Declarative Memory
Semantic memory As noted, semantic memory refers
the recall of general or factual knowledge. Older adults
commonly complain of subjective semantic memory
problems when, for example, they report difficulty
recalling the names of common objects or other welllearned
information. Yet, despite these subjective complaints,
semantic memory is among the more stable
memory systems across the adult life span. The construct
can be operationally defined by requiring subjects
to define words or provide the answers to factual
questions (e.g., the capital of a certain country), such
as on the Vocabulary and Information subtests of the
Wechsler Adult Intelligence Scales. Semantic memory
is often included as part of the definition of ‘crystallized
intelligence,’ which reflects an accumulation of
information acquired over time and that is relatively
impermeable to the effects of normal aging or mild
brain disease.
It is well established that semantic memory shows
very little decline in normal aging. In fact, semantic
knowledge accumulation and memory increase into
the sixth and seventh decades of life and may show
only a gradual decline afterward. Much of our understanding
of the impact of age on semantic memory has
come from large-scale longitudinal and cross-sectional
studies of normal aging. For example, longitudinal
data from the Canberra Study , which followed a random
sample of adults over the age of 70 years, demonstrated
that crystallized abilities remained stable over
approximately 8 years. This pattern of stability was
apparent when age-associated differences (i.e., crosssectional
analysis) were considered as well. Similarly,
Denise Park and her colleagues measured knowledgebased
verbal ability, including three semantic memory
tasks tapping word knowledge, in a large sample of
healthy adults ranging in age from 20 to 92 years, and
found a gradual increase in performance across the age
groups. This finding is again consistent with the idea
that semantic knowledge accumulates across the life
span with little or no deleterious effects of normal aging.
Although there is little cross-sectional or longitudinal
evidence to suggest that semantic memory changes significantly
with normal aging, why are subjective complaints
of semantic recall so common among older
adults? One phenomenon, termed ‘tip-of-the-tongue’
(TOT), may explain this occurrence. TOT refers to
Aging and Memory in Humans 175
Encyclopedia of Neuroscience (2009), vol. 1, pp. 175-180
Author's personal copy
the common experience in which individuals have the
feeling that they know the correct information (e.g., a
person’s name, a relatively low-frequency word), yet
they are unable to recall it explicitly. The frequency
of the TOT phenomenon increases with advancing
age and may underlie the perceived difficulty with
semantic recall.
Episodic memory The contrast between episodic
memory and semantic memory was introduced by
Endel Tulving in the early 1970s. Episodic memory
refers to the explicit recollection of events, the ‘what,’
‘where,’ and ‘when’ of information storage, and
though it is conceptually distinct from semantic
memory, the two memory systems interact. Episodic
memory binds together items in semantic memory to
form conceptually related time-based events. For
example, the explicit recall of a learned story about
a cowboy requires episodic memory for the story and
semantic memory, or prior knowledge, of the items
contained within the story.
Unlike semantic memory, episodic memory declines
considerably with age. Older adults, for example,
when prompted, have more difficulty recalling what
they had for breakfast than do younger adults. The
formal observation that episodic memory is affected
by aging has existed for decades and has been documented
numerous times.
Episodic memory is typically tested by requiring
persons to learn information explicitly (e.g., a list or
story) and recall it after a delay period. The three
aspects of episodic memory include the encoding
phase, the storage phase, and the retrieval of the
encoded and stored information. These three phases
show differential aging effects. Older adults’ overall
difficulty on tasks of episodic memory may be partially
accounted for by a more shallow depth of
encoding, as compared to younger adults – that is,
older adults recall less information because of more
limited processing of the initial study stimuli. This
idea is supported by findings of greater age-related
decline in acquisition or early retrieval of new information
than in the degree of forgetting (i.e., the
amount of information lost relative to the amount
of information encoded). However, on free-recall
tasks of list-learning paradigms, older adults recall
fewer absolute words than do their younger adult
counterparts; when given the correct stimulus in a
recognition paradigm, older adults tend to incorrectly
endorse more distractor stimuli, or foils.
Several observations about episodic memory and
aging have emerged from the recent literature. First,
the pattern of age-related differences in episodic
memory appears to be similar across several modalities
and domains, such as story recall, paired
associate learning, face and word recall, and recognition
paradigms of verbal and nonverbal information.
Second, cross-sectional data suggest that age-associated
episodic memory decline begins as early as age 20 years
and decline linearly until about age 60, at which time
there is a more precipitous decline. Third, whether
the amount of interindividual variability increases
systematically as a function of age is still somewhat
unclear. Greater variability with aging would suggest
that episodic memory decline might be a marker
for insipient brain pathology, rather than a primary
aspect of normal aging.
There are several competing theories postulating
the potential mechanisms for age-associated declines
in episodic memory. They can generally be divided
into four areas, including an age-related failure of
memory monitoring, or metamemory; age-associated
decreases in the depth of initial encoding; age-related
impairment in processing of contextual information;
and age-associated decline in a number of processing
resources. While these theories have not been fully
substantiated empirically, the latter has received the
greatest amount of support in the cognition literature.
Proponents of a ‘resource reduction hypothesis’ argue
that central to age-associated changes in episodic
memory is a reduction in primary cognitive resources
such as attention or working memory, or a reduction
in the ability to engage due to an age-associated diminution
in attentional inhibitory control. Others argue
that age-associated memory decline is not due to a
reduction in available attentional resources per se, but
rather to age-related declines in perceptual processing
speed.
Source memory Episodic memory comprises the
information that is being recalled as well as the context
in which the information was learned. This latter
aspect is referred to as source memory, and there is
increasing evidence that even when older adults successfully
recall information, they may have difficulty
recalling the source in which the information was
acquired. The phenomenon has been demonstrated
in the identification of the temporal context in
which an item was learned, as well as the spatial
and perceptual context. Studies by Daniel Schacter
and colleagues required older adults to listen to different
speakers read different blocks of declarative
statements, and found that memory for the source of
the declarative sentences was disproportionately
worse than was memory for the statements. This
general finding has been well replicated and has been
extended to demonstrate age-associated impairment
in both specific-source memory and partial-source
memory. For example, older adults have difficulty,
relative to younger adults, remembering which of
176 Aging and Memory in Humans
Encyclopedia of Neuroscience (2009), vol. 1, pp. 175-180
Author's personal copy
four people spoke a word (i.e., specific-source memory)
as well as remembering partial information about the
person who spoke the word, such as his or her gender
(i.e., partial-source memory).