Maxime Donadieu1,2,3, Yann Le Fur1,2, Sylviane Confort-gouny1,2, Arnaud Le Troter1,2, Maxime Guye1,2, and Jean-Philippe Ranjeva1,2
1CRMBM UMR 7339, Aix Marseille Université CNRS, Marseille, France, Metropolitan, 2CEMEREM Pole d'Imagerie, AP-HM CHU Timone, Marseille, France, Metropolitan, 3Siemens Healthcare, Saint-Denis, France, Metropolitan
Synopsis
Using
2D-semilaser 1H-MRSI
sequence centered on thalamus and acquired at 7T in 10 healthy
volunteers, we
demonstrate that the neurochemical profiles (relative
NAA, Cr and Cho levels)
are different between pulvinar, ventral-lateral, dorsal-medial and
anterior nuclei. Moreover, left/right differences in neurochemical
profiles, especially for NAA levels, showed a left NAA lateralization
for the ventral-lateral nucleus and the pulvinar and in contrast
higher right NAA levels in the anterior nucleus. These results
suggest that the various neurochemical profiles of these thalamic
nuclei may be related to their functional specificity.Introduction
The
thalamus is composed by more than thirty functional nuclei, each
connected to specific subcortical and cortical areas (1). This
complex hub structure is involved in a wide range of various brain
functions such as motor, memory, attention and emotion (2), and has
been shown to be affected in most of neurological diseases such as
Alzheimer, Schizophrenia, or Multiple Sclerosis (3).
1H-MRSI
is a non-invasive method to explore brain metabolism but it
suffers for a poor SNR inducing low spatial resolution which prevents
to sample accurately the thalamic nuclei.
Nevertheless, 7T MR scanners offer drastic increase
in SNR, allowing to reach in time compatible with a
clinical exams, 2D-MRSI with voxel size less than 0.1cm3,
a spatial resolution sufficient to characterize the neurochemistry of
the largest thalamic nuclei.
To
determine if thalamic nuclei have specific metabolic fingerprints, we
compare in 10 healthy volunteers, the metabolic
patterns (relative NAA, Cr and Cho levels) of pulvinar,
ventral-lateral, dorsal-medial and anterior nuclei using a
high-resolution 2D-semilaser 1H-MRSI
sequence (4, 5) acquired at 7T. We hypothesize that the metabolic
profiles of the various thalamic nuclei should be different relative
to their functional specificity.
Materials
and Methods
10
healthy volunteers (mean age = 24.5 years± 3, range = 22-32, 4
women, 6 men) gave
their written consent to participate to the study, which was approved
by the local ethics committee.
MRI:
7T MR explorations were performed on a 7T Magnetom step 2 system
(Siemens, Erlangen, Germany), using a volumic transceiver/32-channel
receiver head coil (NOVA). After localizer, third order B0 shimming
and B1 calibration, sagittal T1-weighted
3D-MP2RAGE
was acquired as anatomical reference (TE/TR=3.13/5000ms,
TI1/2=900ms/2750ms,
FOV=240mm, voxel size=0.6*0.6*0.6mm3,
flip angle1/2
= 6°/5°, partitions = 256, Tacq=10.12min).
MRSI:
After manual shimming optimization (water linewidth <
40Hz in VOI ), one
axial 2D semi-laser 1H-MRSI was acquired in the AC-PC plane, centered
on thalami (TE/TR=59/2820ms,
flip angle=90°, voxel size=3.1*3.1*10 mm3,
FOV = 50*50*10 mm3,
number of excitation = 8, bandwidth = 4000 Hz, matrix = 16*16, Tacq =
21.42 min) (Figure
1).
Post-processing:
To locate more accurately the subparts of thalami, the Oxford
thalamic connectivity atlas (6) was projected onto the T1
weighted images of each subject (Figure
2).
MRSI
data were post-processed using the CSIAPO software (7). Eight voxels,
each totally included in one region composed by a single type of
cortical projection according to the atlas were selected and labeled
as left or right i) pulvinar (areas projecting toward
Parietal-temporal-occipital cortex) ii) ventral-lateral
(areas projecting toward pre-central cortex), iii) dorsal-medial
(area projecting toward frontal cortex) and iv) anterior (area projecting toward cingulate cortex). Peak fitting was
performed using the AMARES (8) (CSIAPO). NAA, Cr and Cho
peak areas were normalized by the sum of the three areas (NAA+Cr+Cho)
and corrected for chemical shift artifact determined on water phantom
measurements with shifted slice selection frequency corresponding to
differences in frequency resonances of NAA, Cr and Cho.
Results
ANOVA
looking at differences in neurochemical profiles according to nuclei
and hemispheres showed a significant global effect (F=5.49, p<0.0001)
with significant (metabolism x hemisphere) interaction (p<0.0001),
and (metabolism x nuclei) interaction (p=0.031).
Post-hoc
Wilcoxon test (p<0.05 corrected for multiple comparisons) showed
i) significant higher NAA, Cho and lower Cr levels within the
right anterior nuclei relative to the left, ii) higher NAA, Cr
and lower Cho levels in the left Pulvinar relative to the right and
iii) higher NAA and lower Cho levels within the left ventral-lateral
nucleus relative to the right (Figure
3abc).
The neurochemical profiles were different Figure
4 (a-f)
between the four nuclei, especially in the left hemisphere where NAA
pulvinar
> NAA ventrolateral
> NAA dorsal-medial
& NAA anterior
(Kruskall Wallis p<0.05 corrected by Steel Dwass).
Discussion/Conclusion
1H-semi-laser
2D-MRSI at 7T (4)
enabled to acquire high quality spectra with voxel size below 0.1 ml.
For the first time, we demonstrate that the neurochemical profiles
are different between the pulvinar, the anterior nucleus, the
dorsal-medial nucleus and the ventral-lateral nucleus. Interestingly,
the presence of left/right differences in neurochemical profiles,
especially for NAA levels, showed a left NAA lateralization for the
ventral-lateral nucleus and the pulvinar mostly connected to left
lateralized prefrontal and temporal-parietal-occipital cortices
respectively, and in contrast higher right NAA levels in the anterior
nucleus mostly connected to the cingulate cortex. These results
suggest that neurochemical profiles of nuclei may be related to their functional specificity.
Further
works will aim at confirming these results on a larger cohort of
healthy subjects accounting for age, sex and hand lateralization
before exploring the specific metabolic disorders of each nucleus in
patients suffering for various neurological diseases.
Acknowledgements
MD
is a recipient of a CIFRE grant from the ANRT and Siemens healthcare.
This project is supported by the French IA Equipex 7T-AMI (2011) and
the A*MIDEX 7T-AMISTART (2013) programs.
We thank the Center for Magnetic Resonance Research (CMRR), University of Minnesota for the access to the 2D-semilaser 1H-MRSI sequence.
References
1.
Morel A, Magnin M, Jeanmonod D: Multiarchitectonic and stereotactic
atlas of the human thalamus. J
Comp Neurol
1997; 387:588–630.
2.
Minagar A, Barnett MH, Benedict RH, et al.: The thalamus and multiple
sclerosis Modern views on pathologic, imaging, and clinical aspects.
Neurology
2013; 80:210–219.
3.
Power BD, Looi JC: The thalamus as a putative biomarker in
neurodegenerative disorders. Aust
N Z J Psychiatry
2015:0004867415585857.
4.
Scheenen TWJ, Heerschap A, Klomp DWJ: Towards 1H-MRSI of the human
brain at 7T with slice-selective adiabatic refocusing pulses. Magn
Reson Mater Phys Biol Med
2008; 21:95–101.
5.
Marjanska M, Auerbach EJ, Valabrègue R, Van de Moortele P-F,
Adriany G, Garwood M: Localized 1H NMR spectroscopy in different
regions of human brain in vivo at 7 T. NMR
Biomed
2012; 25:332–339.
6.
Behrens TEJ, Johansen-Berg H, Woolrich MW, et al.: Non-invasive
mapping of connections between human thalamus and cortex using
diffusion imaging. Nat
Neurosci
2003; 6:750–757.
7.
Le Fur Y, Nicoli F, Guye M, Confort-Gouny S, Cozzone PJ, Kober F:
Grid-free interactive and automated data processing for MR chemical
shift imaging data. Magn
Reson Mater Phys Biol Med
2009; 23:23–30.
8.
Vanhamme L, van den Boogaart A, Van Huffel S: Improved method for
accurate and efficient quantification of MRS data with use of prior
knowledge. J
Magn Reson 1997;
129:35–43.