Presentazione di PowerPoint

1
Outline

• Definitions
• Abnormal memory amnesics and mnemonists
• How do we forget?
• Memory stores
• Memory tasks
• Working memory
• Levels of processing
• Semantic and episodic memory
• Cognitive neurosciences notes
• Information encoding
• Oblivion
• Consolidation
• Retrieval
• Constructive theories

2
Naïve definitions

• Main Entry memory
• the power or process of reproducing or recalling
  what has been learned and retained especially
  through nonconscious associative mechanisms
  conscious or unconscious evocation of things past
  
• the process of reproducing or recalling what has
  been learned as manifested in some special way or
  as associated with some bodily process visual
  memory muscular memory
• persistent modification of structure or of
  behavior resulting from an organism's activity or
  from its passively acquired experience
• the totality of what has been learned and
  retained especially as evidenced by recall and
  recognition drew on his memory to supply the
  needed names even birds and animals have an
  ancestral memory
• the function of memory regarded as a compartment
  or chamber in which images, perceptions, or
  learning are stored filling their memory with a
  lumber of words
• a particular act of recalling something learned
  or experienced the fact or a condition of
  recalling REMEMBRANCE, RECOLLECTION, RECALL
• an image, impression, or other mental trace of
  someone or something known or experienced the
  content of something remembered
• a component in an electronic computing machine
  (as a computer) in which information (as data or
  program instructions) may be inserted and stored
  and from which it may be extracted when wanted a
  device external to a computer for the insertion,
  storage, and extraction of information
• a capacity for showing effects recognized as the
  result of past treatment used especially of
  materials the wire begins to turn in the other
  direction corresponding to the first twisting
  a capacity for returning to a former condition ..

(adapted from Merriam-Webster Unabridged
Dictionary)
3
Experts definitions

• Memory is the means by which we retain and draw
  our past experiences to use this information in
  the present (Tulving)
• As a process, memory refers to the dynamic
  mechanisms associated with retaining and
  retrieving information about past experience
  (Crowder, 1976)

4
Abnormal memories

• The study of individuals with abnormal memories
  is fundamental for the scientific understanding
  of memory.
• Individuals with deficitary memory are defined
  amnesic. Individuals with exceptional memory are
  defined mnemonists.

Amnesia

• The most famous case of amnesia is H.M., who
  had neurosurgery with removal of epileptic foci
  at age 29.
• The operation caused a specific pattern H.M.
  could not recall any event occurred after the
  event, whereas the preceding events were left
  intact.
• This form of amnesia in which the memory of
  events following the cause of the amnesia is
  selectively compromised is called anterograde
  amnesia.
• This impairment leaves the patient in un
  permanent present, with a very short conciousness
  window.
• The functional damage seems to be due to an
  interruption in the flowing of information from
  the Short- to the Long-Term Memory.

5
Amnesia

• Another patient, reported by Russel e Nathan
  (1946) underwent to a contusion that caused a
  selective loss of a slice of memories from 1922
  to the traumatic event. Previous memories were
  preserved. Moreover, the functional recovery,
  which eventually was complete, advanced from the
  oldest memories to those closer to the trauma.
• This and other similar phenomena led to thinking
  that LTM strengthen on the base of a temporal
  gradient. Older memories would have longer time
  for consolidation than newer ones, resulting
  stronger.
• This property of memory can be observed in the
  tendency of eldery people of talking more of
  their remote experiences than of newer ones.

6
The mnemonists

• The amount of information that is learned and
  can be recalled is highly variable among
  individuals. If we count how many telephone
  numbers we know by heart, we would reveal
  enormous differences.
• Although memory is normally variable, there are
  exceptional cases in which the capacity of
  learning and recalling is enormously higher than
  for normal subjects.
• Mr Shereshevsky went to Dr. Luria office, in
  Russia, asking to be tested for his memory. The
  result was sensational
• Shereshevsky could recall lists of numbers that
  he had memorized decades earlier, and was
  actually unable to forget the lists he had
  memorized while perfoming as a mnemonist. He
  could memorize nonsense syllables, a challenge
  specifically designed to thwart mnemonic
  associations. S experienced synesthesia,
  responding to stimulation of one sense with a
  perception in one or more different senses. For
  example, he could see sounds and feel their taste
  and texture. His remarkable abilities were
  somewhat disabling. He was not able to read
  poetry or fiction easily, as each word or phrase
  would blossom into an intense visualization that
  might be contradicted by the next one...
  Shereshevsky's pathological memory interfered
  with his ability to hold a regular job, enjoy
  literature, or even seemingly to think in the
  abstract without being distracted by sensory
  association.".

7
The mnemonists

• The secret of mnemonists seems to lay in their
  synesthesia, a phenomenon for which a sensation
  produces a secondary subjective sensation. This
  allowed Mr S to form complex cross-modal
  associations between words and mental images.
• Other mnemonists base their exceptional skills
  upon different, more symbolic cues, like
  associating a list of numbers with dates or sport
  events.
• Independently on the mental representation,
  based on symbols or on mental images, it seems
  that the base for the mnemonists skills is a
  supra-normal ability of forming associations of
  some kind.

8
General Intro

• Memory is a fundamental cognitive skill for the
  cognitive and psychological integrity of
  individuals.
• It has different manifestations, and its basic
  processes can be either automatic and effortless
  or require active effort.
• It is currently thought the memory is diveded in
  sub-abilities each based on different,
  independent systems.
• The cognitive approach to memory makes the
  fundamental division into three basic processes
  encoding, storage and retrieval.
• Encoding has, in turn, two stages acquisition e
  consolidation. Aquisition puts the incoming data
  in a sensory buffer. Consolidation creates a
  stable representation over time.
• Storage, which is the result of acquisition and
  consolidation, creates and holds data in
  permanent records.
• Retrieval uses stored information to create a
  conscious representation or to execute some
  active behavior previously learned, like a motor
  act.

9
Memory stores

• The complexity of human memory is such that it
  requires different levels of categorization.
• The main categorization divides the memory stores
  according to the duration of the trace by
  postulating three stores Sensory Memory (SM),
  Short Term Memory (STM) and Long Term Memory
  (LTM).
• SM (also called iconic, echoic, etc. according
  to the specific sensory modality) is active from
  the sensory event until few seconds from it.
• STM, or immediat memory, holds material within a
  period from few seconds until some minutes after
  its occurence.
• LTM is measured in terms of minutes, hours and
  days, although it can be life lasting.

10
Memory stores

• The three stores were formalized firstly by
  Atkinson e Shiffrin (1968).
• They proposed the so-called modal model of
  memory, in which the transfer of information from
  SM to STM is based upon attentional selection,
  while that from STM to LTM is based upon reharsal.

11
Sensory Memory

• The classical task to probe SM consists in the
  rapid presentation of letter matrices. Subjects
  are asked to recall the letters presented at
  various intervals.
• The matrix is displayed very briefly (50 ms) to
  avoid scanning eye movements.

H Q Z P K R Z E M B A F

• From this study the average span of iconic
  memory ranged around 4 elements.
• However, subjects often reported they saw
  clearly all the letter presented, but the recall
  of the first letters would somehow mask that of
  the following letters.

12
Sensory Memory

• G. Sperling (1960) modified this task asking the
  observers a partial recall.
• In this version, the matrix was presented
  entirely, but the subjects were requested to
  recall only one line signaled after the matrix
  disappearence. The signal of the line to be
  reported was presented at different times, -0.1
  to 1 second from the matrix disappearence. This
  does eliminate the interference in recalling a
  long list of letters.

S Q W P H S L E W R C O
Signal at various intervals

• Since subjects could not predict what line had
  to be reported, the actual span corresponded to
  the number of letters reported multiplied by the
  number of lines (3).

13
Sensory Memory

• Using this technique the measured span grew from
  4 to about 9 items when the cue was synchronous
  or appeared immediately after the matrix. For
  longer lags it decayed at 4-5 items.

• This study has shown the existance of a sensory
  buffer of very short duration (lt 1 second).

14
Short Term Memory

• Write down the name of the scientist who
  discovered the properties of SM

George Sperling
15
Short Term Memory

• While SM is difficult to experience consciously
  (information is seldom available for only 50 ms,
  and often new information overlaps to impede its
  recall), the properties of STM are often evident
  and subjects to continuous -negative-
  verifications.
• In a famous study by G. Miller (1956) it was
  suggested that the amount of items that could be
  held in STM was 72.
• The problem arises when a definition of item has
  to be drawn in fact, if we chunk a list of 21
  numbers in groups of three items, we will
  probably recall the whole list without much
  troubles!
• This is indeed a mnemonic (i.e. a memory
  strategy) known as chunking.

16
Long Term Memory

• LTM constitutes the deposit of informations,
  notions, happenings that form the base of our
  conscious knowledge throughout life and that are
  used for contingent behaviors.
• On these bases, LTM seems to be unlimited and
  permanent, making the task of defining its last
  and span virtually impossible.
• While he was mapping the somatosensory cortex by
  stimulating the brain of epileptic patients,
  Penfield (1969) noticed the side effect that
  subjects recalled events of their first childhood
  that had been apparently lost. This was
  interpreted as a sign of the fact that LTMs are
  permanent.
• LTM has been divided in two sub-types (Tulving,
  1990) esplicit an implicit memory. To express a
  similar concept, Squire (1987) used the terms
  declarative and non declarative memory.
• Esplicit, or declarative memory is the conscious
  content of our long-term knowledge. Its our base
  of knowledge.
• Implicit, or non-declarative memory is all that
  is accessed in an unaware fashion, like the
  procedural knowledges (riding a bycicle), the
  priming and all those behaviors derived from
  conditioning, abituation or sensitization.

17
Memory tasks

• Respecting the complexity of its object, there
  are a variety of tasks to study the properties of
  memory.
• A first category distinguishes between recall
  and recognition tasks.
• In the first case its asked to recover
  information from memory, in the second to select
  or identify some elements previously learned.
• To mention from the recall tasks are
• serial recall tasks, where items must be
  recalled in the exact order they were learned,
• free recall tasks, without any order constraint,
  and
• facilitated recall tasks, in which items are
  learned in couples, but only one member of the
  couple is presented during the test and the
  paired element must be recalled.
• Among the recognition tasks are the familiar
  true-false tests and multiple choice tests.
• Usually recognition is better than recall.

18
Memory tasks

• A second distinction is for implicit vs esplicit
  memory tests.
• Recall and recognition tests are directed to the
  esplicit memory, in that they imply conscious
  recovery of information.
• In the implicit memory tests is assessed to what
  extent knowledge or facts to which we got exposed
  influence the outcome of target tests or
  behaviors.
• A typical implicit memory test is word
  completion ca _ _ _ _ _ _
  fa _ _ _ _ _ _
• The information preceding the test can be
  manipulated in controlled lab conditions. Under
  this optimal setting, the words used most
  frequently to complete are those seen in the
  implicit learning session. This occurs even to
  amnesic patients. This phenomenon is known as
  priming.
• Category, familiar

19
Memory tasks

• A third category is aimed to test procedural
  memory, that is suited for tasks such as motor
  behaviors.
• Among the procedural memory tests we can
  mention
• solving puzzles
• reading mirror writing
• walking with prisms
• In all these tests, the amnesic skills are
  tipically well preserved.

20
Working memory

• The main alternative to the modal model
  (Atkinson and Shiffrin) of memory, which
  distinguishes 3 main temporal stages, is one that
  attributes a key role to the working memory.
• Working memory is similar to the STM, but with
  an active role for the information processing.
• Rather than a simple container of information
  for short time, working memory would have an
  active role in
• selecting and moving material from/to the LTM
  module
• selecting and integrating sensory data
• operating a meta-integration of all the
  information needed for the behavior in progress
• The temporal span and the capacity of this
  module are both limited.

21
Levels Of Processing

• A completely different approach is that of
  Levels Of Processing (LOP) by Craik and Lockart
  (1972).
• Rather than relying on discrete stores, memory
  would work in a continuum based on the depth of
  encoding.
• The LOPs would then be infinite as the point in
  a line.
• The deeper the LOP, the likelier the encoding.
• Generally, tre levels are distinguished
  physical, acustical and semantic.
• The key experiments for the theory asked
  specific questions on words presented to the
  subjects. The questions were formulated to
  activate selectively one of the three levels.

TABLE is it capital or not? PHYSICAL
LEVEL Nature does it rhyme with
future? ACOUSTIC LEVEL sunflower is it a
plant? SEMANTIC LEVEL
22
Levels Of Processing

• The results showed that the recall improved as a
  function of the level.
• Further support to the LOP came from the studies
  of Zinchenko (1962, 1981), who showed that the
  retrieval of semantically related words (dog
  animal) was better than for words that were
  concretely connected (dog leg), and the recall
  for these was in turn better than for unrelated
  words.
• However, many critics were brought to the LOP
  approach
• the definitions of the levels is loopy the depth
  of encoding is both the cause and the effect of
  memorization
• in some cases the recall based on rhymes is
  better than for semantic contents.
• The solution of this seem to lie in the
  relationship between the encoding of the material
  when it was learned and the level of recall
  required by the task.
• In general, it seems that the use of different
  levels when recalling increases the chance of
  successful retrieval.

23
A synthetic view

• Baddeley conciliates the two models by extending
  the LOP approach

visuo-spatial block Note stores mental
images for short times
phonologic- articulatory loop stores
briefly subvocalizzations for the verbal
comprehension and the acoustic reharsal
central executive coordinates attentional
activities and drives the responses
supplementary perceptive and cognitive systems
level of acoustic processing
the STM brain it manages its content by
moving it in and out the store
level of physical processing
24
Semantic and episodic memory

• The patients of Penfield, after stimulation
  recalled particular facts or episodes, but never
  semantic concepts linked to specific events.
• Tulving proposed a dissociation between a memory
  system devoted to specific episodes in a temporal
  location (e.g. what you had for lunch two days
  ago) and one storing semantic information,
  devoted to meanings and notions, without a
  precise temporal location (e.g. who studied SM,
  what is a mnemonist, etc.)
• A number of neuropsychological evidence support
  this distinction. For example, lesions to the
  temporal lobe impair the faculty of recalling the
  when of particular events, but not their
  recognition.

25
Semantic memory concepts and schemas

• Semantic memory would be based on concepts, that
  is ideas to which features are attributed and
  other ideas can be associated in networks. For
  example, a car has a cc, a color, a price, etc..
  Also, the idea of a car is easily associated with
  that of the highway, wheels, pollution, etc..
• The schemas would be organizing structures for
  concepts based upon experience and culture.
• Collins e Quillian proposed that semantic memory
  had a hierarchical structure.
• The farther the two terms in the network, the
  slower the responses. For example, by changing,
  in the second question, mammal with animal,
  reaction times get slower.

C. Q. task (inclusion in classes, true-false)
Is the chair a piece of furniture?Is the tree a
part of a car? Is the cat is a mammal? Is the
square a geometric figure?
26
Semantic memory concepts and schemas

• In the alternative model by Smith (1974) the
  semantic storage is based on feature matching
  rather than semantic hierarchies.
• However, as both models could not explain all
  data, the new approach of sematic networks was
  proposed.
• Connectionist models (Parallel Distributed
  Processing), explain the semantic memory in terms
  of semantic networks.
• According to this approach, working memory would
  be the part of the network working in a given
  moment. A node that activates another is called
  prime, its effect priming.

mammal
feline
primate
chimpanzee
macaque
27
Cognitive neurosciences of memory

• The first attempts to localize memory in the
  brain were quite unsuccessful.
• Indeed, the attempts to localize specific
  memories, called engrams by Lashley brought to
  the conclusion that memories do not have a
  specific location.
• However, there seems to be a correlation between
  brain structures and general memory functions.
• In particular, the finding of double
  dissociations has given fundamental insights. For
  example, lesion to the left parietal cortex
  compromise STM leaving LTM unaffected, while
  medial temporal lesions compromise LTM while
  leaving STM unaffected.
• Double dissociations were found also between
  declarative and non-declarative memory.

28
Cognitive neurosciences of memory

• It seems now that different structures have
  different roles for memory.
• The cortex seems to be involved in the LT storage
  of sensory information.
• The hyppocampus has a key role for the
  declarative knowledge, complex learning and the
  LTM consolidation (perhaps by linking material
  stored in spared areas of the brain).
• The basal ganglia are involved in the implicit
  memory, but not in the priming.
• The cerebellum has to do with the classically
  conditioned responses.
• At the level of single neurons, a repetitive
  stimulation of a neural path increases the
  likelyhood of firing of the post-synaptic
  neurons, effectively decreasing their threshold.
• About the neurotrasmitters, serotonine,
  acetylcoline and noradrenaline increase the
  memory-related neural transmission.
• In the normal hyppocampus there is a high
  concentration of acetylcoline, which is
  substantially reduced in Alzheimer patients.
• In alcoholics, the serotonin level is seriously
  altered and is correlated with diencephalic
  damages. A typical deficit in these patients is
  the Korsakoff syndrome, that involves a severe
  form of amnesia anterograde.

29
Memory processes

• We are now going to focus onto memory processes
  encoding, storage and retrieval.
• These three processes are assumed to be
  sequential in order to start every next stage a
  minimum output from the previous stage needs to
  be provided.
• Encoding has to do with the language the system
  uses to understand the outer world (e.g., binary,
  analogic, etc.).
• Storage has to do with the properties of the
  memory stores.
• Retrieval has to do with the mechanism by which
  specific stored information is selected and used
  for contingent behavior.

30
Read in silence this piece
The procedure is actually quite simple. First you
arrange items into different groups. Of course
one pile may be sufficient depending on how much
there is to do. If you have to go somewhere else
due to lack of facilities that is the next step
otherwise, you are pretty well set. It is
important not to overdo things. That is, it is
better to do too few things at once than too
many. In the short run this may not seem
important but complications can easily arise. A
mistake can be expensive as well. At first, the
whole procedure will seem complicated. Soon,
however, it will become just another facet of
life. It is difficult to foresee any end to the
necessity for this task in the immediate future,
but then, one can never tell. After the procedure
is completed one arranges the materials into
different groups again. Then they can be put into
their appropriate places. Eventually they will be
used once more and the whole cycle will then have
to be repeated. However, that is part of life.
31
this was the original piece presented by
Bransford and Johnson (1972) to the subjects of
their experiment can you recall the key
parts What is it about??? would you agree it
was about washing clothes??? !!!!!
32
Information encoding

• Conrad (1964) did an experiment aimed to
  understand the code for brief, temporary storing
  (like retaining a telephone number before to
  store it in our cell phone).
• A sequence of 6 letters were presented very
  briefly and subjects were asked to report them as
  they saw them immediately after the stimuli.
  Regardless of whether the letters were presented
  visually or acustically, the errors were always
  of phonetic nature (letters with similar sounds
  were confused). Conrad concluded that STM had an
  acustic code.
• Baddeley (1966) found similar acoustic code for
  the semantic memory. He noticed that the recall
  of acustically similar words (i.e. rhymes) is
  much worse than that of acustically uncorrelated
  words, while there is no such difference for
  words semantically related/unrelated.
• However, the use of semantic and visual codes was
  found in later studies by Shulman (1970) and
  Posner (1969).

33
Information encoding

• Whereas STM uses mainly sensory codes, there are
  many converging data that the preferential code
  for LTM is semantic.
• A fundamental study was made by Grossman and
  Eagle (1970).
• Analyzing the false alarms with the 18
  distractors of the control list, the errors with
  sematically related distractors were twice as
  frequent than with unrelated distractors.

Learning list
Test list with distractors
1) house2) clock3) spaghetti4) cup 41)
gondola
1) clock2) train3) cup4) italy 41) boat
Unrelated distractor
5 waiting
Semantically related distractors
34
How soon do we forget?

• A classical experiment by Petersen e Petersen
  (1959) has shown the time by which information
  not reiterated is forgotten.

ISTRUCTIONS

• A string of 3 consonants will be displayed for 1
  second, followed by a number. As soon as you will
  se the number you will have to count backwards
  (with no recall pauses!!!) subtracting 3 numbers
  as fast as you can (e.g. if 17 is displayed, you
  will count 14, 11, 8, 5) while keeping your eye
  on the display. At a certain point a red spot
  will signal to stop. At the stop signal write
  down the string as you recall it. We will do it
  for 7 trials.

gtk
rtm
zxt
klf
qkl
tvm
zps
74
53
72
47
59
101
96
35
How soon do we forget?

• Class data plotted

36
How soon do we forget?

• This experiment is perfect to show how well
  laboratory investigations can investigate the
  properties of memory.
• Once the stimulus set is defined, the lag
  between learning and recall is systematically
  varied and performance is measured.
• In the original P. P. experiment there was a
  drop of accuracy up to about 10 for lags of 18
  seconds.
• The data were interpreted in terms of
  interference new learnings interfere with old
  information by obscuring it.
• An alternative account is that information
  undergoes to a natural decay independently of
  interference.
• The data of the P. P. experiments were taken
  to suggest that counting backwards might have
  interfered with the recall.

37
The oblivion interference or decay?

• Interference and trace decay qre the two main
  explanations of the reasons of forgetting.
• The study of Brown e Peterson was taken to
  support the interference model.
• There are two types of interference proactive
  and retroactive.
• The first is an effect of what happens after the
  material is presented, as in the B P
  experiment, the second occurs before the target
  information is presented.

38
Read these words, presented for 1 each on the
left side of the screen , then write down
all you can recall
telephone cloud book tree shirt cat bulb chalk cha
ir flower clock bat carpet soup pillow
39
Serial Position effect
40
The oblivion interference or decay?

• The serial presentation curve can be explained by
  the interference model the first elements of the
  list endure only the proactive interference, the
  last ones only the retroactive interference, the
  intermediate both types.
• According to the decay account, the trace would
  fade out naturally, without interferences, unless
  it is not consolidated.
• Testing the decay hypothesis is not as easy as
  testing the interference hypothesis, in that its
  not easy to exclude interferences and reharsal.

41
Trasfer to LTM and consolidation

• How do we face oblivion and transfer data in the
  LTM register?
• Consolidation, within this context, is the
  process of integration, by association, of new
  information with previously existing information.
  This process would be the base of the transfer of
  information.
• Another basic mechanism is the attentive
  selection of information attention would
  increase the chance of consolidation.
• During the process of consolidation, information
  would be deformed and degraded. By metamemory are
  meant the strategies used to help and improve our
  memories and their consolidation.
• The most widely used strategy is reharsal, that
  can be overt or covert.
• This practice effect depends mainly on the total
  time employed, even though studies show that
  distributed sessions are more effective than only
  one long session. Moreover, distributed practice
  is more effective when the sessions take places
  in different contexts.

42
Organizing information

• The stored information is organized in different
  ways.
• Specific strategies tho organize memories are
  called mnemonics.
• There are different mnemonics. The most common
  ones block notes and diaries.
• The use of mnemonics is of particular importance
  for the perspective memory. In this case we use
  external,physical cues, or constraints that act
  as reminders by hindering the continuation of the
  activity unless one actively notice them.

43
Retrieval

• Saul Sternberg (1966) did a series of experiments
  of memory scanning to study the retrieval from
  STM.
• Subjects were presented with lists of numbers of
  varying length and, after a short pause, they
  were asked whether a test number was present in
  the list or not.
• They measured RTs of individual subjects

44
Retrieval

• If the mechanisms of recovery were parallel,
  then RTs should not depend on the length of
  the list (RTslist1RTslist2).
• Viceversa if the mechanisms were serial
  (RTslist1gtRTslist2).
• Assuming a serial process, if the matching of the
  test with the items of the list was exaustive
  (including all the items), RTs should not depend
  on the serial position of the item in the list.
• Viceversa if the problem was self-terminating.

AGTKDORSYW
SLKRDZ
AGTKDORSYW
SLKRDZQZAL
AGTKDORSYW
SLKRDZQZAL
45
NYU Florence spring 2007 results
46
Retrieval

• The results point clearly to a serial exaustive
  mechanism

serial
exaustive