Neglect

1
Neglect

• By the end of this lecture you will have learned
• More neglect phenomena
• Outlines of some theoretical accounts of
  neglect
• That no theory currently accounts for all
  neglect phenomena
• That neglect phenomena are informing models
  of how the brain encodes space.
• Essential reading
• Parkin, Ch 5
• Additional reading
• Bradshaw JL Mattingley JB (1995) Clinical
  neuropsychology. Behavioral and brain science. Ch
  6
• Vallar, G. (1998) Spatial hemineglect in humans.
  Trends in Cognitive Sciences, 2, 87-97

2
Neglect - early theories
Focused on the sensory / motoric features of
neglect E.g. neglect is consequence of visual
field deficit (in addition to a generalised loss
of info processing ability) Clearly not the case
in representational or tactile neglect Also cant
explain implicit processing
Is the deficit primarily motoric? Evidence against
Walker et al (1991) Simple target detection
paradigm with 3 conditions Overlap Target and
fixation point overlap 0 Gap Fixation offset
with target onset 100 Gap Target onset 100 ms
after fixation offset
3
Overlap condition
8
5


(Fixation point remains during target
presentation)
4
0 ms gap condition


(Fixation disappears simultaneously with target
presentation)
5
100 ms gap condition
8
5


(Fixation disappears 100ms before target
presentation)
6
Walker et al
PREDICTIONS If neglect is due to a sensory /
motor deficit, target detection in neglected
field should be equally poor in each
condition. E.g. no reason why a central
manipulation should effect target detection in
neglected field.
In fact, performance was far better in the 100ms
gap condition (see Posners attentional
hypothesis for explanation)
7
Theoretical Approaches
A distinction is commonly made between
attentional and representational models of
neglect Attentional models Deficits in
mechanisms involved in maintaining arousal or
vigilance and in controlling orientation toward,
and selection of, stimuli as targets to be
elaborated for further processing. (Kinsbourne,
1987 Humphreys Riddoch, 1993 Posner, 1988
Heilman) Representational models Deficits in
mechanisms involved in reconstructing an internal
map of space, based either on sensory or
internally derived information (Bisiach, 1993
Rizzolatti Berti, 1990)
8
Neglect - attentional theories
The very term neglect implies an attentional
involvement Effects of cueing in line bisection
provides supportive evidence. There are many
other examples of neglect being ameliorated by
cueing patients to attend to the neglected
side BUT - attention is not a unitary
construct Often difficult to establish exactly
what is meant by attentional involvement Stronges
t theories based on reasonably well articulated
models of normal attentional functioning
9
Neglect - Kinsbournes model

• Orientational bias model - hemispheric rivalry
  model -vectorial model. One of the oldest
  models.
• Attention controlled by paired opponent
  processesors controlled by R L hemispheres
• Each processor directs attention (controls an
  orienting response) to contralateral space
• In neglect the disinhibited left processor is
  in control, biasing attention to the right side
  of space

• Predictions
• There is an attentional gradient
• Attention is exaggerated towards the right e.g.
  an ipsilesional bias, not just a contralesional
  impairment

10
Neglect - Kinsbournes model

• There is some evidence for attentional gradients

• Smania et al (1998)
• RTs to detect targets presented at different
  eccentricities
• Triangles contralesional Circles Ipsilesional
• Slower the further left
• Faster the further right
• Kinsbourne (1993) gives other examples.

11
Neglect - Kinsbournes model

• Can explain representational neglect because
  processors assumed to be active in exploration of
  external and internal space.
• Can also explain why left side of objects, or
  left of two objects is neglected, even if in
  right hemispace

Problems - not all unilateral lesions produce
neglect Not all patients have attentional
gradients Cant explain object centred neglect
Not clear why right neglect should be so
rare. Heilman (1995) and Mesulam (1998) assume RH
controls attention to LR space, LH only controls
attention to R space.
12
Neglect - Posners model

• Shifting attention involves
• Disengaging
• Shifting
• Re-engaging
• Evidence from patients with neglect, PSP and
  pulvinar lesions provide some support for this
  model
• Core deficit in neglect is difficulty in
  disengaging attention from an ipsilesional
  stimulus

13
Posners paradigm
14
Posners paradigm
15
Posners paradigm
16
Posners paradigm

Cue is a valid predictor on 80 of trials
Delay between cue and target can be varied
17
Neglect - Posners model
Neglect patients miss catch trials, or are
delayed in responding because cannot disengage
attention from cue
18
Neglect - Posners model
Walker et al results also explained by this model
- performance is best when attention does not
have to disengage (e.g. in gap condition) Mark
et al (1988) Neglect patients perform better in
a cancellation task if they were allowed to erase
the lines rather than cross them. E.g. attention
no longer had to be disengaged from stimuli on
right in erase vs cross condition. Problems - in
dark, neglect patients search mainly on their
non-neglected side - not predicted if no stimuli
are present to capture attention.
19
Neglect - Limiting theories of attention
If it assumed that there is no attention (or
severely impaired attention) on the neglected
side, then data from neglect patients can be used
to inform theories of attention 1. Does object
segmentation occur pre-attentively? Triesman -
Feature Integration Theory Preattentive Simple
features registered in parallel Attentive
Objects formed by combining features Prediction
Objects are not segmented in neglected field BUT
- Demonstrations of object based neglect argue
against this.
20
Neglect - Limiting theories of attention
1. Does object segmentation occur
pre-attentively? Driver et al (1997) Extinction
Patients report seeing only circle on right
Patients report seeing a barbell
E.g. attentional system capable of object
segmentation
21
Neglect - Limiting theories of attention
2. Early vs Late selection Can unattended
information be processed semantically? Berti
Rizzolatti (1992)Speeded Animal / Vegetable
decision task for object in right hemifield
RTs for correct categorisation of right
object Identical 777ms Same Category 795ms
Different Category 890
22
Neglect - Limiting theories of attention
3. What is the object of attention? Does
attention operate on the basis of locations in
space or objects? Neglect phenomena highlight
the variety of reference frames in which
attention appears to operate Also, dissociations
have been observed for neglect in personal,
peripersonal and extrapersonal reference
frames Halligan Marshall (1991) Severe
neglect in peripersonal space (e.g. conventional
line bisection) Little or no neglect in
extrapersonal space (pointing a light, throwing a
dart) Cowey et al, 1994 extrapersonal neglect
only Guariglia Antonucci, 1992 Personal
neglect only
23
Neglect - Limiting theories of attention
3. What is the object of attention? See previous
lecture for evidence that attention can also be
object based. Object centred neglect In an
object centred frame of reference, the objects
parts are computed in relation to its
structure Patient NG (see previous lecture) has
object centred neglect Tipper Behrmann,
1996Target detection task -error rate
Static condition Left 75 Right 25
Rotate condition Left 10 Right 65
24
Neglect - representational theories

• E.g Piazza del Duomo effect
• Assume a construct such as Baddeleys VSS - a
  screen on which spatial information may be
  represented
• In neglect part of screen (e.g. left) is
  dysfunctional
• There are circumscribed brain areas where
  lesions would result in a representational loss
  limited to definite regions of (egocentric)
  space Bisiach et al, 1985 - representational
  scotoma
• Attention is not allocated to objects, but to
  representations of objects provided by the
  perceptual system
• Attentional theories disregard this step.

25
Neglect - Bisiachs model

• Many other examples of representational neglect
• Problems - Not clear how well this model
  accomodates object centred neglect - although
  objects are clearly represented at some level
• Also some patients demonstrate visual but not
  representational neglect - Brain, 1941
• Bisiach concedes representational scotoma
  analogy may have been overkill (Bisiach, 1993).
  Also lists other problems and potential answers.
• Posner data also difficult to accommodate.
• Rizzolattis model is also representational -
  attentional deficits are secondary

26
Neglect - transformational theories
Suggest that processes which translate sensory
input into motor output are impaired
Problems - not all patients show no
displacement of the subjective straight ahead -
Only address ego-centric neglect (not allocentric
or object centred)
27
Neglect - adequacy of explanations
...few current explanations of neglect are
anything more than a description of the phenomena
phrased in terms that insinuate understanding
without actually delivering to say that neglect
is a representational disorder is to describe,
not explain, the fact that some patients show the
Piazza del Duomo effect Marshall et al, 1993
The term neglect should be treated by CNs in
the same way as apahsia is - e.g. it does not
exist as an entity - It is a useful shorthand
for conveying a range of perceptual, attentional,
intentional and representational problems in
dealing with left space - all of which may have
different underlying causes.
28
Neglect
Attentional and representational models of
neglect are not necessarily mutually
exclusive such evidence widens the scope of
the interpretations that can be given of these
disorders in terms of attention to such an extent
as to make attentional explanations logically
indistinguishable from representational
explanations. Bisiach, 1994 Attentional
networks are important in the creation and
scanning of a visual image just as they must work
with object recognition in actual perception
Posner, 1993
29
Neglect - The neural coding of space
The beginnings of a resolution? Recently
investigators have begun to relate neglect to the
neural coding of space (e.g. Pouget Driver,
2000 Pouget Sejnowski, 2001) Neurons in
monkey parietal cortex have retinotopic receptive
fields which can be modulated by non-retinotopic
information (e.g. posture, eye position -
possibly attention?)
30
Neglect - The neural coding of space
Animal work suggests that the same parietal
neurons involved in encoding representations of
objects are also involved in directing attention.
(e.g Colby Goldberg, 1999) E.g.
representational vs attentional debate is not
really relevant. In addition, many neurons in
the parietal lobule are involved in multisensory
integration (e.g. cells exist with congruent
receptive fields for stimuli in different
modalities) So can explain some of the cross
modal effects in neglect (e.g cross modal
extinction)
31
Neglect - Computational Models

• Pouget Sejnowski (2001)
• Neurons in parietal cortex represent space using
  basis functions
• Given eye position and retinal position, basis
  functions allown an objects position can be
  transformed into other reference frames (e.g.
  head centered).
• This allows objects to be encoded in multiple
  frames of reference simultaneously
• Computational model developed entirely on the
  basis of computational principles and
  neurophysiological data.
• NP data can be used as an independent test of the
  model
• Model lesioned by deleting right basis function
  maps
• Appears to account well for most
  neuropsychological data

32
Neglect - Difficult data
Halligan Marshall (1991)
Many neglect patients do not have a problem with
marking the centre of a sqaure
33
Neglect - Outstanding questions

• Given the many different codes in which space is
  represented neurally, why is the human
  phenomonological experience of space unitary?
• What is role of ST / LT memory in neglect
• To what extent is the neglected information
  recreated
• What strategies are most effective for
  rehabilitation? Robertson - general attentional
  problems (reduced capacity) in addition to
  neglect.

34
Neglect - Conclusions

• Given the range of phenomena regarded as
  manifestations of neglect it is unlikely any
  unitary theory will ever provide an adequate
  explanation
• Advances in cognitive neuropsychology tend to
  reflect the introduction of a detailed model of
  normal performance
• There are no detailed models of spatial
  cognition
• Recent research has concentrated less asking
  what is neglect? and more on asking what can
  data from brain damaged individuals tell us about
  how the brain encodes spatial information?.
• Data from neuroscience and computational
  modelling may provide a way forward