Lecture 9A – Foundations of Biological and Cognitive Psychology: Divided Attention and Automaticity

Introduction to Divided Attention

• Concept of Divided Attention: This refers to the attempt to multi-task or attend to more than one source of information simultaneously. While common in daily life (e.g., walking and chewing gum, driving and holding a conversation), its success depends on specific cognitive constraints.

• Historical Context: Lyndon B. Johnson famously mocked President Gerald Ford by questioning his ability to walk and chew gum at the same time.

• Determining Factors of Success: Whether two tasks can be performed concurrently is determined by four primary factors:

  1. Task Difficulty: Harder tasks consume more resources.

  2. Task Similarity: Tasks that share modalities or response types interfere more.

  3. Task Timing: The temporal proximity of stimuli affects processing speed.

  4. Practice: Repetition can lead to task automation, though it may not eliminate interference entirely.

Experimental Evidence for Task Interference

• 1. Task Difficulty (Sullivan, 1976):

  • Method: Used a dichotic shadowing task where participants listened to a primary message while attempting to detect a target word in the non-attended (secondary) message.

  • Findings: When the shadowing task became more difficult (less predictable), target detection in the non-attended message dropped significantly.

  • Implication: This aligns with Perceptual Load theory, suggesting that high-difficulty primary tasks exhaust attentional resources, leaving none for secondary tasks.

• 2. Task Similarity (Treisman & Davies, 1973):

  • Modality Interference: Found that two tasks interfere significantly more when they share the same modality (e.g., two visual tasks vs. one visual and one auditory task).

  • Response Similarity (McLeod, 1977):

    • Task: Manual tracking task combined with a tone identification task.

    • Response Conditions: Participants responded either via manual button press (other hand) or spoken word.

    • Results: Manual responses interfered more with the manual tracking task than spoken responses did.

• 3. Sensory Sensitivity and Imagery (Segal and Fusella, 1970):

  • Sensitivity Measure ($d'$): Sensitivity refers to the threshold required for stimulus detection; high sensitivity means faint stimuli are detectable.

  • Task: Participants imagined either an auditory or visual stimulus while attempting to detect a real, faint auditory or visual signal.

  • Conflict Results: Imagery in the same modality as the real stimulus (e.g., visual imagery during a visual detection task) reduced detection sensitivity significantly more than crossed modalities.

Task Timing and the Cognitive Bottleneck

• Psychological Refractory Period (PRP): A central cognitive bottleneck occurs when two stimuli are presented in rapid succession, each requiring a speeded response.

  • Observation: The Reaction Time ($RT$) to the second target is substantially lengthened the closer it occurs in time to the first target.

  • Explanation (Pashler et al., 2001): Response to the second stimulus is delayed because the brain is still processing the first stimulus at a central bottleneck.

• Comparison to Attentional Blink (AB):

  • AB: Observed in Rapid Serial Visual Presentation (RSVP) streams; involves a failure to perceive the second target.

  • PRP: Involves speeded responses to both targets; the targets are often in different modalities.

  • Mechanisms: There are important differences in underlying mechanisms, though both relate to temporal processing limits (Arnell et al., 2004).

The Role of Practice in Attention

• Spelke, Hirst & Neisser (1976):

  • Study: Two participants (Diane and John) practiced reading a short story while writing down dictated words for $5$ hours per week over $4$ months.

  • Progression: After $6$ weeks, recall for dictated words was still poor. Eventually, they could write down word categories while reading.

  • Conclusion: Human ability to develop specialized skills suggests that general limits on cognitive capacity may be hard to define.

  • Critique: The study focused on accuracy rather than speed ($RT$), and participants might have flexibly shifted attention between tasks.

• Persistent Interference: Practice typically reduces but rarely eliminates the PRP effect. Pashler (1993) found that the PRP effect remained observable even after over $10{,}000$ practice trials.

Theoretical Models of Divided Attention

• Central Capacity Models:

  • Assumption: There is a single pool of central capacity (like a "Central Executive") used flexibly.

  • Mechanism: Interference occurs if the total resource demand of two tasks exceeds the available capacity.

  • Bourke et al. (1996): Used four tasks (random letter generation, pattern learning, manual task, and tone detection). Found that the task requiring the most central capacity interfered the most with variety of secondary tasks.

• Multiple Resource Models (Wickens, 1984):

  • Assumption: Capacity consists of separate modular sources specialized for different processes.

  • Stages of Processing:

    1. Encoding: Perceptual processes (visual vs. auditory).

    2. Central Processing: Codes (spatial vs. verbal).

    3. Response: Output (vocal vs. manual).

  • Selective Interference: Interference is determined by whether tasks compete for the same specific type of resource.

  • Limitations: The model focuses heavily on visual/auditory inputs and lacks detail on how tasks are coordinated.

Automatic vs. Controlled Processing

• Automatic Processes (Shiffrin & Schneider, 1977):

  • Features: Fast, parallel processing, does not require attention/capacity, unavailable to consciousness, unavoidable (occurs automatically given a stimulus), and hard to modify once learned.

• Controlled Processes (Shiffrin & Schneider, 1977):

  • Features: Slow, serial processing, requires attention/capacity, available to consciousness, and can be used flexibly depending on task demands.

• The Stroop Effect:

  • Mechanism: Competing processes between naming the ink color (primary task) and reading the word (automatic process).

  • Result: Incongruent trials (e.g., the word "BLUE" in red ink) lead to significant interference and slower $RT$.

  • Data Ranges: Congruent trials are fastest, followed by control trials (e.g., "XXXX" in color), with incongruent trials being the slowest ($RT$ range approx. $0$ to $400+$ $ms$).

Shiffrin and Schneider (1977) Paradigm

• Memory Set vs. Visual Set: Participants were given characters to remember and then searched for them in a visual display.

• Consistent Mapping: The target categories and distractor categories never overlapped (e.g., consonants as targets, numbers as distractors). This resulted in parallel search and automaticity.

• Varied Mapping: Target and distractor sets swapped or overlapped across trials. This required serial processing and controlled attention.

• Evaluation: While influential, the theory is descriptive and over-simplified. Even "automatic" parallel search shows slight slows as set size increases, suggesting some capacity demand.

Norman & Shallice’s Supervisory Attentional System (SAS)

• Level 1: Fully Automatic: Controlled by "schemas" (organized plans for familiar actions).

• Level 2: Partially Automatic (Contention Scheduling): Processes that resolve conflicts between competing schemas without deliberate conscious direction.

• Level 3: Deliberate Control (SAS): Used for novel tasks, planning, problem-solving, and inhibiting strong habitual responses. It is the mechanism of "executive control."

• Clinical Implications and Frontal Lobes:

  • Dysexecutive Syndrome: Associated with frontal lobe damage; symptoms include difficulty in planning and organizing action.

  • Utilization Behaviour (Lhermitte, 1983): Patients with medial frontal lesions will inappropriately use any object presented to them (e.g., putting on glasses they don't need), suggesting a failure of the SAS to inhibit automatic schemas.

Evaluation and the Homunculus Problem

• Top-Down Control Functions: Monitoring performance, detecting/resolving conflict, switching responses, inhibiting inappropriate actions, and sustaining attention on goals.

• The Homunculus Problem: Proposing a brain area or "system" (like the SAS) that controls attention leads to infinite recursion—asking what "chooses" or "controls" the controller.

Real-World Implications of Divided Attention

• Driving and Phones: Research (e.g., Redelmeier & Tibshirani, 1997) shows hands-free phones are not safer than handheld ones, implying the bottleneck is cognitive (conversation/thought), not just manual motor demands.

• Studying and Productivity: Multiple screens, notifications, and music during study create task-switching costs and resource competition, potentially reducing performance.

Questions & Discussion

• Q1: What factors affect whether two tasks can be successfully carried out simultaneously?

  • Answer: All of the above (Task difficulty, similarity, and similarity of required responses).

• Q2: According to Shiffrin and Schneider (1977), what is a characteristic of automatic processing?

  • Answer: It does not require attention.

• Q3: A student trying to revise while listening to a podcast sees a drop in performance. According to multiple resource theories, what is the best explanation?

  • Answer: Both tasks rely on similar verbal/auditory resources.

• Q4: A participant performs visual tracking and auditory response with mild impairment, but visual-verbal plus auditory-verbal causes a sharp drop. Why?

  • Answer: Tasks interfere more when they compete for the same modality-specific and response resources.