Lecture 3: Attention

Contact Information

• Email: matt.roser@plymouth.ac.uk
• Office: PSQ B207
• Office Appointment:
  • Tuesday 10-11am
  • Thursday 10-11am
• Check in code XX-XX-XX
• Dr Matt Roser

Lecture 3: Part 1. Attention

• Readings:
  • Chapter 8. Baars & Gage. Fundamentals of cognitive neuroscience: a beginner's guide.
  • Chapter 11 of Banich and Compton, 3rd Ed. Useful supplement.
  • Chapter 12 of Gazzaniga, Ivry, Mangun (2009), Cognitive Neuroscience (3rd Ed.) Useful supplement.

Lecture Overview

• What is attention / what does it do?
• Theoretical models of attention
• How to study attention
• Failures of attention
• Brain mechanisms of attention
• Exclusions:
  • Arousal
  • Sustained attention
• Focus: Attention to locations, features, and objects.

What is Attention?

• William James (1890): ‘Everyone knows what attention is.’
• More information is available than we can process at one time.
  • Sensory stimuli (e.g., the socks on your feet)
  • Thoughts
• Attention is the preferential treatment / selection of a subset of that information.
• Involves both:
  • Voluntary/controlled processes
  • Involuntary/reflexive processes (e.g., sounds made by people nearby)

Theories of Attention

• Selective Attending
  • Broadbent’s (1958) model (B&C p306 on)
    • Top-down selection of relevant inputs (e.g., conversation at a party) at an early stage of processing
    • But some unattended information intrudes (e.g., your name)

Theories of Attention: Bottleneck

• At what stage does the ‘bottle-neck’ of limited capacity lie?
  • Early versus late selection:
    • Early: affects perceptual analysis (only process info from select locations?)
    • Late: affects higher-level stages (decisions, semantic/memory encoding)
  • A modified theory: early attenuation with later selection

Attention in Space

• Evidence for early selective attenuation of information from visual cueing of spatial locations
• Posner’s cueing paradigm
  • Fixate centrally
  • Present a central predictive cue
  • Involves voluntary, covert orienting
  • Valid Trials: Target appears at the cued location
  • Invalid Trials: Target appears opposite the cued location
  • Reaction Times (RTs) compared between Valid and invalid conditions

Attention in Space: Results

• RTs were fastest to targets in the Valid condition
• This suggested that stimuli presented at the valid target location were preferentially processed
• Question: Was this selection at an early or late stage?

Attention in Space: ERP Evidence

• ERP evidence for early selective attenuation (B&C p314)
• Early visual evoked potentials, generated in extrastriate cortex, were of larger amplitude for targets at Valid locations than targets at Invalid locations
• Preferential processing of attended input
• See also B&C Fig 11.7 (Mangun & Hillyard, 1991)

Voluntary and Reflexive Orienting

• Central, symbolic, cues evoke voluntary shifts of attention
  • This is slow
  • Validity effects show up with long Stimulus Onset Asynchronies (SOAs)
    • SOA is the delay between the presentation of the cue and the presentation of the target
    • Here it’s 800ms
• Top-down control and Bottom-up processes

RT Table: Voluntary Orienting

• The table shows reaction times (RTs) in milliseconds (ms) for valid and invalid conditions at short (100ms) and long (800ms) Stimulus Onset Asynchronies (SOAs).
• A validity effect is observed only at the long SOA, demonstrating the slow nature of voluntary orienting.

Voluntary and Reflexive Orienting

• Peripheral, non-symbolic, cues evoke reflexive shifts of attention
  • This is fast
  • Validity effects show up with short Stimulus Onset Asynchronies (SOAs)
    • SOA is the delay between the presentation of the cue and the presentation of the target
• Top-down control and Bottom-up processes

RT Table: Reflexive Orienting

• The table illustrates reaction times (RTs) in milliseconds (ms) for valid and invalid conditions at short (100ms) and long (800ms) Stimulus Onset Asynchronies (SOAs).
• A validity effect is apparent usually only at short SOAs, reflecting the rapid nature of reflexive orienting.

Top-down vs. Bottom-up

• Visual search - finding target features amongst distractors
  • Targets defined by single features can be identified preattentively (i.e. perceptually, without applying attention)
    • They ‘pop out’
  • Targets defined by feature conjunctions require serial attention (i.e. searching) to ‘bind’ the properties (features) of objects together
    • E.g. ‘redness’ and ‘O shape’

Visual Search Strategies

• Strategies can increase the efficiency of serial visual search
  • Guided visual search results in a lesser impact of added distractors
    • Search is restricted to a subset of items, based on a single feature (target = red circle; exclude black items from serial search)
• Top-down control and Bottom-up processes

Attention to Objects

• Attention can also be ‘object based’ (B&C p311)
• Objects can influence the orienting of attention in response to cues
• RTs to Invalid-Same-Object trials faster than RTs to Invalid-Different-Object trials (Egly, Driver, & Rafal, 1994, Journal of Experimental Psychology: General, 123(2), 161-177)

Attention to Objects

• Subjects can selectively attend to stimulus class (e.g. face or place) and show correlated changes in inferotemporal regional brain processing
  • Wojciulik, Kanwisher, and Driver (1998). Journal of Neurophysiology, 79,1574-1578.
• Attentional deficits can track objects

Disorders of Attention

• Hemispatial neglect
  • Usually results from right parietal damage
  • Neglect of contralesional space (usually left)
  • Space can be ‘object defined’
  • Lesion overlay map

Disorders of Attention

• Neglect can follow an object (object reference frame)
• Deficit in attention to visual memories (not a memory deficit) (egocentric reference frame)
• Imagine a piazza from either end

Brain Mechanisms of Attention

• What is attention in the brain? (B&C p315-316)
• Attention selects information for preferential processing in a variety of ways:
  • Spatial location - increased activation is observed in sensory areas that are organized with regard to space, such as early visual-processing areas, and regions that provide a spatial map of the world, such as parietal regions.
  • Item attributes - activation is increased in the brain region specialized for processing attended characteristic, such as motion-sensitive area MT.
  • Objects - increased activation is observed in areas that process objects, such as the ventral visual-processing stream.
• Attention does not reside in one particular region of the brain but influences the processing of distinct brain modules, ramping up their activity if they are processing information that is attentionally relevant.
• Attention is a modulatory process.

Brain Mechanisms of Attention

• Attentional control and amplification of widespread processing
• Multiple regions activated in fMRI studies of attentional orienting
• Top-down modulation of early-processing areas via reciprocal connections through the thalamus (B&C p313 – 315)
• Synchronisation of neural firing prior to stimulus input? (suggested by regional fMRI activation)
• Neuron’s preferred stimuli may become more specific with increased attention (i.e. the ‘tuning curve’ narrows – suggested by recordings in monkeys)

Brain Mechanisms: Attentional Networks

• Three main attentional networks in the brain.
  • Alerting - maintaining a high state of sensitivity to incoming stimuli.
  • Orienting - attending to the source of a sensory signal.
  • Executive - directs attention according to an individual’s goals
• B&C Fig. 11.1 (Posner & Rothbart, 2007)

Conclusions

• Selection happens at both early and late stages
• Attention is modulated by both top-down and bottom-up processes
• Attention can be paid to locations, features or to objects
• Multiple cortical and subcortical brain areas are involved