DISCRETE VS CONTINUOUS PROCESSING - THE FATE OF AN INCOMPLETELY PROCESSED PERCEPTUAL DIMENSION

A relevant issue in the debate on continuous vs. discrete processing o f information is whether incompletely processed information does or do es not affect a subsequent reaction. Two stimuli (SL, SR) were present ed on eye level, subtending a visual angle of 45 degrees. SL was alway s inspected first, followed by a saccade to SR and finally by a same/d ifferent response. The fixation time of SL (TL), the saccadic time (TM ) and the time from fixating SR to the response (TR) were separately m easured. SL and SR consisted of two-dimensional stimuli (size and lett er shape) constructed in such a way that encoding size took longer for one group and encoding shape took longer for another group of 10 subj ects. All subjects were tested in three conditions: Shape was relevant in one, size in the second, and both dimensions were relevant in the third condition. TL was less when encoding the relevant dimension was fast. When both dimensions were relevant, TL was about as long as when only the slow dimension was relevant, suggesting parallel and interfe rence-free processing during TL. When only the slow dimension was rele vant, TR (same) was much longer when the fast dimension differed. When the fast dimension was relevant, TR (same) was slightly longer when t he slow dimension differed, which can be handled by either model. The experiment was repeated with three well-practiced and less variable su bjects who carried out sufficient trials to measure TR as a function o f TL. The results of this study were in line with the discrete model: A different slow and irrelevant dimension did not affect the same resp onse regardless of the duration of TL. Interestingly, subjects were ca pable of retrieving the slow dimension, suggesting a code which can be used for retrieval but which does not affect the same/different respo nse.