Human Memory (psc130)

• Short-Term Memory (STM)


  Memory that lasts for several seconds
• Long-Term Memory (LTM)


  Memory that lasts for a long time
• Capacity of STM


  • Accurate recall of 7 +/- 2 items (Miller, 1956)

  
• Digit span study


  1. List of digits, letters or words (2,3,4... items)

  2. Recall items in the correct order
• Chunking can increase the apparent capacity of STM by breaking the string into meaningful parts
• Although there may be individual differences in 'capacity', capacity measures can be influenced by strategies and experience
• Duration of STM


  • Several seconds
• Type of information held in STM

  • Phonological information
• Ockham's Razor


  If there are 2 theories that account for the data then we should prefer the simpler theory
• Evidence for 2-box model of memory

  • Single dissociation
  • Double dissociation
  • Crossover dissociation

  
• Serial Position Curve


  • Probability of recall as a function of study (or serial) position
  • Earlier items reflect LTM
  • Recency effect reflects STM

  
• Increase study time (Murdock, 1962)

  1. Study 15 words at 1 vs. 2 sec/item

  2. Free recall immediately after
• Increasing study duration increases the LTM portion of curve but leaves the STM portion unaffected
• Introduce a delay between study and test (Postman & Philips, 1965)

  1. Study 30 word list

  2. Test: immediate free recall vs delayed recall (10 seconds)
• Delay reduces STM but leaves LTM portion of curve unaffected
• The primacy effect occurs because the early items are rehearsed more, which increases LTM (Rundus, 1971)

  
• Memory Related Neuroanatomy


  • Amygdala
  • Frontal
  • Parietal
  • Occipital
  • Temporal
  • Parahippocampal Gyrus
  • Hippocampus
  • Thalamus
  • Fornix
  • Parahippocampal Gyrus
  • Fusiform Gyrus
• Medial Temporal Lobe (MTL)

  • Hippocampus, amygdala, parahippocampal gyrus
• The hippocampus is part of the medial temporal lobe
• Neuropsychological dissociations between STM and LTM

  1. Milner (1966) - Patient HM: deficit in LTM, normal digit span (STM)

  2. Baddeley & Warrington (1970) - Amnesics had a reduced LTM component, but normal STM component
• Amnesia is a long term memory impairment in the absence of other cognitive deficits
• Dementia patients typically show additional cognitive impairments beyond just memory
• Shallice & Warrington (1970) - Patient K.F.

  1. Damage to left temporal/parietal lobe

  2. Digit span of 1-2 items, normal LTM
• The double dissociations between STM/LTM reflect a problem with the modal model of memory
• What is the capacity of STM?

  Miller(1956) - memory and digit span

  -study: list of digits , letters or words (2,3,4.. items)

  • Test: recall items in the correct order, accurate recall of 7+/ -2 items
  • capacity of STM is about 7 items
• Increasing STM capacity 


  Chunking: breaking the string into meaningful parts

  can increase the apparent capacity of STM

  -although there may be individual differences in "capacity", capacity measures can be influenced by strategies and experience.
• Additional experiments on STM : What is the duration of STM?

  Several seconds

  What type of information does STM hold?

  Phonological information (speech based sound)
• Is there a "real" distintion between STM and LTM?

  Ockham's Razor or "the principle of parsimony": if there are 2 theories that account for the data then we should prefer the simpler theory (prefer one box model over a 2 box model.)
• Evidence for the more complex 2- box model:

  a single dissociation (e.g a variable has an effect on one task/process but not the other: A influences X but not Y) ( dissociates two forms, has an effect on one task but not the other )
• Evidence for the more complex 2- box model:

  A double dissociation would provide very strong evidence (e.g., A influences X but not Y, and B influences Y but not X)
• Evidence for the more complex 2- box model:

  A crossover dissociation would also provide evidence (e.g, A increases X and decreases Y ) ( one variable that does the opposite)
• The serial position curve

  Probability of recall as a function of study (or serial) position