Attention

Attention

1. Early Selection Models

   • Information Processing System (Broadbent, 1958) Gate closed to unattended info
   • EEG of auditory attention- Event related Potentials
     • Hillyard et al (1973) streams of sounds in earphones. Larger ERPs when attended to- N1 wave. 90ms after stimulus onset
     • Woldorf & Hillyard (1991) More participants and trials. Earlier waves 20-50ms processing in primary auditory cortex or even medial geniculate nucleus in thalamus
     • Woldorf & Hillyard (1993) used MEG to localize N1 and P20-50 to auditory cortex
   • EEG of voluntary visual attention
     • Focus on centre point and pay covert attention to one area, attending to stimuli in that area
     • Eason et al (1969) Modulations begin 70-90ms after onset. Amplitude of wave larger when attended
       • Early effects not found when based on other complex features such as colour, orientation, or judging what it is doesn't lead to short latency modulation (greater than 120ms)
   • EEG of reflexive visual attention
     • Attention reflexively attracted by sensory event preceding target, response times are quicker
       • Only quicker when time between cuing and target are short (<250ms)
       • Long periods between cue and target lead to inhibition of return
       • Hopfinger & Mangun (1998) Early occipital P1 wave enlarged for targets quickly following cue
         • Long time after cue causes effect to reverse, P1 response diminishes or is inhibited
   • fMRI study Kastner et al (2006) effects of attention manifest before cortical processing?
     • Attention cued to one side, when target was detected, lateral geniculate nucleus of thalamus was activated and multiple cortical visual areas
       • Greater in amplitude for attended rays on contralateral LGN and visual cortex

2. Covert attention

   • von Helmholtz (1984) focused on a central point while covertly attending elsewhere
   • Cocktail Party Effect (Cherry, 1953) Selective auditory attention and covert auditory attention
     • Dichotic listening task-unattended info ignore

3. Late Selection Model

   • Unattended info can suddenly grab attention
   • All info must be processed, regardless of attending or not (Moray, 1959)
     • Processed equivalently, reaching semantic encoding

4. Neuroimaging of features

   • Subjects paid attention to features of changing stimulus; form, colour, motion or location
     • Increased attention led to increased processing in areas associated with processing feature
       • V4= colour; V5= motion
       • Paying attention to particular feature enhances processing in area of brain responsible
   • Objects
     • O'Craven et al (1999)
       • Images of houses and faces superimposed together, one stimulus moves
       • Subjects told to pay attention to one feature
         • When paying attention to faces, activity in fusiform face area increased
         • When paying attention to houses, activity in parahippocampal place area increased
         • When paying attention to motion, activity in V5 area increased, as well as area associated with the picture being attended to
       • Attention facilitates processing of features of attended object

5. Bias Competition Model

   • When 2 stimuli fall on receptive field of a visual neuron - competition to control firing of neuron
     • Attention resolves competition as neurone fires for attended stimuli
   • Kastner et al (1998)
     • In the absence of focussed attention, stimuli interfere with each other
       • Reduces neural response evoked in comparison to when shown sequentially
     • When attention is used, stimuli no longer interfere
       • Attention attenuates influence of distractor

6. Attentional control networks

   • Hopfinger et al (2000) modified Posner's so target is flickering checkerboard. Indicate when attended board had a grey square
     • 8s gap between cue and target allowed haemodynamic responses linked to attending cue to be analysed separately from responses linked to detection and response to target
       • Areas involved included suerior frontal cortex, inferior parietal cortex, superior temporal cortex, posterior cingulate cortex and insula
         • Areas not involved in sensory processing of cue, as if passively viewed, these areas are not activated

7. Neglect Syndrome

   • Caused by unilateral damage to Inferior Parietal Lobe, causing neglect in space opposite
     • Damage more common on right hemisphere, so patients neglect left hemispace
       • 2 models put forward to explain this preponderance
         • Representational Model: RH has attentional representations for R and L hemispace, LH only for R hemispace, so damage to LH doesn't cause severe neglect as RH can represent both
         • Attentional Bias Model: LH and RH have attentional biases to contralateral hemispace. Bias is asymmetrical, LH is more strongly biased ti R hemispace. Bias usually unnoticeable
           • With RH lesion, attentional bias in LH is unopposed by RH, so bias of LH to direct attention to the R hemispace increases.
             • LH lesions produce less severe neglect
   • Extinction
     • Failure to perceive stimuli in opposite field
       • If stimuli presented one at a time, then normal detection - not a visual problem
     • Tests for Neglect
       • Line bisection - given a sheet with horizontal lines and asked to bisect in the middle of all lines
         • Patients strike to the right of lines, and only in middle and right horizontal lines
       • Bisach & Luzzati (1978) Imagine themselves in local Piazza and describe surroundings
         • Only describe things on one side. When asked to imagine themselves on the opposite side, they name all things on the other side
           • Not a problem of memory
   • Implicit Processing
     • Marshall & Halligan (1988) 2 identical houses but one has flames in window. Patients cannot perceive the difference, but when forced to chose, chose the one without flames
       • Ladavas (1993) Words presented in neglected field to prime another word in good field. Neglected word processed to a semantic level as patients were faster to decide if word was real or not
   • Cortical Mechanisms
     • Heilman et al (1983) CT scans on 10 patients found Right Inferior Parietal Lobule and Temporoparietal Junction were sites of most overlap
       • Vallar & Perani (1986) supported this involvement
         • Karnath et al (2001) Brain sans of a large group with RH lesions, with and without neglect. Regions most associated with neglect were superior temporal gyrus, insula and basal ganglia
           • Differences found could be due to different methods used to analyse the data, or how participants were selected for