Analysis of interpolation effects in the reslicing of functional MR images

Pneumatic artificial muscle-based stimulator for passive functional magnetic resonance imaging sensorimotor mapping in patients with brain tumours

BACKGROUND

Two concerns with respect to pre-operative task-based motor functional magnetic resonance imaging (fMRI) in patients with brain tumours are inadequate performance due to patients' impaired motor function and head motion artefacts.

NEW METHOD

In the present study we validate the use of a stimulator based on a pneumatic artificial muscle (PAM) for fMRI mapping of the primary sensorimotor (SM1) cortex in twenty patients with rolandic or perirolandic brain tumours. All patients underwent both active and passive motor block-design fMRI paradigms, performing comparable active and passive PAM-induced flexion-extensions of the icontralesional index finger.

RESULTS

PAM-induced movements resulted in a significant BOLD signal increase in contralateral primary motor (M1) and somatosensory (S1) cortices in 18/20 and 19/20 (p<.05 FWE corrected in 16/18 and 18/19) patients, versus 18/20 and 16/20 (p<.05 FWE corrected) during active movements. The two patients in whom the PAM-based stimulator failed to induce any significant BOLD signal change in the contralateral M1 cortex differed from the two in whom active motion was conversely ineffective. At the group level, no significant difference in contrast magnitude was observed within the contralateral SM1 cortex when comparing active with passive movements. During passive movements, head motion was significantly reduced. Comparison with existing method(s) As compared to the several robotic devices for passive motion that were introduced in the past decades, our PAM-based stimulator appears smaller, handier, and easier to use.

CONCLUSION

The use of PAM-based stimulators should be included in routine pre-operative fMRI protocols along with active paradigms in such patients' population.

Inpainting as a Technique for Estimation of Missing Voxels in Brain Imaging

Issues with model fitting (i.e. suboptimal standard deviation, linewidth/full-width-at-half-maximum, and/or signal-to-noise ratio) in multi-voxel MRI spectroscopy, or chemical shift imaging (CSI) can result in the significant loss of usable voxels. A potential solution to minimize this problem is to estimate the value of unusable voxels by utilizing information from reliable voxels in the same image. We assessed an image restoration method called inpainting as a tool to restore unusable voxels, and compared it with traditional interpolation methods (nearest neighbor, trilinear interpolation and tricubic interpolation). In order to evaluate the performance across varying image contrasts and spatial resolutions, we applied the same techniques to a T1-weighted MRI brain dataset, and N-acetylaspartate (NAA) spectroscopy maps from a CSI dataset. For all image types, inpainting exhibited superior performance (lower normalized root-mean-square errors, NRMSE) compared to all other methods considered (p's < 0.001). Inpainting maintained an average NRMSE of less than 5% even with 50% missing voxels, whereas the other techniques demonstrated up to three times that value, depending on the nature of the image. For CSI maps, inpainting maintained its superiority whether the previously unusable voxels were randomly distributed, or located in regions most commonly affected by voxel loss in real-world data. Inpainting is a promising approach for recovering unusable or missing voxels in voxel-wise analyses, particularly in imaging modalities characterized by low SNR such as CSI. We hypothesize that this technique may also be applicable for datasets from other imaging modalities, such as positron emission tomography, or dynamic susceptibility contrast MRI.

Synchronization and Alignment of Follow-up Examinations: a Practical and Educational Approach Using the DICOM Reference Coordinate System

This work presents an approach for synchronization and alignment of Digital Imaging and Communications in Medicine (DICOM) series from different studies that allows, e.g., easier reading of follow-up examinations. The proposed concept developed within the DICOM's patient-based reference coordinate system allows to synchronize all image data of two different studies/examinations based on a single registration. The most suitable DICOM series for registration could be set as default per protocol. Necessary basics regarding the DICOM standard and the used mathematical transformations are presented in an educative way to allow straightforward implementation in Picture Archiving And Communications Systems (PACS) and other DICOM tools. The proposed method for alignment of DICOM images is potentially also useful for various scientific tasks and machine-learning applications.

Neurophysiological correlates of age-related changes in working memory capacity

Cognitive abilities such as working memory (WM) capacity decrease with age. To determine the neurophysiological correlates of age-related reduction in working memory capacity, we studied 10 young subjects (<35 years of age; mean age=29) and twelve older subjects (>55 years of age; mean age=59) with whole brain blood oxygenation-level dependent (BOLD) fMRI on a 1.5 T GE MR scanner using a SPIRAL FLASH pulse sequence (TE=24 ms, TR=56 ms, FA=60 degrees , voxel dimensions=3.75 mm(3)). Subjects performed a modified version of the "n" back working memory task at different levels of increasing working memory load (1-Back, 2-Back and 3-Back). Older subjects performed as well as the younger subjects at 1-Back (p=0.4), but performed worse than the younger subjects at 2-Back (p<0.01) and 3-Back (p=0.06). Older subjects had significantly longer reaction time (RT) than younger subjects (p<0.04) at all levels of task difficulty. Image analysis using SPM 99 revealed a similar distribution of cortical activity between younger and older subjects at all task levels. However, an analysis of variance revealed a significant group x task interaction in the prefrontal cortex bilaterally; within working memory capacity, as in 1-Back when the older subjects performed as well as the younger subjects, they showed greater prefrontal cortical (BA 9) activity bilaterally. At higher working memory loads, however, when they performed worse then the younger subjects, the older subjects showed relatively reduced activity in these prefrontal regions. These data suggest that, within capacity, compensatory mechanisms such as additional prefrontal cortical activity are called upon to maintain proficiency in task performance. As cognitive demand increases, however, they are pushed past a threshold beyond which physiological compensation cannot be made and, a decline in performance occurs.

Comparison of fMRI motion correction software tools

Motion correction of fMRI data is a widely used step prior to data analysis. In this study, a comparison of the motion correction tools provided by several leading fMRI analysis software packages was performed, including AFNI, AIR, BrainVoyager, FSL, and SPM2. Comparisons were performed using data from typical human studies as well as phantom data. The identical reconstruction, preprocessing, and analysis steps were used on every data set, except that motion correction was performed using various configurations from each software package. Each package was studied using default parameters, as well as parameters optimized for speed and accuracy. Forty subjects performed a Go/No-go task (an event-related design that investigates inhibitory motor response) and an N-back task (a block-design paradigm investigating working memory). The human data were analyzed by extracting a set of general linear model (GLM)-derived activation results and comparing the effect of motion correction on thresholded activation cluster size and maximum t value. In addition, a series of simulated phantom data sets were created with known activation locations, magnitudes, and realistic motion. Results from the phantom data indicate that AFNI and SPM2 yield the most accurate motion estimation parameters, while AFNI's interpolation algorithm introduces the least smoothing. AFNI is also the fastest of the packages tested. However, these advantages did not produce noticeably better activation results in motion-corrected data from typical human fMRI experiments. Although differences in performance between packages were apparent in the human data, no single software package produced dramatically better results than the others. The "accurate" parameters showed virtually no improvement in cluster t values compared to the standard parameters. While the "fast" parameters did not result in a substantial increase in speed, they did not degrade the cluster results very much either. The phantom and human data indicate that motion correction can be a valuable step in the data processing chain, yielding improvements of up to 20% in the magnitude and up to 100% in the cluster size of detected activations, but the choice of software package does not substantially affect this improvement.

Prospective registration of human head magnetic resonance images for reproducible slice positioning using localizer images

PURPOSE

To facilitate assessing brain tumor growth and progression of stroke lesions by reproducible slice positioning in human head magnetic resonance (MR) images, a method for prospective registration is proposed that adjusts the image slice position without moving the patient and with no additional scans.

MATERIALS AND METHODS

The gradient reference frame of follow-up examinations was adjusted to achieve the same image slice positioning relative to the patient as in the previous examination. The three-dimensional geometrical transformation parameters for the gradients were determined using two-dimensional image registration of three orthogonal localizer images. The method was developed and evaluated using a phantom with arbitrarily adjustable position. Feasibility for in vivo applications was demonstrated with brain MR imaging (MRI) of healthy volunteers.

RESULTS

Standard retrospective registration was used for assessing the quality of the method. The accuracy of the realignment was 0.0 mm +/- 1.2 mm and -0.2 degrees +/- 0.9 degrees (mean +/- SD) in phantom experiments. In 10 examinations of volunteers, misalignments up to 49.2 mm and 21 degrees were corrected. The accuracy of the realignment after prospective registration was 0.1 mm +/- 1.5 mm and 0.2 degrees +/- 1.5 degrees.

CONCLUSION

Image-based prospective registration using localizer images of the pre- and postexaminations is a robust method for reproducible slice positioning.

Interpolation artifacts in multimodality image registration based on maximization of mutual information

Mutual information (MI) is an increasingly popular match metric for multimodality image registration. However, its value is affected by interpolation, which may limit registration accuracy. The purpose of this study was to characterize the artifacts from eight interpolators and to investigate efficient strategies to overcome these artifacts. The interpolators were: 1) nearest neighbor; 2) linear; 3) cubic Catmull-Rom; 4) Hamming-windowed sinc; 5) partial volume; 6) NN with jittered sampling (JIT); 7) NN with histogram blurring (BLUR); and 8) NN with JIT and BLUR. The impact of interpolation on MI was evaluated in two dimensions over different translational and rotational misregistration. Interpolation caused spurious fluctuations in MI whenever the voxel grids had coinciding periodicities and were nearly aligned. The artifacts did not lessen by using intensity interpolators with wider support (e.g., cubic Catmull-Rom, Hamming-windowed sinc). PV could lead to either arch artifacts or inverted-arch artifacts, depending on the relative voxel sizes. Several strategies reduced artifacts and improved registration robustness: JIT, BLUR, avoiding an extreme number of intensity bins, and resampling the images in a rotated orientation with different relative voxel sizes (e.g., pi/3). These findings also apply to related methods, including normalized MI, joint entropy, and Hill's third moment.

Dopaminergic modulation of cortical function in patients with Parkinson's disease

Patients with idiopathic Parkinson's disease suffer not only from classic motor symptoms, but from deficits in cognitive function, primarily those subserved by the prefrontal cortex as well. The aim of the current study was to investigate the modulatory effects of dopaminergic therapy on neural systems subserving working memory and motor function in patients with Parkinson's disease. Ten patients with stage I and II Parkinson's disease were studied with functional magnetic resonance imaging, during a relatively hypodopaminergic state (ie, 12 hours after a last dose of dopamimetic treatment), and again during a dopamine-replete state. Functional magnetic resonance imaging was performed under three conditions: a working memory task, a cued sensorimotor task and rest. Consistent with prior data, the cortical motor regions activated during the motor task showed greater activation during the dopamine-replete state; however, the cortical regions subserving working memory displayed greater activation during the hypodopaminergic state. Interestingly, the increase in cortical activation during the working memory task in the hypodopaminergic state positively correlated with errors in task performance, and the increased activation in the cortical motor regions during the dopamine-replete state was positively correlated with improvement in motor function. These results support evidence from basic research that dopamine modulates cortical networks subserving working memory and motor function via two distinct mechanisms: nigrostriatal projections facilitate motor function indirectly via thalamic projections to motor cortices, whereas the mesocortical dopaminergic system facilitates working memory function via direct inputs to prefrontal cortex. The results are also consistent with evidence that the hypodopaminergic state is associated with decreased efficiency of prefrontal cortical information processing and that dopaminergic therapy improves the physiological efficiency of this region.

Quantifying head motion associated with motor tasks used in fMRI

In functional magnetic resonance imaging (fMRI) studies, long experiment times and small intensity changes associated with brain activation frequently lead to image artifacts due to head motion. Methods to minimize and correct for head motion by restraint, fast imaging, and retrospective image registration are typically combined but do not completely solve the problem, particularly for specific patient populations. As an initial step toward optimizing future designs of head restraints and improving motion correction techniques, the head motion characteristics of groups of stroke subjects, age-matched controls, and young adults were investigated with the aid of an MR simulator and a highly accurate position tracking system. Position measurements were recorded during motor tasks involving either the hand or the foot. Head motion was strongly dependent on the subject group and less upon the task conditions based on ANOVA calculations (P < 0.05). The stroke subjects exhibited approximately twice the head motion compared to that of age-matched controls, and the latter's head motion was about twice that of young adults. Moreover, the range of head motion in stroke subjects over all tasks was approximately 2 +/- 1 mm, with the motion occurring predominantly as translation in the superior-inferior direction and pitch rotation (nodding). These results lead to several recommendations on the design of fMRI motor experiments and suggest that improved motion correction strategies are required to examine such patient populations comprehensively.

Quantitative evaluation of convolution-based methods for medical image interpolation

Interpolation is required in a variety of medical image processing applications. Although many interpolation techniques are known from the literature, evaluations of these techniques for the specific task of applying geometrical transformations to medical images are still lacking. In this paper we present such an evaluation. We consider convolution-based interpolation methods and rigid transformations (rotations and translations). A large number of sinc-approximating kernels are evaluated, including piecewise polynomial kernels and a large number of windowed sinc kernels, with spatial supports ranging from two to ten grid intervals. In the evaluation we use images from a wide variety of medical image modalities. The results show that spline interpolation is to be preferred over all other methods, both for its accuracy and its relatively low computational cost.

Modeling geometric deformations in EPI time series

Even after realignment there is residual movement-related variance present in fMRI time-series, causing loss of sensitivity and, potentially, also specificity. One cause is the differential deformation of the sampling matrix, by field inhomogeneities, at different object positions, i.e., a movement-by-inhomogeneity interaction. This has been addressed previously by using empirical field measurements. In the present paper we suggest a forward model of how data is affected by an inhomogeneous field at different object positions. From this model we derive a method to solve the inverse problem of estimating the field inhomogeneities and their derivatives with respect to object position, directly from the EPI data and estimated realignment parameters. The field is modeled as a linear combination of cosine basis fields, which facilitates a fast way of implementing the necessary matrix operations. Simulations suggest that the solution is tractable and that the fields are estimable given the deformed images and knowledge of the relative positions at which they have been acquired. An experiment on a subject performing voluntary movements in the scanner yielded plausible estimates of the deformation fields and their application to "unwarp" the time series significantly reduced movement-related variance.

MOMS: maximal-order interpolation of minimal support

We consider the problem of interpolating a signal using a linear combination of shifted versions of a compactly-supported basis function phi(x). We first give the expression for the cases of phi's that have minimal support for a given accuracy (also known as "approximation order"). This class of functions, which we call maximal-order-minimal-support functions (MOMS) is made of linear combinations of the B-spline of the same order and of its derivatives. We provide an explicit form of the MOMS that maximizes the approximation accuracy when the step-size is small enough. We compute the sampling gain obtained by using these optimal basis functions over the splines of the same order. We show that it is already substantial for small orders and that it further increases with the approximation order L. When L is large, this sampling gain becomes linear; more specifically, its exact asymptotic expression is 2/(pie)L. Since the optimal functions are continuous, but not differentiable, for even orders, and even only piecewise continuous for odd orders, our result implies that regularity has little to do with approximating performance. These theoretical findings are corroborated by experimental evidence that involves compounded rotations of images.

Effects of dextroamphetamine on cognitive performance and cortical activation

Monoaminergic neurotransmitters are known to have modulatory effects on cognition and on neurophysiological function in the cortex. The current study was performed with BOLD fMRI to examine physiological correlates of the effects of dextroamphetamine on working-memory performance in healthy controls. In a group analysis dextroamphetamine increased BOLD signal in the right prefrontal cortex during a task with increasing working-memory load that approached working-memory capacity. However, the effect of dextroamphetamine on performance and on signal change varied across individuals. Dextroamphetamine improved performance only in those subjects who had relatively low working-memory capacity at baseline, whereas in the subjects who had high working-memory capacity at baseline, it worsened performance. In subjects whose performance deteriorated, signal change was greater than that in subjects who had an improvement in performance, and these variations were correlated (Spearman rho = 0.89, P<0.02). These data shed light on the manner in which monoaminergic tone, working memory, and prefrontal function interact and, moreover, demonstrate that even in normal subjects the behavioral and neurophysiologic effects of dextroamphetamine are not homogeneous. These heterogeneic effects of dextroamphetamine may be explained by genetic variations that interact with the effects of dextroamphetamine.

Interpolation revisited

Based on the theory of approximation, this paper presents a unified analysis of interpolation and resampling techniques. An important issue is the choice of adequate basis functions. We show that, contrary to the common belief, those that perform best are not interpolating. By opposition to traditional interpolation, we call their use generalized interpolation; they involve a prefiltering step when correctly applied. We explain why the approximation order inherent in any basis function is important to limit interpolation artifacts. The decomposition theorem states that any basis function endowed with approximation order can be expressed as the convolution of a B-spline of the same order with another function that has none. This motivates the use of splines and spline-based functions as a tunable way to keep artifacts in check without any significant cost penalty. We discuss implementation and performance issues, and we provide experimental evidence to support our claims.

Automatic measurement of changes in brain volume on consecutive 3D MR images by segmentation propagation

This article presents a technique to automatically measure changes in the volume of a structure of interest in successive 3D magnetic resonance (MR) images and its application in the study of the brain and lateral cerebral ventricles. The only manual step is a segmentation of the structure of interest in the first image. The analysis comprises, first, precise rigid co-registration of the time series of images; second, computation of residual deformations between pairs of images; third, automatic quantification of the volume change, obtained by propagation of the segmentation of the structure of interest through the series of MR images. This approach has been applied to monitor changes in the volume of the brain and lateral cerebral ventricles in a healthy subject and a patient with primary progressive aphasia (PPA). Results are consistent with those obtained by application of the boundary shift integral (BSI) and by stereology in the same subjects.

Characterization and correction of interpolation effects in the realignment of fMRI time series

Subject motion in functional magnetic resonance imaging (fMRI) studies can be accurately estimated using realignment algorithms. However, residual changes in signal intensity arising from motion have been identified in the data even after realignment of the image time series. The nature of these artifacts is characterized using simulated displacements of an fMRI image and is attributed to interpolation errors introduced by the resampling inherent within realignment. A correction scheme that uses a periodic function of the estimated displacements to remove interpolation errors from the image time series on a voxel-by-voxel basis is proposed. The artifacts are investigated using a brain phantom to avoid physiological confounds. Small- and large-scale systematic displacements show that the artifacts have the same form as revealed by the simulated displacements. A randomly displaced phantom and a human subject are used to demonstrate that interpolation errors are minimized using the correction.

Survey: interpolation methods in medical image processing

Image interpolation techniques often are required in medical imaging for image generation (e.g., discrete back projection for inverse Radon transform) and processing such as compression or resampling. Since the ideal interpolation function spatially is unlimited, several interpolation kernels of finite size have been introduced. This paper compares 1) truncated and windowed sinc; 2) nearest neighbor; 3) linear; 4) quadratic; 5) cubic B-spline; 6) cubic; g) Lagrange; and 7) Gaussian interpolation and approximation techniques with kernel sizes from 1 x 1 up to 8 x 8. The comparison is done by: 1) spatial and Fourier analyses; 2) computational complexity as well as runtime evaluations; and 3) qualitative and quantitative interpolation error determinations for particular interpolation tasks which were taken from common situations in medical image processing. For local and Fourier analyses, a standardized notation is introduced and fundamental properties of interpolators are derived. Successful methods should be direct current (DC)-constant and interpolators rather than DC-inconstant or approximators. Each method's parameters are tuned with respect to those properties. This results in three novel kernels, which are introduced in this paper and proven to be within the best choices for medical image interpolation: the 6 x 6 Blackman-Harris windowed sinc interpolator, and the C2-continuous cubic kernels with N = 6 and N = 8 supporting points. For quantitative error evaluations, a set of 50 direct digital X rays was used. They have been selected arbitrarily from clinical routine. In general, large kernel sizes were found to be superior to small interpolation masks. Except for truncated sinc interpolators, all kernels with N = 6 or larger sizes perform significantly better than N = 2 or N = 3 point methods (p << 0.005). However, the differences within the group of large-sized kernels were not significant. Summarizing the results, the cubic 6 x 6 interpolator with continuous second derivatives, as defined in (24), can be recommended for most common interpolation tasks. It appears to be the fastest six-point kernel to implement computationally. It provides eminent local and Fourier properties, is easy to implement, and has only small errors. The same characteristics apply to B-spline interpolation, but the 6 x 6 cubic avoids the intrinsic border effects produced by the B-spline technique. However, the goal of this study was not to determine an overall best method, but to present a comprehensive catalogue of methods in a uniform terminology, to define general properties and requirements of local techniques, and to enable the reader to select that method which is optimal for his specific application in medical imaging.

Petrosal bone: coronal reconstructions from axial spiral CT data obtained with 0.5-mm collimation can replace direct coronal sequential CT scans

PURPOSE

To investigate whether coronal multiplanar reconstruction (MPR) images of the petrosal bone from axial spiral computed tomographic (CT) data obtained with 0.5-mm collimation can replace direct coronal sequential CT scans obtained with 0.5- or 1.0-mm collimation.

MATERIALS AND METHODS

The differences in diagnostic quality between thin-section coronal sequential CT scans of 24 petrosal bones in 12 patients and matched MPR images were assessed by five observers. The matched MPR images were calculated with both trilinear and tricubic interpolation. Image resolution was determined by measuring the three-dimensional point spread function.

RESULTS

All observers preferred tricubically interpolated MPR images over trilinearly interpolated images. Subjective differences in image quality between direct coronal scans and matched tricubically interpolated MPR images were small. Only the direct coronal scans with the highest image quality (0.5-mm collimation, 465 mAs) were judged to be slightly better than the matched MPR images. With regard to direct coronal scans obtained at 245 mAs and/or 1.0-mm collimation, either there was no preference or the MPR images were preferred.

CONCLUSION

Coronal MPR images from axial spiral CT obtained with 0.5-mm collimation can replace direct coronal sequential CT scans.

Improved quality of re-sliced MR images using re-normalized sinc interpolation

This article shows that re-normalizing the interpolation kernel for a constant integral can make a significant improvement in performance of sinc interpolation methods. A comparison was performed between standard and re-normalized sinc kernels of various sizes using data from four commonly used magnetic resonance (MR) imaging sequences. Standard rotations were performed and compared with a "gold standard" data set generated by use of a large (13 x 13 x 13) sinc kernel. Measurements of systematic pixel intensity offset error and variance of generated residuals were used to estimate resultant interpolation error. Theoretical estimates of the consequent savings in computation time were compared with the measured time required for each algorithm and with the automated image registration (AIR) program. The use of a small (5 x 5 x 5) re-normalized kernel produced relative errors comparable to those in the gold standard data set, allowing saving in computation time of up to 30 times in comparison with standard sinc interpolation. This approach brings the implementation of MR volume re-slicing much closer to the demands of a clinical environment. J. Magn. Reson. Imaging 1999;10:582-588.

Technical solution for an interactive functional MR imaging examination: application to a physiologic interview and the study of cerebral physiology

Studies with functional magnetic resonance (MR) imaging produce large unprocessed raw data sets in minutes. The analysis usually requires transferring of the data to an off-line workstation, and this process frequently occurs after the subject has left the MR unit. The authors describe a hardware configuration and processing software that captures whole-brain raw data files as they are being produced from the MR unit. It then performs the reconstruction, registration, and statistical analysis, and displays the results in seconds after completion of the MR image acquisition.