Memory Revised

MEMORY

What is Memory?

“Persistence of learning over time via

storage and retrieval of information”

History: The “Cognitive Revolution”

• Mental processes can’t be explained solely by conditioning (e.g., language) – Noam Chomsky: language acquisition is systematic

• Study of mental process (e.g., memory, thinking, language, decision making) grew in 1960’s

Research Questions

• How many memory “systems” do we have?

– How much information can we hold in memory (capacity)?

– How long does information stay in memory (duration)?

• Is memory always accurate?

• What factors enhance memory?

3-Stage Model of Memory Atkins & Shiffrin (1968)

capacity vs. duration

Sensory Memory

• Initial recording of sensory input.

• Like retaining a glimpse of what you saw or heard.

• Iconic Memory: fleeting visual memory

• Echoic Memory: fleeting auditory memory (echo)

Sensory Memory

• How might we test capacity & duration of iconic memory?

• Sperling’s (1960) full and partial report

– Iconic memory

Short-Term Memory (STM)

• Stores info before it’s forgotten or transferred to Long-Term Memory (LTM).

– Recent view: Working memory

• Storage AND manipulation of information

• Demo: Word recall

STM: Characteristics

• Capacity

– Measured via memory span

• the number of items (digits, words, etc.) that can be correctly recalled in order.

• Limited to 7 +/- 2 “chunks” of info (Miller’s (1962) “Magic Number 7”)

– B V S M T A U I vs. T V U S A I B M

STM: Characteristics

• Duration – Measured via distraction task

• “BKG” – distraction – recall

• Purpose of distraction?

• Limited to 18-30 seconds

Is STM really a separate

“system” from LTM?

Evidence for separation of STM and LTM

• Behavioral Evidence

– Serial position effects

• Demo!

• Neuropsychological Evidence

– Amnesic patients

Classic Serial Position Curve

• The U-shaped relationship between a word’s position

in a list and its probability of recall.

Long-Term Memory (LTM)

• “Archive” of information about past events in our lives and knowledge we have learned

– Large capacity (virtually unlimited) • Ranges from memories stored a few minutes

ago to decades ago

– Semi-permanent duration

Types of LTM

• Who was the first president of the U.S.?

• What does “omnipotent” mean?

• How do you spell “cemetery”?

• What did you have for breakfast this morning?

• Recall your 16th birthday. • Recall what you remember

from 9-11-01.

Rarely do we

remember

episodes verbatim;

much of these

memories is

reconstructed

Recall is fairly

accurate, with

exception of

“temporary

retrieval failures”

Episodic Memory

• Episodic memories are not remembered verbatim

– They are reconstructed

• Fill in the missing blanks with what makes sense or what we expect would happen.

• What happens if we “fill in gaps” with incorrect information?

– Leading questions

– Misinformation

Episodic and semantic

memories can overlap • The knowledge that makes up semantic

memories is initially attained through personal experience

• Semantic memory can be enhanced by association with episodic memory

• Semantic memory can influence episodic memory by directing attention

Types of LTM

Implicit Memory

• Procedural memory

– Mirror drawing

H.M.'s mirror drawing

 H.M.’s performance on mirror drawing

• Classical conditioning

– Pairing one stimulus (such as a loud noise) with another, previously neutral stimulus (rat) causes changes in the response to the neutral stimulus

• Little Albert experiment

Implicit Memory

• Priming

– Presentation of a word (prime) affects a response to the same or related word presented later (target)

• Repetition priming

• Phonological priming

• Conceptual/semantic priming

Implicit Memory

Lexical Decision

 Decide if each item is a real word:

WATERMELON

CLOAM

NURSE or TABLE

DOCTOR

 Semantic priming = faster to say “yes” to DOCTOR after NURSE than after TABLE

Evidence for Implicit Memory

• Amnesia (Graf, Mandler, & Squire, 1984)

– Compared to “normals”, amnesics performed worse on traditional (explicit) memory tests, but similar on implicit memory tests.

Explicit Memory in the Brain Hippocampus –

the brain’s save button  Explicit memory not possible w/o hippocampus

 Other brain regions

─ Frontal lobe

─ Association areas

─ Amygdala

Implicit Memory in the Brain

 Procedural Memory

─ Basil ganglia

─ Cerebellum

 Priming

─ Association areas

 Classical Conditioning

─ Association areas

─ Limbic structures, e.g. amygdala

How our brains store memory

 Long-term Potentiation (LTP)

─ Changes in the excitability of interacting neurons

─ Occurs when neurons fire together repeatedly within a short time period

─ Makes it easier for action potentials to occur

─ Results in increased # of synaptic connections

 Drugs that block LTP reduce/block memory consolidation

 Drugs that facilitate LTP improve memory

LTM Processes

Encoding: getting info into memory

Consolidation: transferring from STM to LTM

 Storage: keeping info in memory

Retrieval: getting info out of memory

Why we forget

• Retrieval failure

• Tip-of-the-Tongue

(TOT)

• Interference

– Retroactive

• Interference with OLD information

– Proactive

interference with NEW information

LTM: Why do we forget?

• Interference

• Other information interferes and makes the forgotten information inaccessible.

• Retroactive

• Proactive

• To overcome interference, over-learn and organize.

• Memory decay

• Forgetting curve

• Failure to encode

• Failure of attention

Why we forget

LTM: Forgetting • Assuming we effectively encoding

information into LTM…then why do we forget?

• Recall 7 dwarfs

• Find the penny:

Neuropsych Examples –

Failure to Encode

Brain lesion patients

• H.M. – Neurosurgery to treat severe epileptic

seizures in 1953 (age 27)

– Removed hippocampus and surrounding tissue

• Clive – Viral encephalitis caused damage to temporal lobes and

hippocampus – http://www.youtube.com/watch?v=Vwigmktix2Y&feature=related

• Both patients have intact STM but impaired LTM (due to damage to hippocampus) – Specifically anterograde amnesia (vs. retrograde)

Amnesia

• Anterograde = cannot acquire new LTMs

• Retrograde = cannot recall LTMs formed prior to accident

• Other kinds of amnesia

– Source amnesia ─ Infantile amnesia

Deja vu

Memory Reconstruction

• Leading questions

– Loftus & Palmer (1974)

• Showed participants scenes of car accident.

• Leading question: “About how fast were the cars going when they ________ into each other?”

Actual accident

Memory reconstruction

Memory Reconstruction

• Leading questions

– Loftus & Palmer (1974)

• Results: Stronger verbs led to higher estimates of speed

Memory Reconstruction

• But did question alter memory representation of the scene? – A week later, asked questions about the original film:

Did you see broken glass?

– Results:

• Note: no broken glass in the film

• 34% of “smashed” participants said “yes”

• 14% of “hit” participants said “yes”

– Conclude:

• Leading questions can alter memory representations

• What if the misleading information is presented subtly and not in question format?

Memory Reconstruction

• Misinformation effect: – Misleading information after an event can also

affect recall

– Loftus, Miller, & Burns (1978) • Showed slides of a red car hitting a pedestrian

• There was a stop sign in the slides

• 20 minutes or a week later, asked whether a second car passed the red car "while it was at the yield sign"

• Testing asked participants to recognize which slides they had seen (had to choose between one with a yield sign, one with a stop sign)

Misinformation Effect

• Results: – 20 minutes: 60% incorrectly recognized the yield

sign

– 1 week: 80% incorrectly recognized the yield sign

– People were quick to make these false judgments – confidence

• Conclusion: – Misinformation can alter memory representations

• Why? – Source memory failure: Misattribute the source of

information

Memory Reconstruction

• Applications

1. Eyewitness Testimony

• 60% of wrongful convictions are from eyewitness (based on 360 cases of wrongful convictions)

2. False Memories

• Hyman et al. (1995) – College students created false memories about family

events when told (falsely) that these events actually occurred.

Example of false recall (Hyman et al.,1995)

LTM: Encoding Strategies

• What types of strategies do you use?

• Many of these strategies are effective because they create multiple associations (and thus connections) in memory…

Type of Rehearsal Examples Effective?

Maintenance Repetition not for long-lasting memories

Elaborative Generation Self-reference Imagery

very effective! •Encourages associations in memory

•Provides many cues for retrieval

•Makes material distinctive

LTM: Encoding Strategies

• Generation effect (Slamecka & Graf, 1978) – Read pairs of related words: sea-ocean

• Involves reading

– Fill in the blank with a word related to the first

word: sea-oc______ • Involves generating

– Also manipulated depth of processing • Deep: words related in meaning

• Shallow: words related in sound (rhymes), e.g., save-cave

– Then gave a recognition test:

diamond ocean light

LTM: Encoding Strategies

• Results: – Generation led to

better memory than reading

• generation effect

– Depth of processing was also important: deeper processing (meaning) led to better memory than shallow (rhyme) processing

LTM: Encoding Strategies

• Self-Reference Effect (Rogers et al., 1977) • Participants evaluated 40 adjectives in one of the following ways:

• Is it in capital letters? (structural)

• Does it rhyme with ____? (phonemic)

• Does it mean the same as ____? (semantic)

• Does it describe you? (self-relevant)

• Results: Recall for self-reference was twice that of the semantic condition.

LTM: Encoding Strategies

• Spacing Effect • Distributing rehearsal over time leads to better

memory than “massing” study sessions

LTM: Encoding Strategies

• Testing Effect – Use 1 (or more) of your study sessions as “testing”

sessions.

– Roediger & Karpicke (2006) • Students studied a text in 4 study periods

• Assigned to one of 3 conditions: – SSSS

– SSSR

– SRRR

• Also asked to predict how well they would remember the material (“judgments of learning”)

– Assesses metacognition

LTM: Encoding Strategies

• Roediger & Karpicke (2006)

– Results: SRRR produced best retention

– Note: Pattern of metacognitive judgments of learning (i.e., predicted recall) is opposite pattern of actual long-term retention.

LTM: Retrieval Strategies

• Context Cues: We have better recall if the contexts at learning and recall are similar to one another

– Godden & Baddeley (1975): deep sea divers

0

10

20

30

40

50

On land Underwater

P e

rc e

n t

c o

rr e

c t

Retrieval Context

Words encoded on land

Words encoded underwater

LTM: Retrieval Strategies

• State-dependent

memory cues: – Lowe (1983): Better recall

if we are in same “state” (e.g., mood, physiological, arousal) at learning and retrieval

• Note: Best Performance = learn Sober, Test Sober!

• You have a variety of options when you study.

• Know which ones have empirical support for improving retention.

• Will you change your study habits, and if so, how?

Conclusions