Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
35 Cards in this Set
- Front
- Back
|
Semantic memory |
Memory for categorical, factual information |
|
|
Characteristics of Semantic Memory (3) |
1. Organized by content (similarity) 2. Allows inferences (logical, hierarchical order) 3. Generalizes beyond a single instance |
|
|
Lexical Decision Task |
Test for semantic relatedness
Words related to a prime were retrieved faster than unrelated words. (implies LTM structure) |
|
|
Hierarchical Model |
for Long Term Memory
Nodes (concepts) Links (directed subordinate --> superordinate) Activation tags (verify inferences) "yellow" "skin" |
|
|
Typicality Effect |
People verify typical instances of a category faster than atypical examples.
Some links stronger than others
Criticism of hierarchical model #1 |
|
|
Violations of Hierarchical Order |
Some activations go around hierarchical order.
"Penguin is an animal" is faster than "penguin is a bird." Criticism of hierarchical model #2 |
|
|
Relatedness Effect |
Speed to respond "false" in hierarchical model should only depend on links. But similarity is considered
"Is a bat a plant" is faster than "is a bat a bird" Criticism of hierarchical model #3 |
|
|
Criticisms of Hierarchical Model (3) |
1. Typicality Effect 2. Violations of Hierarchical Order 3. Relatedness Effect |
|
|
Which type of memory did Tulving call "Mental Time Travel"? |
Episodic memory Association between memory and its source. |
|
|
Retrograde Amnesia |
New information can't be learned Episodic or semantic Patient H.M. |
|
|
Retrograde Amnesia |
Old episodic memories can't be retrieved "TV amnesia" |
|
|
Confabulation |
Fabricated, distorted or misinterpreted memories esp. around time of injury |
|
|
Korsakoff's Syndrome |
After prolonged alcoholism Severe anterograde amnesia Temporally-graded retrograde amnesia Episodic and semantic (they may be similar) |
|
|
Episodic / Semantic Memory Interactions & Similarity |
1. Korsakoff's - affects both. (they are similar) 2. Patient KC and Italian woman - double dissociation (they are distinct) 3. Episodic - builds up, generalizes in childhood, becomes base for semantic. Semantic strengthens with repetition. Episodic is noticed when novel, unusual. |
|
|
Declarative Memory |
Knowing that something is true Semantic Episodic |
|
|
Procedural Memory |
Knowing how to do something A type of implicit memory |
|
|
Patient H.M. |
Anterograde amnesia Declarative knowledge (both episodic/semantic) Procedural memory OK |
|
|
Mirror Tracing Test |
Demonstrated difference between declarative and procedural memory
HM improved on tracing task but had no explicit memory of the task. |
|
|
Explicit Memory |
Conscious awareness of an event or meaning. Can be recalled or recognized
Includes declarative, episodic, semantic |
|
|
Implicit Memory |
Latent learning Classical conditioning Priming Activation without conscious awareness |
|
|
Repetition Priming |
Tested explicit vs. implicit memory Amnesic, regular, alcoholic patients Recall (explicit) Recognition (implicit) Amnesics did worse on recall, equally well on implicit memory task . |
|
|
Propaganda Effect |
People are more likely to rate statements as true if they have seen them before.
Even if they were labelled false
Implicit memory (may not remember seeing it) |
|
|
Adaptive Control of Thought (ACT) Theory |
General, hierarchical model Basic representational unit: proposition Processing: spreading activation Information is discrete and modular Model for LT memory |
|
|
Parallel Distributed Processing (PDP) Model |
Information is distributed, has more plasticity Model for LT memory |
|
|
Spreading activation |
(ACT model of LTM) Nodes activated in sentence processing Most activation -> working memory Spreads along links Stronger links = more activation
|
|
|
Proposition |
Smallest unit of meaning in ACT model of LTM True or False Representational unit, encodes declarative facts 1. Node (concept, idea) 2. Link (association): agent, relation, object |
|
|
Type-Token Distinction |
(ACT model) 1. Type - general concept / node / semantic Class of objects
2. Token - specific instance of type / episodic
|
|
|
Fan Effect |
(ACT model) Longer time needed to recognize (decide True/False) when there are many links, a lots of info retrieval required.
(more distractors = slower retrieval of right info) Amount of activation leaving node is divided by all links exiting the node. |
|
|
Expertise, Plausibility, Fan Effect |
Expertise = lots of info; faster retrieval of general info and plausibility.
Slower to recognize single fact because lots of information retrieval. |
|
|
Spreading Activation : Implications (3) |
1. Relvance = historical activation patterns 2. Computational complexity = lower, not all details activated equally. 3. Confabulation/incompleteness = not all info is retrieved. |
|
|
Parallel Distributed Processing (PDP) |
(Connectionist model)NEURALNET 1. Processing in distributed layers 2. Neurologically inspired 3. Processing = changing weights of inhibitory/excitatory links between units 4. Training = episodic memory 5. Weights = semantic memory
Neural net. Fire/wire together. Add concept = change existing units |
|
|
Valence |
Links between units in PDP are either excitatory or inhibitory. |
|
|
PDP Network |
Same net can = different semantic content Event (episode) = input Weights btwn units change Multiple encodings strengthen connections Semantic (conceptual network) built Context automatically stored Partial input = partial cues to net Resilient to damage |
|
|
Should PDP models have fan effects? |
Yes Initially, many instances, no training, weak connections between units => Slow and incorrect responses.
With training, more inputs = expertise = strong connections => faster, more correct responses. |
|
|
Could there be overwriting in PDP? |
Maybe: If you change the weights between units, erase the original memory. More units = less of a problem? (distributed)
Not a problem in ACT. |
