60 Biological Psychology
landmark in the history of memory research for two reasons.
First, the severe memory impairment (or amnesia) could be
linked directly to the brain tissue that had been removed, sug-
gesting that the medial aspect of the temporal lobe was an
important region for a particular aspect of cognition, that is,
memory function. Second, comprehensive testing of one of
the patients (H. M.) indicated that memory impairment could
occur on a background of otherwise normal cognition. This
observation showed that memory is an isolatable function,
separable from perception and other cognitive and intellec-
tual functions.
The findings from patient H. M. (Scoville & Milner, 1957)
identified a region of the brain important for human memory,
that is, the medial portion of the temporal lobe. The damage
was originally reported to have included the amygdala, the
periamygdaloid cortex (referred to as the uncus in Scoville &
Milner, 1957), the hippocampal region (referred to as the
hippocampus), and the perirhinal, entorhinal, and parahip-
pocampal cortices (referred to as the hippocampal gyrus).
Recently, magnetic resonance imaging of patient H. M. has
shown that his medial temporal lobe damage does not extend
as far posteriorly as originally believed and that damage to
the parahippocampal cortex is minimal (the lesion extends
caudally from the temporal pole approximately 5 cm, instead
of 8 cm, as originally reported; Corkin, Amaral, Gonzalez,
Johnson, & Hyman, 1997).
While these observations identified the medial temporal
lobe as important for memory, the medial temporal lobe is a
large region including many different structures. To deter-
mine which structures are important required that studies be
undertaken in which the effects of damage to medial tempo-
ral lobe structures could be evaluated systematically. Accord-
ingly, soon after the findings from H. M. were reported,
efforts were made to develop an animal model of medial tem-
poral lobe amnesia. During the next 20 years, however, find-
ings from experimental animals with intended hippocampal
lesions or larger lesions of the medial temporal lobe were
inconsistent and difficult to interpret.
In 1978, Mishkin introduced a method for testing memory
in monkeys that captured an important feature of tests sensi-
tive to human memory impairment (Mishkin, 1978). This
method allowed for the testing of memory for single events at
some delay after the event occurred. The task itself is known
as the trial-unique delayed-nonmatching-to-sample task, and
it measures object recognition memory. In Mishkin’s study,
three monkeys sustained large medial temporal lobe lesions
that were intended to reproduce the damage in patient H. M.
The operated monkeys and three unoperated monkeys were
given the delayed-nonmatching-to-sample task in order to as-
sess their ability to remember, after delays ranging from eight
seconds to two minutes, which one of two objects they had re-
cently seen. The monkeys with medial temporal lobe lesions
were severely impaired on the nonmatching task, consistent
with the severe impairment observed in patient H. M. on delay
tasks. Thus, lesions that included the hippocampal region,
the amygdala, as well as adjacent perirhinal, entorhinal, and
parahippocampal cortices caused severe memory impairment.
This work, together with work carried out in the succeeding
few years, established a model of human amnesia in nonhu-
man primates (Mishkin, Spiegler, & Saunders, 1982; Squire &
Zola-Morgan, 1983). Although other tasks have been useful
for measuring memory in monkeys (object discrimination
learning, the visual paired-comparison task; see below), much
of the information about the effects of damage to medial tem-
poral lobe structures has come, until recently, from the
delayed-nonmatching-to-sample task.
Once the animal model was established, systematic and cu-
mulative work eventually identified the structures in the me-
dial temporal lobe that are important for memory. The
important structures are the hippocampal region and the ad-
jacent perirhinal, entorhinal, and parahippocampal cortices
(for reviews, see Mishkin & Murray, 1994; Zola-Morgan &
Squire, 1993). The amygdala proved not to be a component
of this memory system, although it can exert a modulatory
action on the kind of memory that depends on the medial tem-
poral lobe system (Cahill & McGaugh, 1998).
The medial temporal lobe is necessary for establishing one
kind of memory, what is termed long-term declarativeorex-
plicit memory.Declarative memory refers to the capacity for
conscious recollection of facts and events (Squire, 1992). It
is specialized for rapid, even one-trial learning, and for
forming conjunctions between arbitrarily different stimuli. It
is typically assessed in humans by tests of recall, recognition,
or cued recall, and it is typically assessed in monkeys by tests
of recognition (e.g., the delayed-nonmatching-to-sample
task). The medial temporal lobe memory system appears
to perform a critical function beginning at the time of learn-
ing in order that representations can be established in long-
term memory in an enduring and usable form (see also
Eichenbaum, Otto, & Cohen, 1994).
Another important discovery that paralleled in time the
work on the medial temporal lobe system involved the un-
derstanding that there is more than one kind of memory.
Specifically, work with amnesic patients and with experi-
mental animals who sustained lesions to specific brain
regions showed that other kinds of abilities (including skills,
habit learning, simple forms of conditioning, and the phe-
nomenon of priming, which are collectively referred to as
nondeclarative memory) lie outside the province of the me-
dial temporal lobe memory system. Nondeclarative forms of