Chapter 4: Attention

Introduction

Attention is a fundamental cognitive process that allows us to concentrate on specific stimuli while ignoring others. Understanding attention is crucial for comprehending how we process information, make decisions, and interact with our environment. This chapter will explore various conceptualizations of attention, the mechanisms behind it, and the experimental evidence supporting these views.

Conceptualizing Attention

Attention can be understood through multiple frameworks, each offering a unique perspective on how we focus on and process stimuli.

• Spotlight model: Attention can be likened to a spotlight that moves around the world. When someone says, "Look here," they are directing your mental spotlight to focus on a particular area or object in the visual environment.
• Resource allocation to selected stimuli: Attention involves concentrating mental effort on a selected stimulus. This perspective emphasizes that attention is a limited resource that can be directed toward processing specific information. It can be under volitional control, allowing us to choose what to focus on.
• Limited resource: Building on the resource allocation view, this perspective highlights that attention is finite. The more tasks or stimuli we try to attend to simultaneously, the less efficient our processing becomes.
• Inhibition of distraction: Rather than focusing solely on selected target stimuli, this model posits that attention involves inhibiting or blocking out irrelevant distractors. For example, when listening to a lecture in a noisy environment, attention works to suppress the distracting background noise. In some ways, this view of attention is opposite to the view that attention focuses on selected target stimuli.

These models suggest that different types of attention might coexist, each playing a distinct role in how we process information.

Attention as a Spotlight

The spotlight metaphor suggests that attention moves across the visual field, illuminating specific areas while leaving others in the dark. Experiments using visual search tasks have provided insights into how this spotlight functions.

Feature Search vs. Conjunction Search

1. Feature Search: Involves searching for a target that differs from distractors based on a single feature, such as color. For example, finding a red "T" among black "T"s and “S”s (see left-hand side of Figure 4.1). This type of search is fast and efficient because the target "pops out" due to its unique feature.
2. Conjunction Search: Requires finding a target based on a combination of features, such as a red "T" among black "T"s and red "S"s. (see right-hand side of Figure 4.1)This search is slower and more effortful because it involves binding multiple features (color and shape) together.

Figure 4.1. This figure shows a feature search (shown to the left) and a conjunction search (shown to the right). In a feature search people must find a letter that differs on one feature; here people would search for a red T that is embedded in black Ts and black Ss. In a conjunction search people must find a letter that does not differ on any one feature; here people would search for a red T that is embedded in black Ts and red Ss.

"Feature search and conjunction search." by Kahan, T.A. is licensed under CC BY-NC-SA 4.0

Pop-Out Effects

Experiments have shown that reaction time in feature searches is unaffected by the number of items in the display, suggesting parallel processing. In contrast, reaction time in conjunction searches increases with the number of items in the display, indicating serial processing.

Anne Treisman's Feature Integration Theory (FIT) posits that attention is essential for integrating different features of an object (like color, shape, and size) into a coherent perception. She argued that while individual features are processed automatically and in parallel across the visual field, attention is required to combine these features at a specific location to form a unified object. Without attention, features remain unbound and are perceived as free-floating or fragmented. This theory emerged from experiments demonstrating that when attention is diverted, participants often experience "illusory conjunctions," where features from different objects are incorrectly combined, underscoring attention's role in accurate feature binding. According to Treisman attention is the glue that binds features together.

Magic Trick Demonstration

A simple magic trick can illustrate the limitations of attentional resources. When presented with a quick display of cards and asked to remember one card (see Figure 4.2), most people focus on only one card's value and suit but do not bind the value and suit of all distractors (since this requires attentional resources). When shown a new display and told that the magician has read their mind and has removed only their card (see Figure 4.3 on the next page), people often find this impressive.

Figure 4.2. This figure shows a magic trick where people are shown 6 playing cards. People are told to choose a card and then on the next page I will remove the card you selected

"Attention-based magic trick (part 1)" by Kahan, T.A. is licensed under CC BY-NC-SA 4.0

Goal-Directed vs. Stimulus-Driven Attention

Attention can be directed toward objects in the environment either by our goals or by the properties of stimuli themselves.

1. Goal-Directed Capture: Attention is directed based on our goals and intentions. For example, if your goal is to find a friend in a crowded room, you might focus on specific features, like their clothing color.
2. Stimulus-Driven Capture: Certain stimuli capture attention automatically, regardless of our intentions. These stimuli possess inherent properties that make them attention-grabbing. This latter type of attentional capture is evolutionarily advantageous in instances where the attention-grabbing stimulus is harmful.

Stimulus-Driven Pop-Out

Abrupt Onsets

Research by Yantis and colleagues shows that abrupt onsets—stimuli that suddenly appear in the visual field—can capture attention automatically. In a study by Steve Yantis, participants were tasked with identifying a global shape (H or S) formed by smaller local shapes, while sometimes encountering a condition where an incongruent local-level shape abruptly appeared. In the incongruent condition, this local shape (e.g., an "S" when the global shape was an "H") conflicted with the global-level target, while in the neutral condition, no incongruent local shape was present. The critical manipulation involved trials with an abrupt onset, where the incongruent local shape appeared suddenly, compared to no abrupt onset, where no such distracting shape appeared. Results showed that reaction times and accuracy were significantly worse in the incongruent condition with an abrupt onset compared to the no-abrupt-onset or neutral conditions, suggesting that the sudden appearance of the distractor involuntarily captured attention. This led to the conclusion that abrupt onsets are powerful in capturing attention, even when irrelevant to the task, disrupting participants' ability to focus on the global-level target.

Emotional and Facial Stimuli

Emotional stimuli, especially those involving faces, also capture attention. Studies have shown that faces, particularly those expressing negative emotions, slow down task performance when they appear as irrelevant stimuli. For example, in a study by Merikle, Eastwood, and Smilek, participants were asked to count the number of ovals or circles in a display, where these shapes formed parts of face-like emoji patterns with varying emotional expressions (happy, neutral, or negative). Although the task focused solely on identifying and counting geometric shapes that might form the eyes or nose of different faces, the emotional valence of the faces influenced performance. Participants were slower to count the shapes when the faces had negative expressions, compared to neutral or happy faces. The authors concluded that negative emotional stimuli capture attention involuntarily, even when they are irrelevant to the task, demonstrating the automatic and disruptive nature of processing negative emotional information. This pattern was eliminated when the displays were inverted thereby disrupting face processing. This finding underscores the attentional bias towards negative stimuli, likely due to their evolutionary significance in detecting potential threats.

Figure 4.3. In this figure, you will see that I predicted the card you would select and I have removed only this card. (The truth is that this “trick” only seems impressive if a person does not realize that all of the cards have changed). People lack the resources to bind the value and suit of all of the distractor cards and so they may erroneously think the magician has read their mind and has removed only their card.

"Attention-based magic trick (part 2)." by Kahan, T.A. is licensed under CC BY-NC-SA 4.0

Looming Objects

Looming objects, or objects moving directly towards us, are particularly effective at capturing attention. For example, a study by Franconeri and Simons (2003), investigated how looming objects (those that appear to grow larger as they approach) capture attention. In their experiment, participants were asked to perform a task in which they tracked a specific object among several moving items on a screen, while sometimes a looming object was presented among the distractors. The critical finding was that participants were significantly slower to track their designated object when a looming object appeared in the display, suggesting that the looming object involuntarily captured attention. The study demonstrated that looming stimuli, which signal potential threats or important events, are processed with priority, likely due to their evolutionary significance in detecting approaching objects. This research underscores the idea that attention is automatically drawn to stimuli that convey potential danger or importance, even when they are not directly relevant to the task at hand.

The Attentional Blink

The attentional blink is a phenomenon first described by Raymond, Shapiro, and Arnell (1992), in which people experience a temporary lapse in detecting a second target when it appears shortly after a first target in a rapid serial visual presentation (RSVP) task. In this task, participants view a stream of items (e.g., letters or symbols) presented in quick succession in the same location on a computer screen, and they are instructed to identify two targets (e.g., two letters that are intermixed amongst distracting digits). The key finding is that when the second target (T2) appears within 200-500 milliseconds after the first target (T1), participants are significantly less likely to detect it, compared to when it appears earlier or later. This "blink" in attention is thought to result from the limited capacity of attentional resources needed to process T1, temporarily impairing the ability to process T2. The implications of the attentional blink highlight the temporal limits of selective attention and provide insights into how attentional resources are allocated over time. Interestingly, when T2 is a face the blink is reduced and this is reduced further when T2 is a threatening face. This result further supports the conclusion that stimuli with negative capture attention.

Dichotic Listening and Early Selection Theory

Dichotic Listening Experiments

A classic method to study attention is the dichotic listening task, where participants receive different messages in each ear and are asked to shadow (repeat) the message from one ear while ignoring the other. Results from these experiments show that participants are generally unable to report the content of the unattended message, although they can often identify its physical characteristics, such as whether the voice was male or female or if it was speech or music.

Early Selection Theory

Broadbent's early selection theory posits that attention works as a filter, allowing only selected information based on physical characteristics to pass into short-term memory for further processing. This theory suggests that unattended information is filtered out early in the processing stream, preventing it from reaching the stages where meaning is analyzed.

Challenges to Early Selection Theory

Own Name Effect

One notable challenge to early selection theory is the "own name effect." Research has shown that people often notice their own name in the unattended ear during a dichotic listening task. This implies that some level of semantic processing must occur for the unattended information, which contradicts the idea of a strictly early filter.

Treisman's Attenuation Theory

Treisman’s Attenuation Theory (1964) proposes an alternative to Broadbent’s early filter model of attention, suggesting that rather than completely filtering out irrelevant information, unattended stimuli are merely attenuated (weakened) in their processing. According to this theory, attention acts like a filter that reduces the strength of unattended information rather than blocking it entirely, allowing some relevant but weaker information to pass through. The theory was supported by results from experiments showing that people could still detect meaningful information in unattended channels under certain conditions. For instance, in Treisman’s dichotic listening task, participants could report the meaning of an unattended message if it was consistent with the attended message, even though they couldn’t always recall the specific content. These findings support the idea that unattended information is processed to a degree, but at a reduced level compared to attended information, and that meaningful stimuli can "break through" the attenuated filter when they are of personal significance or are highly relevant to material that is attended.

Late Selection Theory

Deutsch and Deutsch's Model

Deutsch and Deutsch's late selection model proposes that all stimuli in the environment are processed to a semantic level, meaning their meanings are analyzed, before any selection occurs. According to this model, attentional selection happens at a late stage, based on the importance or relevance of the information to the individual's goals or needs. This contrasts with early selection models, which propose that irrelevant stimuli are filtered out before semantic processing. The late selection model suggests that even unattended information is processed fully, but only the most relevant information reaches conscious awareness or influences behavior.

Evidence from MacKay's Study

MacKay's 1973 study provides strong evidence supporting late selection models of attention. In this study, participants were presented with ambiguous sentences like "They were throwing stones at the bank" in one ear while words like "money" or "river" were played in the unattended ear. Later, participants' interpretations of the sentence (e.g., financial institution or muddy creek) were influenced by the unattended word they heard, even though they were unaware of it. This finding fits with the late selection model because it shows that the meaning of the unattended word (e.g., "money" or "river") was processed semantically and influenced participants' comprehension of the ambiguous sentence. The study demonstrates that even stimuli not consciously attended to can affect understanding and decision-making, consistent with the idea that all stimuli are processed for meaning before selection.

Inattentional Blindness and Theories of Attention

Inattentional blindness occurs when individuals fail to notice a fully visible object or event because their attention is engaged elsewhere. This phenomenon highlights the limits of human perception and attention.

One of the most famous demonstrations of inattentional blindness is the experiment conducted by Simons and Chabris (1999). In this study, participants were shown a video of two teams passing basketballs—one team dressed in white and the other in black. Participants were instructed to count the number of times the players in white passed the ball, focusing on their task to ensure accuracy. During the video, an unexpected event occurred: a person dressed in a gorilla suit walked into the scene, stopped in the middle, beat their chest, and then walked off-screen, spending about 9 seconds on camera.

Surprisingly, many participants failed to notice the gorilla, even though it was in plain sight. Their attentional resources were so focused on counting the passes by the team in white that they missed this unexpected and highly conspicuous event.

Early Selection View: According to early selection theories (e.g., Broadbent, 1958), attention acts as a filter that selects information for further processing based on physical characteristics like location or color. In this view, participants were so focused on the team in white that information from other parts of the scene, including the gorilla, was filtered out at an early stage of processing. The gorilla was not processed at a deeper level because attention was already committed to the basketball passes, making the unexpected event essentially invisible to the participants. From this perspective, the gorilla was never processed because it was filtered before it reached short-term memory.

Late Selection View: Late selection theories (e.g., Deutsch & Deutsch, 1963) suggest that all stimuli are processed to some extent, but only a subset reaches conscious awareness. According to this view, participants did process the gorilla but the information didn’t reach the level of conscious awareness due to the focus on the counting task. In other words, the participants "saw" the gorilla, but because their attention was elsewhere, the gorilla's presence remained outside their conscious perception.

Load Theory of Attention

Nilli Lavie's load theory offers a reconciliation of early and late selection theories. Lavie proposes that whether selection occurs early or late depends on the perceptual load of the task at hand. In environments with low perceptual load, all stimuli are processed to the semantic level, supporting late selection. However, in high-load environments, our cognitive resources are limited, necessitating early selection to filter out irrelevant information.

Evidence for Load Theory

An experiment illustrating the load theory of attention involves participants identifying a target letter from an array presented in a circular arrangement (ring) on a screen. For example, participants might be instructed to find a specific letter, such as X or N, among other letters in the array. Simultaneously, a distractor letter appears outside the ring. This distractor can be congruent with the target (e.g., also an X) or incongruent (e.g., a different letter, like N). The goal is to assess how the distractor affects the participant's ability to identify the target, particularly under conditions of varying load.

In low-load conditions, the task is easy, such as when the target appears among non-similar letters (e.g., the target is an X among Os). In high-load conditions, the task is more demanding, such as when the target appears among letters that are visually similar or when the array contains a greater number of distractors.

The congruency effect—the difference in reaction time (RT) between congruent and incongruent trials—is typically larger in low-load conditions. In these conditions, participants have spare attentional resources, allowing the irrelevant distractor to be processed, thereby interfering with the task. In high-load conditions, attentional resources are fully consumed by the demanding task of finding the target in the array. As a result, the distractor is less likely to be processed, reducing its impact on performance.

This pattern supports the load theory of attention, which posits that attentional resources are limited, and the extent to which distractors are processed depends on whether the primary task depletes those resources.

Attention as a Limited Resource

A third perspective on attention is that it is a limited mental resource. This viewpoint emphasizes the finite nature of our cognitive capacity, suggesting that multitasking can severely impair performance. The next section will delve deeper into this idea, exploring how dividing attention across multiple tasks can lead to decreased efficiency and increased error rates.

Automaticity in Reading

In the realm of cognitive psychology, automaticity refers to the process whereby tasks are performed with little to no conscious effort. This phenomenon is a crucial aspect of understanding attention, as it highlights the automatic nature of certain cognitive processes through extensive practice.

One illustrative example is the Stroop task. In this task, individuals are asked to name the color of the ink in which a word is printed, rather than reading the word itself. For instance, if the word "RED" is printed in green ink, participants must say "green." However, they often find it difficult to ignore the word and this slows reaction times when the word and color are incongruent (RED) and speeds reaction times when the word and color are congruent (GREEN). This automatic reading occurs because people have a significant amount of practice with reading words, making it an involuntary process.

Characteristics of Automaticity

Automaticity can be characterized by several key criteria:

1. Involuntary Activation: Automatic processes are initiated despite a person’s intentions (i.e., it occurs without intention). For example, if you are told not to read the word in a Stroop task (described above) you will still have difficulty naming the color of the word if the word and color are incongruent. This indicates that word reading occurs despite intentions.
2. Lack of Awareness: Automatic processes can occur even when individuals are not aware of the stimuli. This can be seen in studies involving masked priming, where a word is shown very briefly and then masked by another stimulus. Despite not being aware of the initial word, participants are quicker to recognize related words, indicating that the initial word was processed automatically. For example, if people are shown the word “CAT” rapidly and this word is masked with visual noise they may be faster to identify a related word, DOG, presented immediately afterward.
3. Resource Efficiency: Automatic processes consume few cognitive resources, allowing individuals to perform other tasks simultaneously without interference. This capacity-free nature means that automatic tasks can be performed effortlessly.
4. Rapid Processing: Automatic tasks are typically executed quickly. This is a byproduct of extensive practice and consistent exposure to the task.

Practice and Automaticity

One of the seminal studies in this area was conducted by Shiffrin and Schneider, who demonstrated that practice leads to automaticity. They found that consistent mapping of stimuli and responses is crucial. In their experiments, participants were given a memory set of one, two, or four letters, followed by a target frame containing one, two, or four letters. Their task was to determine if any letter from the memory set appeared in the target frame.

In the varied mapping condition, the memory set and target frame items changed on every trial. This inconsistency prevented the development of automaticity, as participants had to process each trial afresh. Conversely, in the consistent mapping condition, the memory set items were always drawn from the same set of letters. This consistency allowed participants to develop automatic responses, significantly reducing their reaction times regardless of the number of items in the memory set or target frame.

The results from Shiffrin and Schneider's experiments indicated that consistent practice with the same set of stimuli leads to automaticity. This explains why individuals can respond to their own names automatically; they have a lifetime of consistently mapped practice with their names.

Stated differently, people respond automatically to their own name, not simply because they hear and respond to a name daily, but because their name remains the same from one day to the next. If your name changed every day, responding to it would not become automatic. Similarly, driving a car is automatic because the actions required to drive remain consistent. If the way you drove changed daily—such as the location of the gas pedal or the rules of the road—driving would not be automatic. Automaticity develops through consistent practice, where you repeatedly respond to the same stimulus in the same way.

Advantages and Disadvantages of Automaticity

Automaticity is beneficial in many contexts, as it allows for the efficient execution of tasks without taxing cognitive resources. However, it also has its drawbacks:

1. Relearning Difficulties: Once a task becomes automatic, relearning or adjusting the automatic response can be challenging. For instance, switching from typing on a typewriter to a computer keyboard may cause individuals to automatically perform actions suited to the typewriter, leading to errors.
2. Error Overlooking: Automaticity can lead to overlooking errors. Experienced pilots, for example, may miss critical checklist items due to automatic processing, potentially leading to severe consequences, such as landing a plane without deploying the landing gear.
3. Inappropriate Responses: Automatic responses may be inappropriate in certain contexts. For example, if a person is accustomed to responding to a specific name and then changes partners, they might inadvertently use the wrong name due to automaticity.
4. Safety Hazards: In some professions, automaticity can pose safety risks. Factory workers familiar with machinery may instinctively react in ways that are dangerous if the machinery's operation changes.

Negative Priming: Attention as an inhibitory mechanism

Another perspective on attention posits that it functions primarily to inhibit irrelevant stimuli rather than solely focusing on relevant ones. This view suggests that attention works to block out distractions, thereby allowing individuals to concentrate on the task at hand.

Evidence for this inhibitory mechanism comes from studies on negative priming. Negative priming occurs when individuals are slower to respond to a stimulus they previously ignored. For example, if a person is asked to name the color of a word and ignore the word itself, they will be slower to name the color of a word they just ignored. This effect indicates that attention not only focuses on relevant stimuli but also actively inhibits irrelevant information.

Negative priming is measured by looking at differences in performance in an Ignored Repetition Condition and a Control Condition.

Ignored Repetition Condition

In one trial, the participant sees the word RED written in blue ink. The task is to name the ink color (blue), so the word RED is ignored. In the following trial, the target color is red (e.g., the word YELLOW is written in red ink). Here, participants are slower to respond to the ink color (red) because it matches the previously ignored word, RED.

Control Condition

In the control condition, the distractor word from the previous trial does not reappear in any form. For example, the participant sees the word YELLOW written in blue ink, and in the next trial, the target is red (e.g., the word GREEN written in red ink). Since the previously ignored stimulus (YELLOW) is unrelated to the current trial, it does not interfere with the response.

There are several types of negative priming, each revealing different aspects of how our attentional system functions.

Identity-Based Negative Priming

Identity-based negative priming occurs when individuals are slower to respond to the identity of a stimulus because they ignored that identity before (so, “what” had been ignored now appears as the target). For instance, when participants are asked to respond to the color of a word, they are slower to respond if the word was previously ignored but in a different color. Similarly, if a person ignores a picture of a house, they are slow to respond to the word house. This shows that ignoring a specific identity (in this case, the word) can affect the processing speed of that same identity when it reappears.

Location-Based Negative Priming

Location-based negative priming involves slower responses to a location that was previously ignored (so, “where” a distractor had been shown affects reaction times if the target later appears at that location). For example, if a participant is tasked with identifying the position of an 'X' on a screen and the 'X' appears in the far-left position on the prime trial (first trial) but in the far-right position on the probe trial (second trial), the participant's response will be slower if the far-right position was previously occupied by a distractor relative to a situation where the far-right position had not previously be occupied by a distractor. This demonstrates that ignoring information in a specific spatial location can hinder the processing speed when information appears in that same location later on.

Temporal Negative Priming

Recent research by Kahan and colleagues (2020; 2022) has uncovered a third type of negative priming based on time (so, “when” a distractor had been shown affects reaction times if the target later appears at that same relative point in time). Temporal negative priming occurs when the processing of a stimulus is slowed down because it appears at a point in time that was previously associated with a distractor. This discovery was made through experiments involving sequences of stimuli presented at different speeds.

Experimental Setup

In these experiments, participants view a sequence of boxes on a computer screen, where each box may contain an 'X', an 'O', or be blank. The task is to press a key corresponding to the temporal position of the 'X'. For example, if an 'X' appears first in a sequence of four, the participant presses '1'. If it appears third, they press '3'. The sequences are presented at varying speeds, creating different temporal contexts.

Figure 4.4. This figure shows the sequence of events in a temporal negative priming experiment where time flows from top to bottom. The person’s task is to indicate when in the sequence the X was shown. In the prime trial the person would indicate that the X was shown first since the first box contained an X and they would ignore the O in the 3rd position. On the probe trial the person would indicate that the X appeared in the second position. Results show that reaction times are slow on the probe trial if the X appears at a point in time that previously contained the distractor letter O.

"Temporal negative priming sequence." by Kahan, T.A. is licensed under CC BY-NC-SA 4.0

Findings

The research found that participants are slower to respond when an 'X' appears at a temporal position that had previously been occupied by a distractor. In the figure showing this (see Figure 4.4), time flows downward, with the prime trial shown to the left of the subsequent probe trial. On the prime trial, both a target (X) and a distractor (O) are presented, while only the target appears on the probe trial. Larger gaps between the boxes indicate longer delays, meaning a slower rate of presentation. Temporal negative priming was observed in this experiment regardless of whether the response required was the same or different from the previous trial. For example, if an 'X' appears in the second position on the probe (requiring a response of '2') following a sequence where a distractor appeared at that same relative point in time on the prime (even if it had been the 3rd item in that sequence), the response is slower compared to a situations where a distractor had not been presented at that temporal position previously.

Neural Mechanisms

While the brain regions responsible for identity-based and location-based negative priming

have been studied extensively, the neural underpinnings of temporal negative priming remain largely unknown. However, the identification of these three distinct types of negative priming—identity, location, and temporal—provides a more comprehensive understanding of how selective attention operates across different dimensions.

Conclusion

Negative priming is a vital phenomenon in the study of attention, illustrating the selective nature of our cognitive processes. Identity-based negative priming shows how ignoring specific stimuli affects future processing of those stimuli. Location-based negative priming reveals the impact of ignoring spatial locations. The newly discovered temporal-based negative priming adds a temporal dimension to our understanding of attentional selectivity. Together, these findings highlight the complexity and adaptability of our attentional system in managing and processing the vast array of information we encounter daily.