Davidson, R.D. (1998)
EEG measures of cerebral asymmetry: Conceptual and
methodological issues
J. Neuroscience, 39, p 71-89
Comparisons with other neuroscience methods
cf FMRI
fewer constraints on task presentation and performance
Critical
periods can be flagged ie examine epoch during which facial expression is
recorded
Distinctions between Hemisphere specialisation and
hemisphere activation
Hemisphere
Specialisation
Preparedness
to do a task, one hemisphere less accomplished than the other for a particular
task
Hemisphere
Asymmetries in Activation
p87 '
Activation refers to the degree to which a particular hemisphere is working or
engaged'
p 73 '
Activation is typically defined operationally on the basis of the methods used
in a particular study to measure this construct'
p 75 ' A
hemisphere may become selectively activated as a function of priming' this means that the task parameters may
influence the activation level e.g. The requirement to respond with one hand or
another hand (KRO think about cortical activity at a primary level).
' testing
must be performed on a sample assumed to be homogenous with respect to
individual differences in hemisphere specialisation'
Within
subject designs therefore preferred although there can be floor/ceiling effects
ie task effect not apparent because baseline measure already at ceiling/ floor
Dissociations between specialisation and activation
'Dissociations
between these two aspects is common' ' The hemisphere specialised to process a
particular stimulus is not necessarily the one most activated'
p73 ' it
is not necessarily the case that differences in activation are paralleled by
differences in specialisation'
rostral -caudal differences in hemisphere specialisation
and activation
Anterior
hemisphere regions associated with
affective processing
Posterior
hemisphere regions associated with cognitive processing
I.e.
activations are relatively orthogonal
p76 ' stimuli which differ in affective
valence systematically influence activation asymmetries in frontal brain electrical
activity in the absence if any simultaneous measures in parietal asymmetry,
while tasks which are designed differentially to require verbal versus visual
spatial produce changes in parietal and temporal asymmetry in the absence of
any modifications in frontal asymmetry processing produce changes in parietal
and temporal asymmetry'
This has
been checked by the author by taking both frontal and parietal/temporal
measures at the same point in time and performing across(rostral-caudal) (e.g.
r frontal v r parietal/temporal correlations) p87 ' it makes little sense to
talk of a whole hemisphere being activated or specialised for a particular
function'
Methodological Issues
1. metrics of asymmetry
Require
the sites to be homologous
Choose
the dependent measure; oscillation, ratio/raw
( note for ratio measures (R-L)/(R+L) higher
ratios are produced by less alpha in the left hemisphere and/or more alpha in
the right hemisphere, both of which are associated with more relative left
sided activation ( less alpha is taken to be indicative of more activation'
Between
group as opposed to within group studies
Small
differences ( between hemisphere) may be more significant than overall
activations
2.
Artefact
Blink
& eye movement
EMG What
is the spectrum characteristics, how to detect
p78'
facial expression , there are reports of asymmetry. ' in our experience most
severe in the beta range' ( but doesn't
evidence this) ' the presence of muscle artefact is sufficiently likely in
certain scalp locations ( e.g. Temporal leads) as to make meaningful assessment
of beta almost impossible' ' in our laboratory, we sample the EEG at 250 Hz and
compute power density in a high frequency band (70-80Hz) which does not contain
any neurogenic activity and is presumably a function of muscle activity
exclusively' ( KRO does that view still hold?)
3
Reference electrode
Laplacian
operator for dealing with questions that arise concerning the contribution of the
reference electrode
Or
Record A1
and A2 with same reference, then average these channels to form a new reference