DEAD ZONE OF VISUAL ATTENTION REVEALED BY CHANGE BLINDNESS
Igor S. UtochkinHigher School of Economics, Moscow, Russia
Two patterns of attentional allocation around a focus
Focus
Periphery Periphery
Gradient pattern (Castiello & Umiltà, 1990; Downing, 1988; Downing & Pinker, 1985; Eriksen & Hoffman, 1972; Eriksen & St. James, 1986; LaBerge, 1983; Mack & Rock, 1998, etc.)
Focus
Periphery Periphery
Inhibitory-surrounds pattern (Bahcall & Kowler, 1999; Cave & Zimmerman, 1997; Caputo & Guerra, 1998; Cutzu & Tsotsos, 2003; Mounts, 2000a, b; Müller et al., 2005; Thakral & Slotnick, 2010, etc.)
What does determine pattern of allocation?
Task difficulty (Thakral & Slotnick, 2010);
Attentional saliency of a centrally attended item (Müller, Mollenhauer, Rösler & Kleinschmidt, 2005);
Attentional set towards central and peripheral items (focused vs. divided attention) (Müller et al., 2005)
Limitation of previous results
Gaze fixation;
Simplistic arrays of discrete objects;
Brief trials;
Attentional salience is controlled through external attentional capture
What about natural perception?
Active looking;
Complex continual layouts of various objects;
Prolonged observation;
Internal control of attentional salience through interests is available
Change blindness paradigmSpatial allocation of attention appears necessary to perceive changes (Kahneman et al., 1992; Rensink, 2000; Rensink et al., 1997; Tse, 2004);
Direct manipulations with interests are available (Rensink et al., 1997; Jones et al., 2003; Werner & Thies, 2000)
Experiment 1
Standard change detection task under flicker conditions: find a single changing detail in a picture as soon as possible and identify it (miss responses are allowed)
Stimulation
12 natural scenes with one attractive object or its part
Central interests (CI)
Marginal interests (MI)
Three change instances per a picture
CENTRAL CHANGE
CENTRAL CHANGE
NEAR CHANGE
Three change instances per a picture
CENTRAL CHANGE
NEAR CHANGE FAR CHANGE
Three change instances per a picture
Design
Independent Variable: Change location (Central vs. Near vs. Far)
Dependent Variables:Search time;Miss rate;Misidentification rate (successful detection but incorrect recognition)
Results
Miss rate Misidentification rateA ‘dead zone’ surrounds center of interest
Is attention to CI what actually causes dead zone pattern?
CIMI MI
If this is the case…
Experiment 2
Looking for a marginal change (near or far) in the presence of a once noticed central change
Results
A ‘dead zone of attention’ surrounds center of interest
Experiment 3
Both marginal changes (near AND far) compete for prior detection
Results
Thank you for your Attention!
For more information about this study see:Utochkin, I.S. (in press). Hide-and-seek around center of interest: Dead zone of attention revealed by change blindness. Visual Cognition.