Home | Business News | Browse by Publication | H | Human Factors

Stimuli fixation and manual response as a function of expectancies.

Article Excerpt
INTRODUCTION

Under some conditions, only a small part of the available attentional resources is directed to the task at hand. In these situations people perform their task fairly well, on a more or less automated level, while keeping resources available for other tasks or thoughts. This phenomenon is typically present in skill-based tasks or vigilance tasks. Jerison (1970) and Warm (1977) stated that prolonged and continuous vigilance tasks require sustained attention. In their vigilance tasks, targets occurred infrequently in terms of time and were usually clearly perceivable when the observer was alerted to them. Compared with the number of distractors, targets occurred aperiodically and without forewarnings. The observer's response typically has no effect on the probability of appearance of the targets in vigilance tasks. Typical vigilance tasks would be inspecting individual items fox" classification (manufactured goods) or monitoring an ongoing process, such as in a radar air traffic control center and vehicular control (Wiener, 1984).

These tasks, although requiring sustained attention, may induce a low level of arousal (Parasuraman, 1984). Especially in an environment that is highly predictable (with a person having strong expectations about what will happen), relatively little conscious attention may be paid to the task at hand. Mermall (1970) concluded that driving can be a task with minimal commitment, referring to the situation in which drivers are occupied with themselves (e.g., when daydreaming), thereby reducing the interaction with the outer world. He stated that drivers "somehow" learn to drive without thinking about it or without being consciously aware of the driving situation. In these situations, people passively use their mental model to select important information from the environment, rather than actively scanning the visual stimuli in the surroundings and updating the mental model.

The background of this report is this kind of passive information processing, referring to a state in which expectations have replaced a large part of the active information intake, as compared with the state of a highly attentive person. This automatic information processing requires little attention, leaving attentional resources to be distributed to other areas such as other tasks or thoughts. It is a very economic state of information processing, but the state is adequate only if the person's expectations are correct. If something unexpected happens (that does not correspond with the expectations), one could either miss crucial information completely (fail to look, or look but fail to perceive) or experience a delay in manual response time caused by the process of reallocating attentional resources to the visual information in the outside world and interpreting this information in order to make the appropriate manual response. If this happens in a professional situation, the cost of the economic state of information processing may be failing to notice an unexpected detail or responding too late, possibly resulting in accidents.

Some evidence for the effect of expectations on searching for targets has been provided in static experiments. Meyers and Rhoades (1978) looked at the effect of expectancy on visual scan patterns and showed that searching for an object at a nonpredictable location was much slower than searching for an object at a likely location. This implies that people direct their visual scan pattern according to where they expect information to be, which results in an effective search if their expectations are correct.

The literature suggests that the pattern of eye movements somehow indicates the goals of the observer and possibly the area of interest (Liu, 1999; Stark & Ellis, 1981). Theeuwes (1991) investigated the effect of expectation on top-down (active) visual search of everyday traffic scenes presented on slides. The study showed that expectancies about the location of the target (traffic signs are normally located on the right side of the road) had an effect on scan behavior (searching on the right side of the road). Search behavior seemed to be based on the meaning and content of a scene in combination with the object one searches for. In contrast to Meyers and Rhoades (1978), Theeuwes did not find that participants adjusted their search strategy when they learned that expected location was not always a good cue for responding; search strategies remained the same even in conditions in which the target object was consistently presented at unlikely positions. Participants always tended to search at the most likely locations, which led them to respond that there was no target if it did not appear in those places.

Most studies dealing with visual search concentrate on static environments, such as pictures or presentations of scenes on a computer monitor. However, in tasks with a presentation at one discrete moment in time, visual search is assumed to be more active than it may be in dynamic scenes, in which there is a continuous stream of stimulus material and not a sequence of pictures. Few studies have looked at the kind of visual information that is perceived in a dynamic environment when people have strong expectations and direct little attention to actively scanning the environment. How do these expectations influence fixations in the visual environment? Is expected information perceived better because it is in accordance with what people expect, or is inconsistent information perceived better because it has a high informational value? The questions that will be asked are related to these issues: (a) Will eye fixations be different if participants are able to develop expectations of when to expect targets, compared with when participants are without these expectations? (b) Will targets that are presented in a way that is in contradiction with expectations (irregularities) be missed and, if not, will manual response times to these targets be higher?

EXPERIMENT 1

Method

The assumption behind the idea of passive or automatic information processing is that if people do not have strong expectations about the environment, they are less certain of where important information may be located and spend more time fixating all stimuli. If people develop expectations because of sequential exposure to a certain scene, they feel more certain and do not feel the urge to fixate all stimuli. In Experiment 1, the hypothesis was tested that people with expectations about what information is relevant and what information is not will spend more time looking at information that is assumed to be relevant. A second hypothesis is that if information is presented that does not correspond with expectations, the information will be missed.

In order to test these assumptions, a lab task was developed that had some correspondence with a simulated driving environment. The task environment consisted of a road with grass along the sides and relevant objects (the stimuli) and irrelevant objects (corresponding to houses or trees in the surroundings). Scenes from the task are shown in Figure 1.

[FIGURE 1 OMITTED]

Manual responses and eye fixations were recorded in order to test the hypotheses. Eye-movement studies have yielded some valuable results in exploring visual scan patterns, especially in dynamic environments. If participants move their eyes toward a new...

View this article FREE - Now for a Limited Time, try Goliath Business News
Free for 3 Days!