Flashcards in Week 4 Visual Memory Deck (21):
What is the partial report method and how did Sperling (1960) use it to obtain evidence for the existence of iconic memory?
In the partial report method the observer reports only the part of the stimulus that is cued. For example, Sperling (1960) arranged a number of characters into three rows and had observers report only one row (partial report method), as signalled by a tone, played after the characters were presented. Since observers did not know in advance which row they would need to report, he could estimate their total memory capacity based on their performance on reporting the cued row. He found that immediately after presentation, observers had an essentially unlimited memory capacity but this memory decayed very rapidly. He called this type of memory iconic.
What are the principle characteristics of iconic memory Sperling (1960)?
2. Not perfect
3. Lasts for a very short period of time (about a quarter of a second)
4. Not robust, can be destroyed by eye blinks, flashing something on the screen, light in the background, movement of eyes.
How did Luck & Vogel (1997) use a change blindness paradigm to estimate the capacity of visual short-term memory (VSTM)? How large did they find the capacity to be? Did this capacity depend on the complexity of the items?
(Experiment 1 &2 )
Luck and Vogel (1997) presented observers with an array of coloured squares, followed by a blank interval, after which a test array was shown. On half of the trials, the colour of one element of the test array was changed. Observers were asked to indicate whether a change had occurred. By measuring the percentage correct as a function of array size, Luck and Vogel were able to estimate the capacity for visual short-term memory (VSTM). They found this capacity to be approximately 4-5 items, irrespective of the complexity of the items.
Luck & Vogel (1997) Experiment 3, 4 & 5 conclusion on VSTM:
VSTM is object-based (performance decreased with increasing sample size, and not increasing features), and VSTM is more efficient in remembering features belonging to same object. There is an OBJECT ADVANTAGE.
How did Alvarez & Cavanagh (2004) measure object complexity?
Alvarez and Cavanagh (2004) had observers perform a series of visual searches for each item and measured how the reaction times varied as a function of set size. They found that the more complex the object, the steeper this search slope and they used these search slopes as a measure of object complexity.
How did Alvarez & Cavanagh (2004) determine whether object complexity affects VSTM capacity and what did they find? What did they conclude?
They then repeated the change detection VSTM paradigm of Luck and Vogel, but used different items. They found that, in general, VSTM capacity decreased as the complexity of the objects increased. They concluded that the objects used by Luck & Vogel were too simple to show the effect of object complexity on VSTM capacity. However, object advantage still stands.
What is the main conclusion of Todd & Marois (2004) study?
Their main conclusion is that activity in the posterior parietal cortex is tightly correlated with the number of objects stored in VSTM. The more objects that are stored, the higher the BOLD fMRI activity. It is likely that the posterior parietal cortex is involved in VSTM (along with others).
How did Brady et al. (2008) measure visual long-term memory? What did they conclude?
Brady et al. (2008) presented observers with pictures of 2,500 real world objects, one at a time, 3s each, whole experiment lasted for 5 hours. The observers’ task was to indicate identical repeats. When finished, 300 test images were shown.
There were 3 conditions: Novel (93%), Exemplar (88%), and State (87%).
Even for state pairs (i.e. the same object but in different states), performance was very high (87%). They concluded that visual long-term memory has a huge capacity and is highly detailed.
What are the other brain areas involved in VSTM?
Activity in bilateral Intraparietal/Intraoccipital Sulcus correlated with number of items remembered (act of encoding) — thus seems to play a key role in VSTM
Activity in ventral-occipital (VO) cortex correlated with complexity of stimulus and decision process.
Activity in anterior cingulate (AC) cortex correlated with just the decision process (retrieval process).
How did Cohen et al. (2009) measure auditory long-term memory? What did they conclude?
Participants listened to 64 5-second sound clips. Immediately afterwards, they listened to 64 additional clips, half old and half new. They were asked to indicate which had heard before. Performance was poor (78% accurate, chance performance = 50%). Cohen et al. concluded that auditory long-term memory is much worse that visual long-term memory.
What is the difference between Sperling's (1960) full report and partial report?
Full report: A full array of characters is shown and observers have to recall everything. # of items reported around 3.8-5.2 varying in each person regardless of amount of items displayed.
Partial report: Reporting a row of characters that has been cued. Through this partial report, he found that capacity of iconic memory is endless (although participants were forgetting it rapidly as they write it down).
Sperling's full report experiments: (1) and (2)
Despite varying the presentation time of the stimulus, the average number of items recalled is the same , which is around 4-5 items.
What are the characteristics of VSTM?
- lasts for several seconds
- High capacity limited (note affected by lengthening presentation duration)
- Robust, not easily destroyed by eye movements, etc.
- Requires time too form 20-50ms per item
- Adding verbal load did not decrease VSTM capacity showing that verbal memory is separate from VSTM
What was the confound Luck and Vogel (1997) had to address and how did they fix this?
Confound: Performance decreases with increasing set-size as the number of comparisons between remembered sample array and test array also increases. This would result in an increased chances of error.
Confound addressed by cuing just one object and asking if this object had changed. Thus, number of comparisons then does not vary with set- size.
What experimental manipulation did Todd & Marois use in their iconic memory experiment to show that the BOLD fMRI activity? What were their results?
They performed an identical VSTM experiment with 2 conditions, one with a probe display (VSTM experiment) and one without a probe display (iconic experiment) where subjects had to make an immediate judgement about the presence/absence of a disk (thus removing the memory requirement). By subtracting the activity of the 2nd condition from the 1st condition they removed the visual component and retained the memory component. This manipulation removes the visual component and only measures the BOLD fMRI activity when exposed to the memory requirement. Performance were 96-98% regardless of set size.
What did Brady et al. (2008) conclude about why the subjects did not code the objects VLTM in Luck & Vogel (1997)?
In the Luck & Vogel study, observers were required to notice changes in orientation. VLTM is not sensitive to rotations. This is why the subjects in Luck & Vogel could not use their VLTM to perform the task.
How was Konkle et al. (2010)'s different from Brady et al. (2008) detection experiment and what did she conclude?
She performed a repeat detection task of 2,800 images where subjects are presented with max 16 different exemplars of the same category of object. She measured interference slope and found that performance decreased with an increasing presentation of exemplars. She also found that conceptual distinctiveness correlated with interference slope and that perceptually distinctiveness has no effect at all.
Conclusion: People can remember a vast number or conceptually distinct images, even when they're perceptually similar. Perceptual distinctiveness does not aid memory.
Whats the difference between conceptual distinctiveness and perceptual distinctiveness? Give an example.
Conceptual distinctiveness: What does this mean to me?
Perceptual distinctiveness: does this look different?
Example: a royal gala apple vs. fuji apple. perceptually they are different but conceptually they are the same. because they are both apples and they mean the same to me.
What was a caveat of Konkle et al. (2010)'s findings?
This finding seems to be specific to visual memory. It does not hold for auditory memory.
What did Cohen et al. (2009) found in the 2nd experiment about auditory and visual long term memory?
Same as experiment 1 except better sound quality and 5 conditions:
– Sound = just heard sound (i.e. same as Exp 1)
– Sound + Picture = In training phase heard sound and
saw picture. In test phase just heard sound.
– Sound + Name = In training phase heard sound + saw name. In test phase just heard sound.
– Name = just saw name.
– Picture = just saw picture (i.e. similar to Brady et al.)
They found that performance was best in the picture condition (visual memory!!).