BCS 111 Lecture 6.4/7.1

Long-term memory retrieval

• Bousfield (1953): free recall

  • Recall a list of 60 words

  • Ss recall words in clusters

    • Similar to foraging (as discussed in Ecological Approach)

Ecological approach to the study of LTM retrieval

Memory search: verbal fluency task: recall as many words as possible in a given category.

Long-term memory - forgetting

• How do you ”flatten the curve”?

• Study, study and study!

• Mechanisms underlying forgetting?

  • Interference vs. decay (as in STM)

Proactive interference

stronger if old pair has stronger connection than new

• (old pairing) Science – experiment

• (new pairing) Science – dinner

  • Q: Recall the new pair.

  • A: Science - experiment

Retroactive interference

new replaces old

• (old pairing) Science – cat

• (new pairing) Science – rain

  • Q: Recall the old pair?

  • A: Science - rain

Retroactive interference PLUS (old) information decay

stronger if the old one learned a long time ago

• (old) Science – cat

… after 2 weeks

• (new) Science – rain

  • Q: Recall the old pair?

  • A: Science - rain

Our knowledge and concept are organized like a hierarchy of categories in semantic memory

Hierarchical network model (a connectionist model)

• Two types of nodes

  • (Sub-)Category labels

  • Features

Results supporting the hierarchical model

• Slower responses to feature description than category name

• Slower responses to features from the top level

• Slower responses to category members from the top level

Concept and categorization

• Classical view

• Knowledge-based

• Similarity comparison

  • Prototype

  • Exemplar

Classical view

• List of features

• Binary distinction: either belongs or doesn’t belong to the category

• No goodness-of-fit: no good or bad member --> No prototype!!

Knowledge-based view of categorization

• New years eve example

Concept and imagery

• Do you have a mental image for every concept?

• Can you form an image of an abstract idea?

  • Your mental image for the concept “abstract”?

Imagery and words

How do we link word with the thing that the word represents?

• Concreteness of the word: hard to form mental images for abstract concepts

• Language-specific: pictorial vs. non-pictorial words

Brain activation of concrete vs. abstract words

Jessen et al. (2000) fMRI study of German words

• Concrete vs. Abstract words

• Task: simply remember each word while being scanned in MRI

Jessen et al. fMRI result: activation in parietal lobe for concrete words: image of object in space (the “where” stream)

Implications:

Mentally “sense” the concrete objects even if the actual object is not presented

Mental rotation: Cooper & Shepard (1973)

Letter identification with various types of cues (conditions)

• Main task: identify whether the letter in the Testing trial is a normal letter or mirror image of the letter

(cont)

Condition type B (both identity and orientation cue) : with an additional factor- 4 different durations of the cue

Mental rotation: Cooper & Shepard (1973)

Time needed to make judgments in 5 different conditions

• No cue: most difficult (slowest RT)

• Both cues available and combined cues (condition B & C): easiest (faster RT)

(cont)

Time needed to make judgments in Condition B (both cues available) with 5 different cue durations

• The longer the duration, the faster the RT

Pardo-Vazquez & Hernandez-Rey (2012)

• The same letter judgment task (but only Condition N, aka. no cues available)

• Working memory task divides Ss into high vs. low span groups

What factors may contribute to the performance on mental rotation task?

• In both groups: The larger the rotation angle, the higher the error rate and RT.

• HighSpan group outperformed the LowSpan group

  • Mental rotation highly correlated with working memory capacity (related to visual rehearsal)

Localization of mental rotation in the brain (Lamm 2007)

parietal lobe and premotor cortex

1. Encode: remember what the shape looks like

2. Indicate: showing the direction of rotation

(the side with the orange bar should be on the bottom after rotation)

3. Rotate: mentally rotate the shape

4. Match: judge if the rotated result is correct

Mental rotation: Shepard & Metzler (1971)

First condition:

• Picture-plane pairs

Second condition:

• Depth pairs

(cont)

• The larger the rotation angle, the larger the RT (slower)

• Rotating by depth a bit easier when rotating by a large angle (although the difference is not very significant)

  • Implication: we mentally rotate not only 2D but also 3D images!