Short Term Memory to combine
Sperling and Sensory Memory Whole report:
Letters flashed on screen, participants asked to report as many as could be seen Average of 4.5 out of 12 letters (37.5%) The problem here is that how many letters subjects can remember will be limited by the next step, short-term memory, which has limited store (~7+/-2).
Semantic Representations
Meaning-based code that is important for both STM and LTM - but mainly LTM □ Sort of the dictionary definition of something. A dog is an animal that has four legs and goes "woof" A plane has wings and an engine and flies around A chair has four legs and is something that you sit on
Evidence for a Phonological Loop Word-length effect:
Memory for lists of words is better for short words than for long words Takes longer to rehearse long words and to produce them during recall
Irrelevant Speech Effect
Memory is best if material is learned in a quiet setting, without background noise. □ The WM model explains this in terms of phonological interference. Irrelevant speech is acoustic and therefore goes directly into the phonological store where it can interfere with other information. □ Irrelevant visual information doesn't interfere as much because the articulatory control process doesn't translate it into a speech-like code.
•Phonological Similarity Effect:
Memory is worse for items that sound alike than for items that sound different. F M N S X hard to remember R H X K Y easier to remember □ Baddeley explained this in terms of interference in the phonological store; more interference for more similar items. □ He also predicted that the phonological similarity effect would exist for letter strings that are presented visually. □ Once the articulatory control process translates information into a speech-like code, it becomes susceptible to phonological interference. □ The phonological similarity effect shows up regardless of whether stimuli are presented visually or auditorially.
Sperling's Conclusions
Sensory memory (iconic memory) exists Sensory memory decays rapidly Fades by 50% within approximately 150 ms Almost entirely gone within 500 ms Thus, evidence to support sensory memory, and interaction with short-term memory, is strong.
Memory
processes involved in retaining, retrieving, and using information about stimuli, images, events, ideas, and skills after the original information is no longer present
memory capacity
refers to the number of things that can be correctly remembered (encoded and retrieved), with this limit depending on the type of information and its susceptibility to interference.
Persistence of vision:
retention of the perception of light info being transferred to visual cortex from photo receptors. (latency) Sparkler's trail of light Frames in film
Sensory Memory (Iconic Memory)
- Short-lived sensory memory registers all or most information that hits our visual receptors(large capacity) - there is a brief register so info doesn't disappear stored Information decays very quickly: hundreds of milliseconds
Working Memory (WM)
A limited-capacity system for the temporary storage and manipulation of information for complex tasks (e.g., comprehension, language, reasoning). • Differs from STM in that it emphasizes the manipulation and coordination of this information, not just how it is stored (the capacity and duration questions). □ Efficient task performance requires that information be used in a very specific way or sequence; WM addresses this.
Modal Model of Memory
Atkinson and Shiffrin (1968) Computer as a model for human cognition Memory is an integrated system that processes information Acquire, store, and retrieve information Components of memory do not act in isolation Memory has a limited capacity Limited space Limited resources Limited time
Chunking to develop expertise
Chase and Simon (1973) Memory for chess pieces on a board Chess masters and beginners Pieces positioned for a real chess game or randomly positioned Results of Chase and Simon's (1973a, 1973b) chess memory experiment. (a) The chess master is better at reproducing actual game positions. (b) Master's performance drops to level of beginner when pieces are arranged randomly.
Reducing interference by Chunking
Chunking: small units can be combined into larger meaningful units Chunk is a collection of elements strongly associated with one another but weakly associated with elements in other chunks
Neuropsychological evidence to support modal model
HM (patient) has damage to long-term but not short-term memory. Short-term memory has different properties (e.g., limited capacity, short time scale) compared to long-term memory (e.g., "unlimited", much longer time scale).
Working Memory Baddeley Hitch model Key Concept:
If the information is decaying faster than it can be refreshed, then information will be lost; if information can be refreshed faster than it decays, then the information will be maintained.
Decay Theory -Petersen and Petersen experiment
On each trial, subjects first learn group of letters (which changes every trial) and then count backward from the number indicated. Finding...number of letters remembered decreases based on how long you count backward. Implying there is a decay in memory because of time
Decay Reconsidered
Peterson & Peterson's (1959) data are caused by proactive interference rather than decay Keppel & Underwood (1962) showed that there is no decay on the first trial Decay on subsequent trials is due to interference with stimuli presented on previous trials
Evidence Phonological Loop Articulatory suppression
Prevents one from rehearsing items to be remembered Reduces memory span Eliminates word-length effect Reduces phonological similarity effect for reading words
Evidence support Persistence of Vision
Prior to 1960, many scientists noted that visual information seems to persist for a few hundred milliseconds after the offset of a stimulus This was based on introspection, with no data Example: Sparklers, camera flash
Problems with decay theory
Proactive interference (PI): occurs when information learned previously interferes with learning new information
Wikens et al (1976)
Proactive interference can be reduced by employing different semantic categories.
Visual Representations
Processed visual information; colors, shapes, and their spatial inter-relationships. □ Useful in representing scenes, layouts of interiors, and maps.
Word Length Effect
Short words ("man") are easier to remember than long words ("gentleman"). □ The longer a word is, the more time it takes to pronounce. □ The faster you can pronounce a word, the more times you can rehearse it through the phonological loop within a given period of time. □ The more times a word is rehearsed through the phonological loop the better it is remembered. □ The shorter word "cat" should be better remembered than the longer word "hippopotamus"- the word length effect.
Acoustic Representations:
Speech-like code important for reading and thinking. □ Inner Speech: The "voice" inside your head when you read or think (also called "subvocalization").
Keppel & Underwood (1962)
Study that supports the interference interpretation; suggested that some of the forgetting that was observed in the Brown-Peterson Paradigm was due to interference from items learned on previous trials. They only used three trials to depict that interference was the cause for forgetting. They found that there was virtually no forgetting on the first trial, yet on the second and third trials forgetting was observed. Thus, they concluded that when there was no source of interference, there was no STM forgetting.
Central executive
The part of working memory that is responsible for monitoring and directing attention and other mental resources. - allocating resources, manipulating auditory or visual efforts. -Attention controller -Focus, divide, switch attention - Controls suppression of irrelevant information
Working Memory - WM
WM consists of 3 separate but interacting components. • Central Executive: The "boss", coordinates the slave systems to perform some task (a lot like attention) • Visuospatial Sketch Pad: Important for visual imagery and the mental manipulation of visual information. • Phonological Loop: Important for remembering and using auditory information. Consists of two components.
Sperling and Sensory Memory Partial report tone delay:
When the cue is presented shortly after the array of letters and numbers, the subjects transfer information about the cued row from the iconic image into short-term memory, and then report it Because each row contains only 4 items and subjects must store only 1 row, performance is not limited by the fact that only 5 items can be stored in short-term memory If subjects can report 90% of the letters in the cued row, they presumably had a memory of 90% of the entire array at the time of the cue If the cue is too late, they store randomly-selected items rather than storing only the items in the cued row
Working Memory Baddeley Hitch model
Working memory differs from STM STM holds information for a brief period of time WM is concerned with the processing and manipulation of information that occurs during complex cognition in different degrees of time WM is set up to process different types of information simultaneously WM has trouble when similar types of information are presented at the same time
Articulatory Suppression Effect
Worse memory when an irrelevant word is repeated during a retention interval. □ Repeating a word ("the", "the", "the"...) after the presentation of a memory list makes it harder to recall the list items. □ Articulatory suppression ties up the articulatory control process and prevents information from being recycled through the phonological loop. In the absence of recycling, information fades after 2 secs, leading to poor memory. □ Articulatory suppression blocks the phonological similarity effect for visual stimuli, but not for auditory stimuli. □ Repeating a word prevents the articulatory control process from translating visual codes into phonological codes; this translation step isn't needed for auditorially presented items.
Control processes
active processes that can be controlled by the person: Rehearsal Strategies used to make a stimulus more memorable Strategies of attention, attending
Sperling and Sensory Memory Partial report:
cue people with tones to remember different parts gets around capacity limits of STM participants heard tone that told them which row of letters to report Average of 3.3 out of 4 letters (82.5%) Participants could report any of the rows
Digit Span
how many numbers you can remember in short term memery normally 7 plus or minus 2
Phonological loop
the auditory-based part of working memory that allows for the verbal rehearsal of sounds or words repeating info to make sense of it
Memory Code
the form in which information is represented in memory. □ Info in your computer is represented in bits, but how is information represented in your STM? □ Sensory memory uses an iconic or echoic code. This works well for recognition, but it wouldn't help us to do STM tasks.
Visuospatial sketch pad
the part of working memory that holds and processes visual and spatial information. rotating image to get a mental image
Modal Model of Memory: Short-Term Memory STM
• Holds a very small amount of processed (recognized) info for a short time (but longer than sensory memory). • This is both new info coming from sensory memory and recognition, and info retrieved from long-term memory. • limited by interference to only about 20 seconds, in absence of rehearsal STM integrates info from both sources to do a task.
The Digit Span Task (Miller, 1956)
• Measures the number of digits that a person can recall. • Found that people had good recall up to about 7 digits, but after this errors were common. • Concluded that our memory span is 7 +/- 2 (5-9 things) • But memory capacity can be increased by getting help from long-term memory via a process called "chunking". □ Chunking is the grouping of info to form meaningful units. □ Chunk: a set of things that are strongly associated with each other but weakly associated with things in other chunks. • STM capacity should be measured in "chunks", not individual units. Whereas 9 letters is beyond our capacity limit, 3 3-letter chunks is not.
Coding information in STM
• Memory Code: the form in which information is represented in memory. • Semantic Representations: Meaning-based code that is important for both STM and LTM (but mainly LTM) • Visual Representations: Processed visual information; colors, shapes, and their spatial inter-relationships. • Acoustic Representations: Speech-like code important for reading and thinking.
Chase and Simon (1973)
• Studied STM capacity and chunking using chess. • On average, master chess players recalled 16 out of 24 pieces correctly; novices recalled only 4 of the 24 pieces. • When chess pieces were randomly arranged on the board, master and novice subjects recalled the same number. • They concluded that chess masters chunk pieces into more meaningful units, and therefore process chess information in a qualitatively different way. • Although STM capacity is limited to only 4 things, if these things are chunks then vastly more information can be processed.
Forgetting from STM
• The rapid loss of information from memory in the absence of rehearsal. • By rehearsing information - a memory strategy or "control process", you can retain it indefinitely.
Peterson & Peterson (1959)
• To measure the duration of STM in the absence of rehearsal they used a "counting backwards" recall task. □ Presented 3 random consonants (e.g., C J P) for subjects to remember. □ A number was then presented, and subjects had to count backwards from this number by 3s for some variable duration retention interval (time between the presentation of the to-be-remembered material and the memory test). This prevented subjects from rehearsing the letters. □ Following the retention interval (and counting), subjects had to recall (report back) the three letters.
Peterson & Peterson (1959): conclusion
• if rehearsal is prevented, information in STM will fade away after about 20 seconds.
Components of the phonological loop:
□ Phonological Store: holds speech-like info for about 2 sec. □ Articulatory Control Process: (1) translating visual information into a speech-like code and moving this to the phonological store, (2) refreshes the memory traces being held in the phonological store.
