Download Memory

Memory
Atkinson and Shiffrin (1968) multi-store model of memory.
Multi-store model:
 Two distinct stores – short-term and long-term.
 Stimuli from environment is held for less than one second – sensory register.
 Sensory Register:
o Modality-specific – information is stored in the same way in which it is
registered – taste as a taste, etc.
o Capacity is large but only a brief duration – about half a second.
o Processing is mostly unconscious and has to be deduced from
experiments.
o Sperling (1960) indicated it holds at least 9 items but only for a brief
period of time.
 Short-term Memory:
o Sensory register can be lost or is passed on to STM.
o (1956) Miller published “Magical Number Seven” – showed limit of
STM to be between five and nine, on average 7.
o Can be extended through “chunking”.
o If information in STM is rehearsed it can be retained in STM storage
and transferred to LTM for potentially indefinite storage.
 Long-term Memory:
o Refers to anything that we can remember for a period of time lasting
from minutes to a life-time.
o Has a huge capacity and seems to be able to store memories in a
number of forms.
o Mostly abstract/semantic and there is evidence it is stored in an
organised and meaningful way.
 Evaluation (Research):
o Duration of STM was investigated by Peterson and Peterson (1959) –
gave participants trigrams (e.g. MXT) which they had to recall after 018 seconds.
o Only 10% were recalled after 18s therefore showing duration of STM
is about 18 seconds.
o Can be extended through rehearsal and it is as a result of rehearsal that
information is transferred to LTM.
o Proved functional separation of STM and LTM
o We rarely rehearse in everyday life yet we still remember a lot of
things. This can be better explained through levels of processing
theory.
o Over-simplistic view that STM and LTM structures operate in a single,
uniform fashion. Working model is a more active model of human
processing.
o Many studies use laboratory experiments and can be criticised in terms
of ecological validity and demand characteristics. May tell us little
about how memory works in real life.
Levels of Processing Theory of Mind – Craik and Lockhart (1972)
 Proposed process rather than structure was important and that words can be
processed at a number of different levels:
o Shallow/structural level – a visual level, e.g. physical properties of
stimulus.
o Intermediate/phonetic level – an auditory level, e.g. the sound of the
stimulus.
o Deep/semantic level – the meaning of the stimulus and connections to
other stimuli.
 What is done with the material during rehearsal determines if it is
remembered, not simply the repetition of the words.
 Any stimulus can be processed in a number of ways from low levels of
analysis to higher levels.
 The higher levels give more durable memories; rehearsal just prolongs the
memory trace without on its own leading to a permanent memory.
 Used test that we did in class. Thought it preferable for participants not to
know beforehand or they would be tempted to perform additional analyses on
the words.
 Argued Atkinson and Shiffrin’s model gave a simplified view of rehearsal.
 When material is encoded, there are 2 types of rehearsal that can take place; it
is the type of rehearsal that is important for long-term retention:
o Maintenance rehearsal – simple rote repetition, repeating the words
as they have been presented.
o Elaborative rehearsal – analysing the meaning of the rehearsed
material, perhaps linking it with stored knowledge in LTM.
 Assumption is that the depth of processing determines the persistence of a
memory trace in the long-term store, and elaborative rehearsal will entail
deeper processing and thus a more durable memory.
 Evaluation (Research):
o Experiment we did but by Craik and Tulving – significantly more
words recognised at deep level (65%) than phonetic (37%) or shallow
(17%). Relies on incidental learning – advantage of being more trueto-life. Necessary level of deception raises ethical issues.
o Argued perception, attention and memory are interdependent because
memory traces are formed as a result of perceptual and attentional
processes – major contribution to understanding of memory.
o Open to empirical testing and generally results proved predictions,
worked on incidental learning, which is more ecologically valid.
o Problem – difficulty determining what level of processing actually
occurs in a particular case.
o Craik (2002) accepted that original levels were over-simplified and has
accepted new findings – Morris et al found recall depends on relevance
of the processing to the memory test and that somatic processing is not
always superior to phonetic processing.
The Working Memory Model – Baddeley and Hitch (1974)








Multi-store model proposed single and passive short-term store, whereas
working model emphasises the active nature of STM which has more than one
component.
Three components, allowing for temporary storage of verbal and visuo-spatial
material.
The phonological loop is concerned with auditory and speech based
information.
Visuo-spatial sketchpad allows storage of visuo-spatial material – visual and
spatial information, known as the inner eye.
Coordination over stores is an attentional control system known as central
executive.
Central Executive:
o Performs a number of tasks including:
 Focus and switch attention,
 Co-ordinate the sub-systems – phonological loop and VSS,
 Connect working memory with long-term memory.
o Not unlike a business executive.
o Flexible system which can process information in any sensory
modality in a variety of ways.
o Store information for a brief period of time.
Visuo-spatial Scratchpad:
o Divided into two components:
 A visual component – deals with objects and features like shape
and colour,
 A spatial component – deals with locations and movements in
space, e.g. planning your way through a shopping centre.
The Phonological Loop:
o Two components:
 A phonological store – holds auditory memory traces for a few
seconds before they fade.
 An articulatory rehearsal process, which is essentially subvocal speech – limited capacity of about 3-4 items.
o Auditory information enters automatically into the phonological store.



o Visually presented language – text – can be transformed into
phonological code by silent articulation and placed in the phonological
store.
o “Inner ear” – it remembers vocal sounds in their temporal order.
o “Inner voice” – repeats words in order to prevent them from fading
away after about 2 seconds.
o Research and evidence comes from three areas:
o Similarity of sound: similar sounding letters – VPBG – recalled less
well than dissimilar letters – XAYW. Suggests code is phonological.
o Word-length effect: short words remembered better than long words.
They can be articulated faster than longer words, so more of them can
be articulated in the time before the trace disappears.
o The effect of articulatory suppression: if asked to say something
aloud while rehearsing material in phonological loop, the memory for
the rehearsed material is impaired.
The three part structure was recently (2000) given a fourth component by
Baddeley – episodic buffer. It’s considered necessary because problems
emerged with the original model:
o Had difficulty explaining the interaction between the working memory
and long-term memory.
o No mechanism for allowing the sub-systems to interact.
Episodic Buffer:
o Limited capacity store that binds together information from a number
of fields – verbal, visual, spatial and chronological information.
o It can use access to the different sub-systems and form conscious
awareness.
Evaluation (Research):
o Evidence for a general attentional processor in addition to the subcomponents in the model, is provided by dual task experiments:
 If the two tasks are using the same sub-system, performance is
poorer when performed at the same time than when performed
separately. If they require different sub-systems they are
usually preformed equally well together as separate.
 Articulatory suppression – saying something aloud makes use
of articulatory loop thereby making it difficult to perform a
simultaneous task using the same sub-system due to its limited
capacity.
o Increasing amount of evidence to support relationship between
phonological loop and acquisition of language. Clinical study – patient
with phonological loop deficit, but normal verbal LTM. Baddeley
found a failure to acquire vocabulary of a new language. Also PL is
proving to be a good predictor of the ability of children to learn a
second language.
o Key findings from studies indentifying brain regions associated with
use of PL provided evidence for functional separation of phonological
and articulatory processes. Phonological – Wernicke’s area.
Articulatory – Broca’s area. Verbal memories appear to be in left
hemisphere – spatial memories is in the right hemisphere of the brain.
o Central executive is said to have limited capacity but there is no
evidence as to what the capacity is.
o Information from laboratory studies with high level of experimental
control – artificial environment gives theoretical models but may not
generalise to complex real-life situations.
o Still being researched and modified – provides valuable insight into the
complexity of STM processes.
Key Points:
 The multi-store model of memory consists of sensory register, STM and LTM.
 Each component of the multi-store model has different properties – code,
capacity duration.
 According to multi-store model, the key process for transferring information
from STM to LTM is rehearsal.
 Levels of processing theory emphasises that depth of processing is the key to
long-term retention.
 The working model is a model of short-term memory only.
 The working model has a central executive and a number of sub-systems,
including the visuo-spatial scratchpad and phonological loop; more recently
episodic buffer was added.