Functional imaging with near-infrared spectroscopy (NIRS) : correlation between brain response, apoE genotype, and neuropsychological test performance

Near infrared spectroscopy (NIRS) is a non-invasive optical technique which measures
concentration changes of oxygenated (HbO), deoxygenated (HbR), and total hemoglobin
(HbT) in brain tissue. In conjunction with the ApoE genotype and neuropsychological
measures it might reveal new insights into brain-behaviour relationships.
We studied 240 non-demented elderly individuals selected from the project BASEL
cohort (69 females, 171 males; mean age = 72.3 ± 7.03 years) and 21 patients with probable,
mild Alzheimer’s Disease (12 females, 9 males; mean age = 76.2 ± 6.37 years; mean MMSE
= 25.3 ± 2.08) with multi-channel (left and right frontal, left and right parietal, left and right
temporal) NIRS during performance of either a verbal (letter) fluency task or a computerized
labyrinth test. On the same day, non-demented elderly subjects obtained a comprehensive
neuropsychological assessment - the patients with AD had their neuropsychological
examination first, followed by the NIRS examination maximally 10 days later. In all study
participants the ApoE genotype was determined. The proportion of ApoE-ε4-positives and –
negatives was evenly distributed in both study groups (non-demented elderly/patients with
AD: 45% ApoE-ε4 positives, 55% ApoE-ε4 negatives).
Mean concentration changes of HbO and HbR during cognitive activation with either
task revealed a typical hemoglobin oxygenation response pattern in normal elderly subjects:
an increase in HbO and a decrease of HbR over brain areas activated by the task. The NIRS
parameter HbR turned out to be a specific marker of brain activation, as significant decreases
in HbR were only seen with the verbal fluency task and over the frontal optode positions. In
contrast, the parameter HbO exhibited a pattern of general activation across the whole brain
and for both activation tasks and thus seems to reflect a state of arousal with a very high
sensitivity but no specificity.
Using multiple regression analysis we found that neither the sociodemographic factors
age, years of formal education, and gender, nor the ApoE genotype act as predictors of the
NIRS response in normal elderly subjects. Specifically, ApoE-ε4 carriers did not differ from
ApoE-ε4 non-carriers with respect to concentration changes of HbT at the left parietal
position.
Investigating the relationship between covert and overt behaviour, we found a
statistically significant correlation between the NIRS response during performance of the
verbal fluency task and the NIRS task performance: Subjects able to bring the brain areas
critical for successful performance in a cognitive task on-line, show better task performance.
Further, subjects with a typical NIRS response pattern at the left-frontal position (Left-frontal
NIRS responders) performed significantly better in the NIRS verbal fluency task than subjects
who did not show a typical NIRS response at the left-frontal position (Left-frontal NIRS nonresponders).
In conjunction with the neuropsychological data collected on the same day, we
investigated the relationship between the NIRS response during performance of the verbal
fluency task and neuropsychological test performance. We found that the NIRS parameters
HbO and HbR at the left and right frontal and left temporal position were able to predict 42%
of the performance in a test of executive function (Trail Making Test, quotient of Form B/A).
On another route, we compared the neuropsychological test performance of left frontal NIRS
responders with the performance of left-frontal NIRS non-responders. The results of these
analyses were largely in line with our hypothesis: With the exception of the Wisconsin Card
Sorting Test, left-frontal NIRS responders scored significantly better in tests of executive
function. Further, the subjects with a typical NIRS response at the left frontal position
performed markedly better in the WAIS Block Design and slightly better in the Boston
Naming Test and the WAIS Similarities.
In comparison to our sample of non-demented elderly subjects, the patients with mildstage
AD did not show a reduced NIRS response during cognitive activation as compared to
rest. However, in comparison to the cognitively normal subjects, they showed a clearly
distinguishable profile in the relationship between covert and overt behaviour and between
NIRS response and neuropsychological test performance: The NIRS task performance of the
patient group was negatively correlated with the brain response. Further, left-frontal NIRS
responders in the patient group showed a statistical trend towards lower scores in tests of
executive function. In other words, the patients with AD – yet at a mild stage of the disease -
could not make use of the cortical network required for an efficient task performance as did
the group of cognitively normal elderly subjects.
In conclusion for future functional imaging studies, the analysis of brain-behaviour
relationships might add valuable benefit to the investigation of the brain response alone.