COMUNES DEL AREA GENERAL GPO. K - 51681 TURNO MATUTINO
I. C. OPTATIVAS GRUPO C - 51359 TURNO MATUTINO
This experiment examined, whether covert attention is influenced by referring language and if so, whether that is dependent on volitional control or internal goals. Participants
Figure 3.1.: Display of item 108, the words used are wheel and boot
had to detect a target object appearing in one of two boxes. Preceding the target, they first saw two object photographs in the boxes and subsequently heard a spoken word that referred to one of those objects. The cue thus consisted of the object photograph that was referred to, and the word that was referring. In a compatible cue trial, the target appeared in the box that was previously occupied by the object that was referred to, whereas in an incompatible cue trial, the target appeared in the opposite location. In case of a cueing effect, responses are expected to be faster in the compatible condition.
In addition to trial type, the stimulus-onset-asynchrony (SOA) was varied between 200, 500 and 800 ms. On the one hand, this manipulation aimed to maximize the chance of detecting an effect, because it is unclear how the SOAs from previous research would translate to the use of spoken referring words. On the other hand, this manipulation enabled us to assess how quickly an effect emerges with automatic processes expected earlier then those under volitional control. To directly test whether an orienting effect is dependent on volitional control the predictiveness of the cue was manipulated between participants: In the predictive condition, the cue was compatible in 75% of the trials while in the uninformative condition the cue was equally likely to be compatible or incompatible.
In the predictive condition, the optimal strategy in order to perform well on the task was thus to actively orient towards the cue, whereas in the uninformative condition, the best strategy was ignoring the spoken word. In addition, the temporal delay between the presentation of picture and spoken word further discouraged the volitional forming of a connection between the two.
Figure 3.2.: Procedure of Experiment A1 & A2
3.1.1. Method
3.1.1.1. Participants
Twenty-eight students from the University of Edinburgh participated in the experiment for course credit. They were all native speakers of British English. Age ranged from 18 to 32 with a mean of 19.18. Seven participants were male.
3.1.1.2. Materials
192 experimental items, 48 catch items and 10 practice items were created (see A.1 for a list of all experimental items). An experimental item consisted of two object photographs taken from the commercial collection Hemera Photo Objects and two pre-recorded words referring to them. The photographs measured 120 x 120 px corresponding to 3.4◦ of visual angle. They were displayed 250 px to the left or right of fixation (7.1◦ of visual angle). The words were all British English monosyllabic picturable nouns and they were matched within items for frequency (Leech, Rayson & Wilson, 2001) and length. In a given trial, only one of the word recordings was used, there were thus two versions of each item naming either one
of the objects. Those two versions were used in separate lists for counterbalancing. In order to construct a sufficient number of items, all pictures and part of the words were repeated once during the experiment in new pairings. Practice items were similar to experimental items, except they were not matched for frequency and length. Each experimental item was randomly assigned to one of the six conditions resulting from crossing TrialType (compatible, incompatible) with SOA (200, 500, 800 ms). In each of the six conditions, the named object was equally often on the left or right side of fixation and also the target appeared on both sides of fixation equally often. Predictiveness was manipulated between participants. In the lists used in the predictive condition, 75% of experimental trials were compatible and 25% were incompatible. In the uninformative condition, 50% of the experimental trials were compatible and 50% incompatible. In a catch trial, there was no target. Those were included to prevent participants from responding habitually.
The lists were randomized individually for every participant.
3.1.1.3. Procedure
Participants were seated in front of an Eye-link 1000 remote eye-tracker with a viewing distance of 80 cm to a 20 inch monitor. As illustrated in Figure 1, each trial started with a fixation cross in the middle of the screen and two empty boxes on the left and right side of the fixation cross. After 500 ms the pictures appeared in the boxes and stayed there for 2000 ms. 500 ms later the word was played back over loudspeakers. Depending on the SOA condition, 200, 500, or 800 ms after word onset the target, a small grey circle, appeared for 100 ms in one of the boxes and participants were asked to press the space bar on the keyboard as fast as they could when they detected it. The trial terminated with the participant’s response or after 1500 ms. After a delay of 1000 ms the next trial started automatically. After every 24 trials, 9 times during the experiment in total, participants were given the chance to have a short rest before going on.
Participants were instructed to keep their eyes fixed on the fixation cross throughout the trial. To make sure they attended to the pictures and tried to identify them, they were told there would be a memory test for the pictures after the experiment. They were informed about the probability of the target appearing in the cued location (50-50, 75-25 respectively) and suggested to use the optimal strategy: In the uninformative condition they were encouraged to ignore the word and only pay attention to the target whereas in the predictive condition they were encouraged to pay attention to the word and use it as a hint of where to expect the target.
At the beginning of the experiment there was a practice phase including 10 trials.
Figure 3.3.: Mean RTs of Experiment A1 by Predictability, SOA, and TrialType
During practice, subjects were provided with feed-back if they looked away from the fixation cross and if they made a response in a catch trial.
3.1.2. Results
Saccades away from the fixation cross were infrequent (15%) and did not affect RTs.1 The overall error rate was very low (< 5%) and was not affected by the experimental manipula-tions. RTs further than 2 standard deviations from each individual participant’s mean were removed (< 2%). We conducted repeated measures ANOVAs on the averaged RTs with the with-in participants factors TTYPE(compatible, incompatible) and SOA(200, 500, 800) and the between participants factor PRED(predictive, uninformative) (Figure 3.3). The compound cue of referring word and previously viewed object triggered an orientation to the cued location which was evidenced by shorter RTs in the compatible cue condition (F(1,26) = 40.18, p < .001). There was also a main effect of SOA (F(2,52)
= 45.79, p < .001) due to shorter RTs with longer SOAs, but no interaction between TTYPE and SOA, indicating that the orienting effect was not dependent on any specific SOA. The main effect of PRED was not significant (F < 1), but there was an interaction between PRED and TTYPE (F(1,26) = 11.21, p < .01) due to a smaller difference between compatible and incompatible trials in the uninformative condition (8 ms) than in the predictive condition (24ms). A follow-up ANOVA on the subset in the uninformative condition confirmed the main effect of TTYPE (F(1,13) = 12, p < .01) for that group.
1We conducted a separate analysis which excluded those trials in which participants executed a saccade away from the fixation cross. It confirmed the analyses reported here.
Although the graph suggests a trend towards a more pronounced cueing effect with increas-ing SOA for the predictive group only, the three-way interaction of TTYPE, PRED and SOA was not significant (F < 1).
The results show that participants oriented their attention covertly in response to referring language in a situation where eye movements were suppressed. This orienting effect was neither entirely dependent on volitional control, nor on the intention to process the linguistic stimulus: Finding the effect in the uninformative group rather suggests that picture and word were integrated automatically although they were not co-present and the task was entirely independent of picture and word. In addition to this automatic effect, the results suggest a modulation of the orienting effect by volitional control or internal goals arising from the processing of the word. The interaction revealed that the cueing effect was stronger in the situation where it was relevant to the task and participants were encouraged to make use of it. In summary, the results suggest an automatic influence of a referring word on covert visual attention which is enhanced by either volitional control or the presence of internal goals arising from linguistic processing.