Impact of prospective motion correction, distortion correction methods and large vein bias on the spatial accuracy of cortical laminar fMRI at 9.4 Tesla

Bause, J; Polimeni, Jon̈athan R.; Stelzer, Johannes; In, Myung‐Ho; Ehses, Philipp; Kraemer-Fernandez, Pablo; Aghaeifar, Ali; Lacosse, Eric; Pohmann, R; Scheffler, Klaus

doi:10.1016/j.neuroimage.2019.116434

Cited by 32 publications

(35 citation statements)

References 81 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As large vessels produce field inhomogeneities, spins precessing in the neighboring parenchyma would produce changes in the phase component of the signal; in contrast, small randomly oriented vessels, e.g., in the capillary network within the gray matter, do not produce a net phase and hence, their derived signal remains intact. Although this method helps to improve the local specificity of BOLD, several biases could remain in the corrected data, such as those derived from motion correction or geometric distortions [84];…”

Section: Discussionmentioning

confidence: 99%

Cortical depth-dependent human fMRI of resting-state networks using EPIK

Pais-Roldán

Yun

Palomero-Gallagher

et al. 2020

Preprint

View full text Add to dashboard Cite

Recent laminar-fMRI studies have provided substantial understanding of the evoked cortical responses in multiple sub-systems; in contrast, the laminar component of resting-state networks remains largely unknown. Here, we used EPIK, which offers unprecedented coverage at sub-millimetre resolution, to investigate cortical resting-state dynamics with depth specificity. After verifying laminar activation profiles in response to a task, we investigated whole-brain laminar dynamics. During rest, signal oscillations were stronger in the superficial layers of the cortical ribbon, and functional connectivity analysis indicated that most resting-state cortical connections occurring at rest involve supra-granular layers. Coherence assessed in a lateral motor network suggested a modified distribution of layer-specific responses during task performance compared to resting-state. Whole-brain evaluation of laminar-fMRI encompasses unprecedented computational challenges; nonetheless, it enables a new dimension of the human cerebral cortex to be investigated from a global view, which may be key in the characterization of neurological disorders from a novel perspective.Significance statementResting-state networks sustain conscious awareness and diverse cognitive processing in the absence of tasks. In contrast to cortical areas, the cortical depth-specific signals across different networks have been poorly investigated, mainly due to the small brain coverage enforced in high-resolution imaging methods. Here, we demonstrate the different involvement of the cortical layers in the maintenance of the resting-state dynamics at near whole-brain level, using an optimized fMRI sequence. Given the cytoarchitectonics of the human neocortex, and based on our results, the cortical thickness constitutes an important dimension to characterize the resting-state oscillations in the healthy brain and its functional study may facilitate the identification of novel targets in neurological diseases.

show abstract

Section: Discussionmentioning

confidence: 99%

Cortical depth-dependent human fMRI of resting-state networks using EPIK

Pais-Roldán

Yun

Palomero-Gallagher

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…First, regions with large BV , such as cortical veins, will produce a significant BOLD contrast, as reported experimentally in Scheffler and Ehses 10 . This might considerably bias signals from deeper cortical structures, similar to gradient‐echo EPI 5,51 . However, the imaging point spread function of bSSFP is much narrower compared with EPI, due to the absence of

T_{2}^{*}

decay along the phase‐encoding direction.…”

Section: Discussionmentioning

confidence: 92%

“…10 This might considerably bias signals from deeper cortical structures, similar to gradient-echo EPI. 5,51 However, the imaging point spread function of bSSFP is much narrower compared with EPI, due to the absence of T * 2 decay along the phaseencoding direction. Therefore, with a sufficiently high resolution and without spatial distortions, the large signal from cortical vessels might be effectively suppressed in deeper regions with bSSFP.…”

Section: Discussionmentioning

confidence: 99%

“…3T: TR = [4,5,6,8,10] These measurements were repeated for bSSFP flip angles of α = [5,10,20,30]° at 3 T and α = [5,10,20]° at 9.4 T/14.1 T, resulting in 10 scan sessions: four at 3 T and three at 9.4 T/14.1 T. The MR protocol of each session was completed by a B + 1 mapping sequence and the acquisition of standard SE-based reference R 1 and R 2 data. The B + 1 scaling factor c B1+ = α act /α nom (≡ actual/nominal flip angle) was obtained using the vendor's default implementation at 3 T/9.4 T (ie, TurboFLASH imaging with and without a preconditioning RF pulse 25,26 ) and actual flip angle imaging 27 at 14.1 T. Standard reference R 1 and R 2 values were calculated based on single-slice inversion-recovery SE scans with variable TIs and single-slice single-echo SE scans with variable TEs, respectively.…”

Section: Magnetic Resonance Experimentsmentioning

confidence: 99%

“…Currently, high‐resolution layer‐resolved fMRI is receiving increased interest, 3 implicating the need for dedicated acquisition techniques that are able to provide images with high SNR, no spatial distortions, and reduced sensitivity to large veins 4,5 . Balanced SSFP (bSSFP) is an interesting alternative sequence for fMRI, 6‐10 especially for laminar BOLD experiments, as it enables distortion‐free imaging with high SNR efficiency and is in this respect superior to standard EPI.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Intravascular BOLD signal characterization of balanced SSFP experiments in human blood at high to ultrahigh fields

Pérez-Rodas

Pohmann

Scheffler

et al. 2020

Magnetic Resonance in Med

Self Cite

View full text Add to dashboard Cite

show abstract

Vessel distance mapping: a novel methodology for assessing vascular‐induced cognitive resilience

Garcia-Garcia

Mattern

Vockert

et al. 2022

Alzheimer's & Dementia

View full text Add to dashboard Cite

BackgroundThe association between cerebral blood supply and cognition has gained increasing interest, considering the remarkable anatomical variability of the circle of Willis. Thus, qualitative classifications of the arteries contributing to the hippocampal supply has been performed in previous studies to determine whether the additional presence of vessels might translate into cognitive differences and cerebral structural changes when coexisting with vascular pathology, to the extent of constituting a cognitive reserve. Nevertheless, the promising results in these regards are not without controversy. Hence, Vessel Distance Mapping (VDM) is here introduced, in order to ascertain the correlation of VDM‐metrics with each other and with cognitive status, as well as to elucidate which metrics have the greatest impact on cognition.MethodA battery of cognitive tests was used (including ADAScog and MoCA) in 51 subjects (71 ± 8.5yrs) with (n=20) and without (n=31) cerebral small vessel disease. Their hippocampal‐related vessels were manually segmented by using high‐resolution 7T Time‐of‐Flight‐MRI. Vessel distance maps were generated by computing the distances of each voxel to its nearest vessel, obtaining 3 VDM‐metrics: global‐VDM (vessel density surrogate), VS‐VDM (vessel‐specific supply), and COMD (distance to the center of mass, reflecting vessel distribution). Hippocampal masks were used to assess the whole and subregional hippocampus. Regression models were applied for evaluating the interrelation among the metrics, and their association with cognition.ResultHigh correlations were observed for intra‐metric comparisons, as well as when comparing whole hippocampal vs. sub‐hippocampal metrics. Greater values of VDM‐metrics reflecting higher distances among vessels were associated with poorer cognitive outcomes only in subjects affected by vascular pathology (global‐VDM R²=0.7248, p=0.0054; COMD R²=0.7248, p=0.0013). Metrics revealing vascular distribution (p=0,0053) and density (p=0,0169) were the most influential on cognition, and in this context those for the whole hippocampus predominated over the subregional ones.ConclusionA mixed contribution of vessel distribution and density is proposed to confer cognitive resilience, in a way consistent with anatomical fundamentals and with previous research findings. VDM provides a new, structure‐driven approach on which to raise new questions supported by the statistical robustness of a quantitative method with potential clinical implications.

show abstract

Impact of prospective motion correction, distortion correction methods and large vein bias on the spatial accuracy of cortical laminar fMRI at 9.4 Tesla

Cited by 32 publications

References 81 publications

Cortical depth-dependent human fMRI of resting-state networks using EPIK

Cortical depth-dependent human fMRI of resting-state networks using EPIK

Intravascular BOLD signal characterization of balanced SSFP experiments in human blood at high to ultrahigh fields

Vessel distance mapping: a novel methodology for assessing vascular‐induced cognitive resilience

Contact Info

Product

Resources

About