Behavioural Brah{Research, 49 (1992) 189-196 9 1992 Elsevier Science Publishers B.V. All rights reserved. 0166-4328/92/$05.00

189

BBR 01304

The spatial distribution of attention in S - R compatibility G.P. A n z o l a a n d G . B . F r i s o n i Clinica Neurologica, Universith di Brescia, Brescia (Italy) (Received 13 November 1990) (Revised version received 20 March 1992) (Accepted 15 April 1992)

Key words: Stimulus-response compatibility; Selective attention

In certain choice reaction time experiments the subjects, although not specifically instructed to do so, perform a parcellation of space into right and left components. Previous research has shown that when the stimulus locations are marked on the screen, the subdivision of space into right and left halves is bound to the point where the subject has been instructed to focus selective attention. It is not known whether, in the absence of specific instructions, subjects are able to focus selective attention between stimulus locations when no marker is present. We tried to clarify this problem .by introducing a 'probe' stimulus that could map the spatial distribution of attention in subjects engaged in a compatibility task. The results s u r e s t that, in the absence of visual cues marking the stimulus positions, attention is kept in the disengaged modality before the presentation of the stimuli. In this state of disengagement a powerful tendency to orient to the rightmost side of the display, within each visual field, is apparent. Various interpretations of this finding are discussed.

INTRODUCTION

In choice reaction time (RT) experiments the subject is instructed to respond to a particular stimulus by selecting the appropriate response among a set of predetermined alternatives. In a number of pioneering papers, Fitts and associates observed that some stimulusresponse (S-R) pairings yield faster responses than others and the term compatibility was coined to designate this phenomenon 3'6'29. The compatibility effects occur in quite a large variety of experimental conditions in which the task demands may vary considerably. The recent taxonomy by Kornblum, Hasbroucq and Osman 13 provides simple rules to classify in a general frame the various types of compatibility hitherto defined as spatial, symbolic and Simon effect 1'~5'23'2s. In the experimental paradigm that brings about the Simon effect (the type III ensemble in Kornblum et al.'s taxonomy), subjects are typically presented with two geometric figures or two coloured lights and are instructed to press the left key in response to one figure or colour and the right key in response to the other figure or colour. The stimuli are randomly shown in the left or right side of the display. Despite the fact that the

Correspondence: G.P. Anzola, Clinica Neurologiea, Universit5 di Brescia, P. le Spedali Civili 1, 25125 Brescia, Italy.

side is irrelevant, the reaction time (RT) is faster when the side of the response key corresponds to the side where the stimulus appears than when stimulus and responses occur on opposite sides 4'24'25. The same holds when both stimuli are presented in two different locations that can be labeled right or left in the same hemifield. In this case the relative positions of the stimuli bring about the compatibility effect2JI'26 Although the location of the stimulus is irrelevant for the selection of the key, it provides a directional cue that the subject is unable to ignore and that is liable to affect the speed of the response. In principle, when a stimulus appears in any one of two locations in space there are two ways by which the subject can code its position: the first is by reference to some body coordinates, such as the vertical meridian of the retina or the body midline, the second is to select an environmental frame of reference based on spatial landmarks. The available evidence rules out the egocentric frame of reference, since compatibility effects are still obtained when both stimuli appear on the same side of the body midline or of the fixation point 2'11'27. By contrast, Nicoletti and Umilfft have recently found that selective attention may be used to split the representation of space into left and right components 16. In their experiment stimuli could appear in one of six boxes in the same visual hemifield, divided into two groups of three, with a gap separating the two groups. In the

190 absence of specific instructions, the boundary marking the passage from compatible to incompatible presentations was the gap between the two groups of boxes. If, however, the subject was instructed to focus selective attention between the boxes of a group, the boundary now marking the passage between compatible and incompatible presentations was represented by this attentional point. This finding suggests that in the experiments in which the Simon effect is obtained, subjects could make use of an attentional reference axis located between the two locations where stimuli are presented. However a major difference exists between the paradigm employed by Nicoletti and Umilt~. and that of the classical literature on the Simon effect. In the former case, boxes marking the locations of the stimuli were present throughout the session, whereas in most experiments on the Simon effect, stimuli were delivered in an empty field 26"27. This may be crucial, as the available evidence suggests that in an empty field, attention is distributed over the whole visual display instead of being focused on a point in space 5"7'8. In the present study we tried to directly test the interpretation of Nicoletti and Umilt~ with a visual display that allowed no visible marker of the locations of the stimuli. The main question was: when a subject engaged in a compatibility task is expecting a stimulus without visual cues of its locations, is he/she still able to focus attention between the two stimuli? With the aim of clarifying this issue we employed a procedure derived from Posner and associates ~8-2~ and we mapped the spatial attention of subjects engaged in a compatibility task of the Simon effect type. The Posner paradigm allows for different predictions according to whether attention is focused or diffused. In the former case, if the subjects allocate their attentional focus between the two presentation sites, the RT to a 'probe' stimulus flashed in that location - - center of the attentional focus - - will be shorter than the RT to the same stimulus presented anywhere outside the attentional focus. In the latter case, if attention is kept diffused on the whole visual display, the location of the 'probe' stimulus is irrelevant and RT to various probe locations will not differ from each other.

EXPERIMENT 1

Materials and methods Subjects. Sixteen young subjects (8 males and 8 females) took part in Experiment 1. Their age ranged between 17 and 36 years (average 26.7), and the educational level was high school or University. All subjects were right-handed, had normal or corrected-to-

normal visual acuity and no history of neurological disease. They were naive as to the aim of the experiment, and did not receive any monetary reward. Apparatus. The experiment took place in a quiet, dimly illuminated room. Stimuli were presented using an Apple lI/e personal computer and an Apple A2M2010P video. The centre of the video was marked by the appearance of a cross (0~ ' x 0~ ') which was the subject's visual fixation point during the trials. The stimuli were a square (0~ ' x 0 ~ a rectangle (0020 ' x 0~ which represented the stimuli for the compatibility task, and a vertical bar (0~ ' x 1 ~ which represented the 'probe' stimulus for the attentional task; all stimuli could appear in four locations in each visual field, as shown in Fig. 1. The four locations, called location I, II, III and IV, were placed on the horizontal meridian at a distance of 0 ~ 2~ ', 4~ 14'30 ", 5040 ' from the cross. Location I was the centre of the cross itself, i.e. the visual fixation point. The compatibility stimuli could appear either in location II or in location IV, the probe stimulus could appear either in location I (Fix) or in location III (Mid). The subject sat with both eyes open in front of the screen and his head was positioned in an adjustable head-and-chin rest so that the distance between the eyes and the video was 60 cm. According to the exper-

500 490 480 470 460 450 '

440,

LEFT FIELD

RIGHT FIELD

od IV

I11

II

I

Fig. 1. Experiment 1. Top: reaction times to the probe. Ordinate: RTs in ms. Bottom: sketch of the display showing the probe locations in the left visual field with respect to the compatibility stimuli and the fixation point. The locations of probe and compatibility stimuli in the right visual field are not shown. Roman numerals refer to the location where stimuli are presented: locations II and IV for compatibility stimuli, locations I and III for the probe.

19I imental situation, the subject placed the index finger and middle finger of either the right hand (RH) or left hand (LH) on two contiguous keys of the computer board, '.' and ']' for the RH and 'Z' and '\' for the LH, so that the keys pressed by the RH were in the right hemispace, while the keys pressed by the LH were in the left hemispaee. The experimenter sat close to the subject in order to monitor the subject's eyes and to delete reaction times on trials when the eyes moved from the cross. Procedure. The subjects were instructed to keep their gaze on the cross and press the key corresponding to their rightmost finger (middle finger for the RH and index finger for the LH) when the rectangle appeared, and the key corresponding to their leftmost finger (index finger for the RH and middle finger for the LH) when the square appeared, irrespective of the location of the stimulus. When the vertical bar appeared, the subjects were left free to press either key. While instructing subjects, we however avoided giving any spatial or rightleft cue; accordingly, effectors of the response were not called rightmost and leftmost fingers when talking to the subjects, but simply index finger and middle finger. The subjects were then informed that the compatibility stimuli would be presented on 80~0 of the trials, while the probe stimulus would appear on 20 ~o of the trials. The four possible locations were clearly shown onto the screen by the experimenter. Lastly, we asked the subjects to respond as fast as possible, but without making any errors. Then the subjects underwent an informal session of practice (80 compatibility trials and 20 probe trials) which was not used in the final statistical analysis. Stimuli appeared in the four possible locations (I, II, III and IV) in a quasi-random sequence; each presentation could not appear more than three times consecutively. Overall ratio of presentation between the compatibility stimuli and the probe stimulus was 8 0 ~ and 2 0 ~ , respectively, 4 0 ~ being compatible presentations, 4 0 ~ incompatible presentations, 10~ Fix and 10~ Mid presentations. A single trial began with the presentation of the cross in the centre of the otherwise empty screen and a simultaneous acoustic warning signal (duration of the warning signal = 250 ms); 750 ms after the appearance of the cross the stimulus was flashed for 100 ms, an interval too short to allow the oculomotor reflex. Immediately after the response, the cross disappeared and in the same place the RT (expressed in ms) or, if it was the case, the Italian words meaning 'KEY ERROR', 'ANTICIPATION', ' O U T OF TIME' were presented. This lasted 200 ms, and after a delay of 500 ms, the cross reappeared and another trial began. RTs under 150 ms and over 1,100 ms

were automatically discarded and replaced later in the sequence. The subjects underwent a single session of 4 blocks of 200 trials each. In each block the stimuli appeared in the same hemifield, either right field (RF) or left field (LF). The subjects had to respond with the same hand in one of the four possible combinations: RF-RH, RFLH, LF-LH, LF-RH. Between blocks there was a pause of a few minutes. Situations were balanced between subjects and within session. Four each subject the median of correct RTs to each type of stimulus (square, rectangle, probe) in each location (II and IV for compatibility stimuli, I and III for the probe) was calculated; the median was then the value used in all the following computations. Resttlts The overall error rate amounted to less than 4 ~ ; hence error scores were not subjected to statistical analysis. For each subject, compatible and incompatible RTs in each of the four hand-field situations were calculated, by averaging the medians of left square and right rectangle RTs (compatible RTs) and right square and left rectangle RTs (incompatible RTs). An ANOVA was then performed with fieM (RF and LF), hand (RH and LH), and compatibilio, (compatible and incompatible) as primary source variables. A compatibility main effect was statistically significant (F~,~5= 88.22, P < 0.0001). The average of compatible RTs was 412 ms and the average of incompatible RTs was 434 ms. The field x compatibility interaction was not significant, indicating that the compatibility effect was the same for the two visual fields. The data for the probe presentations were entered into an ANOVA with main effects of position (Fix and Mid), hand (RH and LH) and field (RF and LF). The position was significant (FI.15 = 12.311, P < 0.005), because overall the Mid position yielded faster RTs than the Fix position (447 vs. 467 ms respectively). However, the field x position interaction was also highly significant (Fl,~5 = 33.94, P < 0.0001). Post-hoe comparisons with the N e w m a n Keuls method showed that while in the RF Mid RTs were faster than Fix RTs (443 vs. 492 ms), in the LF Mid RTs tended to be slower than Fix RTs, although unsignificantly (450 vs. 44I ms respectively). Of the two other sources a marginally significant field main effect emerged (F1,15=5.98, P

The spatial distribution of attention in S-R compatibility.

In certain choice reaction time experiments the subjects, although not specifically instructed to do so, perform a parcellation of space into right an...
794KB Sizes 0 Downloads 0 Views