Marieke van den Kerkhof1,2, Jacobus F.A. Jansen1,2,3, Lisanne P.W. Canjels1,2,3, Robert J. van Oostenbrugge2,4,5, Benedikt A. Poser6, and Walter H. Backes1,2
1Department of Radiology and Nuclear Medicine, Maastricht University Medical Center, Maastricht, Netherlands, 2School for Mental Health and Neuroscience, Maastricht University, Maastricht, Netherlands, 3Department of Electrical Engineering, Eindhoven University of Technology, Eindhoven, Netherlands, 4Department of Neurology, Maastricht University Medical Center, Maastricht, Netherlands, 5Cardiovascular Research Institute Maastricht, Maastricht University, Maastricht, Netherlands, 6Maastricht Brain Imaging Center, Faculty of Psychology and Neuroscience, Maastricht University, Maastricht, Netherlands
Synopsis
Resting-state
fMRI with a short TR enables unaliased sampling of the BOLD-signal, by disentangling
cardiorespiratory pulsation signals. In this study, we aimed to explore the
feasibility of obtaining a high-frequency spectrum and determined the influence
of aging. Structural and high-frequency resting-state fMRI was performed using
7T MRI on 5 young and 5 elderly subjects. The power spectra, calculated for
different brain regions, showed a clear separation of spontaneous BOLD
fluctuations and the respiratory and cardiac pulsations. This pilot study demonstrated the
feasibility of acquiring high-frequency spectra using fMRI. Furthermore, initial
results confirm that the BOLD effect attenuates with aging.
Introduction
Resting-state
(rs-)functional MRI is mostly used to detect spontaneous low-frequency
fluctuations (10-100mHz). Dynamic scanning with a high sampling rate(>2Hz)
enables unaliased sampling of physiological pulsations in the fMRI signal1,2.
Therefore, it will be possible to discern the respiratory and cardiac signals
from the fluctuations originating from the neuronal activity. In
(cerebro)vascular disorders and ageing, these fluctuations may be affected and
reflect abnormal neurovascular coupling due to vessel wall stiffening3,4.
Using fMRI, previous studies found an increased cardiac pulsatility with aging
in normal appearing white matter5,6. The first aim of this
study was to explore the feasibility of obtaining an extended frequency spectrum
in different brain regions, including the brainstem, using high-frequency rs-fMRI
at 7T. The brainstem was also included, since this region is challenging for
fMRI-measurements due to the pulsations of the circulating cerebrospinal fluid
(CSF)7. Secondly, this technique was applied in a small cohort to
study the influence of age on the amplitude of fluctuations.Methods
Subjects: 5 young healthy subjects (21-30
years,2 males) and 5 elderly healthy subjects (60-69 years,3 males).
MRI acquisition: Images were acquired using a 7T MRI system
(Magnetom, Siemens Healthineers, Erlangen, Germany) with a 32-channel
phased-array head coil. To improve B1+ field homogeneity across the brain,
dielectric pads were placed proximal to the temporal lobe. An MP2RAGE (TR/TE=5000/2.47ms,TI1/TI2=900/2750ms,α1/α2=5°/3°,cubic voxel size=0.7mm) was applied to
acquire 3D whole-brain T1-weighted images for anatomical reference. The rs-fMRI time-series were acquired using a whole-brain
multiband echo-planar-imaging sequence (TR/TE=383/17ms, multiband factor 4, pixel
size 2.5x2.5mm2, sagittal slice thickness 2.5mm,48 slices,1000 volumes,
duration 6:32min:sec). The short-TR achieves an effective sample frequency of
2.6Hz, obtaining a maximum frequency in the power spectrum of 1.3Hz, which
enables unaliased sampling of the cardiac pulsations. Simultaneously,
physiological signals were recorded using a pulse oximeter and a respiratory
belt.
Image analysis: For preprocessing of the structural
and the functional images, bias field correction, brain extraction and
distortion correction were performed (FSL and SPM12). Furthermore, B1
correction8 was applied for the structural images, and these were segmented
with FreeSurfer (v6.0.5)9, after which regions of interest (ROI) were
defined in the brainstem and the gray matter of the superior frontal and temporal
cortex. The latter two were chosen for their different proximity of large blood
vessels. On the functional time-series, slice timing correction, deregression
of white matter and CSF signal, and spatial alignment was performed.
After preprocessing, the discrete power (Fourier)
spectrum for the predefined ROIs was calculated. Furthermore, the amplitude of
fluctuations (AF) was calculated in: the (standard) resting-state low frequency
(10-100mHz), respiratory (200-400mHz) and cardiac range (750-1250mHz)10.
These frequency ranges were defined based on the corresponding minimum and
maximum frequency found in the spectra of the recorded physiological monitoring
signals.
Statistics: For comparison of the AF between young
and elderly subjects, a non-parametric Mann-Whitney U-test was performed (SPSS Statistics 25,IBM Inc.,Chicago,Illinois),
with a significance level of 0.05.Results
Figure 1
shows an example of the power spectrum of the BOLD-signal in the temporal
cortex and the simultaneous recordings from the pulse oximeter
and respiratory belt. The cardiac and respiratory peak can be observed in the BOLD-spectrum,
and are supported by the physiological monitoring results.
Comparison
between young and elderly subjects reveal a significant lower AF in the low
frequency range for the superior temporal cortex (pole) (p=0.016) for the
elderly (Fig.2). Furthermore, a trend of lower AF in elderly subjects can be
observed in the brainstem and temporal cortex for this frequency range (p=0.056).
For the
respiratory and cardiac frequency range, no significant differences were found
between young and elderly subjects. However, it can be observed that the
brainstem, superior temporal and frontal cortex appear to have a higher AF in the
elderly compared to the young adults. The superior temporal pole shows a significantly
higher AF in the young adults (p=0.016) compared to the elderly.
Figure 3 depicts an example in the difference
for the AF distribution for the low frequency range. The maps of the young
adult show higher AF values than the maps of the older subject.Discussion
The multiband fMRI sequence with a short
TR applied in this study was able to discern the respiratory and cardiac
pulsations from the low frequency fluctuations originating from neurovascular
coupling in all predefined brain regions.
Significant trends in the low
frequency range (10-100mHz) were found, where young adults had a higher
amplitude for the fluctuations compared to the elderly subjects. To further confirm
this observed effect, more subjects are currently being included.
The different AF values found in the
frontal and temporal cortex can be explained by the proximity and enhancement
by freshly (i.e. unsaturated) inflowing spins from large vessels in these
regions. Regions more closely located to the large vessels will experience more
enhanced cardiac pulsations by blood flow.
Conclusion
This study
shows the feasibility to acquire unaliased high-frequency spectra in different
brain regions, including the brainstem, using a whole-brain fMRI protocol with
short TR (383ms). Initial results show that the native BOLD effect attenuates
with aging. Future applications will be in hypertensive patients, which may
gain more insights in the cerebral regulation of blood circulation and
functional (cerebrovascular) abnormalities observed in hypertension.Acknowledgements
No acknowledgement found.References
1Trapp C, Vakamudi K, Posse S. On the
Detection of High Frequency Correlations in Resting State fMRI. NeuroImage. 2018;
164: 202–213
2Tong Y, Hocke L, Frederick B. Short repetition time multiband echo-planar imaging with simultaneous pulse recording allows dynamic imaging of the cardiac pulsation signal. Magn.
Reson. Med. 2014; 72(5): 1268–1276.
3O’Rourke
M, Hashimoto J. Mechanical factors in arterial aging: a
clinical perspective. J Am Coll Cardiol. 2007;
50: 1–13
4Yang A,
Thai S, Lin C, et al. Frequency and amplitude modulation of resting-state fMRI
signals and their functional relevance in normal aging. Neurobiology of aging.
2018; 70: 59-69
5Makedonov I, Black S, MacIntosh B. BOLD fMRI in
the White Matter as a Marker of Aging and Small Vessel Disease. PLoS ONE. 2013;
8(7): e67652
6Viessmann O, Möller H, Jezzard P. Dual
regression physiological modeling of resting-state EPI power spectra: Effects
of healthy aging. NeuroImage. 2019; 187:68-76
7Sclocco R, Beissner F, Bianciardi M, et al. Challenges
and opportunities for brainstem neuroimaging with ultrahigh field MRI.
Neuroimage. 2018;168: 412-426
8Marques J, Gruetter R. New
Developments and Applications of the MP2RAGE Sequence - Focusing the Contrast
and High Spatial Resolution R1 Mapping. PLoS One. 2013; 8(7): e69294
9Fischl B, Salat D, Busa E, et al. Whole brain segmentation: automated
labeling of neuroanatomical structures in the human brain. Neuron. 2002; 33(3):341-55
10Zou Q, Zhu C, Yang Y, et al. An improved approach to detection of
amplitude of low-frequency fluctuation (ALFF) for resting-state fMRI:
fractional ALFF. J. Neurosci. Methods. 2008; 172(1):137-41