Page 1 of 14
Chapter 11: Learning, Memory, and Amnesia –
How Your Brain Stores Information
• Learning and memory are both neuroplastic processes that deal with the brain’s ability to change its
functioning in response to experience.
• Learning: how experience changes the brain
• Memory: how these changes are stored and subsequently reactivated
• Amnesia: pathological loss of memory
11.1 AMNESIC EFFECTS OF BILATERAL MEDIAL TEMPORAL
The case of H.M., the man who changed the study of memory
• In 1953, at the age of 27, Henry Molaison (1926-2008) had bilateral medial temporal lobectomy, the
removal of the medial portions of both temporal lobes, for the treatment of his severe epilepsy. This
included the removal of most of the hippocampus, amygdala, and adjacent cortex.
• Lobectomy: a lobe or a major part of one is removed from the brain; lobotomy: a lobe or the major
part of one is separated from the rest of the brain by a large cut, but is not removed.
• H.M.’s generalized seizures were eliminated and his incidence of partial seizures was reduced to 1-2 a
day. He also retained normal perceptual and motor abilities, and his IQ increased from 104 to 118.
• Retrograde amnesia: backward-acting amnesia, forgetting things learned before the surgery
• Anterograde amnesia: forward-acting amnesia, forgetting things learned after the surgery. Analyzed
for whether problem is in short-term memory (storing new information for brief periods of time while the
person attends to it) or long-term memory (storing new information once the person stops attending to
it), or both.
• H.M. had mild retrograde amnesia for events that occurred in the 2 years before his surgery, but could
remember well remote events like those of his childhood.
• H.M. had normal short-term anterograde memory; his digit span was 6 digits
• H.M. had severe long-term anterograde amnesia; became suspended in time on the day of his surgery
Formal assessment of H.M.’s anterograde amnesia: discovery of unconscious memories
• H.M. was given 7 tests to assess his long-term memory. The first 2 document his severe deficits in
long-term memory, whereas the last 5 indicate he was capable of storing long-term memories without
• Digit Span +1 Test: In this test of verbal long-term memory, H.M. was asked to repeat 5 digits that
were read to him at 1-second intervals. At the next trial, he was presented with the same sequence of 5
digits with 1 new digit added on the end, and so on. After 25 trials, H.M. could not move past the 8-digit
sequence; normal subjects can repeat about 15 digits.
• Block-Tapping Memory-Span Test: H.M. had global amnesia, amnesia for
information presented in all sensory modalities. This test demonstrated his
amnesia was not limited to verbal material: 9 blocks spread out on a board and
H.M. asked to watch the experimenter touch a sequence of them and then to
repeat the same sequence of touches. H.M. had a block-tapping span of 5 blocks,
which is normal, but he could never learn to correctly touch a sequence of 6
• Mirror-Drawing Test: Task was to draw a line within the boundaries of a star-
shaped target by watching his hand in a mirror, and the number of times he went outside the boundaries Page 2 of14
on each trial was recorded. His performance improved over the 3 days of 10 trials each, indicating
retention of the task; however, he could not recall ever having seen the task before.
• Rotary-Pursuit Test: Subject tries to keep the tip of a stylus in contact with a
target that rotates on a revolving turntable. H.M.’s performance improved
significantly over 9 daily practice sessions, despite his claim that he had never
seen the task before.
• Incomplete-Pictures Test: A non-sensorimotor test of memory with 5 sets of
fragmented drawings, each set containing drawings of the same 20 objects but
differing in their degree of fragmentation (set 1 most fragmented, set 5 complete
picture). The subject moves down the sets 1-5 until all 20 items have been
identified. H.M.’s performance improved when he retook the test 1 hour later, despite his inability to
• Pavlovian Conditioning: A tone was sounded just before a puff of air was administered to his eye, and
trials repeated until the tone alone elicits an eye blink. H.M. learned this conditioning task and retained
the conditioned response 2 years later, with no conscious recollection of the training.
3 major scientific contributions of H.M.’s case
• 1) Showed the medial temporal lobes play an especially important role in memory, challenging the
previous view that memory functions are diffusely and equivalently distributed throughout the brain.
This renewed efforts to relate individual brain structures to specific mnemonic processes.
• 2) Multiple memory systems: different modes of storage for short-term, long-term and remote memory
(memory for experiences in the distant past). H.M.’s specific problem appeared to be in memory
consolidation, the translation of short-term memories into long-term memories.
• 3) Learning without awareness: Two categories of long-term memories: explicit memories (conscious
memories) and implicit memories (improved test performance without conscious awareness).
Medial temporal lobe amnesia
• Preserved intellectual function, loss of ability to form explicit long-term memories while retaining ability
to form implicit long-term memories of the same experiences.
• Repetition priming tests: used to assess implicit memory, especially memory for words. Subjects are
first shown a list of words and not asked to learn or remember anything; later they are shown a series
of fragments of words from the original list and asked to complete them. Amnesic subjects perform as
well as control subjects even without the explicit memory of seeing the first list.
• Why are there two parallel memory systems? The implicit system likely evolved first since it is simpler,
but there is an advantage to having a second conscious system: flexibility of using their implicit
knowledge in a different way or a different context.
Semantic and episodic memories
• Semantic memories: explicit memories for general facts or
• Episodic memories (autobiographical memory): explicit
memories for particular events of one’s life.
• Patients with medial temporal lobe amnesia have particular
difficulty with episodic memories – cannot “time travel” into his
personal past or the future, cannot imagine what they’ll be doing
for the rest of the day, the week, or their life. However, they can
progress through mainstream schools and acquire reasonable
levels of language ability and factual knowledge (semantic
• This problem is difficult to spot, due to the experimenter’s
unfamiliarity with the patient’s life, and the patient becoming
effective at providing semantic answers to episodic questions. Page 3 of14
Effects of cerebral ischemia on the hippocampus and memory
• Cerebral ischemia: an interruption of blood supply to the brain. Can result in medial temporal lobe
• Brain damage was restricted largely to the pyramidal cell layer of a part of the hippocampus, the CA1
subfield. This suggests that hippocampal damage by itself can produce medial temporal lobe amnesia.
11.2 AMNESIA OF KORSAKOFF’S SYNDROME Hippocampus: CA1, CA2,
CA3, A4 subfields and
• Korsakoff’s Syndrome: a disorder of memory common in alcoholics, dentate gyrus. CA stands
attributable to brain damage associated with thiamine (vitamin B1) deficiency. for cornu ammonis,
another name for the
• Lesions in the medial diencephalon (medial thalamus and hypothalamus), hippocampus.
and diffuse damage to other brain structures including the neocortex,
hippocampus, and cerebellum.
• Specific damage to a pair of medial diencephalic nuclei, the mediodorsoal nuclei of the thalamus
(which project to the frontal cortex). However, likely other structures involved, such as the mammillary
• Korsakoff amnesia is a type of medial diencephalic amnesia; characterized by anterograde amnesia
for episodic memories in early stages of disorder, and progressive retrograde amnesia into the later
stages of the disorder, even into remote early childhood memories. Similar to H.M.’s medial temporal
lobe amnesia, but Korsakoff’s is progressive.
• In its advanced stages, is characterized by sensory and motor problems, extreme confusion,
personality changes, and a risk of death from liver, GI, or heart disorders.
11.3 AMNESIA OF ALZHEIMER’S DISEASE
• Alzheimer’s disease: a progressive disorder that starts with mild deterioration of memory but develops
into dementia so severe that the patient becomes incapable of simple activities, like eating, speaking,
recognizing a spouse, or bladder control. Characterized by anterograde episodic and semantic amnesia,
and progressive retrograde loss (starts with most recent info: in reverse direction in which memory was
gained); relatively intact implicit memory.
• Predementia Alzheimer’s patients: general anterograde and retrograde deficits in explicit memory, as
well as some deficits in short-term memory and some types of implicit memory – for verbal and
perceptual material, but not for sensorimotor implicit learning.
• Neuropathology is severe loss of brain tissue with enlarged ventricles and sulci, and pathological
markers of amyloid plaques and neurofibrillary tangles in the neurons. There is also diffuse brain damage
in many other areas, including the medial temporal lobe and the prefrontal
• Levels of ACh are greatly reduced due to degeneration of the basal
forebrain (nucleus basalis) and the pedunculopontine nucleus; this ACh
depletion is likely the cause of Alzheimer’s amnesia.
• Brain cholinergic systems in the rat (double-y maze): In first half, rat starts
off in one of the chambers on the left Y and obtains food in the middle of the
maze. Then the middle has the left door closed and the right door opens
(sometimes after a delay), and the rat must go to the chamber in the right Y that is not the one it went to
in the last trial (nonmatched-to-position). It needs 2 kinds of memory: reference memory for the first
half to reach the food (just has to recall semantic information of where the food is), and working
memory for the second half to choose the correct nonmatched chamber (declarative memory, has to
remember where it went last time). Changing delay time before rat proceeds to 2 half of the test and
seeing its success at remembering.
• Rats with lesions to the basal forebrain, and destruction of cholinergic neurons, do much worse in the
second half of the test because their working memory is destroyed – we see a delay effect as their
success rate drops to 50% (just guessing) as delay increases to 30s vs. open right away. Page 4 of14
11.4 AMNESIA AFTER
CONCUSSION: EVIDENCE FOR
• Blows to the head that do not penetrate the
skull can be severe enough to produce
concussions, a temporary disturbance of
consciousness, and can cause
posttraumatic amnesia (PTA).
• PTA usually involves permanent retrograde
amnesia for the events that led up to the
traumatic blow, and permanent anterograde
amnesia for many of the events that occur
during the subsequent period of confusion. During the period of confusion, the patient may seem lucid
because short-term memory is normal, but later will have no recollection. Risk of later developing
Alzheimer’s or Parkinson’s disease also increases significantly.
• Duration of period of confusion and anterograde amnesia is longer than the coma, which is longer than
the period of retrograde amnesia. More severe blows produce longer comas, longer periods of confusion
and of amnesia.
• Islands of memory: surviving memories for isolated events that occur during periods for which other
memories have been wiped out.
Gradients of retrograde amnesia and memory consolidation
• Gradients of retrograde amnesia after concussion provide evidence for memory consolidation:
concussions preferentially disrupt recent memories, so it suggests that the storage of older memories
have been strengthened, i.e. consolidated.
• Hebb’s Theory: Memories of experiences are stored short-term by neural activity reverberating
(circulating) in closed circuits, and are susceptible to disruption (like a blow). This eventually induces
structural changes of involved synapses that provide stable long-term storage.
• Electroconvulsive shock (ECS) is an intense, brief, diffuse, seizure-inducing current administered to
the brain through large electrodes. They are used to study memory consolidation: would disrupt neural
activity an erase from storage only those memories that have not yet been converted to long-term
structural storage. The length of period of retrograde amnesia produced would provide estimate of
amount of time needed for consolidation.
• In one study, thirsty rats had a water spout placed in the niche of their cage and are allowed to drink.
Then, 10 seconds, 1 minute, 10 minutes, 1 hour or 3
hours later, each rat receives one ECS. The retention
of each rat’s memory of the water spout was assessed
based on how many times the rat explores the niche,
indicating they remember the water spout’s location.
• Control rats explored 10 times, indicating they
remembered the discovery of water the previous day;
rats with ECS 1 hour and 3 hours after learning also
did this. However, rats with ECS 10 seconds, 1 minute
or 10 minutes after the learning trial explored the niche
significantly less. The longer the rat had to consolidate
the memory (amount of time before ECS
administration), the less the memory is affected by
ECS. Consolidation of memory of learning trial took
between 10 minutes and 1 hour. Page 5 of 14
• Other studies observed long gradients of ECS-
produced retrograde amnesia: patients measured for
memory of television shows that had played for only
one season in different years prior to the ECS
treatment. The difference between the before-and-
after-treatment scores serve as estimate of memory
loss for the events of each year. Results showed
disruption of memories of TV shows in the last 3 years,
but not in prior years, due to having less time to
consolidated, and thus being more disrupted by ECS.
• This is incompatible with Hebb’s theory of
consolidation: gradients covering days, weeks or years
not easily accounted for by the disruption of
reverberatory neural activity. Memory consolidation
can continue for a very long time after learning,
Hippocampus and consolidation
• H.M.’s bilateral medial temporal lobectomy disrupted only the retrograde memories acquired just before
his surgery. Standard consolidation theory: memories are temporarily stored in the hippocampus until
they can be transferred to a more stable cortical storage system. Medial temporal lobe lesions in animals
produce temporally graded retrograde amnesia.
• Multiple-trace theory (Nadel & Moscovitch): Hippocampus and other structures involved in memory
storage store memories for as long as they exist, not just in the period immediately after learning. When
a conscious experience occurs, it is rapidly and sparsely coded in a distributed fashion throughout the
hippocampus and other structures.
• Retained memories become progressively more resistant to disruption because each time a similar
experience occurs or the original memory is recalled, a new engram (change in brain that stores a
memory) is established and linked to the original engram, making the memory easier to recall and the
original engram more difficult to disrupt. This is compatible with the finding of long gradients of retrograde
• Reconsolidation: Each time a memory is retrieved from long-term storage, it is temporarily held in
labile (changeable, unstable) short-term memory, where it is once again susceptible to PTA until it is
• Nader, Schafe, and LeDoux (2000): protein-synthesis inhibitor anisomycin injected into the amygdala of
rats that had just been trained for fear conditioning. The infusion produced retrograde amnesia for the
fear conditioning, even though the conditioning trial had occurred days before.
• Only certain kinds of memories may be susceptible to reconsolidation; more research involves fear
11.5 NEUROANATOMY OF OBJECT-RECOGNITION MEMORY Page 6 of14
• To study the neural bases of amnesia, we must use controlled
experiments with precise lesions in various structures, to control
what and when the subjects learn, and how and when their
retention is tested.
• Initial difficulty in developing animal model of medial temporal
lobe amnesia: 1) animal tests did not account for specificity of
damage to explicit long-term memories; 2) it was incorrectly
assumed that amnesic effects were largely attributable to
hippocampal damage (and so lesions focused on the
Monkey model of object-recognition amnesia: delayed
• Gaffan, Mishkin and Delacour: showed that monkeys with
bilateral medial temporal lobectomies have major problems
forming long-term memories for objects encountered in the
delayed nonmatching-to-sample test. A monkey is presented
with a distinctive object (the sample) under which it finds food;
then, after a delay, the monkey is presented with 2 test objects,
the sample and an unfamiliar object. The monkey must remember
the sample object so that it can select the unfamiliar object to
obtain food concealed beneath it.
• Normal, well-trained monkeys perform correctly on 90% of trials
when the retention intervals were a few minutes long; monkeys
with BMTL lesions
had major deficits.
was normal at short
delays but fell close
to 50% (choosing at
random) after delays
of several minutes;
due to working
memory deficit of
• Temporal lobe
hippocampus, amygdala, rhinal cortex.
Delayed nonmatching-to-sample test for rats
• Rat model is valuable: making hippocampal lesions in monkeys
involve aspiration (suction) of large portions of rhinal cortex in addition
to the hippocampus, leading to a lot of extraneous damage. In rats, the Mumby box: the
size and location of the hippocampus improves ability to limit apparatus used to
extraneous damage to a small area of parietal neocortex; lesions can also be test nonmatching-to-
done electrolytically or with neurotoxin injections.
sample tests for rats.
Similar to double-Y
maze, also testing
Similar results to test
for monkeys: more
errors made as delay
time increases. Page 7 of14
Neuroanatomical basis of the object-recognition deficits resulting
from medial temporal lobectomy
• Removal of the rhinal cortex consistently produces severe and
permanent deficits in performance on delayed nonmatching-to-
sample test and other tests of object recognition. In contrast, bilateral
removal of hippocampus produces either moderate deficits or no
deficits, and bilateral destruction of amygdala has no effect.
• For R.B.’s amnesia case of cerebral ischemia leading to damage of
pyramidal cells in his CA1 hippocampal subfield has been replicated
in monkeys and rats. They showed severe deficits in delayed
• How can ischemia-produced hippocampal damage of one small area be associated with severe deficits
in performance on the delayed nonmatching-to-sample test, when deficits associated with total removal
of the hippocampus are only slight?
1. Ischemia-produced hyperactivity of CA1 pyramidal cells damage neurons outside the hippocampus
(like the enterorhinal) due to excessive release of excitatory NTs: glutamate neurotoxicity –
2. This e