1. Explicit Memory

2. Procedural Memory

3. Sensory Memory

1. Recalling a fact/event

2. recalling how to do a certain motoric task

3. Inititial memory processor...must be differentiated from physiological processes

PHYSIOLOGICAL PROCESSES

1. Color-After Effect

2. Auditory Masking

3. Perceptual Contrast

1. After staring at a blue image for extended period of time w/o mvmt., the retina constantly fires 'blue' signals and when contrasted with a neutral color (white), the shape of image remains on neutral surface but in it's opposite color (yellow)

2. When played a tone over and over again, the auditory neurons fire constantly (lowering the threshold for that pitch). When tone stops, the neurons need time to recover therefore the same tone played during recovery period will not be perceived because it requires more energy than available. 

3. when an extreme sensory input is followed by a neutral sensory input, the perception leans toward the opposite extreme.

**These three physiological processes must be distinguished from sensory memory

Modal Model of Memory

(Atkinson & Shiffrin -- 1960s)

first model to explain memory: 

Sensory Input=> STM=> LTM

*There is a rehearsal period during STM which encompasses control processes*

Sensory Memory Icon Examples

1. Shepard Tone

2. Sparkler's trail

3. Flip-books/old movie reels

1. Continuously descending tone which never gets any lower

2. Fireworks which create an illusion of trail, but the trail is not there, it is Sens. Mem. creating the image

**Significance: raises question of what the sensory memory icon looks like ==> sperling's study

Sperling's study of sensory memory

condition 1:

- present card w/12 letters for 50ms.

- asks pts. what was on card

- pts have clear memory of card, but can only recall ~4 letters

Condition 2:

- present cards for 50ms. 

- plays either high/med/low tone, which is to be associated with a particular row on the card, after the card is gone

- pts. can report rows accurately==> *suggests that the icon has all info from card, but still can only recall ~4 letters therefore there is some kind of decay apparent*

Condition 3: What is decay timeline?

-plays tones after card but delays the tones by 500ms. (pts. remember much less)

-Delay tone by 1 second and the icon is completely gone

**Significance: this study really supports the idea that sensory memory is the starting point where ALL information is processed** 

Evidence for having 2 memory processors (STM/LTM)

1. Anterograde Amnesia

2. Retrograde Amnesia

-Ribot's Law

1. Inability to remember anything after a traumatic event

• * ppl with anterograde amnesia like so can play a game and if left alone & reintroduced, they do not remember meeting a person
• ** This suggests that they can store info into one box but not the other which enforces the need for 2 memory processors**

2. Memory of events before a brain injury is lost, but new memories can be formed 

- lack of episodic memory (biographical, event related info)

-They do not forget procedural memories (physiological)

-RIBOT's LAW: W/respect to retrograde amnesia things that happen early in life are recovered easier than events that occur closely before the accident.

Serial Free Recall

(Ebbinghaus)

1. Primacy Effect

2. Recency Effect

-pts. study list of words #'d 1-10

[image]

1. When words @ beginning of list of better recalled

2. When words @ end of list are better recalled

STM

I. Capacity

1. Miller

-Digit span task

2. Chase & Simon (1973)

-Chunking

-Experimental Process

1. determining how long phone #'s should be in order to be memorable

Digit Span Task: determines that memorization is optimal with 7 +/- 2 digits

2. Chunking: a chunk of info represented by multiple digits  [i.e. (631) 555-5839 - each group of digits is a chunk]

-In their experiment: they measure task-recall positions using a chess board w/pieces in specific positions

Condition 1: 

-Pieces are arranged for a game

[image]

Condition 2: 

-Pieces are arranged randomly

[image]

Results suggest that the capacity of STM depends on how informative/relative the information is to an individual*

**STM Capacity is very variable because of DURATION

STM

II. Duration

1. Peterson & Peterson (1959)

1. Give pts. list of words to remember & at the end of list they give pts a 3-digit #

-pts. asked to count backwards by threes

-@ 3 seconds and 18 seconds after # task pts. are asked to recall the word list

[image]

[image]

**By the 10th trial, everyone is worse overall therefore STM duration dissipates over time

1. Proactive Interference

2. Retention Interval

3. Interpresentational Interval

1. Information from previous trials (word lists) interfering with recollection. 

-Incorrect info from past makes the probability of incorrect recall in future higher

2. Interval after encoding of target info is over

3. period of time (pauses) between presentation of stimuli

Glanzer & Cunitz (1966)

(long-term recency effect)

Experimental Process & Findings

-Present pts with 12 word list

-impliment a retention interval by counting from 1-9 after the word list is taken away

-Implementing this retention interval results in the recency effect being erased therefore they **find that: if ppl do not immediately recall, then the recency effect is gone**

Watkins, Neath, & Sechler (1989)

(long-term recency effect)

1. Experimental process/finding

2. Ratio Rule

-suggests that there is a difference in the way we encode a word list as opposed to the way we encode info like the presidents

1. Has pts. memorize a 12 word list

-makes the Interpresentational interval AND the retention interval a count between 1-9

-Finds: that the recency effect is restored.

2. the IPI and the RI should be approximately equal to each other to show the recency effect. 

-**This finding suggests that the recency effect is not a function of STM**

Anderson & Schooler

1. Content Analysis

1. Looking through periodicals/current events

2. They Suggest that the recency effect of our memory is linked to how we experience/deal with the world.

-Through content analysis they find the same curved pattern as recency effect when looking at Freq. of mentioning a certain story/event over time

**This turns away from the Modal Model of Memory by including social context with the way we remember things**

Baddley & Hitch (1986)

1. Working Memory

2. Working Memory Model

1. Type of memory process that allows us to play around w/info we are working w/@ the moment. (i.e. remembering the beginning of a sentence by the time it is over)

2.[image]

**Baddeley works off of Brooks' suggestion for 2 slave systems and suggests that we need something to manage them (CE)**

Evidence for Phonological Store

1. Phonological Similarity Effect

2. Articulatory Suppression Effect

3. Irrelevant Speech Effect

4. Word-Length Effect

1. Conrad (1964)

-Remembering a list of letters that phonologically similar (i.e. PGTCD all end in /ee/) leads to lower accuracy as opposed to letters that are not phonologically similar (i.e. RXKWT)

-As a result of this finding B&H suggest the phono. store as a place where the phono. info that is stored is sensitive to this similarity among speech sounds

_______________________________________________

2. Murray(1968) Finds that Repeating function words like 'The' While remembering a word list leads to decreased accuracy therefore we need a phonological loop b/c repeating function words suppresses rehearsal of target info.

______________________________________________

3. Colle & Welsh(1976)

-Has pts. remember a word list while listening to someone speak and finds that this leads to decreased accuracy

**This is evidence for Phono store because the phonological info from the speech is interfering with words on the list getting into the store**

_______________________________________________

4. MacWorth(1963)

-has pts. remember a word list of varying syllable structure.

-Finds that as the number of syllables increased, the more difficult it is to remember

Evidence for Visuo-Spacial Sketch Pad

1. Mental Representation

2. Propositional Representation

3. Analog Representation

1. What we can know of the world

2. single simple statement that can be either True or False (proposition) (i.e. NP is N of NYC, NP is N of Philly)

3. Representation that reassures spacial & physical characteristics of whatever is being represented (i.e. a map)

*Significance: We are capable of storing both types of info therefore we need a mental process that can separate and organize spacial information while we experience it**

Evidence for Visuo-Spacial Sketch Pad

1. Roger Shepard (1973) 

-Mental Rotation Experiment

2. Kosslyn, Ball, & Reiser (1978)

-Mental distance

1. looks @ how ppl mentally rotate objects

-Argues that we have a spacial representation of things in the world suggesting that mental rotation should occur the same way they do in the world.

-Has pts. analyze complex 3D objects and report whether or not it is reversed

-Finds a nice linear relationship between RT and angle from vertical position therefore concludes that mental rotation functions same as rotation in the real world **Evidence for Visuo-spacial sketch pad**

_______________________________________________

2. have ppl. memorize a map of a fictional place

-pts. asked to imagine a light @ pt. of map and move it from pt. A to pt. B

-Find that TY is linearly related to distance on map therefore **evidence for visuo-spacial sketch pad**

Baddely (2000)

-Revision of Working Memory Model to account for situations that need to be integrated

1. Interference

2. Episodic Buffer

3. Revised Model

1. Memory theorists notice that phonological & visuo-spacial info. interfere with each other therefore Baddely suggests a 3rd processor to explain.

2. Third memory processor that integrates info. from LTM & slave systems 

-accounts for info. associated w/specific event (i.e. seeing sheet music & hearing the piece is an integration of phono & visuo-spacial info that is linked to a certain event or memory)

-episodic buffer is responsible for chunking as well

3.[image]

LTM

-infinite in time and space therefore very complex in terms of processing

Contents:

1. Declarative Memory (explicit)

2. Implicit Memory

3. Maintenance Rehearsal

1. The kind of memory of which we are aware

2. memory info of which we are not aware

3. Keeping info in working memory (does not explain all encoding)

How do we know if we know something? 

-we can find out by differentiating how info is stored (encoded) and retrieved.

1. Declaritive Memory

a) Episodic Memory

b) Semantic Memory

How do we test Declarative Memory?

1. Memory of which we are aware

a) Memory of specific moments or events

-Tulving uses 'mental time travel' as retrieval analogy for episodic mem.

b) Facts-specific bits of info

we test DM by 1. serial-recall

2. List recognition

3. Lexical Decision Task

Implicit Memory

a) Procedural Memory

b) Repetition Priming

c) Evidence for RP

-Memory of which we are not aware of upon encoding

a) How to perform certain motoric tasks (i.e. ride bike, play video games, use language)

b) response to 1 stimulus is improved by having seen a previous stimulus (i.e. if shown a list of words and then given a LDT afterwards, you would have a faster RT to the words that were on the word list given earlier)

c) Evidence for Repetition Priming: Word Fragment Completion: pts. given a list of words and asked to evaluate in some way but not explicitly asked to remember the words. 

-Afterwards, pts. given a partial word completion task and look @ RT

-Pts. fill in words from list even though they have no explicit memory of words being on the list

Craik & Tulving (1972)

-move away from modal model

1. Levels of Processing

develop Levels of Processing in order to understand why some info is available to some ppl & not to others

-Shallow processing: phonological info, sound level of speech

-Deep Processing: semantics; meaning & associations of words

*their idea is that the more deeply you process info, the better your recall 

_______________________________________________

Hyde & Jenkins (1973)

-Tests Levels of Processing further

-give word list & many incidental encoding conditions 

-Each encoding condition is encoded deeper than last (part of speech, word freq., pleasantness)

[image]

**Performance for control condition is slightly worse than intentional semantic encoding therefore suggesting that there is a difference btwn shallow and deep encoding**

Problems with Levels of Processing

1. Deep processing provides a circular argument

2. Transfer-Appropriate Processing

1. how do we know if something is encoded deeply?==> B/c % recall is better. Why is % recall better? ==> B/c it is encoded deeply.

2. To define encoding as good or bad is useless because it depends on the retrieval process suggesting significance between interaction of encoding and retrieval circumstances

This concept was developed by Morris, Bransford, & Franks (1977)

-Fill in blank task and provide a possible word. Does it fit? 

-the possible words make up the word list that pts. will need to recall later.

-had 2 conditions: a shallow rhyming task, and a deep semantic task

When results were compared: pts. from cond.1 are awful at rhyming recall when they encoded semantically therefore suggesting that deep processing does not always result in better recall

***This realization of the importance of the interaction btwn encoding and retrieval suggests that context matters and leads to ===> Context Effects

Context Effects

1. Encoding Specificity

2. State-Dependent Learning

1. There should be interaction btwn encoding & retrieval

Goddard & Baddely (1975)

-look @ scuba divers 

-have them remember word list on land & in water

-Find: that divers are equally as bad at opposite tasks

[image]

_______________________________________________

2. emotional state at encoding leads to improved recall if pts. is in same emotional state @ recall

Watson & Tulving 

Cued recall experiment

- Participants have 2 tasks: 

1. They study weak paired-associate word lists

(i.e. Glue--Chair)

2. They are given associates and asked for associations from task 1

3. Test: Give a cued recall

-When asked if chair was on the list pts. have poor recall

-when given glue, recall of chair (it's paired associate) is good

**Suggests that having glue as context for remembering chair is so useful that it goes beyond a recognition test**

Raises question: Why does having extraneous info @ encoding lead to better retrieval?===>Retrieval cues

1. Retreival Cues

2. Evidence for Retrieval Cues

1. idea that a single memory is more accessible (easily retrieved) if it contains "pointers": bits of info associated to the target memory

Note: these associations can be made over a lifetime or during the short time of a study

_______________________________________________

2. Mantyla (1986)

-Incidental encoding w/long word list

-Pts. asked to come up with 3 associates for each word on list but not told that they need to remember the list itself

Condition 1: cue words w/pts. own associates

Condition 2: cue words w/ a different pts' associates

Condition 3: cue words for pts. who have not seen the word list (no training phase)

[image]

**if ppl are 90% accurate when they have the correct info in association web, why aren't they always @ 90% and where is the missing 10%? **

To answer this question we must focus on retrieval AND forgetting

Forgetting: 

1. Ebbinghaus' % savings function

1. Every forgetting process seems to happen with a steady decline & a leveling out over time regardless of the time measure (seconds/mins/hrs/days/yrs etc.)

-this suggests that there is less importance in the distinction btwn STM & LTM

-Ebbinghaus sort of discovers a universal law for forgetting

[image]

*Sometimes we get forgetting functions which break this 'universal' falling function*

1. Massed Practice

2. Distributed Practice

1. "Cramming" within a certain time frame when you know the task is approaching soon

2. Same time frame as massed practice, but spread out

- This results in a lower performance level overall, but a much less steep forgetting function

[image]

Possible Reasons behind Forgetting:

1. Encoding Failure

2. Storage Failure

3. Retrieval Failure

1. Info is interrupted upon encoding

2. Info decays in storage

3. Info is gone upon retrieval

Encoding Failure Theories

1. Consolidation Theory (1900: pre-modal model)

(proposed by Müller & Pilzecker)

2. Perservation

___________________________________________

3. Jenkins & Dallenbach (1924)

______________________________________________

4. Ribot's Law

1. idea that active processes for encoding (studying/repetition etc.) are not all that is involved in encoding process

2. idea that encoding processes continue after explicit studying 

-This persistence leads to a consolidated memory

_______________________________________________

3. -Have pts. memorize word list and wait 2, 4, 6, 8 hrs. before testing pts. 

-Cond. 1: have pts wait during the set retention intervals

-Cond. 2: Have pts. nap during the set retention intervals

[image]

**Better recall in condition 2 b/c it allows perservation to go on uninterrupted**

_______________________________________________

4. events learned early on in life are recovered easier than right before a traumatic accident/event

Storage failure & Retrieval failure theories

1. Mcgeooh (1932)

2. Interference Theory

a) Proactive Interference

b) Retroactive Interference

3. Paired Associates Task

-Response Interference

1. argues that saying memories decay because time has passed is like saying time causes metal to rust

2. memory does not decay over time, but dissipates as a result of interference 

a)when previously learned info interferes w/recall performance on later tasks

b) new info interferes w/recall performance of previously learned info.

3. Mem. an associated word list

-Dog(A) - Tree (B) = task 1

-Dog(A) - Water (D) = task 2

-pts. are tested on recall of A-B 

Response Interference: the idea of retrieval competition meaning B & D are competing upon retrieval suggesting that the stronger association will beat the weaker one

=====> Melton&Irwin's research

FORGETTING

Melton & Irwin

-if Response Interference(retrieval competition) is true then the stronger association should win every time

-use word list recognition to test response interference & retroactive interference

-have pts. memorize 2 associated word lists (A-B:task 1) and (A-D:task 2)

-5 conditions for task 1 consist of 5 practices each condition

-5 conditions for task 2 consist of: relaxation condition, 5 practices, 10 practices, 20 practices, and 40 practices

-Participants are tested on (A-B) list

* If response interference is valid, there should be differences in results from condition 3 on: ppl. should respond with D b/c it is the stronger association*

Findings: @ condition 5(40 A-D practices), recall is better for B suggesting that response interference is not that prevalent

-They conclude that not all memory errors are due to interference and not all stronger associations beat out weaker ones

FORGETTING:

1. Hickens, Born, & Allen (1963)

2. Brown-Peterson Task

1. show release from proactive interference

- use Brown-Peterson Task: where pts. are asked to remember non-words

-Then given distraction task 

-Find: w/ each individual trial, % correct goes down due to proactive interference

**When pts. are asked to remember a 3-digit number after non-words, they show release from proactive interference** suggesting that not all memory errors are due to interference

Evidence for Decay from Storage

Big problem: 

-the only way we can access storage info is via retrieval

-We must keep in mind that Absence of evidence =/= evidence of absence

-We can test pts. implicitly, explicitly, and monitor over extended period of time, but we still cannot just assume that the info is not stored

Target info that is encoded, is not encoded alone; It comes alone w/context therefore

1. Forgetting is failure to be in "proper state" @ retrieval

2. Forgetting is failure to discriminate target information from "irrelevant contextual info"