Psychology 144: Memory & Amnesia
Professor Anne Yilmaz
1/10/24
What is Memory?
● The process by which we encode, store, and retrieve information – involves many brain systems
● Clive Wearing: unique amnesic – severe brain damage resulted in the inability to form any new
memories; “this is the first time I’ve ever spoken, seen you, been here, etc.”
○ Lives in one singular moment, no past or future; moment-to-moment consciousness
Brain Structures
● Temporal lobe important because it contains the hippocampus (medial temporal lobe)
● Hippocampus responsible for putting things into long-term memory
○ Bilateral (one hippocampus on each side); classic amnesics have damage limited to this
region & area immediately surrounding tissue
● Multimodal Brain = different regions of the brain are responsible for processing different sensory
inputs (sound, vision, taste, etc.)
● An “experience” is the simultaneous activation of all these brain regions
○ There is not one singular processing area in the brain; this ability is distributed across the
brain depending on the sensory input
Models of Memory
● Atkinson-Shiffrin Model (1968):
○ Environment → sensory memory → short-term memory → long-term memory
(simplified)
■ Information from environment enters our sensory memory, our attention puts it
into STM and we “keep it alive” (rehearsal), encode into LTM via hippocampus
Where are long-term memories stores?
● They are stored where they are made = when we remember something, the same areas of the
brain will be activated again “memory engram”
1/12/24
Sensory Memory
● Sperling’s (1964) Iconic Memory Experiment: iconic memory = sensory memory
○ Presented matrix of letters for 50 ms; asked participants to report as many letters as
possible – found that subjects only recall about half of the presented letters
● Follow-up → sounded a low, medium, or high tone immediately after matrix disappeared (tone
signaled which row to report)
○ Recall of that row was almost perfect when the tone was added
● Partial report method allows us to determine how many items are in an individual’s icon
○ Calculation: size of icon = (number of remembered items in the row) x (number of rows)
■ (# of letters reported from a row) x (# of rows) = estimated letters in iconic image
○ If the tone was delayed, the icon will already start to fade from memory and recall
accuracy will decrease
Hypothesis: perception of Continuous Motion
, ● allows us to smoothly perceive motion as fluid, rather than staccato stop motion
Atkinson-Shiffrin Model
● Murdock (1962) “word list experiments”: primacy and recency effect; easier to remember words
presented at beginning/end of a sequence
○ Read list of words and then immediately asks participants to free-recall any words they
remember
○ Primacy Effect believed to be a long-term memory phenomenon
○ Recency Effect believed to be a short-term memory phenomenon
● Overt Rehearsal: want participants to rehearse out loud to keep items in memory
○ First few items receive the most rehearsal and therefore are more easily recalled
● Incidental Learning = learning by accident; subjects unaware of the impending memory test and
therefore do not rehearse
○ No rehearsal → expect to see primacy effect no longer present
○ How do we test whether incidental vs intentional learning causes a primacy effect?
● Speeded Lists: less opportunity to rehearse; reduced primacy effect but unaffected recency effect
● Amnesia: rehearsal does not help; primacy effect reduced or eliminated (no mechanisms to retain
first items presented, but recency effect remains)
Conclusion = the primacy effect is a long-term phenomenon (the more rehearsal, the more retention)
● Recency Effect (STM):
○ Immediate Recall vs Delayed Recall – immediate = ask participants to immediately recall
items; delayed = gap allows participants to rehearse, so researchers present a distractor
task for 15 seconds (to prevent rehearsal)
■ No distractor task after last item → strong recency effect
■ 15s distractor task after last item → no recency effect
● Final Free Recall: immediate recall experience plus one more final memory test
○ Participants recall words, do something else afterwards, and then are presented w/ a final
memory task (tell me again how many words you remember); participant is unprepared
■ Everything that is remembered is recalled from LTM; “negative recency effect” –
the words at the end have the least amount of rehearsal and therefore are less
likely to be remembered than words in beginning/middle
Conclusion = the negative recency effect is a long-term memory phenomenon (the less rehearsal an item
receives, the less likely it is to be remembered)
1/17/24
Recap from last week…
➢ If someone has hippocampal damage, this will not affect their sensory memory
➢ Recap → recency effect is STM phenomenon; primacy effect is LTM phenomenon
➢ Immediate recall vs final free recall tests (Craik, 1970)
● Craik and Watkins (1973): last few items of a list usually receive the least amount of rehearsal;
explains the “negative recency effect” in final free recall
○ Change procedure so that last few items are rehearsed more, which means we should see
positive recency effect; still use immediate and final free recall design
, ○ However, researchers still saw a negative recency effect during the final free recall
despite prompting participants to rehearse more
● Craik and Tulving (1975): incidental memory task; manipulated depth of processing for a list of
words – shallow (surface features), intermediate (phonemic processing), deep (semantic)
● We remember info better when we relate it to yourself/something you care about!
● Maintenance Rehearsal keeps info alive in STM, and Elaborative Rehearsal puts it into LTM
○ Maintenance Rehearsal – repeating word to yourself over and over
○ Elaborative Rehearsal – process the info on a deeper level (semantically)
● Primacy Effect = long term memory phenomenon; stays in memory due to depth of processing
How does depth of processing relate to the primacy effect?
● Fewer items to weave together; easier to remember the first words because elaborative rehearsal
becomes more difficult w/ more words to “link together”
1/19/24
● Typical STM tasks overestimate the storage capacity of STM (7 +/- 2)
○ In reality, it is actually much closer to the magical number 4
What happens when steps are taken to minimize chunking?
● Example – use unattended lists; while playing video game, ignored digits are presented through
headphones, followed by unexpected recall test for the digits
● Baddeley’s Working Memory Model (revision of STM concept):
○ Input via sensory memory → visuospatial (vision) WM and phonological loops (speech
sounds) → interacts w/ central executive back and forth → LTM
● Chess Positions: study chess positions for 10s; interfering activity for 10s
○ Manipulated Variable = type of distractor task
■ Phonological loop suppression (ALS): repeating a word aloud
■ Visual-sketchpad suppression (VSSP): push keys on a calculator
● Results: there was no change between control and ALS/verbal suppression group; however,
performance dropped significantly is the VSSP/vision suppression group
● STUDY EPISODIC BUFFER (will be on exam)
● Baddeley’s WM Model can be inserted in the middle of the Atkinson-Shiffrin Model
○ Sensory memory → Baddeley’s WMM → long term memory
● Patient PV (Baddeley):
○ PV had a left-hemisphere stroke in 1977, causing her to have a very short digit span (2)
■ L-Hemisphere = phonological loop/verbal WM
■ NO hippocampus damage
○ She had normal recall of meaningful material, normal performance on visuo-spatial tasks,
and normal functioning in everyday life
Why do you think PV can perform as well as controls in this AUDITORY task even though she doesn’t
have a phonological loop?
● Meaningful information through word-association allowed this information to be directly sent to
LTM – “understanding IS memorizing”; did not require phonological rehearsal
● In a second experiment (non-word task including not real words), patient PV could not recall a
single variable = fails at a paired-associated non-word task
Professor Anne Yilmaz
1/10/24
What is Memory?
● The process by which we encode, store, and retrieve information – involves many brain systems
● Clive Wearing: unique amnesic – severe brain damage resulted in the inability to form any new
memories; “this is the first time I’ve ever spoken, seen you, been here, etc.”
○ Lives in one singular moment, no past or future; moment-to-moment consciousness
Brain Structures
● Temporal lobe important because it contains the hippocampus (medial temporal lobe)
● Hippocampus responsible for putting things into long-term memory
○ Bilateral (one hippocampus on each side); classic amnesics have damage limited to this
region & area immediately surrounding tissue
● Multimodal Brain = different regions of the brain are responsible for processing different sensory
inputs (sound, vision, taste, etc.)
● An “experience” is the simultaneous activation of all these brain regions
○ There is not one singular processing area in the brain; this ability is distributed across the
brain depending on the sensory input
Models of Memory
● Atkinson-Shiffrin Model (1968):
○ Environment → sensory memory → short-term memory → long-term memory
(simplified)
■ Information from environment enters our sensory memory, our attention puts it
into STM and we “keep it alive” (rehearsal), encode into LTM via hippocampus
Where are long-term memories stores?
● They are stored where they are made = when we remember something, the same areas of the
brain will be activated again “memory engram”
1/12/24
Sensory Memory
● Sperling’s (1964) Iconic Memory Experiment: iconic memory = sensory memory
○ Presented matrix of letters for 50 ms; asked participants to report as many letters as
possible – found that subjects only recall about half of the presented letters
● Follow-up → sounded a low, medium, or high tone immediately after matrix disappeared (tone
signaled which row to report)
○ Recall of that row was almost perfect when the tone was added
● Partial report method allows us to determine how many items are in an individual’s icon
○ Calculation: size of icon = (number of remembered items in the row) x (number of rows)
■ (# of letters reported from a row) x (# of rows) = estimated letters in iconic image
○ If the tone was delayed, the icon will already start to fade from memory and recall
accuracy will decrease
Hypothesis: perception of Continuous Motion
, ● allows us to smoothly perceive motion as fluid, rather than staccato stop motion
Atkinson-Shiffrin Model
● Murdock (1962) “word list experiments”: primacy and recency effect; easier to remember words
presented at beginning/end of a sequence
○ Read list of words and then immediately asks participants to free-recall any words they
remember
○ Primacy Effect believed to be a long-term memory phenomenon
○ Recency Effect believed to be a short-term memory phenomenon
● Overt Rehearsal: want participants to rehearse out loud to keep items in memory
○ First few items receive the most rehearsal and therefore are more easily recalled
● Incidental Learning = learning by accident; subjects unaware of the impending memory test and
therefore do not rehearse
○ No rehearsal → expect to see primacy effect no longer present
○ How do we test whether incidental vs intentional learning causes a primacy effect?
● Speeded Lists: less opportunity to rehearse; reduced primacy effect but unaffected recency effect
● Amnesia: rehearsal does not help; primacy effect reduced or eliminated (no mechanisms to retain
first items presented, but recency effect remains)
Conclusion = the primacy effect is a long-term phenomenon (the more rehearsal, the more retention)
● Recency Effect (STM):
○ Immediate Recall vs Delayed Recall – immediate = ask participants to immediately recall
items; delayed = gap allows participants to rehearse, so researchers present a distractor
task for 15 seconds (to prevent rehearsal)
■ No distractor task after last item → strong recency effect
■ 15s distractor task after last item → no recency effect
● Final Free Recall: immediate recall experience plus one more final memory test
○ Participants recall words, do something else afterwards, and then are presented w/ a final
memory task (tell me again how many words you remember); participant is unprepared
■ Everything that is remembered is recalled from LTM; “negative recency effect” –
the words at the end have the least amount of rehearsal and therefore are less
likely to be remembered than words in beginning/middle
Conclusion = the negative recency effect is a long-term memory phenomenon (the less rehearsal an item
receives, the less likely it is to be remembered)
1/17/24
Recap from last week…
➢ If someone has hippocampal damage, this will not affect their sensory memory
➢ Recap → recency effect is STM phenomenon; primacy effect is LTM phenomenon
➢ Immediate recall vs final free recall tests (Craik, 1970)
● Craik and Watkins (1973): last few items of a list usually receive the least amount of rehearsal;
explains the “negative recency effect” in final free recall
○ Change procedure so that last few items are rehearsed more, which means we should see
positive recency effect; still use immediate and final free recall design
, ○ However, researchers still saw a negative recency effect during the final free recall
despite prompting participants to rehearse more
● Craik and Tulving (1975): incidental memory task; manipulated depth of processing for a list of
words – shallow (surface features), intermediate (phonemic processing), deep (semantic)
● We remember info better when we relate it to yourself/something you care about!
● Maintenance Rehearsal keeps info alive in STM, and Elaborative Rehearsal puts it into LTM
○ Maintenance Rehearsal – repeating word to yourself over and over
○ Elaborative Rehearsal – process the info on a deeper level (semantically)
● Primacy Effect = long term memory phenomenon; stays in memory due to depth of processing
How does depth of processing relate to the primacy effect?
● Fewer items to weave together; easier to remember the first words because elaborative rehearsal
becomes more difficult w/ more words to “link together”
1/19/24
● Typical STM tasks overestimate the storage capacity of STM (7 +/- 2)
○ In reality, it is actually much closer to the magical number 4
What happens when steps are taken to minimize chunking?
● Example – use unattended lists; while playing video game, ignored digits are presented through
headphones, followed by unexpected recall test for the digits
● Baddeley’s Working Memory Model (revision of STM concept):
○ Input via sensory memory → visuospatial (vision) WM and phonological loops (speech
sounds) → interacts w/ central executive back and forth → LTM
● Chess Positions: study chess positions for 10s; interfering activity for 10s
○ Manipulated Variable = type of distractor task
■ Phonological loop suppression (ALS): repeating a word aloud
■ Visual-sketchpad suppression (VSSP): push keys on a calculator
● Results: there was no change between control and ALS/verbal suppression group; however,
performance dropped significantly is the VSSP/vision suppression group
● STUDY EPISODIC BUFFER (will be on exam)
● Baddeley’s WM Model can be inserted in the middle of the Atkinson-Shiffrin Model
○ Sensory memory → Baddeley’s WMM → long term memory
● Patient PV (Baddeley):
○ PV had a left-hemisphere stroke in 1977, causing her to have a very short digit span (2)
■ L-Hemisphere = phonological loop/verbal WM
■ NO hippocampus damage
○ She had normal recall of meaningful material, normal performance on visuo-spatial tasks,
and normal functioning in everyday life
Why do you think PV can perform as well as controls in this AUDITORY task even though she doesn’t
have a phonological loop?
● Meaningful information through word-association allowed this information to be directly sent to
LTM – “understanding IS memorizing”; did not require phonological rehearsal
● In a second experiment (non-word task including not real words), patient PV could not recall a
single variable = fails at a paired-associated non-word task
