Perceptual load theory

Perceptual load is a theory of selective attention proposed by Nilli Lavie & Yehoshua Tsal in 1994. It explains the degree of distractor interference in processing a visual display as a function of the task-relevant processing load. The theory assumes that this processing capacity is limited, automatic and proceeds from relevant to irrelevant items.

The theory was proposed as a solution to the long standing debate in attention between early-selection and late-selection attentional accounts, and claimed that diverging results in previous attention research could be accounted for by the level of perceptual load in the various paradigms. Load theory puts aside the disagreement regarding the locus of the ‘attentional bottleneck’ of selection (characterising the ‘early’ vs ‘late’ selection debate) and instead addresses the efficiency of selection.

Key Assumptions

Perceptual load theory is a limited-capacity model, assuming there is a maximum amount of perceptual processing available at a given time, and also proposing that a perceiver will automatically allocate attention until the entire processing resource is exhausted. The theory proposes that task-irrelevant stimuli will be processed to a different degree as a function of the total amount of perceptual load in a visual display.

In order for the priority of allocation to be established, relevant and irrelevant stimuli must be adequately distinguished from each other. The relevance of a particular item may be determined by its location, by top-down factors such as task instructions or expectancy, or by bottom-up factors such as salience.

‘High’ versus ‘Low’ Load

The theory rests on a distinction between ‘low’ and ‘high’ perceptual load. The actual distinctions between ‘low’ and ‘high’ load displays are relative, rather than absolute. The level of perceptual load can be raised by increasing the number of items that serve as relevant response alternatives in a task, or increasing the difficulty of processing each of these items. For example, a low-load task may involve searching for a target that has a distinguishing feature (such as colour), whereas a high-load task may involve a conjunction search, where the target is defined by a combination of features (such as colour and shape) which makes it harder to detect.

Under conditions of ‘low’ perceptual load, the theory predicts that any remaining capacity that has not been allocated to the processing of relevant stimuli will ‘spill over’ to task-irrelevant stimuli. This ‘spillover’ under low-load conditions is seen as automatic and inevitable, thus not under voluntary control. The allocation of attention and subsequent perceptual processing is prioritised so that stimuli designated as task-relevant are attended before task-irrelevant stimuli, continuing in this order until the capacity is exhausted. Therefore, the theory asserts that irrelevant items such as distractors will only be perceived, and cause interference, under conditions of low load, when perceptual capacity has not been used up in the processing of relevant items.

Conversely, ‘high’ perceptual load displays involve either a larger set of relevant items to search, or require more information to process each item. These increased processing demands prevent irrelevant, low-priority items from consuming scarce processing capacity, which results in less distractor interference, as there is no remaining processing capacity for them to be perceived. According to the theory, this results in the effective rejection of task-irrelevant distractors in high-load displays.

Typical Experimental Paradigm

A typical experiment investigating perceptual load effects involves measuring distractor interference in a flanker-task display with a target and a nearby distractor, either alone (‘low load’ condition) or alongside number of non-target items (‘high load’ condition). Participants are typically instructed to locate and respond to the target (i.e. indicate whether it is an ‘X’ or a ‘Z’) whilst ignoring the peripheral distractor which is either congruent (the same as the target), incongruent (the opposite response of the target, i.e. ‘Z’ if the target is ‘X’) or neutral (unrelated to the target or response, i.e. an ‘N’). The time taken to respond to the target in the presence of incongruent distractors, minus the response time in the presence of neutral or congruent distractors yields a measure of distractor influence (congruence effect). According perceptual load theory, this congruence effect is larger under ‘low load’ conditions than under ‘high load’ conditions.

High load effects are generally found with around 6+ non-target items, when capacity is exhausted and the interference by distractors plummets. Low load effects are typically found with one or two distractors.

Counter-evidence

A number of researchers have demonstrated results that are inconsistent with the predicted perceptual load effect, of high-load displays yielding less distractor interference than low-load displays. Research has shown that the typical perceptual load effect may be reduced, eliminated, or even reversed under certain conditions: when the salience of targets and distractors is manipulated by altering their onset and offsets, when the task-relevant features of the target and the irrelevant features of the distractor appear in the same object, when the likely target location was pre-cued and thus the spatial extent of attention was narrowed, when the spatial proximity of the distractor to the target is increased, when the number of possible target locations is increased, or when low- and high-load trials are mixed within an experimental block.

These results, challenging the key tenets perceptual load theory, suggest that perceptual load may be important, but it is not the only factor influencing the selectivity of attention and degree of distractor interference.

Criticisms

A main focal point for criticism of perceptual load theory is that it revolves around a concept that has only been vaguely outlined, making it hard to operationalise. Critics also point out that the relative ‘high’ and ‘low’ load categories are often drawn from the results (response time and distractor interference) which has been criticised as engaging in circular reasoning. It has been suggested that the absence of definitional criteria for what constitutes load based on theoretical reasoning, rather than relying on empirical observations, means that the theory cannot adequately stand up to refutation.

Critics also argue that reaction time should not be treated as a pure measure of load because it also reflects task difficulty and cognitive demands. They also suggest that increasing the complexity of the task as a way of increasing perceptual load makes it very difficult to extricate the effects of perceptual load and cognitive load or working memory. Similarly, Tsal & Benoni suggest the frequent use of set size to indicate the level of load confounds perceptual load with dilution of the distractor, which they address in an alternate dilution account of attentional selection. This dilution account asserts that an increased set size merely dilutes the distractor by filling the display with items that have similar features to the distractor, thus causing them to compete for attention and response. From a dilution perspective, decreased distractor interference is a result of ‘noise’ or competition rather than due to the unavailability of resources suggested by perceptual load theory. It has been suggested that when dilution is controlled for, a reverse load effect occurs.