3 Introduction to Brain Imaging
57
Perhaps even more problematic is the sparse knowledge of the predictive
value of various fMRI paradigms and analysis procedures for patient postsurgical
outcome. Given the complex representation of function in the brain, the large
intraindividual variability in brain anatomy and function, and the potential for
functional reorganisation and plasticity in healthy brains and due to disease, there
is not a simple relationship between focal brain lesions and functional deficits.
Therefore, the predictive value of fMRI presurgical mapping for patient outcome
must be estimated based on experimental data and in the context of specific
activation paradigms and analysis procedures. Systematic large-scale studies to
determine the value of fMRI presurgical mapping of different activation paradigms
in various patient populations are still rare. Nevertheless, recent efforts have yielded
white papers with recommended paradigms for universal presurgical fMRI mapping
and systematic assessment of the diagnostic accuracy and prognostic value (D. F.
[16]), for determining hemispheric lateralisation and predicting postsurgical deficits
of cognitive functions such as language and memory [121].
Nevertheless, it is important to remember that alternative mapping methods such
as the gold standard cortical stimulation method suffer from similar limitations. Cortical stimulation is considered to identify regions that are functionally essential, and
their damage during surgery would therefore result in neurological deficit. However,
this technique cannot account for possible postsurgical functional reorganisation
that could prevent permanent loss of function [56, 102]. Cortical stimulation can
also in some cases indirectly interfere with brain areas connected with, but removed
from, the stimulated site. Thus, there is a risk of false-positive results with cortical
stimulation, and like fMRI, the predictive power for patient outcome needs to be
estimated based on systematic analysis of experimental data.
3.2.2.1 Language and Memory Hemispheric Lateralisation
An important application of fMRI mapping is for assessing hemispheric dominance
for language and memory (Fig. 3.2). The rationale for the test is the assumption
that surgery on the dominant hemisphere carries a higher risk for postsurgical
cognitive decline. The clinical gold standard to assess hemispheric dominance is
the Wada test [131], which involves intracarotid injection of a barbiturate to sedate
one hemisphere at a time and carries risk and discomfort for the patient. Several
factors may confound the interpretation of the Wada results, including agitation or
somnolence of the patient related to the anaesthetisation, inadequate disabling of
certain brain regions in the anaesthetised hemisphere, or uncontrolled crossover
of the anaesthetic to the examined hemisphere [76, 112]. In comparison, fMRI
is noninvasive and performed under conditions of close-to-normal language processing. fMRI also provides sub-hemispheric spatial resolution. For these reasons,
language fMRI has the potential of being more sensitive and accurate than the Wada
test. Concordance between fMRI and Wada language lateralisation has generally
been reported as very high, in the range of 80–90% [14, 64, 68], although this
evaluation depended on an assignment to dominance categories (left, right, bilateral)
57
Perhaps even more problematic is the sparse knowledge of the predictive
value of various fMRI paradigms and analysis procedures for patient postsurgical
outcome. Given the complex representation of function in the brain, the large
intraindividual variability in brain anatomy and function, and the potential for
functional reorganisation and plasticity in healthy brains and due to disease, there
is not a simple relationship between focal brain lesions and functional deficits.
Therefore, the predictive value of fMRI presurgical mapping for patient outcome
must be estimated based on experimental data and in the context of specific
activation paradigms and analysis procedures. Systematic large-scale studies to
determine the value of fMRI presurgical mapping of different activation paradigms
in various patient populations are still rare. Nevertheless, recent efforts have yielded
white papers with recommended paradigms for universal presurgical fMRI mapping
and systematic assessment of the diagnostic accuracy and prognostic value (D. F.
[16]), for determining hemispheric lateralisation and predicting postsurgical deficits
of cognitive functions such as language and memory [121].
Nevertheless, it is important to remember that alternative mapping methods such
as the gold standard cortical stimulation method suffer from similar limitations. Cortical stimulation is considered to identify regions that are functionally essential, and
their damage during surgery would therefore result in neurological deficit. However,
this technique cannot account for possible postsurgical functional reorganisation
that could prevent permanent loss of function [56, 102]. Cortical stimulation can
also in some cases indirectly interfere with brain areas connected with, but removed
from, the stimulated site. Thus, there is a risk of false-positive results with cortical
stimulation, and like fMRI, the predictive power for patient outcome needs to be
estimated based on systematic analysis of experimental data.
3.2.2.1 Language and Memory Hemispheric Lateralisation
An important application of fMRI mapping is for assessing hemispheric dominance
for language and memory (Fig. 3.2). The rationale for the test is the assumption
that surgery on the dominant hemisphere carries a higher risk for postsurgical
cognitive decline. The clinical gold standard to assess hemispheric dominance is
the Wada test [131], which involves intracarotid injection of a barbiturate to sedate
one hemisphere at a time and carries risk and discomfort for the patient. Several
factors may confound the interpretation of the Wada results, including agitation or
somnolence of the patient related to the anaesthetisation, inadequate disabling of
certain brain regions in the anaesthetised hemisphere, or uncontrolled crossover
of the anaesthetic to the examined hemisphere [76, 112]. In comparison, fMRI
is noninvasive and performed under conditions of close-to-normal language processing. fMRI also provides sub-hemispheric spatial resolution. For these reasons,
language fMRI has the potential of being more sensitive and accurate than the Wada
test. Concordance between fMRI and Wada language lateralisation has generally
been reported as very high, in the range of 80–90% [14, 64, 68], although this
evaluation depended on an assignment to dominance categories (left, right, bilateral)