Purpose. We measured observers' temporal resolution for binding visual features presented in the same location and in different locations.
Methods.
We varied the temporal frequency at which conjunction information
was presented to determine observers' temporal threshold for
reporting conjunctions of orientation and color and orientation
and brightness. In the case of orientation and brightness, for
the same-location condition we alternated a bright
leftward-tilted gabor with a dim rightward-tilted gabor or a dim
leftward-tilted gabor with a bright rightward-tilted gabor and
asked observers to report whether the leftward-tilted item was
bright or dim. The luminances of the gabors were chosen such
that no conjunction information was available in their sum. In
the different-location condition, we presented a leftward-tilted
gabor adjacent to a bright patch, alternated it with a
rightward-tilted gabor adjacent to a dark patch, and asked
observers to report which features occurred together.
Results.
When the features were presented in the same spatial location,
observers could report conjunctions of brightness and orientation
accurately even at stimulus temporal frequencies of 30 Hz, and
conjunctions of color and orientation at stimulus temporal
frequencies of 15 Hz. When the features were presented in
different locations, the mean threshold frequencies fell to 3 Hz
for brightness and orientation and 3 Hz for color and
orientation.
Conclusions.
The process that binds features presented in the same location
operates very quickly, faster than many estimates of the temporal
resolution of visual attention. In contrast, a slow attentional
process is required to detect the cooccurrence of features
presented in different locations.
|
|
