Chapter 6: Memory - Ashton Southard

muscleblouseΤεχνίτη Νοημοσύνη και Ρομποτική

19 Οκτ 2013 (πριν από 3 χρόνια και 7 μήνες)

70 εμφανίσεις

Chapter 6

Memory

Three Processes of Memory


Memory



an active system that receives information from
the senses, puts that information into a usable form, and
organizes it as it stores it away, and then retrieves the
information from storage


Although there are several different models of how memory
works, all of them involve the same 3 processes


Getting the information into the memory system


Storing the information


Retrieving the information

Encoding: Putting Information In


The first process in the memory system is to get sensory information (ex.
Sight, sound, etc.) into a form that the brain can use


Encoding



the set of mental operations that people perform on sensory
information to convert that information into a form that is usable in the
brain’s storage systems


Ex. When people hear a sound, their ears turn the vibrations in the air into
neural messages from the auditory nerve (transduction), which make it possible
for the brain to interpret that sound


Encoding is not limited to turning sensory information into signals for the
brain


It is accomplished differently in each of the three different storage systems of
memory


In one system, encoding may involve rehearsing information over and over to
keep it in memory, whereas in another system, encoding involves elaborating on
the meaning of the information


Storage: Keeping Information In


The next step in memory is to hold on to the information for
some period of time


Storage



holding onto information for some period of time


The period of time will actually be of different lengths,
depending on the system of memory being used


Ex. In one system of memory, people hold on to information
just long enough to work with it, about 20 seconds or so; but in
another system, people hole on to information more or less
permanently

Retrieval: Getting Information Out


The biggest problem many people have is retrieval


Retrieval


getting information that is in storage into a
form that can be used


Ex. You have taken an essay exam, later after the test is over,
you then remember several other things you could have said

Models of Memory: Information
-
Processing


Many researchers feel the
information
-
processing model
of memory is the most comprehensive and has been the most
influential


Assumes the processing of information for memory storage is
similar to the way a computer processes memory in a series of
three stages


Focuses on the way information is handled, or processed,
through three different systems


The processes of encoding, storage, and retrieval are seen as part
of this model

Models of Memory: Parallel Distributed
Processing (PDP) Model


It is common to refer to the three memory systems of the
information
-
processing model as stages, but that implies a sequence of
events


While many aspects of memory may follow a series of steps or stages,
there are those who see memory as a simultaneous process, with the
creation and storage f memories taking place across a series of mental
networks “stretched” across the brain


Parallel distributed processing (PDP) model


memory
processes are proposed to take place at the same time over a large
network of neural connections


Derived from work in the development of artificial intelligence (AI)


In the AI world, PDP is related to
connectionism

(the use of artificial neural
networks to explain the mental abilities of humans)

Models of Memory: Levels
-
of
-
Processing Model


The information
-
processing model assumes that the length of time that a memory
will be remembered depends on the stage of memory in which it is stored


Levels
-
of
-
processing model


assumes information that is more “deeply
processed,” r processed according to its meaning rather than just the sound or
physical characteristic of the word or words, will be remembered more efficiently
and for a longer period of time


Ex. If the word
BALL

is flashed on a screen and people are asked to report whether
the word was in capital letters or lowercase, the word itself does not have to be
processed very much


only its visual characteristics need to enter into conscious
attention


But, if those people were asked to use that word in a sentence, they would have to
think about what a ball is and how it can be used


They would have to process its meaning, which requires more mental effort than processing just
the word’s “looks”


Numerous experiments have shown that thinking about the meaning of something is
a deeper level of processing and results in longer retention of the word


Which Model Explains Which Aspects
of Memory


All of three models can be used to explain some, if not all, research findings
about memory


Information
-
processing model


Provides a “big picture” view of how the various memory systems relate to each
other


how the “memory machine” works


PDP model


Is less about the mechanics of memory and more about the connections and
timing of memory process


Levels
-
of
-
processing model


The depth to which information is processed addresses the strength of those
parallel connections within each of the three memory systems, with strength and
duration of the memory increasing as the level of processing deepens


Although the information
-
processing model takes center stage for
explaining memory for now, it is important to remember the concepts of
the levels at which information is processed and the way that those processes
may take place


Information
-
Processing Model:

3 Memory Systems


Information
-
processing theory bases its model for human
though on the way a computer functions


Data are encoded in a manner that the computer can
understand and use


The computer stores that information on a disk, hard drive, or
memory stick


Then, the data are retrieved out of storage as needed


Information
-
processing theorists were the first to propose
that there are three stages or types of memory systems


Sensory memory, short
-
term memory, and long
-
term memory

Sensory Memory


Sensory memory


the first stage of memory, the point at which
information enters the nervous system through the sensory systems


Information is encoded into sensory memory as neural messages in the
nervous system


As long as those neural messages are traveling through the system, it can be said
that people have a “memory” for that information that can be accessed if needed


Example: the “double take”


Imagine you are driving down the street, looking at the people and cars on either
side of your car. All of a sudden you think “Wait a minute, was that guy not
wearing pants?!” then you look back to check.


How did you know to look back? Your eyes had already moved past the possible
pants
-
less guy, but some part of your brain must have just processed what you
saw.


This can only be explained by the presence, however brief, of a memory for what
you saw


There are 2 kinds of sensory memory that have been studied extensively


Iconic (visual) and echoic (hearing) sensory memories


Iconic Sensory Memory: Capacity


Iconic sensory memory


visual sensory memory, lasting only a fraction of a second


George
Sperling

(1960)


Partial report method
: presented subjects with a grid of letters and sounded either a
high, medium, or low tone after the grid of letters was taken away, which signaled
which row subjects were to report


The tone was sounded after the grid was taken away to that subjects couldn’t just
memorize one row of letters










Found that subjects could accurately report any of the three rows, meaning that they
entire grid was in iconic memory and available to the subjects


Thus, the capacity of iconic memory is everything that can be seen at one time


Rows of Letters

Tone Signaling

Which Row
to Report

LHTY

High

tone

EPNR

Medium tone

SBAX

Low tone

Iconic Sensory Memory: Duration


Sperling

also found that if he delayed the tone for 1 second, subjects could
no longer accurately report letters from the grid


The iconic information had completely faded out of memory in that brief time


In real life, information that has just entered iconic memory will be pushed
out very quickly by new information, a process called
masking


Research suggests that after only a quarter of a second, old information is
replaced by new information


Eidetic imagery


the ability to access a visual memory for 30 seconds or
more


Very rare condition, often called
photographic memory
, causes is unknown


People with eidetic imagery ability might be able to look quickly at a page in a
book, then by focusing on blank wall or piece of paper, “read” the words from
the images that still lingers in their sensory memory


More common in children and tends to diminish by adolescence or young
adulthood


Iconic Memory: Function


Iconic memory actually serves a very important function in
the visual system


Remember that the eyes make tiny movements called
saccades

that keep vision from adapting to a constant visual stimulus,
so that what is stared at steadily doesn’t slowly disappear


Iconic memory helps the visual system to view surroundings
as continuous and stable in spite of the saccadic movements


It also allows enough time for the brain stem to decide if the
information is important enough to be brought into
consciousness


Like the possibly pants
-
less guy from the example earlier

Echoic Sensory Memory


Echoic memory


the brief memory of something a person has just heard


Ex. The “What?” phenomenon


When you are reading, watching TV, or concentrating on something and someone
walks up to you and asks you if your ready for dinner. You sit there for a second or
2, and then say “What? Oh … yes I’m ready to eat now”


You didn’t really process the statement from the other person as he or she
said it, you heard it, but your brain didn’t interpret it immediately.


Instead, it took several seconds for you to realize that


1) something was said


2) it may have been important


3) you’d better try and remember what it was


If you realize all this within about 4 seconds (the duration of echoic
memory), you will likely be able to “hear” an echo of the statement in your
head


Like a kind of “instant replay”

Echoic Sensory Memory


Capacity is limited to what can be heard at any one moment
and is smaller than the capacity of iconic memory, although it
lasts longer (about 2
-
4 seconds)


Useful in conversations, allows a person to remember what
someone said long enough to recognize the meaning of a
phrase


And like iconic memory, it allows people to hold on to
incoming information long enough for the lower brain centers
to determine whether or not processing by higher brain centers
is needed


Short
-
Term and Working Memory


Short
-
term memory (STM)


the memory system in which information is
held for brief periods of time while being used


Like a sort of “desk” where people hold and work on information for a short time


If an incoming sensory message is important enough to enter consciousness, that
message will move from sensory memory to short
-
term memory


This is accomplished through the process of
selective attention


the ability to
focus on only one stimulus from among all sensory input


Dr. Donald E. Broadbent’s original filter theory of memory suggests a kind of
“bottleneck” occurs between sensory memory and STM


Only a stimulus that is “important” enough will be passed on to be analyzed for
meaning in STM


What is important is determined by a kind of “pre
-
analysis” in attention centers in the brain
stem


Other stimuli are filtered out and will not reach consciousness


Also, when a person is thinking actively about information, that information is said
to be conscious and in STM

Short
-
Term and Working Memory


But, it is difficult to use this original selective attention filter theory
to explain the “cocktail
-
party effect”


Ex. When you’re at a party somewhere and there’s a lot of noise and
several conversations going on in the background but you are still able to
notice when someone says your name


In this situation, the areas of the brain that are involved in selective
attention had to be working


even though you were not consciously
aware of it


Then, when that important bit of information (your name)
“appeared,” those areas somehow filtered the information into your
conscious awareness


in spite of the fact that you weren’t paying
conscious attention to the background noise


In the original theory, selective attention only operates after you are
consciously aware of a stimulus


Short
-
Term and Working Memory


A newer theory, proposed by Dr. Anne M.
Treisman
, suggest that selective
attention operates in a two
-
stage filtering process


1
st

stage: incoming stimuli in sensory memory are filtered on the basis of simply
physical characteristics


This is similar to Broadbent’s original idea


However, the filtering in this case is not an “all
-
or
-
nothing” event as in Broadbent’s theory,
rather it is a lessening or decrease (
attenuation
) of the signal strength of those unattended
sensory stimuli when compared to attended stimuli


In other words, we lessen the signal strength of stimuli that are not important and not
attended to


2
nd

stage: involves the processing of only the stimuli that meet a certain threshold
of importance


Attenuated stimuli are present at this 2
nd

stage so something as subjectively important as
one’s own name may be able to be “plucked” out of the attenuated incoming stimuli


Ex. Selective attention is still working even when we’re asleep. This is why when
a mother is asleep she can be awakened by her baby’s cries but sleep through the
noise of the nightly train passing by (the noise of the train is unimportant and
has been attenuated while the baby’s cries are important)

Short
-
Term and Working Memory


What happens when information does pass through the selective
attention filter and into STM?


STM tends to be encoded primarily in auditory (sound) form


That means that people tend to “talk” to themselves inside their heads


Although some images are stored in STM in a kind of visual
“sketchpad,” auditory storage accounts for much of STM encoding


Research in which participants were asked to recall numbers and
letters showed that errors were nearly always made with numbers of
letters that
sounded similar
than errors with those that
looked similar

Short
-
Term and Working Memory


Some memory theorists use the term working memory as
another way of referring to STM


This usage is not entirely correct: STM has traditionally been
thought of as a thing or a place into which information is put


Working memory
is an active system that processes the
information in STM and is thought to consist of 3 interrelated
systems


The central executive (“CEO” or “Boss”): that controls and
coordinates


The visual “sketchpad”: for visual information


A kind of auditory “recorder”: for auditory information



Short
-
Term and Working Memory


The central executive interprets visual and auditory
information and the visual and auditory information is itself
stored in STM


Ex. When someone is reading a book


The sketchpad will contain images of the people and events of
the particular passage being read


The recorder “plays” the dialogue in the person’s head


The central executive helps interpret the information from both
systems and pulls it all together


So in a sense, then, STM can be seen as being a part of the
working memory system

Short
-
Term and Working Memory


For instance, if you see someone familiar at the mall


You pull that person’s name from your more permanent
memory and visualize that name along with the memory of the
last time you saw the person, almost like you’re viewing it on a
screen in your head


At the same time, you will hear the name in your head


The central executive pulls these different types of information
together and you are able to successfully greet the person


Where
you see and hear this information is in STM


The
process

that allows this to happen and coordinates it all is
working memory

Capacity: The Magic Number 7


George Miller (1956): reviewed research on memory, including studies using a memory test
called the
digit
-
span test


A series of number is read to subjects, then, subjects are asked to recall the numbers in order


Each series of numbers gets longer and longer, until subjects cannot recall any of the numbers in
order


Most everyone gets past the fist 2 sequences of numbers


But some people will make mistakes on the 6 digit span, about half will make mistakes on the 7
digit span, and very few people will be able to get past the 9 digit span without making any
mistakes


This lead Miller to conclude that the capacity of STM is about seven items or pieces of
information


He called this the magic number 7, plus or minus 2 items (from 5
-
9 bits of information)


Current research suggests younger adults can hold 3
-
5 items of information at a time if a
memory strategy is not being used


When the information is in the form of longer, similar
-
sounding words or unfamiliar words,
that capacity reduces until it is only about 4 items

Capacity: Chunking


Chunking

is a way to sort of “fool” STM memory into holding
more information than usual


It is a process of recoding or reorganizing information


If the bits of information are combined into meaningful units,
or chunks, more information can be held in STM


Ex. Remembering the digit span: 4,8,3,9,4,3,7,1,6,2


Would be easier if you grouped the numbers together into
chunks: 483
-
943
-
7162


Instead of 10 separate bit of information, there would only be 3
“chunks” that read like a phone number

“Short”
-
Term Memory: How Long Does
it Last?


Research shows that STM last from about 12
-
30 seconds
without rehearsal


After that, the memory seems to rapidly “decay” or disappear


The findings of one recent study with mice suggest that in order
to form new memories, old memories must be “erased” by the
formation of newly formed neurons


The hippocampus only has so much storage room


M
any of the memories formed there will be transferred to more
permanent storage in other areas of the brain


But, some memories, without rehearsal, will decay as new neurons
(and newer memories) are added to the already existing neural
circuits

STM: Maintenance Rehearsal


Maintenance rehearsal


practice of saying some information to be remembered over and over
in one’s head in order to maintain it in STM


Ex. Like repeating a phone number over and over again just long enough for you to dial it


A

person is simply continuing to pay attention to the information to be held in memory, and
since attention is how that information got into STM in the 1
st

place, it works quite well


Information will stay in STM until rehearsal stops, then the memory will rapidly decay and is
forgotten


If anything interferes with maintenance rehearsal, memories are also likely to be lost


Ex. If you are trying to count a stack of dollar bills by reciting each number out loud while
counting, and someone asks you what time it is, you will probably forget what the last number
was and have to start all over again


Interference in STM can also happen if the amount of information to be held exceeds capacity
(5
-
9 bit of information)


This is why it’s possible to remember the 1
st

few names of people you meet at a party, but as
more names are added, they displace the older names


FYI: a better way to remember people’s names is to associate the name with
s
omething about
the person’s appearance, this may help move the name from STM into more permanent storage

Long
-
Term Memory


Long
-
term memory (LTM)



the system of memory into which all the
information is placed to be kept more or less permanently


The capacity of LTM seems to be unlimited for all practical purposes


Think about it, would there ever really come a time when you couldn’t fit one
more piece of information into your head?


As for duration of LTM, there is a relatively permanent physical change in
the brain itself when a memory is formed


That means that many of the memories people have stored away for a long time
(even ones from childhood) may still be there


BUT, that doesn’t mean that people can always retrieve those memories, they
may be
available

but not
accessible
, meaning they are still there, but for various
reasons people cannot “get to” them


Ex. Its like knowing that there is a certain item on the back of the top shelf of the kitchen
but having no ladder to reach it, the item is there (available) but you cant get to it
(accessible)

Long
-
Term Memory


Information that is rehearsed long enough may actually find its way into LTM


Most people tend to learn poems and the multiplication tables by maintenance rehearsal, also
known as rote learning


Rote

is like “rotating” the information in your head, saying it over and over again


But maintenance rehearsal is not the most efficient way of putting information into LTM,
because to get the information back out, your have to remember it almost exactly as it went in


Ex. Try this: what is the 15
th

letter of the alphabet?


Did you have to recite or sing the alphabet song to get to that letter? I bet so.


Although, many long
-
term memories are encoded as images, sounds, smells, or tastes, in
general, LTM is encoded in meaningful form


A kind of mental storehouse of the meanings of words, concepts, and all the events that people
want to keep in mind


Even the images, sounds, smells, and tastes involved in these events have some sort of meaning
attached to them that gives them enough importance to be stored long term


If STM can be thought of as a working “surface” or desk, then LTM can be thought of as a huge
series of filing cabinets behind the desk


Files are stored in an organized fashion, according to meaning


Files have to be placed in the cabinets in a certain organized fashion to be useful




LTM: Elaborative Rehearsal


Elaborative rehearsal


a method of transferring information from STM
to LTM by making that information meaningful in some way


The easiest way to do this is to connect new information with something
that is already well known


Ex. The French work
maison

means “house.”


A person could try to memorize that (using maintenance rehearsal)by saying over and
over “
Maison

means house.”


But, it would be much easier and more efficient if that person simply thought, “
Maison

sounds like mason, and masons build houses.” That makes the meaning of the word tie
in with something the person already knows (masons who lay bricks to build houses)


The levels
-
of
-
processing approach discussed earlier proposes that
information that is more “deeply processed,” or processed according to
meaning will be remembered more efficiently and for a longer period of
time


This is true, elaborative rehearsal is a deeper kind of processing than
maintenance rehearsal and also leads to better long
-
term storage

Types of Long
-
Term Information


Long
-
term memories include general facts and knowledge,
personal facts, and even skills that can be performed


Procedural (nondeclarative) memory is memory for skills


Because it usually involves a series of steps or procedures


Declarative memory is memory for facts


Because facts are things that are known and can be declared
(stated outright)



Procedural (Nondeclarative) LTM


Type of long
-
term memory including emotional associations,
memory for skills, procedures, habits, and conditioned
responses


These memories are not conscious but are implied to exist because
they affect conscious behavior


Ex. Tying shoes or riding a bicycle


There is evidence that different areas of the brain are
responsible for procedural memories and declarative memories


The amygdala is the most probably location for emotional
associations, such as fear


T
he cerebellum in the hindbrain is responsible for storage of
memories of conditioned responses, skills and habits


Procedural (Nondeclarative) LTM


Evidence that separate areas of the brain control memory comes from studies of
people with damage to the hippocampus


This damage causes
anterograde amnesia


the loss of memory from the point of
the injury or trauma forward, or the inability to form new long
-
term
declarative

memories


In one study, patients with this disorder were taught how to solve a particular puzzle


P
atients were able to learn the sequence of moves necessary to solve the puzzle


But, when brought back into the testing room at a later time, they could not
remember ever having seen the puzzle (or the examiner) before


Each trial was like the first one ever for these patients, as they were unable to store
the long
-
term memory of having been in the room or having previously met the
examiner


Yet, they were able to solve the puzzle even while claiming that they had never seen it
before


Their procedural memories for how to solve the puzzle were evidently formed and
stored in a part of the brain separate from the part controlling the memories they
could no longer form

Procedural (Nondeclarative) LTM


The patients in the previous study had the kind of memory problems
that people with Alzheimer’s disease have


Yet even people with Alzheimer’s disease do not forget how to walk,
talk, fasten clothing, or even shoes (although they do lose motor ability
because the brain eventually fails to send the proper signals)


These are all procedural, nondeclarative memories


They may not be able to tell someone that they know how to do these
things, but they can still do them


Alzheimer’s disease affects the hippocampus and the frontal cortex
(involved in decision making and planning) and eventually affects
other areas of the brain after it has progressed nearly to the end


In fact, it would be rare to find someone who has lost procedural
memory


Literally, these are the kind of memories people “never forget”

Procedural (Nondeclarative) LTM


Procedural memory is similar to the concept of
implicit memory



memory that is not easily brought into conscious awareness


Because memories for these skills, habits, and learned reflexes are not also not
easily brought into conscious awareness


Ex. The fact that people have the knowledge of how to tie their shoes is
implied
by
the fact that they can actually tie them


Such knowledge is in people’s memories because they use this information,
but they are often not consciously aware of this knowledge


Although procedural memories are very often implicit, not all implicit
memories are necessarily procedural


A memory from one’s early childhood of being frightened by a god may not be a
conscious memory in later childhood but may still be the cause of that older
child’s fear of dogs


Conscious memories for events in childhood, on the other hand, are usually
considered to be a different kind of long
-
term memory called
declarative memory

Declarative LTM


Declarative memory


type of LTM containing information
that is conscious and known


This type of memory is about all the things that people
know


Including general facts such as the names of the planets in the
solar system, that adding 2 and 2 makes 4, and that a noun is the
name of a person, place, or thing


Also includes what people know about the things that have
happened to them personally like what you ate for breakfast this
morning and what you saw on your drive to class


There are 2 types of declarative long
-
term memories


Semantic and episodic

Declarative LTM: Semantic


Semantic memory


type of declarative memory containing
general knowledge, such as knowledge of language and
information learned in formal education


This is the general knowledge that anyone has the ability to know


Most of this information is what is learned in school or by reading


Includes the awareness of the meanings of words, concepts, and
terms as well as names of objects, math skills, and so on


Type of knowledge use on games shows like Jeopardy


Semantic memories, like procedural memories, are relatively
permanent


But it is possible to “lose the way” to this kind of memory

Declarative LTM
: Episodic


Episodic memory


type of declarative memory containing personal information not
readily available to others, such as daily activities and events


Includes memory of daily life and personal history, like a kind of autobiographical
memory


Ex. Certain birthdays or anniversaries that were particularly special, and childhood
events


Unlike procedural and semantic LTM, episodic memories tend to be updated and
revised more or less constantly


Ex. You can probably remember what you had for breakfast today, but what you had for
breakfast 2 years ago on this date is most likely a mystery


Episodic memories that are especially
meaningful
, such as the memory of the first day
of school or your first date, are more likely to be kept in LTM


Although these memories may not be as exact as people sometimes think they are


The updating process is a kind of survival mechanism, because although semantic and
procedural memories are useful and necessary on an ongoing basis, no one really needs
to remember every little detail of every day


The ability to forget some kinds of information is very necessar
y

Declarative LTM:
Episodic & Semantic


Episodic and semantic memories are examples of declarative or
explicit
memory



memory that is consciously known


Explicit memories are easily made conscious and brought from long
-
term storage
into short
-
term memory


The knowledge of semantic memories such as word meanings and episodic
memories such as what you ate for breakfast can be brought out of the “filing
cabinet” and placed on the “desk” where that knowledge becomes
explicit
, or
obvious


The difference between implicit memories (like how to ride a bike) and
explicit memories (like the names of all the planets) is that it is impossible or
extremely difficult to bring implicit memories into consciousness


Explicit memories can be forgotten but always have the potential to be made
conscious


Ex. When someone reminds you what you had for breakfast the day before, you will
remember and realize that you had that information all along

LTM

Declarative memory

(explicit)

Episodic memory

Events experienced by a
person

Semantic memory

General facts and knowledge

Procedural Memory

(implicit)

Motor skills, habits,
classically conditioned
reflexes

LTM Organization


LTM has to be fairly well organized for retrieval to be so quick


Research suggests that LTM is organized in terms of related
meanings and concepts


In one study (Collins &
Quillian
, 1969) had subjects respond
“true” or “false” as quickly as possible to sentences such as “a
canary is a bird” and “a canary is an animal”


Found that responses to sentences such as “canary is an animal” took
longer than responses to sentences such as “a canary is a bird”


Why? Because information exists in a kind of network, with nodes
(focal points) of related information linked to each other in a kind
of hierarchy



To verify the statement “a canary is a bird” requires moving to
only one node


But, “a canary is an animal” would require moving through
two nodes and, therefore, takes longer

LTM Organization


The results of this study lead the researchers to develop the
semantic network model


assumes information is stored
in the brain in a connected fashion, with concepts that are
related stored physically closer to each other than concepts
that are not highly related

LTM Organization


The parallel distributed processing model can be used to explain how rapidly the
different points on the networks can be accessed


Although the access of nodes within a particular category (ex. Birds) may take place
in a serial fashion, access across the entire network may take place in a parallel
fashion


Allowing several different concepts to be targeted at the same time


Ex. You might be able to think about birds, cats, and trees simultaneously


Maybe, the best way to think of how information is organized in LTM is to think
about the internet


A person might go on one website and from that site link to many other related sites


Each related site has its own specific information but is also linked to many other
related sites


And you can have more than one site open at the same time


This may be very similar to the way in which the mind organizes the information
stored in LTM