shuffled scenes real textures contact: jennifer.dimase@yale

The role of attended objects in picture recognition memory DiMase, J.S.1, Oliva, A.2, Horowitz, T.S.3,4, Wolfe, J.M.3,4

1 Yale University, 2 Michigan State University, 3 Harvard Medical School, 4 Brigham and Women's Hospital Is attention to pictures required for later recognition?

PURPOSE

Do different attentional demands have different effects?

STIMULI

S C R A M B L E D S C E N E S & " R E A L " T E X T U R E S

T E X T U R E M E M O R Y & S E A R C H

CONCLUSIONS

When scenes are fixated butnot attended, memory isreduced to the level of recognition for attendedmeaningless textures.

What is the nature of the costs totexture recognition? Would search and tone costs be distinguishable if they did not reducememory performance to chance?

Attention to the visual search task markedlyreduced recognition memory for scenes.

Single task memory for meaningless textures is comparable to memory for meaningful scenes under dual task conditions. But there was also a search task cost for textures.

Is the cost dependent on the "meaning" of the scene?Does the "preattentive" information come for free?

Performance on the tone task was ~90%

What if the meaning is lost but the preattentive information isleft intact?

Recognition of scrambled scenes is the same as with textures.But again, the search task interfered.

SCENES TEXTURES

SHUFFLED SCENES

SHUFFLEDREAL TEXTURES

methodS

contact: jennifer.dimase@yale.eduThis research was supported by a grant from NIMH.

Present stimulus (scene or texture) with search display (and tones if Tone Task)

Search: 0, 1, or 2 Fives among Twos?Tone: 0, 1, or 2 High among Low Tones?Memory: Press Key to Continue

Present stimulus (scene or texture)Memory Task: Old or New Image?

+

--Unlimited Response Time--

--Unlimited Response Time--

--500ms--

S C E N E M E M O R Y & S E A R C H

How well are scene photographs remembered whenattention is focused on another task?

SCENES TEXTURES

MEMORY ALONE

MEMORY + SEARCH

0.3

0.45

0.6

0.75

0.9

1

Search Accuracy

single dual(chance)

0.3

0.45

0.6

0.75

0.9

1

Search Accuracy

single dual(chance)

What would happen if the textures were even more basic (e.g., single features like "red" or "vertical")?

What if we used a dynamic attention task and manipulated the attentional load by varying the number of target tracking items?

For robust picture recognition,both preattentive visual properties and meaningfulinformation are required.

Could the difficulty recognizing images be due to a dual task cost?

M E M O R Y & T O N E M O N I T O R I N G

FUTURE

DIRECTIONS

MEMORY ALONE

MEMORY + SEARCH

SCENES TEXTURES

D'

Picture Recognition Accuracy

The visual search task imposes a cost above and beyond the dual task cost in scenes. Both search and tone tasksreduce texture memory to chance.

While this performance is poor,it is still better than chance, suggesting that some preattentive information is encoded.D'

SCENES TEXTURES

D'

MEMORY ALONE

MEMORY + SEARCH

MEMORY & TONE

TRAINING/SEARCH TRIALS (32)

PICTURE MEMORY TRIALS (32)

Single Task Search (N=14) Single Task Picture Memory (N=14) Dual Task Search & Picture Memory (N=14)Dual Task Tone & Picture Memory Control (N=10)

CONDITIONS