Course outline
Module 1
Module 1: What Attention Is
What Attention Is
Knowledge check
Q1. In the context of neural networks, what is the primary purpose of an attention mechanism?
- A. To compress all input data into a single fixed-length vector
- B. To let a model weigh how much each part of its input should influence the output ✓
- C. To randomly drop connections during training to prevent overfitting
- D. To normalize the activations of each layer to zero mean and unit variance
Why: Attention lets a model compute a weighted combination of its inputs, where the weights reflect how relevant each input element is to the current output being produced.
Q2. What is the intuitive analogy often used to explain attention in machine learning?
- A. A filing cabinet that stores data in alphabetical order
- B. A spotlight that focuses on the most relevant parts of a scene ✓
- C. A river that carries information in one direction only
- D. A thermostat that regulates the temperature of a processor
Why: Like a spotlight, attention concentrates processing on the most relevant elements while the rest receive less emphasis, which is why the spotlight analogy is so commonly used.
Q3. Why was attention introduced into sequence-to-sequence models for machine translation?
- A. Early encoder-decoder models forced the entire source sentence into one fixed-size vector, creating a bottleneck for long inputs ✓
- B. Recurrent networks could not process text at all without attention
- C. Attention was needed to convert words into numerical embeddings
- D. Translation required a model to ignore word order entirely
Why: Compressing a long sentence into a single fixed vector loses information, so attention allowed the decoder to look back at all encoder states and pick the relevant ones at each step.
Q4. In the basic attention computation, what do the attention weights represent?
- A. The learned parameters of the output layer
- B. Normalized scores indicating how much focus each input position receives, typically summing to one ✓
- C. The total number of tokens in the input sequence
- D. The loss value computed after each training batch
Why: Attention scores are usually passed through a softmax, producing a probability-like distribution over input positions that sums to one and reflects relative focus.
Q5. Which statement best describes how attention differs from a standard recurrent hidden state?
- A. Attention discards all information from earlier positions in the sequence
- B. Attention allows direct access to any position in the input, rather than relying only on information passed step by step ✓
- C. Attention can only be applied to images, not to sequences
- D. Attention removes the need for any learned parameters
Why: A recurrent state must carry information forward sequentially, while attention provides a direct connection to every input position, making distant relationships easier to capture.
Module 2
Module 2: Selective Focus and Filtering
Selective Focus and Filtering
Knowledge check
Q1. In the context of attention, what is 'selective attention'?
- A. The ability to perform several tasks at the same time without any loss of performance
- B. The process of focusing on relevant stimuli while filtering out irrelevant ones ✓
- C. The automatic response to sudden loud or bright stimuli
- D. The ability to remember information over long periods of time
Why: Selective attention refers to the mental process of concentrating on specific relevant information while suppressing or ignoring competing, irrelevant input.
Q2. What is the 'cocktail party effect'?
- A. Feeling tired after attending a large social gathering
- B. Remembering a conversation you were not paying attention to
- C. The ability to focus on one voice in a noisy room while tuning out others ✓
- D. Becoming distracted by background music at a party
Why: The cocktail party effect describes how people can concentrate on a single conversation amid many competing voices and background noise, demonstrating selective filtering of auditory input.
Q3. Which of the following is an example of a bottom-up (stimulus-driven) attentional capture?
- A. Deliberately searching a crowd for a friend's face
- B. Noticing a sudden flashing light in your peripheral vision ✓
- C. Focusing on a book to prepare for an exam
- D. Choosing to ignore a notification while studying
Why: Bottom-up attention is driven by salient properties of the stimulus itself, such as sudden movement, brightness, or abrupt onset, which capture attention automatically regardless of current goals.
Q4. What does 'filtering' mean in attention research?
- A. Removing all sensory input so the brain can rest
- B. Blocking or attenuating processing of unattended information so that limited capacity is not overwhelmed ✓
- C. Storing information permanently in long-term memory
- D. Increasing the volume of attended sounds
Why: Filtering is the mechanism by which the brain reduces the processing of unattended input, protecting limited processing capacity from being overloaded by irrelevant information.
Q5. Which early theory proposed that attention acts as a filter that allows only physical features of unattended messages to be processed, before meaning is analyzed?
- A. Broadbent's filter model ✓
- B. Miller's magical number seven
- C. Atkinson and Shiffrin's multi-store model
- D. Baddeley's working memory model
Why: Broadbent's filter model (1958) proposed an early selection filter based on physical characteristics such as pitch or loudness, with semantic analysis occurring only for attended messages.
Module 3
Module 3: Sustained Attention and Vigilance
Sustained Attention and Vigilance
Knowledge check
Q1. Which term best describes the ability to maintain focus on a task or stimulus over an extended period of time?
- A. Selective attention
- B. Sustained attention ✓
- C. Divided attention
- D. Orienting attention
Why: Sustained attention, also called vigilance, refers to the capacity to keep focused on a task or stimulus over prolonged periods, as opposed to briefly directing attention to a single event.
Q2. What is the vigilance decrement?
- A. An increase in reaction time caused by practice effects
- B. A decline in detection performance over the course of a prolonged monitoring task ✓
- C. A temporary improvement in focus after a short break
- D. The loss of working memory capacity under stress
Why: The vigilance decrement is the well-documented drop in the ability to detect rare or subtle signals as time on task increases, typically observed within the first 15 to 30 minutes of a monitoring task.
Q3. Which task is the classic laboratory measure of sustained attention developed by Norman Mackworth in the 1940s?
- A. The Stroop task
- B. The Posner cueing task
- C. The Continuous Performance Test or Mackworth Clock Test ✓
- D. The Wisconsin Card Sorting Test
Why: Mackworth's Clock Test, studied for radar operators in World War II, and later Continuous Performance Tests were designed to measure vigilance by requiring participants to detect infrequent target events over long periods.
Q4. Which factor is most likely to IMPROVE sustained attention during a long monitoring task?
- A. Monotonous, unchanging stimuli
- B. Very low signal rate with no feedback
- C. Brief breaks and knowledge of results ✓
- D. Excessive caffeine consumed immediately before the task
Why: Brief rest breaks and feedback on performance help maintain alertness and motivation, while monotony, low signal rates, and lack of feedback tend to worsen vigilance.
Q5. Which brain network is most strongly associated with maintaining sustained attention and vigilance?
- A. The visual cortex alone
- B. The cerebellum
- C. Fronto-parietal networks together with the noradrenergic arousal system ✓
- D. The hippocampus
Why: Sustained attention relies on fronto-parietal control networks working with arousal systems, particularly the locus coeruleus noradrenergic system, which regulates alertness and the ability to stay engaged over time.
Module 4
Module 4: Divided Attention and Multitasking
Divided Attention and Multitasking
Knowledge check
Q1. What is divided attention?
- A. Focusing on a single stimulus while ignoring all others
- B. Simultaneously attending to two or more tasks or stimuli ✓
- C. The ability to remember information over long periods
- D. A state of deep sleep in which the brain filters out sound
Why: Divided attention refers to the capacity to process two or more tasks or stimuli at the same time, as opposed to selective attention, which focuses on one source while filtering out others.
Q2. Which finding from dichotic listening research best supports the idea that attention has limited capacity?
- A. Participants could always repeat both messages perfectly
- B. Participants often failed to notice content in the unattended ear ✓
- C. Participants noticed their own name in the unattended channel with no difficulty
- D. Participants performed better when the two messages were identical
Why: In dichotic listening studies, people attending to one message typically failed to report the content of the unattended message, suggesting that limited attentional resources filter out much of the unattended input.
Q3. What does the term 'task switching cost' describe?
- A. The money spent on training people to multitask
- B. The slower and less accurate performance that occurs when moving between tasks ✓
- C. The energy the brain uses while resting
- D. The fee charged for simultaneous phone calls
Why: Task switching cost refers to the reduction in speed and accuracy that results from shifting attention from one task to another, reflecting the time and cognitive effort needed to reconfigure goals and rules.
Q4. According to research on multitasking, which pairing is most likely to interfere with performance?
- A. Listening to music while walking on a familiar route
- B. Two tasks that both rely heavily on verbal processing ✓
- C. Folding laundry while listening to a podcast
- D. Breathing while reading silently
Why: Tasks that draw on the same cognitive resource, such as two verbal tasks, compete more strongly and produce greater interference than tasks using different modalities or well-practiced routine actions.
Q5. What is the effect of practice on divided attention performance?
- A. Practice has no effect because attention capacity is fixed
- B. Practice can reduce interference, making some combined tasks feel automatic ✓
- C. Practice always makes multitasking worse by overloading memory
- D. Practice only helps when tasks are performed in silence
Why: With extended practice, some tasks become automatic and require fewer attentional resources, which can allow people to perform two tasks together with much less interference, as seen in skilled pianists or drivers.
Module 5
Module 5: Attention Capture and Distraction
Attention Capture and Distraction
Knowledge check
Q1. Which type of attention capture occurs automatically when a sudden flash of light or loud noise grabs your focus, regardless of your current goals?
- A. Top-down attention
- B. Bottom-up (stimulus-driven) attention ✓
- C. Sustained attention
- D. Selective attention
Why: Bottom-up attention is driven by salient features of the stimulus itself, such as sudden changes, bright colors, or loud sounds. It operates automatically, pulling focus without requiring a conscious goal.
Q2. What is the term for the phenomenon where a person fails to notice a fully visible but unexpected object, such as a gorilla walking through a basketball game video, because their attention is focused elsewhere?
- A. Change blindness
- B. Inattentional blindness ✓
- C. Attentional blink
- D. Sensory adaptation
Why: Inattentional blindness describes failing to perceive a clearly visible stimulus when attention is occupied with another task. The famous gorilla study by Simons and Chabris demonstrated this effect.
Q3. What is the attentional blink?
- A. A reflex that closes the eyes when a bright light appears
- B. A brief period of reduced ability to detect a second target shortly after identifying a first target in rapid serial presentation ✓
- C. A decline in attention that occurs after 20 minutes of continuous work
- D. The involuntary shift of gaze toward a moving object
Why: The attentional blink is a short window, typically around 200 to 500 milliseconds, during which processing a first target makes it harder to detect a second target that appears soon after.
Q4. Which of the following is the best example of distraction caused by a stimulus that is irrelevant to your current task but captures attention because it is personally meaningful?
- A. Hearing a background hum from a refrigerator
- B. Noticing your own name spoken in a crowded room ✓
- C. Seeing a uniform gray wall
- D. Feeling a steady breeze on your skin
Why: The cocktail party effect shows that personally significant stimuli, such as your own name, can break through attentional filters even when you are focused on something else.
Q5. Research on task switching suggests that frequently switching between tasks primarily results in which outcome?
- A. Improved multitasking ability with no cost
- B. Increased switch costs, meaning slower and less accurate performance after each change ✓
- C. Permanent elimination of distractibility
- D. Faster processing speed without any loss in accuracy
Why: Switching tasks incurs measurable switch costs in both time and accuracy, because the mind must reconfigure goals and rules each time. Frequent switching therefore tends to reduce overall efficiency.
Module 6
Module 6: Attention in the Brain
Attention in the Brain
Knowledge check
Q1. Which brain region is most strongly associated with the top-down control of attention, including planning and selecting goal-relevant information?
- A. Primary visual cortex
- B. Prefrontal cortex ✓
- C. Cerebellum
- D. Brainstem medulla
Why: The prefrontal cortex supports executive functions such as goal maintenance and the top-down biasing of sensory processing toward task-relevant stimuli.
Q2. The parietal cortex, particularly the intraparietal sulcus, is known to be involved in which aspect of attention?
- A. Storing long-term memories
- B. Controlling voluntary spatial orienting of attention ✓
- C. Regulating heart rate during stress
- D. Producing speech sounds
Why: The parietal cortex, especially the intraparietal sulcus, plays a key role in directing attention voluntarily to locations or features in space.
Q3. What is the term for the brain's selective filtering mechanism in the thalamus that helps determine which sensory signals reach the cortex?
- A. Thalamic gating ✓
- B. Synaptic pruning
- C. Neurogenesis
- D. Myelination
Why: Thalamic gating refers to the thalamus regulating the flow of sensory information to the cortex, acting as a filter that enhances attended inputs and suppresses others.
Q4. Which neurotransmitter system is strongly implicated in regulating arousal and the alerting component of attention?
- A. Noradrenergic (norepinephrine) system ✓
- B. Cholinergic system only in the spinal cord
- C. GABA-only inhibitory pathways in the retina
- D. Serotonergic system in the hippocampus alone
Why: The locus coeruleus releases norepinephrine widely in the brain, modulating arousal and the alerting function of attention.
Q5. Neuroimaging studies found that when attention is directed to a particular visual feature such as color or motion, what happens in the corresponding sensory cortex?
- A. Activity is suppressed across all areas
- B. Activity is enhanced for the attended feature's processing region ✓
- C. Activity shifts entirely to the auditory cortex
- D. No change occurs in any region
Why: Attention boosts neural responses in the sensory areas specialized for the attended feature, such as area V4 for color or area MT for motion, which is a core finding of attentional modulation research.
Module 7
Module 7: Improving and Managing Attention
Improving and Managing Attention
Knowledge check
Q1. Which term describes the ability to concentrate on one stimulus while ignoring other competing stimuli?
- A. Divided attention
- B. Selective attention ✓
- C. Sustained attention
- D. Automatic processing
Why: Selective attention is the process of focusing on a particular stimulus while filtering out irrelevant information. Divided attention involves splitting focus across tasks, and sustained attention refers to maintaining focus over time.
Q2. What is the name of the phenomenon where people fail to notice an unexpected object in their visual field because their attention is focused elsewhere?
- A. The cocktail party effect
- B. The Stroop effect
- C. Inattentional blindness ✓
- D. Change blindness
Why: Inattentional blindness occurs when attention is directed toward one task so strongly that unexpected stimuli go unnoticed, as famously shown in the 'invisible gorilla' study. Change blindness is a related but distinct effect involving failure to detect changes between scenes.
Q3. Which of the following is a well-supported strategy for improving sustained attention during study?
- A. Multitasking across several subjects at once
- B. Working in short, focused intervals followed by brief breaks ✓
- C. Studying for as long as possible without any pauses
- D. Keeping your phone nearby so you can check notifications
Why: Short focused work intervals with breaks, such as the Pomodoro technique, help manage the natural decline in attention over time. Multitasking and constant notifications tend to fragment attention and reduce performance.
Q4. What does the term 'attentional bottleneck' refer to?
- A. A limit on how much information can be fully processed at one time ✓
- B. A physical delay in the eyes' ability to move between objects
- C. A memory storage problem that prevents recall of learned material
- D. A tendency to become fatigued after prolonged physical exercise
Why: An attentional bottleneck describes the limited capacity of the mind to process multiple inputs simultaneously, so only a subset of incoming information receives full processing. It is a cognitive limit, not a physical or memory issue.
Q5. Which practice is most associated with training the ability to voluntarily direct and sustain attention?
- A. Passive television watching
- B. Mindfulness meditation ✓
- C. Speed reading without comprehension checks
- D. Listening to music at high volume
Why: Mindfulness meditation repeatedly trains people to notice when their attention wanders and gently return it to a chosen focus, strengthening voluntary attentional control. The other options do not systematically train this skill.