Circadian variability is negligible in primary visual cortices as measured by fNIRS.

Neural activation leads to an increase of regional cerebral blood flow. Most of the functional imaging studies implicitly assume that variability of the hemodynamic response throughout a single day is negligible. To test this assumption we measured brain activation by functional near-infrared spectroscopy (fNIRS) in the visual cortex of ten subjects six times throughout the day, from 0800-1800 h, during an event-related checkerboard paradigm. Concentration of oxygenated hemoglobin increased, whereas concentration of deoxygenated hemoglobin decreased at each time point examined, without significant influences of daytime. Variability of the hemodynamic response was higher across subjects than for single subjects across day. In conclusion, our study is the first one supporting the common practice of ignoring circadian variability in functional imaging studies.

Spontaneous slow hemodynamic oscillations are impaired in cerebral microangiopathy.

Small-vessel disease or cerebral microangiopathy (CMA) is a common finding in elderly people. It is related to a variety of vascular risk factors and may finally lead to subcortical ischemic vascular dementia. Because vessel stiffness is increased, we hypothesized that slow spontaneous oscillations are reduced in cerebral hemodynamics. Accordingly, we examined spontaneous oscillations in the visual cortex of 13 patients suffering from CMA, and compared them with 14 age-matched controls. As an imaging method we applied functional near-infrared spectroscopy, because it is particularly sensitive to the microvasculature. Spontaneous low-frequency oscillations (LFOs) (0.07 to 0.12 Hz) were specifically impaired in CMA in contrast to spontaneous very-low-frequency oscillations (0.01 to 0.05 Hz), which remained unaltered. Vascular reagibility was reduced during visual stimulation. Interestingly, changes were tightly related to neuropsychological deficits, namely executive dysfunction. Vascular alterations had to be attributed mainly to the vascular risk factor arterial hypertension. Further, results suggest that the impairments might be, at least partly, reversed by medical treatment such as angiotensin-converting enzyme inhibitors/angiotensin II receptor blockers. Results indicate that functional near-infrared spectroscopy may detect changes in the microvasculature due to CMA, namely an impairment of spontaneous LFOs, and of vascular reagibility. Hence, CMA accelerates microvascular changes due to aging, leading to impairments of autoregulation.

Towards a standard analysis for functional near-infrared imaging.

Functional near-infrared spectroscopy (fNIRS) allows the ability to monitor brain activation by measuring changes in the concentration of oxy- and deoxy-hemoglobin. Until now no standardized approach for fNIRS data analysis has been established, although this has to be regarded as a precondition for future application. Hence, we applied the well-established general linear model to optical imaging data. Further, fNIRS data were analyzed in the frequency domain. Two visual tasks were investigated with optical imaging: a checkerboard paradigm supposed to activate the primary and secondary visual cortex, and a paradigm consisting of moving colored stimuli (rotating 'L's) additionally involving the motion area V5. Analysis with the general linear model detected the activation focus in the primary and secondary visual cortex during the first paradigm. For the second paradigm, a second laterally localized activated brain region was found, most likely representing V5. Spatially resolved spectral analysis confirmed the results by showing maxima of power spectral density and coherence in the same respective brain regions. Moreover, it demonstrated a delay of the hemodynamic response in the motion area. In summary, the present study suggests that the general linear model and spatially resolved spectral analysis can be used as standard statistical approaches for optical imaging data, particularly because they are almost independent of the assumed differential path length factors.