Stimulus control of visual fixation duration in a single saccade paradigm.

Subjects performed a task in which they successively inspected two stimuli presented at an equal distance of the left (SL) and the right (SR) of the visual meridian and subtending a 45 degrees visual angle. This was followed by a joint response on the basis of the status of both stimuli. The instruction was to fixate the position of SL at the start of a trial, followed by a single saccade to SR. In earlier studies on this paradigm it was suggested that the fixation time of SL (TL) may serve as a modern version of the (d)-reaction in which the time for achieving perceptual identification is measured uncontaminated by decision. This suggestion has the implicit assumption that, at least in this single saccade paradigm, the saccade from SL to SR is triggered when perceptual identification has been completed. There is the potential alternative that the duration of TL is programmed in advance, depending on the general processing demands of the stimuli in a block of trials. These two options were tested in two experiments. In the first, presentation of SL was delayed with a fixed or with a variable interval (200-400 ms) during a block of trials. In the second, the stimulus quality of SL was varied between and within blocks of trials. The results of either manipulation argue against preprogramming TL, and they are consistent with the hypothesis that the saccade is triggered upon completion of perceptual processing. Thus, TL in the single saccade paradigm appears a serious candidate for the (d)-reaction.

Signal processing during and across saccades.

Visual perceptual processing has been found to occur exclusively during fixations of the eye (Sanders and Houtmans, 1985; Sanders and Rath, 1991). Does fixation time also reflect postperceptual processes such as target classification and response selection as well, or can these processes continue during a saccade? In a series of experiments on this question two signals were presented at an angular distance of 100 degrees. At the start of a trial subjects fixated the left signal (S1), which was always a letter that did or did not belong to a predefined target set of variable size. Then they moved to the right signal (S2) which indicated the mapping of target and no target responses onto the response keys. Subjects were capable of starting the saccade upon identification of S1, so that target classification of S1 might occur either during the saccade or while fixating S2. Although the duration of the saccade was long enough to complete target classification, the effect of set size was still partially reflected in the fixation duration of S2. This could either be due to incomplete processing during the saccade or to interference while integrating S1 and S2. The results of two additional conditions were in support of the latter hypothesis, since they showed similar effects while no saccade was involved. Subsequent experiments confirmed the conclusion that target classification can continue during a saccade. The final experiment showed that response selection can continue during a saccade as well: The effect of S-R compatibility was reduced in case of a small saccade (8 degrees) and disappeared altogether when a large saccade was carried out (100 degrees).