Cognition
• STM and Working Memory
• Memory comes in multiple forms and covers many types of behaviors
• Definitions: Process involved in retaining, retrieving and using information about stimuli, images, events, ideas and
skills after the original information no longer exists
• Also: Memory is active whenever some past experience has an affect on ongoing thought or behavior or in the future
• Multiple ways of Remembering:
• In a night, a flare causes a brief image of a face
• Perception continues for a fraction of second after the dark. This ability makes ‘movies’ possible
• Sensory memory
• You meet a person and ask for their number and memorize it by repeating it. Information that lasts for 10-15 seconds
with rehearsal
• STM or WM
• LTM: Information retained for long periods of time
• Experiences – Episodic memory
• Skills – Procedural memory
• Facts – Semantic Memory
Cognition
• People underestimate the necessity for STM / WM
• Exams / Daily schedule / Names / Phone numbers /
Directions
• Everything that is in awareness is actually STM
• Modal Memory
• Atkinson and Shiffrin (1968)
• Boxes are structural features
• Arrows are control processes e.g., rehearsal /
Elaborative rehearsal / Attentional strategies
Cognition
• Sensory memory
• Sparklers trail
• -Persistence of vision
• Films involve opening and closing of a shutter
but people see a continuous transition
Cognition
• Sperling (1966)
• 50 ms Presentation of an array of letters
• Subjects could report 4.5 / 12 letters
• Partial report method: Subjects could report
any row 3.3 / 12 or 82%
Cognition
• Delayed Partial report: After a delay of even 1 second, subjects could report 1 letter / row
• This is iconic memory with an auditory equivalent of Echoic memory
• STM: What we know, reading think of is in STM. Brief duration but of central importance
• Recall method
• Stimulus -> Delay -> Recall
• We may analyze the amount recalled, patterns in the recalled items e.g., categories
• Content of recall may also vary e.g., events, or facts etc.
• Brown-Peterson task (1959) People over estimate duration of STM. Showed that we are dealing
with a very short term system
• NKR 45 (count backwards in 3s till asked to recall)
• 80% recall if asked after 3 second delay
• 12% recall after 18 second delay
• GHT 29
Cognition
• Conclusion was that STM is a very high decay rate memory
mechanism
• However, Keppel and Underwood (1962): Performance on
the first trial is high even after an 18 second delay
• Suggested that the mechanism of forgetting is not a high
decay rate but Proactive interference
• Proactive vs. retroactive interference
• Whether due to decay or interference, it was agreed that
STM retention is limited to 15-20 seconds unless rehearsed
• Also limited is the capacity
Cognition
• Digit Span usually 7 +- 2
• Miller (1956)
• More recently Cowan 2001 have suggested that it is actually 4 items
• Dual task performance: Baddeley and Hitch (1974): Reasoning was impaired when digit
load was increased, also sentence comprehension or word list. 6 random digits produced
more interference than counting from 1-6
• However, there was no catastrophic breakdown
• Daneman and Carpenter (1980): Loading STM and also taxing processing capacity
• Reading Span:
• When his eyes opened, there was no gleam of triumph, no shade of anger
• The taxi turned on Michigan avenue where they had a view of the lake
• Luck and Vogel (1997): Change detection in simple squares.
• 100MS display -> 900MS delay -> 2000MS recall
• Perfect performance for 3 squares. Starts deteriorating after 4 items
Cognition
• Chunking: Combination into meaningful units
• When units are more closely associated with each other than with other units,
they form a chunk
• NNCBCBBIFNCI vs. CIAFBIBBCCNN
• Ericsson (1980)
• S.F. after 900+ hours of training could repeated 79 digits without error. Used STM
LTM connections to create efficient chunks.
• Instead of asking how many units, can we ask how much information can be held?
• Alvarez and Cavanagh (2004)
• When they introduce squares, cubes and more complex shapes, change detection
goes down
• Squares 4.4
• Cubes 1.4
Cognition
Cognition
• So far STM has been conceptualized primarily as a storage system and we are asking
its capacity and duration. However, it also involves lots of dynamic processes e.g.,
chunking. Also its usage indicates it is used in complex cognition.
• Baddeley and Hitch (1974): Concept of Working Memory WM
• Limited capacity system for storage AND MANIPULATION of information for complex
tasks such as comprehension, learning and reasoning
• E.g.,
• Can I have Brocolli & Mushroom Pizza
• We’re out of mushroom. Would you like to substitute spinach?
• Or
• Add numbers from 1-100
• Noted that people are capable of certain kinds of multitasking
• 7149 + Reading a paragraph for understanding
Cognition
• Already saw that loading STM effects reasoning, comprehension and list learning
• We also saw effects of phonemic similarity which shows in related tasks also
e,.g.,
• A is not preceded by B – AB (as compared with MC). Here phonemic similarity
shows a mild impact
• These phenomenon were explained with reference to an Phonological loop, a
system with phonemic representation
• Separately, a central executive controls and coordinates tasks which has control
memory capabilities depending on tasks
• Data further suggested that combining 2 visuo-spatial tasks e.g., mental imagery
and tracking target location were more difficult that a visual + verbal task. This
suggested separate subsystems but with some overlapping components
Cognition
• Phonological Loop
• Phonological Store + Articulatory Rehearsal process
• Visuo-spatial Sketch pad
• Central Executive: Pulls information from LTM, coordinated activity and given attention during
multitasking
• Phonological loop
• Similarity Effect: Conrad (1964). Letters projected on screen. Subjects asked to write down later
in order
• Errors: F-> S but not E i.e., phonological rather than visual similarity
• Word Length Effect
• Beast, Bronze, wife, gold vs. Alcohol, property, amplifiers, officer
• Baddeley et al (1984): Short words 77% vs. Longer words 60%
• People remember the no. of words they can articulate in ~ 2 secs
• Articulatory Suppression
Cognition
• Dishwasher, humming bird, engineering, hospital
• Vs. Articulatory suppression +
• Automobile, apartment, basketball, mathematics
• The retention is reduced and also the word length + phonological similarity effect is
removed (but only for visual presentation)
• Somewhat curious result is explained if articulatory suppression continued for recall as well
presentation
• Now word length effect removed for auditorily presented items but not phonemic similarity
• This suggests automatic access of verbal material to Phonological loop. Visual material need
recoding by subvocalization
•
• Visuospatial sketchpad:
• Involved also in creation of visual imagery
• Shepard and Metzler (1971)
Cognition
Cognition
• 40 degrees -> RT 2 secs
• 140 degrees->RT 4 secs
• Della Sala et all (1999): Pattern span
• 2x2 matrix with 2 shaded squares to 5x6
matrix with 15 shaded squares
Cognition
Cognition
• Subjects could handle upto 9 shaded squares before starting to
make mistakes
• This is larger than the supposed size 4 of WM but chunking
strategies are being used. Also seems to be s subtle distinction
between visuo and spatial
• Lee Brooks (1968): Interference in the visuospatial sketchpad
• Capital letter ‘F’
• Start clockwise and state whether the angle is ‘internal’ or
‘external’
• OR
• Point to the correct choice
• Much easier to say than to point
Cognition
• Central Executive
• Attention controller: How attention is divided, switched
• E.g., speaking on the phone and driving
• CE-> Phonological loop + VS sketchpad
• Frontal / Prefrontal lobes play key role in CE
• Damage associated with WM dysfunction e.g.,
• Perseveration
• Vogel et al (2005)
• High capacity group vs. low capacity group
Cognition
Cognition
• Change detection procedure
• Cue 1/10 sec display & then blank screen. Followed by a test display
• Event related potentials measured
• Greater response with a larger amount of WM utilized
• Adding blue bar distractors caused a much larger ERP in the low capacity
group
• Episodic Buffer
• WM can hold more information than the Baddeley Model suggests
• E.g., sentences of 15-20 words. May be due to chunking but also seems
assisted by LTM information
• Consists of sequential information across input modes. Dense Amnesics
can remember stories
Cognition
• WM and the Brain
• Techniques:
• Lesion studies
• Single cell recording
• Brain imaging
• Recording of electrical signals
• Memory action consist of delay and waiting
• Prefrontal Cortex damage:
• Goldman-Rakic 1992 Delayed response task
• With PFC lesioning, performance at chance level in monkeys
• Children under 8 months ‘Out of sight, out of mind’
•
Cognition
• Funahashi (1989); Neurons in PFC that
continue to fire when stimulus removed
• Fixation point X. Square is flashed. After delay,
monkey moves eyes to square location
Cognition
• Particular cells fire for particular locations
• However, the identification pf PFC with WM is that we know that WM can
contain detailed information and that this is processed in other brain regions
• Mark Stokes (2015)
• Activity silent WM: Short term synaptic strengthening may also be responsible
for WM
• Harrison and Tong (2009)
• Flashed 2 oriented gratings. Delay, after which the subjects are asked to judge
if grating has been rotated to the left or the right
• fMRI: Neural mind reading. Map brain activation for different gratings
• Afterwards algorithm predicted with 83% accuracy which grating was being
‘held’ in WM
Cognition
Cognition
• Maths and WM
• Ashcraft and Kirk (2001)
• Dual task condition for maths problems that were either
carry (27+17) or no carry (7+2)
• Poorer performance for ‘carry’ problems
• Ramirez and Beilock (2011) Worrying causes choking. If so,
easing worries should ease choking.
• Expressive writing group: Increase performance if writing
addressed feeling of anxiety
