The Status of Semantic and Episodic
Memory in Amnesia
Elizabeth A Kensinger and Kelly S Giovanello
Athinoula A. Martinos Center for Biomedical Imaging
and Department of Psychology, Harvard University
Abstract
Since Scoville and Milner's landmark report of the amnesic patient referred to as H.M., it
has been known that the medial temporal lobes play a critical role in our ability to form
new memories. Importantly, studies of amnesia have highlighted the fact that not all
forms of memory are equally affected by damage to the medial temporal lobes.
Nondeclarative or implicit memory is largely intact, allowing patients to acquire
perceptual and motor skills, conditioned responses, and to demonstrate priming.
Declarative or explicit memory, in contrast, is disrupted. While this distinction between
nondeclarative and declarative memory in amnesia has been explored in detail over the
past five decades, more recent studies have begun to examine possible dissociations
within the domain of declarative memory. One issue regards the extent to which new
semantic information can be learned following damage to the medial temporal lobe. The
present chapter will review evidence suggesting that (a) the medial temporal lobe is not
required for normal retrieval of semantic knowledge but is necessary for normal
acquisition of factual information and that (b) while semantic learning is not normal
following damage to the hippocampal formation, acquisition of some new semantic
knowledge can be supported by regions outside of the hippocampus proper (e.g.,
perirhinal, parahippocampal, or lateral temporal cortices). Another issue concerns whether
the cognitive processes contributing to episodic memory performance are uniformly
disrupted following damage to the medial temporal lobe. The current chapter will review
data suggesting that (a) amnesia disrupts recollection-based mechanisms more so than
familiarity-based mechanisms and that (b) when amnesic patients are encouraged to rely
on their feelings of familiarity, their performance on recognition tasks can be enhanced.
Key words: semantic memory, episodic memory, amnesia, hippocampus, medial temporal
lobe, neuroimaging
Introduction2 Elizabeth A Kensinger and Kelly S Giovanello
To function in daily life, we must be able to retrieve facts about the world (semantic
knowledge) and to remember the specific spatial and temporal details of prior experiences in our
lives (episodic memory). In addition, we must be able to learn new facts and to record new
experiences. Insight into the neural processes underlying these essential abilities came in 1953,
when the neurosurgeon William Beecher Scoville performed an experimental operation to treat
intractable epilepsy. Scoville removed the medial temporal lobes of a man known by the initials
H.M. As reported in the landmark paper coauthored with Brenda Milner (Scoville and Milner,
1957), this surgery resulted in a dense anterograde amnesia. Since the time of his operation, H.M.
has shown a dramatic deficit in his ability to acquire new, long-term declarative memories (see
Corkin, 2002, for review). This impairment in declarative learning extends across all types of
information (verbal and nonverbal, semantic and episodic) and exists regardless of the way in
which memory is assessed (using recall, recognition, or learning to criterion). Thus, it has been
clear since the early studies with H.M. that amnesic patients show decrements in their declarative
(episodic and semantic) memory ability, indicating that the medial temporal lobes are critical for
the normal acquisition of this information.
Retrieval of Semantic Knowledge
A more recent debate has concerned the extent to which medial temporal-lobe structures are
required for accessing established semantic memories (e.g., Nadel and Moscovitch, 1997; Squire
and Zola, 1998; Tulving and Markowitsch, 1998; Vargha-Khadem et al., 1997). On the one hand,
it is plausible that the same medial temporal-lobe structures that are required to learn new facts
would also be required to retrieve information that has already been learned. On the other hand,
these abilities (the acquisition and retrieval of semantic knowledge) could be supported by distinct
processes, with acquisition relying on medial temporal-lobe structures and retrieval depending on
structures outside of the medial temporal lobe. One way to adjudicate between these alternatives is
to examine the semantic retrieval ability of amnesic patients who have extensive damage to the
medial temporal lobe. If these structures are critical for semantic retrieval, than these patients
should be impaired on tasks requiring retrieval of world knowledge. Conversely, if semantic
retrieval can be supported by regions outside of the medial temporal lobe, than these patients
should not show impairments on these tasks.
To distinguish these alternatives, Kensinger, Ullman and Corkin, 2001 examined H.M.’s
ability to retrieve lexical knowledge (stored information about words, including their meanings
and proper forms; Tulving, 1972; Ullman et al., 1997; Ullman 2001), believed to be one
component of semantic memory. H.M. performed a comprehensive series of tasks assessing his
lexical memory ability. On a variety of tasks, including picture naming and spelling, H.M.
performed within 1 standard deviation of the mean of age- and education- matched control
patients. He also performed within the normal range on tasks that required him to generate
irregular word forms (e.g., “Every day I dig a hole. Yesterday I ____ a hole.”), an ability believed
to rely on the lexical memory system.
Additional evidence that H.M.’s ability to retrieve semantic information was unaffected by
his surgery was gleaned from a longitudinal examination of his performance on the Information,
Similarities, Comprehension, and Vocabulary subtests of the Wechesler-Bellevue tests and the
Wechsler Adult Intelligence Scale (Kensinger, Ullman, and Corkin, 2001). These subtests assess
memory for general world knowledge (e.g., from the Comprehension task: “What is the thing toThe Status of Semantic and Episodic Memory in Amnesia 3
do if you find an envelope in the street that is sealed, addressed, and has a fresh stamp on it?”;
from the Information task: “Whose name is typically associated with the theory of relativity?”)
and word meanings (e.g., on the Similarities task, participants are asked what two words, such as
eye and ear, have in common; on the Vocabulary task, participants are asked to orally define
words). H.M. had been tested on these subtests 20 times between 1953 (preoperatively) and 2000.
When his performance across these test sessions was analyzed, there was no main effect of time,
suggesting that H.M.’s semantic knowledge had remained consistent over time, and was not
negatively affected by his surgery.
These examinations with H.M. suggest that the medial temporal-lobes are not required for
normal retrieval of semantic knowledge, but they do not speak to what neural processes do
support semantic retrieval. To address this question, researchers have examined the neural
processes affected in individuals who show semantic memory deficits. Much of this research has
been focused on patients with semantic dementia. Semantic dementia is characterized by a
progressive loss of fluent speech and degradation of world knowledge (Hodges et al., 1992; 2000;
Lambon Ralph et al., 1998, 2001). Critically, patients with semantic dementia typically have
atrophy and dysfunction in the antero-lateral temporal cortex, with relative sparing of the medial
temporal lobes (Hodges et al., 1992; Garrard et al., 1997). Moreover, across a series of patients
with semantic dementia, the most consistent locus of atrophy appears to be in the left anterolateral
temporal lobe (Mummery et al., 2000). Thus, these studies provide convergent evidence that
regions of the temporal neocortex likely are critical for normal semantic retrieval.
Further evidence to support this hypothesis has come from patients with focal (non-
progressive) lesions of the anterotemporal lobe. Particularly when the lesions are left-lateralized,
these patients often show semantic retrieval deficits. For example, Siri, Kensinger, Cappa, Hood,
and Corkin (2003) reported a patient with a deficit retrieving knowledge about particular categories
of objects (e.g., fruits, vegetables, birds, musical instruments) following damage to the anterior
temporal lobe (as a result of herpes encephalitis). This deficit was stable over a number of years,
and existed across a variety of tasks. When the patient’s semantic priming performance was
assessed, it became apparent that while he showed normal priming for some specific items within
these categories, there were other items for which he did not show normal priming (Kensinger,
Siri, Cappa, and Corkin, 2003). In other words, for some items, the representation appeared
intact, but it could not typically be accessed (those items for which the patient showed priming),
whereas other items no longer appeared to have a semantic representation (those items for which
the patient did not show priming). Thus, it appeared that the anterior temporal damage caused a
destruction of the representations of some items, and a disruption in the ability to access
representations of other items.
These data converge on the conclusion that regions of the temporal neocortex support
semantic retrieval. Additional evidence to support this hypothesis came from a study of a variety
of amnesic patients, some with damage restricted to the hippocampus proper, and others with
damage extending to other regions of the medial temporal lobe and to the temporal neocortex
(Schmolck, Kensinger, Squire, and Corkin, 2002). This investigation revealed that patients with
damage limited to the medial temporal lobe showed no deficits on semantic learning tasks. In
contrast, patients with damage to the lateral temporal lobe did show semantic retrieval deficits,
and the magnitude of their deficits correlated with the extent of their damage in lateral temporal
cortex.
These studies, therefore, provide strong evidence that medial temporal lobe structures are not
required for normal retrieval of semantic knowledge. Instead, it seems that the ability to retrieve
(or to store) semantic representations relies on the anterior and lateral temporal cortices.4 Elizabeth A Kensinger and Kelly S Giovanello
Acquisition of Semantic Knowledge
Although the medial temporal lobes do not appear critical for the retrieval of already-
established semantic knowledge, numerous studies have provided evidence that these structures are
required for the normal acquisition of new semantic information. Thus, amnesic patients with
damage to the medial temporal lobes show profound deficits in acquiring factual information
(Milner, 1959), and damage circumscribed to the hippocampus proper seems sufficient to disrupt
normal semantic learning (Manns et al., 2003). Moreover, the amount of damage to the medial
temporal lobe appears to correspond with the magnitude of the deficit in learning new semantic
information (e.g., Verfaellie et al., 2000).
These studies, therefore, highlight the fact that the medial temporal lobes (and, more
specifically, the hippocampus proper) are required for normal semantic memory acquisition, and
that amnesic patients with damage to those regions show impaired semantic learning.
Nevertheless, it is important to note that, even with extensive medial temporal lobe damage,
some semantic learning does appear to occur. For example, patients with dense anterograde
amnesia have been able to remember the words used to complete meaningful sentences
(Shimamura and Squire, 1988), and to learn new computer-related vocabulary and computer
commands (e.g., Glisky and Schacter, 1988; van der Linden and Coyette, 1995). H.M. has also
shown the ability to learn new factual information when he can anchor it to premorbidly-acquired
knowledge (e.g., learning that the Salk vaccine protects against polio, given that polio was a
familiar disease to H.M.; Skotko et al., 2004).
It is plausible that some of this semantic acquisition is supported by nondeclarative memory
mechanisms, such as habit learning. For example, Bayley and Squire (2002) found that an
amnesic patient, E.P., with extensive medial temporal-lobe damage, as well as damage to
temporal neocortex (Stefanacci et al., 2000), was nevertheless capable of acquiring semantic
knowledge. However, this knowledge tended to be rigid and inflexible, showing no characteristics
of declarative memory.
There have been demonstrations of semantic learning in amnesia, however, in which the
semantic learning appears to reflect declarative memory. For example, H.M performs well above
chance at selecting famous names among nonfamous foils, even when these individuals became
famous following the onset of his amnesia (O’Kane, Kensinger and Corkin, 2004). He also is able
to freely recall information about a small number of people who have become well-known after
1953. For example, H.M. indicated that Woody Allen was a “comic, in movie pictures” and that
Michael Gorbachev was “famous for making speeches…head of the Russian parliament” (examples
from O’Kane et al., 2003). H.M. also is able to draw the floor plan of the house where he lived
for many years, despite the fact that he did not move into the house until after his surgery
(discussed in Corkin, 2002).
It is important to note that this semantic knowledge differs from that of control participants
in a number of ways. Most notably, while control participants typically are better at generating
knowledge about individuals who became famous recently as compared to remotely, amnesic
patients often show the reverse pattern of performance (O’Kane, Kensinger, and Corkin, 2004;
Verfaellie et al., 2000; Westmacott and Moscovitch, 2001). In addition, there seems to be a less
consistent access to the semantic representations in patients with extensive medial temporal lobe
damage. For example, while H.M. is often able to indicate that Ronald Reagan was President asThe Status of Semantic and Episodic Memory in Amnesia 5
well as an actor, that J.F.K. was assassinated, and that Margaret Thatcher was a British politician
(Kensinger et al., unpublished observations; cited in O’Kane et al., 2004), he is not always able to
generate this information. Moreover, in one demonstration of semantic learning in H.M. (Skotko
et al., 2004), the learning was not permanent, but rather had decayed after a relatively short period
of time (a few months).
This pattern of results suggests that the semantic learning that proceeds in the absence of a
functioning hippocampus requires multiple exposures and extended repetitions, and even with
such repetitions it does not reach the same level of detail and stability as is the hallmark of
semantic learning in the presence of intact hippocampal function. Thus, it seems likely that while
the hippocampus is essential for the one-trial learning that can often underlie the accumulation of
rich semantic knowledge. In contrast, regions outside of the hippocampus proper (either posterior
parahippocampal gyrus or lateral temporal cortices) may be able to support the gradual
accumulation of semantic knowledge over multiple repetitions. Even after these repetitions, the
knowledge learned may not be as stable or accessible as information learned via hippocampal
mechanisms.
This set of conclusions would be consistent with prior failures to demonstrate semantic
learning in H.M. and in other amnesic patients when information is taught over a restricted period
of time in a laboratory setting (Gabrieli et al., 1998) or when memory is assessed for infrequently
encountered new vocabulary words (Postle and Corkin, 1998). It is plausible that encoding
variability, and encoding repetitions, are critical for new semantic learning to occur following
extensive medial temporal lobe damage. There may not have been sufficient repetitions, or
sufficient encoding variability, to allow for the non-hippocampal mechanisms to build up
semantic memories in these other situations.
Clarification of the neural processes supporting this gradual acquisition of semantic
knowledge may come from comparisons of the amnesic patient H.M. who, as discussed above,
shows evidence of semantic learning with the amnesic patient E.P., who has not shown evidence
of such learning. E.P. has more diffuse damage than H.M., in both the caudal parahippocampal
gyrus and in the lateral temporal cortex (Stefanacci et al., 2000). Thus, it may be that these
regions damaged in E.P., but spared in H.M., are those that allow some gradual learning of
declarative, semantic knowledge.
Taken together, the investigations of semantic learning in amnesia suggest a couple of
conclusions. First, some new semantic learning can proceed in the absence of a functioning
hippocampus. This learning may be supported either by parahippocampal cortex or by the lateral
temporal cortex. Second, while these extra-hippocampal mechanisms can support limited semantic
knowledge acquisition, they do not appear capable of supporting the rapid and flexible acquisition
of semantic knowledge that individuals with intact hippocampi engage.
Semantic Memory In Amnesia: Conclusions
In summary, studies of semantic memory in medial temporal-lobe amnesia have confirmed
that damage limited to the medial temporal lobes does not disrupt the ability to retrieve already-
acquired semantic knowledge. Rather, it appears that damage to the temporal neocortex is required
for semantic retrieval deficits to emerge. In contrast, the medial temporal-lobe is necessary for
normal semantic learning. Damage to the hippocampal formation impairs semantic learning, and
the magnitude of the impairment is correlated with the extent of the damage. However, even with6 Elizabeth A Kensinger and Kelly S Giovanello
extensive damage to medial temporal-lobe structures (as with the patient H.M.), some new
semantic learning is possible (though the learning is clearly not normal). It may be that the
structures of the medial temporal-lobe spared in H.M. (i.e., the caudal portion of the
parahippocampal gyrus) are sufficient to support new semantic learning. It also may be the case
that lateral temporal regions are capable of the gradual acquisition of declarative semantic
knowledge when a sufficient number of encoding episodes occur.
Episodic Memory in Amnesia
A pervasive deficit in episodic memory is dramatically exemplified in patients with
anterograde amnesia, who are unable to acquire and retrieve any events or episodes from their
personal life that have occurred since the onset of their amnesia. Such an impairment is typically
revealed on explicit tests of memory, which ask patients to consciously retrieve recent experiences
in the form of recall or recognition (Giovanello and Verfaellie, 2001a). Although both recall and
recognition tests require conscious retrieval of recent experiences, their processing demands are not
identical. Dual-process models of recognition memory postulate two distinct bases of
performance: recollection and familiarity (Jacoby and Dallas, 1981; Mandler, 1980). Recollection
is thought to be a conscious, attention demanding process in which prior aspects of an experience
are retrieved. Familiarity, in contrast, is thought to be an unconscious, automatic process that
arises when fluent processing of a stimulus is attributed to prior experience with that stimulus.
Whereas recall tasks require conscious recollection of contextually appropriate information,
recognition tasks can be supported either by recollection or by familiarity.
Amnesic patients' impaired performance on tasks of recall and recognition suggests that both
recollection and familiarity rely on the integrity of the medial-temporal lobes. Under certain
conditions, however, amnesic patients can show surprisingly good recognition memory (Bowers,
et al., 1988; Johnson and Kim, 1985). Based on a review of the literature, as well as their own
data, Yonelinas and colleagues (1998) concluded that amnesic patients, although impaired in both
processes, evidence a disruption in familiarity that is consistently smaller than the disruption in
recollection. In contrast to this view, however, others have argued that recollection and familiarity
are equally impaired in amnesia (Knowlton and Squire, 1995; Reed and Squire, 1997; Reed and
Squire, 1998).
Evidence for a Disproportionate Deficit in
Recollection in Amnesia
At present it remains controversial as to whether amnesic patients evidence a disproportionate
disruption in recollection. One line of research that is relevant to this debate concerns the
comparison of amnesic patients’ recall and recognition performance. It has been hypothesized that
if recollection and familiarity-based processes are equally impaired in amnesia, then proportionate
deficits in recall and recognition should be seen. Alternatively, if familiarity-based recognition is
relatively preserved in amnesia, then amnesics’ recall should be more severely impaired than their
recognition.
It is not possible, however, to compare directly the severity of impairments across recall and
recognition tasks because these tasks use different scales of measurement. To overcome thisThe Status of Semantic and Episodic Memory in Amnesia 7
problem, studies examining whether amnesic patients show disproportionate deficits in recall
typically equate the recognition performance of amnesics and controls, and then assess whether,
under these conditions, recall performance is equated as well. Recognition has been equated by
delaying the testing for control subjects or by giving the amnesic patients additional study
exposure. Perhaps because of the complexities inherent in this approach, the results of such
studies have been inconsistent. Some studies have reported that when recognition was equated
between amnesic patients and control participants, the patients’ recall remained impaired (Hirst et
al., 1986, 1988). Others have found that once recognition memory was matched across groups,
recall was similar across groups as well. (Haist, Shimamura, and Squire, 1992; Kopelman and
Stanhope, 1998; MacAndrew, Jones, and Mayes, 1994; Shimamura and Squire, 1988).
In a more recent study, Giovanello and Verfaellie (2001b) provided evidence that the
relationship between amnesics’ recall and recognition performance depends, at least in part, on the
method by which amnesics’ recognition is matched to that of controls. In one experiment, both
amnesics and controls were tested after the same delay, but amnesics were provided with
additional study exposures. It was hypothesized that even with additional study exposures, the
patients’ recognition would be mediated primarily by familiarity, while control participants would
able to use both recollection and familiarity as a basis for recognition. Consistent with this
hypothesis, amnesic patients recall remained impaired even when recognition was matched to that
of controls. In a second experiment, recognition memory was equated by giving the amnesic and
control groups equal study exposure, but by testing controls following a 24-hour delay. Because
recollection declines more rapidly than does familiarity over a delay (Gardiner and Java, 1991;
Tulving, 1985), it was hypothesized that the performance of controls would be mediated mainly
by familiarity, as is that of amnesic patients. As hypothesized, recall was equated across groups
under these equating conditions. These findings highlight the importance of analyzing the
performance of amnesic patients and control participants in term of underlying processes. When
such an analysis was conducted, two apparently contradictory patterns of performance could both
be explained with reference to the notion that amnesic patients have a disproportionate deficit in
recollection.
Another line of research relevant to the debate over whether recollection is disproportionately
disrupted in amnesia, concerns the comparison of amnesic patients’ item and associative
recognition memory. The study of associative memory in amnesia has been motivated by
cognitive studies in young adults that provide empirical evidence for the distinction between item
and associative recognition. For example, employing unrelated word pairs in a study with young
control participants, Hockley and Consoli (1999) showed that item and associative recognition are
differentially associated with familiarity and recollection; associative recognition was accompanied
by a greater proportion of "remember" responses (assumed to reflect recollection) than was item
recognition, whereas item recognition was accompanied by a greater proportion of "know"
responses (assumed to reflect familiarity) than was associative recognition. Based on their
findings, Hockely and Consoli (1999) concluded that associative recognition is based on
conscious recollection to a greater extent than is item recognition, and conversely, that item
memory is based on familiarity to a greater extent than is associative recognition.
These findings suggest that amnesics’ associative recognition performance should differ
depending upon whether or not recollection is disproportionately impaired. If amnesia disrupts
recollection-based recognition more than familiarity-based recognition, amnesic patients should be
disproportionately impaired in associative, relative to item, recognition. Alternatively, if
recollection and familiarity-based processes are equally impaired in amnesia, then proportionate
deficits in item recognition and associative recognition should be observed.8 Elizabeth A Kensinger and Kelly S Giovanello
To determine whether amnesic patients evidence a disproportionate impairment in associative
memory, several studies have equated item recognition between amnesic patients and control
participants and have then examined whether, under these conditions, associative recognition is
matched as well (Giovanello, Verfaellie, and Keane, 2003b; Stark, Bayley, and Squire, 2002;
Turriziani, et al., 2004). While item recognition involves discriminating between studied and
unstudied stimuli, associative recognition involves distinguishing between studied associations
(referred to as “intact pairs”) and novel associations that consist of stimulus components
previously seen, though not seen together (referred to as “recombined pairs”). These patient studies
have revealed that associative recognition remains impaired in amnesia, even when performance on
item recognition is matched (Giovanello, Verfaellie, and Keane, 2003b; Turriziani, et al., 2004)
but see (Stark, Bayley and Squire, 2002). Specifically, relative to controls’ performance, amnesic
patients’ hit rates (i.e., correct endorsements) for intact pairs were lower, while their false alarm
rates (i.e., incorrect endorsements) for recombined pairs were higher. Such a pattern of performance
is consistent with a disproportionate deficit in recollection in amnesia.
Taken together, these findings are inconsistent with the view that recollection and familiarity
are equally impaired in amnesia (Knowlton and Squire, 1995; Reed and Squire, 1997; Reed and
Squire, 1998). Rather, these findings lend support to the notion that amnesia disrupts
recollection-based mechanisms to a greater extent than familiarity-based mechanisms (Aggleton
and Shaw, 1996; Verfaellie and Treadwell, 1993; Yonelinas et al., 1998).
Enhancing Familiarity-Based Processing in
Amnesia
If amnesic patients have a relative preservation of familiarity-based recognition, the question
arises as to whether conditions can be created that enhance patients reliance on familiarity, thus
boosting their recognition performance. Verfaellie, Giovanello, and Keane (2001) explored this
possibility by manipulating amnesic patients’ response criterion in a recognition task. Their
interest in this manipulation arose from a study that compared the recognition performance
patients with ECT-induced amnesia under a high-criterion condition and a low-criterion condition
(Dorfman et al., 1995). In the high- criterion condition, patients were told to endorse a test word
only if they were relatively sure that it had appeared on the study list. In the low-criterion
condition, patients were encouraged to say yes to a test word if it at all seemed familiar. Patients
endorsed more studied, but not unstudied, words in the low- than in the high-criterion condition,
indicating a change in discriminability. The patients’ enhanced recognition was attributed to the
contribution of familiarity to performance in the low-criterion condition.
Using a similar manipulation in patients with global amnesia, Verfaellie, Giovanello, and
Keane (2001) did not find that changes in response criterion affected amnesic patients’ recognition
accuracy (see also Reber and Squire, 1999). This direct manipulation on response criterion,
however, was generally weak and failed to affect overall performance either in amnesic patients or
in control participants. In a second experiment, response criterion was manipulated indirectly by
providing information about the alleged base-rate of study items on the recognition test (30%
versus 70%). In the 30% base-rate condition, participants were told that 3 out of every 10 words
would be from the study list. In the 70% base-rate condition, participants were told that 7 out of
every 10 words would be from the study list. The base-rate manipulation was successful in
affecting the performance of control participants, as controls endorsed more items (both studiedThe Status of Semantic and Episodic Memory in Amnesia 9
and unstudied) in the 70% than in the 30% condition. Most importantly, while this manipulation
affected control participants’ response bias, it enhanced amnesic patients’ discriminability.
Furthermore, an analysis of patients’ “remember” and “know” responses indicated that that the
improved accuracy in amnesia was associated with enhanced familiarity-based recognition. These
findings demonstrate that encouraging amnesic patients to rely on familiarity can boost their item
recognition performance.
More recently, Giovanello, Keane, and Verfaellie (2003a) examined whether conditions exists
in which familiarity can support associative recognition, and if so, whether amnesics’ associative
recognition is less impaired under these conditions. In one experiment, using the remember/know
paradigm with control participants, they compared the contribution of familiarity to associative
recognition for stimuli that were meaningfully integrated pre-experimentally (compound words)
with stimuli that were not (unrelated words). It was hypothesized that familiarity may contribute
to a greater extent to associative recognition for compound stimuli than to associative recognition
of unrelated word pairs. This hypothesis arose from Yonelinas and colleagues’ (1999) finding that,
in the nonverbal domain, when stimuli are encoded as a whole, familiarity makes a larger
contribution to associative judgments, whereas when stimuli are encoded as a set of features, the
contribution of familiarity is greatly reduced. It was demonstrated that familiarity makes a greater
contribution to associative recognition of compound stimuli than to associative recognition of
unrelated word pairs.
In a further experiment, Giovanello, Keane, and Verfaellie (2003a) examined associative
recognition memory in amnesic patients and in control participants for the stimuli employed in
their previous experiment. It was hypothesized that if amnesic patients’ deficit in associative
memory is due to the fact that associative memory typically depends on recollection, amnesics’
performance should be less impaired under conditions in which associative memory can be
supported by familiarity for the association, as is the case for compounds. Alternatively, if the
associative memory deficit occurs regardless of the mechanism that supports recognition, than
amnesics’ associative recognition performance should be equally impaired for compound stimuli
and for unrelated stimuli. The findings were consistent with the former hypothesis: Whereas
associative recognition for compounds and unrelated words was nearly identical in control
participants, associative recognition was greater for compounds than for unrelated words in
amnesic patients. These results demonstrate that amnesics’ associative recognition is enhanced
when associative recognition can be supported by familiarity.
Taken together, these findings demonstrate that conditions can be created that enhance
patients reliance on familiarity, thus boosting their recognition performance. Furthermore, these
results show that familiarity can support item recognition performance, as well as associative
recognition performance, in amnesia.
Episodic Memory in Amnesia: Conclusions
In summary, recent findings from studies of episodic memory in amnesia have been
interpreted in the context of dual process models of recognition memory which postulate two
distinct bases of performance: recollection and familiarity. While neither process is normal in
amnesia, several findings suggest that the disruption in familiarity is consistently smaller than the
disruption in recollection. Consistent with this notion, comparisons of free recall and item
recognition, as well as comparisons of item recognition and associative recognition, in amnesia10 Elizabeth A Kensinger and Kelly S Giovanello
demonstrate a disproportionate deficit in recollection-based mechanisms. In light of the fact that
amnesic patients may have a relative preservation of familiarity-based mechanisms, several studies
have examined whether conditions can be created that enhance patients’ reliance on familiarity,
thus boosting their recognition performance. The studies show that encouraging patients to rely on
their feelings of familiarity can boost their performance.
General Conclusions
While memory deficits are the hallmark of amnesia, the research discussed in this chapter has
highlighted the fact that not all aspects of declarative memory are equally affected. This research
has opened the door for research into the neural processes that continue to operate in amnesia,
allowing patients to gradually acquire at least some fragmentary semantic knowledge, and to gain
a sense of familiarity about information they have previously encountered.
References
Aggleton JP, Shaw C. 1996. Amnesia and recognition memory: A re-analysis of psychometric
data. Neuropsychologia 34:51-62
Bayley PJ, Squire LR. 2002. Medial temporal lobe amnesia: gradual acquisition of factual
information by nondeclarative memory. J Neurosci 22:5741–5748.
Bowers D, Verfaellie M, Valenstein E, Heilman KM. 1988. Impaired acquisition of temporal
information in retrosplenial amnesia. Brain and Cognition 8:47-66.
Corkin S. 1984. Lasting consequences of bilateral medial temporal lobectomy:clinical course and
experimental findings in H.M. Semin Neurol 4:249–259.
Corkin S. 2002. What’s new with the amnesic patient H.M.? Nat Rev Neurosci 3:153–160.
Dorfman J, Kihlstrom JF, Cork RC, Misiaszek, J. 1995. Priming and recognition in ECT-
induced amnesia. Psychonomic Bulletin and Review 5:244-248.
Gardiner JM, Java RJ. 1991. Forgetting in recognition memory with and without recollective
experience. Memory and Cognition 19: 617-623.
Garrard P, Perry R, Hodges JR. 1997. Disorders of semantic memory. J Neurol, Neurosurg
Psychiatry 62:431–435.
Giovanello KS, Verfaellie M. 2001a. Memory Systems of the Brain: A Cognitive
Neuropsychological Analysis. Seminars in Speech and Language 22:107-116.
Giovanello KS, Verfaellie M. 2001b. The relationship between recall and recognition in amnesia:
Effects of matching recognition between amnesic patients and controls. Neuropsychology
15:444-451.
Giovanello KS, Verfaellie M, Keane MM. 2003a. Familiarity-based associative recognition
memory in amnesia. Journal of Cognitive Neuroscience, C179:107.
Giovanello KS, Verfaellie M, Keane MM. 2003b. Disproportionate deficit in associative
Recognition relative to item recognition in global amnesia. Cognitive Affective Behavioral
Neuroscience 3:186-194
Haist F, Shimamura AP, Squire LR. 1992. On the relationship between recall and recognition
memory. J Exp Psychol: Learn Mem Cogn 18:691-702.The Status of Semantic and Episodic Memory in Amnesia 11
Hirst W, Johnson MK, Kim JK, Phelps EA, Risse G, Volpe BT. 1986. Recognition and recall in
amnesics. J Exp Psychol: Learn Mem Cogn 12:445-451.
Hirst W, Johnson MK, Phelps EA, Volpe, BT. 1988. More on recognition and recall in amnesics.
J Exp Psychol: Learn Mem Cogn 14:758-762.
Hockley WA, Consoli A. 1999. Familiarity and recollection in item and associative recognition.
Memory and Cognition 27:657-664.
Hodges JR, Patterson K, Oxbury S, Funnell E. 1992b. Semantic dementia: progressive fluent
aphasia with temporal lobe atrophy. Brain 115:1783–1806.
Hodges JR, Bozeat S, Lambon Ralph MA, Patterson K, Spatt J. 2000. The role of conceptual
knowledge in object use: Evidence from semantic dementia. Brain 123:1913–1925.
Jacoby LL, Dallas M. 1981. On the relationship between autobiographical memory and perceptual
learning. J Exp Psychol Gen 110:306-340.
Johnson MK, Kim JK, 1985. Recognition of pictures by alcoholic Korsakoff patients. Bulletin of
the Psychonomic Society 23:456-458.
Kensinger EA, Ullman MT, and Corkin S. 2001. Bilateral medial temporal lobe damage does not
prevent the retrieval and use of grammatical or lexical information: evidence from the amnesic
patient H.M. Hippocampus 11:347-360.
Kensinger EA, Siri S, Cappa SF, and Corkin S. 2003. Role of the anterior temporal lobe in
repetition and semantic priming: Evidence from a patient with a category-specific deficit.
Neuropsychologia 41:71-84.
Knowlton B, Squire LR. Remembering and knowing: Two different expressions of declarative
memory. J Exp Psychol: Learn Mem Cogn 21:699-710.
Kopelman, MD, Stanhope N. 1998. Recall and recognition memory in patients with focal frontal,
temporal lobe and diencephalic lesions. Neuropsychologia 36:785-796.
Lambon Ralph MA, Graham KS, Ellis AW, Hodges JR. 1998. Naming in semantic
dementia—what matters? Neuropsychologia 8:775–784.
MacAndrew, SBG, Jones GV, Mayes AR. 1994. No selective deficit in recall in amnesia.
Memory 2: 241-254.
Mandler, G. 19980. Recognizing the judgment of previous occurrence. Psychological Review
87:252-271.
Mummery CJ, Patterson K, Price CJ, Ashburner J, Frackowiak RS, Hodges JR. 2000. A voxel
based morphometric study of semantic access from pictures and words. Bury St. Edmunds:
Thames Valley dementia: the relationship between temporal lobe atrophy and semantic
memory. Ann Neurol 47:36–45.
Nadel L, Moscovitch M. 1997. Memory consolidation, retrograde amnesia and the hippocampal
complex. Curr Opin Neurobiol 7:217–227.
O’Kane G, Kensinger EA, and Corkin S. 2004. Evidence for semantic learning in amnesia: A
study with the amnesic patient H.M. Hippocampus 14:417-425.
Reber PJ, Squire LR. 1999. Relaxing decision criteria does not improve recognition memory in
amnesic patients. Memory and Cognition 27:501-511.
Reed JM, Squire LR. 1997. Impaired recognition memory in patients with lesions limited to the
hippocampal formation. Behavioral Neuroscience 111:1163-1170.
Reed, JM, Squire LR. 1998. Retrograde amnesia for facts and events: Findings from four new
cases. Journal of Neuroscience 18:3943-3954.
Schmolck H, Kensinger EA, Corkin S, and Squire LR (2002). Semantic knowledge in patient
H.M. and other patients with bilateral and medial and lateral temporal lobe lesions.
Hippocampus 12:520-533.12 Elizabeth A Kensinger and Kelly S Giovanello
Scoville WB, Milner B. 1957 Loss of recent memory after bilateral hippocampal lesions. J Neurol
Neurosurg Psychiatry 20:11–21.
Shimamura AP, Squire LR. 1988. Long-term memory in amnesia: cued recall, recognition
memory and confidence ratings. J Exper Psychol: Learn Mem Cogn 14:763-770.
Siri S, Kensinger EA, Cappa SF, Hood KL, and Corkin S (2003). Questioning the living-
nonliving dichotomy: Evidence from a patient with an unusual semantic dissociation.
Neuropsychology, 17, 630-645.
Skotko B, Kensinger EA, Locascio JJ, Einstein JG, Rubin DC, Tupler LA, Krendl AC, and
Corkin S (2004). Puzzling thoughts for H.M.: Can new semantic memories be anchored to
old semantic memories? Neuropsychology, 18, 756-769.
Squire LR, Zola SM. 1998. Episodic memory, semantic memory, and amnesia. Hippocampus
8:205–211.
Stark CEL, Bayley PJ, Squire LR. 2002. Recognition memory for single items and for
associations is similarly impaired following damage to the hippocampal region. Learning and
Memory 9:238-242.
Stefanacci L, Buffalo EA, Schmolck H, Squire LR. 2000. Profound amnesia following damage to
the medial temporal lobe—a neuroanatomical and neuropsychological profile of patient E.P. J
Neurosci 20:7024 –7036.
Tulving E. 1972. Episodic and semantic memory. In: Tulving E, Donaldson W, editors.
Organization of memory. New York: Academic Press.
Tulving E. 1985. Memory and consciousness. Canadian Psychology 26:1-12.
Tulving E, Markowitsch HJ. 1998. Episodic and declarative memory: role of the hippocampus.
Hippocampus 8:198 –204.
Turriziani P, Fadda L, Caltagirone C, Carlesimo GA. 2004. Recognition memory for single items
and for associations in amnesic patients. Neuropsychologia 42:426-433.
Ullman MT. 2001. The declarative/procedural model of language. J Psycholinguist Res 30:57–
69.
Ullman MT, Corkin S, Coppola M, Hickok G, Growdon JH, Koroshetz, WJ, Pinker S. 1997. A
neural dissociation within language: evidence that the mental dictionary is part of declarative
memory, and that grammatical rules are processed by the procedural system. J Cogn Neurosci
9:266 –276.
Vargha-Khadem F, Gadian DG, Watkins KE, Connelly A, Van Paesschen W, Mishkin M. 1997.
Differential effects of early hippocampal pathology on episodic and semantic memory.
Science 277:376 –380.
Verfaellie M, Giovanello KS, Keane MM. 2001. Recognition memory in amnesia: Effects of
relaxing response criteria. Cognitive Affective Behavioral Neuroscience 1:3-9.
Verfaellie M, Koseff P, Alexander MP. 2000. Acquisition of novel semantic information in
amnesia: effects of lesion location. Neuropsychologia 38:484–492.
Verfaellie M, Treadwell J. 1993. Status of recognition memory in amnesia. Neuropsychology 7:5-
13.
Westmacott R, Moscovitch M. 2001. Names and words without meaning: incidental postmorbid
semantic learning in a person with extensive bilateral medial temporal damage.
Neuropsychology 15: 586 –596.
Yonelinas AP, Kroll NEA, Dobbins I, Lazzara M, Knight RT. 1998. Recollection and familiarity
deficits in amnesia: Convergence of remember-know, process dissociation, and receiver
operating characteristic data. Neuropsychology 12:323-339.The Status of Semantic and Episodic Memory in Amnesia 13
Yonelinas AP, Kroll NEA, Dobbins IG, Soltani M. 1999. Recognition memory for faces: When
familiarity supports associative recognition judgments. Psychonomic Bulletin and Review
6:654-661.
Acknowledgements
E.A.K. is supported by NIH grant MH 070199 and K.S.G. is supported by NIH grant AG 023439You can also read
