Robert Turner - Profile on Academia.edu (original) (raw)

Papers by Robert Turner

Diffusion tensor MRI can anatomically segment human amygdaloid subregions in vivo

A comparative study of different image analysis packages for the processing of fMRI data

Effects of Tuning Condition, Head Size and Position on the SAR of a MRI Dual-row Transmit Array at 400 MHz

We have developed a workflow for numerical investigation of magnetic resonance imaging RF coil ar... more We have developed a workflow for numerical investigation of magnetic resonance imaging RF coil arrays. Use of fully parameterized 3-D and RF circuit models, and customized scripts for several time-consuming post-processing steps resulted in fast and continuous (24/7) generation of useful and important results. The high simulation speed allowed sensitivity analysis to be conducted for the most important dependencies - array diameter, human model position and tuning condition. For a dual-row array of 280 mm in diameter, as compared with a dual-row array of 250 or 230 mm in diameter, the safety excitation efficiency is higher and a lower margin of error is needed to maintain safety in operation. The impact of the tuning conditions and head position on the spatial-average 10-gram SAR significantly increases when the distance between the human model and the array decreases.

SAR in interleaved excitation of an MRI RF array

Proceedings of the 2012 IEEE International Symposium on Antennas and Propagation, 2012

ABSTRACT We have numerically investigated the safety implications of interleaved RF excitation, a... more ABSTRACT We have numerically investigated the safety implications of interleaved RF excitation, as proposed by Avdievitch (2011), when used with multi-row near field arrays at 300 MHz. In most configurations, for both circular polarized excitation and static RF shimming, interleaved excitation resulted in multiplication of power deposited in the entire human tissue by a factor approximately equal to the number of interleaves, and a significant increase of specific absorption ratio averaged over 10 gram of tissue (SAR10g). However, for a specific configuration, in which interleaved excitation at each time step resulted in significant power reflected by the entire array, and RF excitation profiles that were well separated in space, the SAR10g remained practically the same as with simultaneous excitation.

Is it possible to achieve sufficient inversion efficiencies in CASL experiments at high B0 and low B1 field strengths?

Introduction Perfusion imaging using magnetically labeled water as an endogenous tracer is capabl... more Introduction Perfusion imaging using magnetically labeled water as an endogenous tracer is capable of measuring cerebral blood flow (1). Several studies at different main magnetic field strengths B0 were performed over the last decade (2-4). Perfusion measurements at 7 T are expected to be more sensitive because of the increased signal-to-noise ratio (SNR) at higher field strengths. For quantification of perfusion in continuous arterial spin labeling (CASL) experiments, however, an exact estimation of the inversion efficiency α is required. As CASL experiments are based on the adiabatic inversion of the flowing spins several simulation studies were performed focusing on different experimental and physiological parameters of the adiabatic fast passage, such as the B1 field amplitude, the strength of the applied gradient and the blood flow velocity (5-7). The presented work, however, centers on issues arising with higher field strengths. At such field strengths (e.g., 7 T) the specific absorption rate (SAR) limits the application of radio-frequency pulses so the B1 field should be as low as possible. Therefore, we investigated whether it is possible to achieve sufficient inversion efficiencies in CASL experiments at high B0 and low B1 field strengths. Method The determination of the inversion efficiency was based on a solution of the Bloch equations using the hard-pulse approximation (8,9). Spin relaxation was included in the simulation. The magnetization was calculated recursively while the actual values of the frequency offset (determined by the applied gradient) and the B1 field were inserted into the solution at every integration step. The simulation of the time dependent magnetization was started far below resonance and ended above resonance at a distance of 3 cm. The step size for the simulation was decreased until the results were stable. The simulation was performed in dependence on the gradient strength and the amplitude of the B1 field assuming a main magnetic field strength B0 of 7 T. For comparison, the inversion efficiencies were also calculated at B0 field strengths of 3 T and 1.5 T. As an estimate, T1 and T2 relaxation times of 2000 ms and 250 ms for arterial blood at 7 T, of 1700 ms and 275 ms at 3 T and of 1200 ms and 300 ms at 1.5 T were assumed. It should be noted that slightly different values for T1 and T2 do not affect the results of the simulations substantially. As the inversion efficiency is relatively insensitive to the blood flow velocity within the physiological range (5) and as the influence of the cardiac cycle can be neglected as long as the labeling pulse comprises at least one cardiac cycle (7), a constant blood flow velocity of 20 cm/s was assumed. For better comparison the inversion efficiencies were also plotted against the adiabaticity factor β which is described elsewhere (7). As for high field strengths the SAR limits the application of RF pulses a low B1 field is desirable. Therefore, gradient strengths producing maximum inversion efficiencies at low B1 field amplitudes were calculated. Results Tab. 1: Inversion efficiencies at a mean velocity of 20 cm/s. Inversion efficiencies α in dependence on the amplitude of the B1 field and on the gradient strength are summarized in Tab. 1. The dependence of α on the adiabaticity factor β is shown in Fig. 1. Whereas the absolute values of the inversion efficiencies vary slightly with field strength B0 a very similar relationship between α and β was found for the different field strengths. An optimum adiabaticity of 3...4 was obtained (Fig. 1). Fig. 2 depicts the dependence of α on the amplitude of applied gradient strength at 7 T. As dictated by the SAR limit low B1 field amplitudes of 1.2 and 0.6 μT were assumed for the simulation. Maximum inversion efficiencies of about 88% (B1 = 1.2 μT) and 85% (B1 = 0.6 μT) were found at gradient strengths of 0.8 and 0.25 mT/m, respectively. Discussion and Conclusion Inversion efficiencies of CASL experiments were calculated numerically under realistic experimental and physiological assumptions. It was shown that the main magnetic field strength and the resulting T1 and T2 relaxation times have an influence of the adiabatic inversion of flowing spins. Hence, for an exact estimation of the inversion efficiency in CASL experiments the field strength B0 has to be considered. However, for all B0 field strengths a B1 field amplitude and a gradient strength resulting in an adiabaticity factor β of about 3...4 can be considered as an optimum. Low B1 field amplitudes as desired at high main magnetic field strengths such as 7 T are still able to produce a sufficient efficiency of the adiabatic spin inversion. This is achievable if a corresponding low gradient strength is applied. However, it has to be mentioned that the simulations were performed assuming a perfectly homogeneous main magnetic field B0. A realistic, more inhomogeneous main magnetic field could markedly influence the shape and the…

NeuroImage, 2015

The position of cortical areas can be approximately predicted from cortical surface folding patte... more The position of cortical areas can be approximately predicted from cortical surface folding patterns. However, there is extensive inter-subject variability in cortical folding patterns, prohibiting a one-to-one mapping of cortical folds in certain areas. In addition, the relationship between cortical area boundaries and the shape of the cortex is variable, and weaker for higher-order cortical areas. Current surface registration techniques align cortical folding patterns using sulcal landmarks or cortical curvature, for instance. The alignment of cortical areas by these techniques is thus inherently limited by the sole use of geometric similarity metrics. Magnetic resonance imaging T1 maps show intra-cortical contrast that reflects myelin content, and thus can be used to improve the alignment of cortical areas. In this article, we present a new symmetric diffeomorphic multi-contrast multi-scale surface registration (MMSR) technique that works with partially inflated surfaces in the level-set framework. MMSR generates a more precise alignment of cortical surface curvature in comparison to two widely recognized surface registration algorithms. The resulting overlap in gyrus labels is comparable to FreeSurfer. Most importantly, MMSR improves the alignment of cortical areas further by including T1 maps. As a first application, we present a group average T1 map at a uniquely high-resolution and multiple cortical depths, which reflects the myeloarchitecture of the cortex. MMSR can also be applied to other MR contrasts, such as functional and connectivity data.

Engineering of 7T transmit multi-row arrays

In evaluation of RF transmit array coils, realistic estimation of losses were included in simulat... more In evaluation of RF transmit array coils, realistic estimation of losses were included in simulations with the aim of obtaining a better match between experimental results and numerical predictions. This required customized design of lossy circuit components, to overcome the limitation of the available built-in capabilities of current simulation tools. Some of the more time-consuming post-processing stages were relocated into Matlab, speeding post-processing by up to a factor of 100. The resulting numerical data can support the fabrication of dual row array with as many as 8 elements in each row, and elements overlapped in the Z direction.

7T-MRT bei Morbus Wilson und Messung von cerebralen Kupferablagerungen mittels Quantitative Susceptibility Mapping (QSM)

RöFo - Fortschritte auf dem Gebiet der Röntgenstrahlen und der bildgebenden Verfahren, 2014

In-vivo-Bestimmung der Kollagenstruktur im Kniegelenkknorpel mittels 7-Tesla MRT und erste Ergebnisse zu deren Altersabhängigkeit

RöFo - Fortschritte auf dem Gebiet der Röntgenstrahlen und der bildgebenden Verfahren, 2012

NeuroImage, 2011

In transcranial magnetic stimulation (TMS), knowledge of the distribution of the induced electric... more In transcranial magnetic stimulation (TMS), knowledge of the distribution of the induced electric field is fundamental for a better understanding of the position and extent of the stimulated brain region. However, the different tissue types and the varying fibre orientation in the brain tissue result in an inhomogeneous and anisotropic conductivity distribution and distort the electric field in a non-trivial way. Here, the field induced by a figure-8 coil is characterized in detail using finite element calculations and a geometrically accurate model of an individual head combined with high-resolution diffusion-weighted imaging for conductivity mapping. It is demonstrated that the field strength is significantly enhanced when the currents run approximately perpendicular to the local gyral orientation. Importantly, the spatial distribution of this effect differs distinctly between gray matter (GM) and white matter (WM): While the field in GM is selectively enhanced at the gyral crowns and lips, high field strengths can still occur rather deep in WM. Taking the anisotropy of brain tissue into account tends to further boost this effect in WM, but not in GM. Spatial variations in the WM anisotropy affect the local field strength in a systematic way and result in localized increases of up to 40% (on average~7% for coil orientations perpendicular to the underlying gyri). We suggest that these effects might create hot spots in WM that might contribute to the excitation of WM structures by TMS. However, our results also demonstrate the necessity of using realistic nerve models in the future to allow for more definitive conclusions.

Magnetic Resonance Materials in Physics, Biology and Medicine, 2012

The Larmor frequency of water protons has reliably linear temperature dependence. Since this freq... more The Larmor frequency of water protons has reliably linear temperature dependence. Since this frequency shift is easily measurable using relatively simple MRI techniques, a remarkable opportunity arises for uniquely non-invasive and accurate temperature evaluation, deep within any water-containing object. Major applications are appearing in the field of image-guided surgery. The cutting-edge papers collected in this Special Issue demonstrate both the versatility and the power of MRI thermometry.

Synopsis The thickness of the cerebral cortex can provide valuable information about normal and a... more Synopsis The thickness of the cerebral cortex can provide valuable information about normal and abnormal neuroanatomy. Accurate cortical thickness measurements from brain MRI require precise segmentation of the grey matter. Here we specifically address the problem of extracting the deep cortical folds or sulci, which can be difficult to resolve in MRI. We propose a method that iteratively solves Laplace's equation for adjacent sublayers of the cortex. This approach preserves the laminar structure of the cortex and provides clear definition of deep sulci. The implementation is computationally efficient. We present inter-subject and intra-subject results that are consistent with the literature.

Areas within and beyond the visual cortex differentially activated by coherent visual motion and dynamic noise

ABSTRACT

In favor of general probability distributions: lateral prefrontal and insular cortices respond to stimulus inherent, but irrelevant differences

Brain Structure and Function, 2014

A key aspect of optimal behavior is the ability to predict what will come next. To achieve this, ... more A key aspect of optimal behavior is the ability to predict what will come next. To achieve this, we must have a fairly good idea of the probability of occurrence of possible outcomes. This is based both on prior knowledge about a particular or similar situation and on immediately relevant new information. One question that arises is: when considering converging prior probability and external evidence, is the most probable outcome selected or does the brain represent degrees of uncertainty, even highly improbable ones? Using functional magnetic resonance imaging, the current study explored these possibilities by contrasting words that differ in their probability of occurrence, namely, unbalanced ambiguous words and unambiguous words. Unbalanced ambiguous words have a strong frequency-based bias towards one meaning, while unambiguous words have only one meaning. The current results reveal larger activation in lateral prefrontal and insular cortices in response to dominant ambiguous compared to unambiguous words even when prior and contextual information biases one interpretation only. These results suggest a probability distribution, whereby all outcomes and their associated probabilities of occurrence-even if very low-are represented and maintained.

How to see the mind

Physics World, 1994

The brain is the most complicated organ in the body. Indeed, the great Harvard neurologist Gerald... more The brain is the most complicated organ in the body. Indeed, the great Harvard neurologist Gerald Fischbach described the human brain as "the most complex structure in the known Universe". How are we to make sense of it? And how are we to connect operations which we perceive as activities of our minds with empirical processes that are observable in the laboratory?

Human Brain Mapping, 2014

The subthalamic nucleus (STN) is an important node of the cortico-basal ganglia network and the m... more The subthalamic nucleus (STN) is an important node of the cortico-basal ganglia network and the main target of deep brain stimulation (DBS) in Parkinson's disease. Histological studies have revealed an inhomogeneous iron distribution within the STN, which has been related to putative subdivisions within this nucleus. Here, we investigate the iron distribution in more detail using quantitative susceptibility mapping (QSM), a novel magnetic resonance imaging (MRI) contrast mechanism. QSM allows for detailed assessment of iron content in both in vivo and postmortem tissue. Twelve human participants and 7 postmortem brain samples containing the STN were scanned using ultrahigh field 7 Tesla (T) MRI. Iron concentrations were found to be higher in the medial-inferior tip of the STN. Using quantitative methods we show that the increase of iron concentration towards the medialinferior tip is of a gradual rather than a discrete nature.

NeuroImage, 2012

There is ongoing debate whether using a higher spatial resolution (sampling kspace) or a higher a... more There is ongoing debate whether using a higher spatial resolution (sampling kspace) or a higher angular resolution (sampling q-space angles) is the better way to improve diffusion MRI (dMRI) based tractography results in living humans. In both cases, the limiting factor is the signal-to-noise ratio (SNR), due to the restricted acquisition time. One possible way to increase the spatial resolution without sacrificing either SNR or angular resolution is to move to a higher magnetic field strength. Nevertheless, dMRI has not been the preferred application for ultra-high field strength (7 Tesla). This is because single-shot echo-planar imaging (EPI) has been the method of choice for human in vivo dMRI. EPI faces several challenges related to the use of a high resolution at high field strength, for example, distortions and image blurring. These problems can easily compromise the expected SNR gain with field strength. In the current study, we introduce an adapted EPI sequence in conjunction with a combination of ZOOmed imaging and Partially Parallel Acquisition (ZOOPPA). We demonstrate that the method can produce high quality diffusionweighted images with high spatial and angular resolution at 7 Tesla. We provide examples of in vivo human dMRI with isotropic resolutions of 1 mm and 800 µm. These data sets are particularly suitable for resolving complex and subtle fiber architectures, including fiber crossings in the white matter, anisotropy in the cortex and fibers entering the cortex.

Transparent thin shield for radio frequency transmit coils

Magnetic Resonance Materials in Physics, Biology and Medicine, 2014

To identify a shielding material compatible with optical head-motion tracking for prospective mot... more To identify a shielding material compatible with optical head-motion tracking for prospective motion correction and which minimizes radio frequency (RF) radiation losses at 7 T without sacrificing line-of-sight to an imaging target. We evaluated a polyamide mesh coated with silver. The thickness of the coating was approximated from the composition ratio provided by the material vendor and validated by an estimate derived from electrical conductivity and light transmission measurements. The performance of the shield is compared to a split-copper shield in the context of a four-channel transmit-only loop array. The mesh contains less than a skin-depth of silver coating (300 MHz) and attenuates light by 15 %. Elements of the array vary less in the presence of the mesh shield as compared to the split-copper shield indicating that the array behaves more symmetrically with the mesh shield. No degradation of transmit efficiency was observed for the mesh as compared to the split-copper shield. We present a shield compatible with future integration of camera-based motion-tracking systems. Based on transmit performance and eddy-current evaluations the mesh shield is appropriate for use at 7 T.