Effects of spatial smoothing on fMRI group inferences

The analysis of functional magnetic resonance imaging (fMRI) data involves multiple stages of data pre-processing before the activation can be statistically detected. Spatial smoothing is a very common pre-processing step in the analysis of functional brain imaging data. This study presents a broad perspective on the influence of spatial smoothing on fMRI group activation results. The data obtained from 20 volunteers during a visual oddball task were used for this study. Spatial smoothing using an isotropic gaussian filter kernel with full width at half maximum (FWHM) sizes 2 to 30 mm with a step of 2 mm was applied in two levels — smoothing of fMRI data and/or smoothing of single-subject contrast files prior to general linear model random-effects group analysis generating statistical parametric maps. Five regions of interest were defined, and several parameters (coordinates of nearest local maxima, t value, corrected threshold, effect size, residual values, etc.) were evaluated to examine the effects of spatial smoothing. The optimal filter size for group analysis is discussed according to various criteria. For our experiment, the optimal FWHM is about 8 mm. We can conclude that for robust experiments and an adequate number of subjects in the study, the optimal FWHM for single-subject inference is similar to that for group inference (about 8 mm, according to spatial resolution). For less robust experiments and fewer subjects in the study, a higher FWHM would be optimal for group inference than for single-subject inferences.     

A conditional Granger causality model approach for group analysis in functional magnetic resonance imaging

Granger causality model (GCM) derived from multivariate vector autoregressive models of data has been employed to identify effective connectivity in the human brain with functional magnetic resonance imaging (fMRI) and to reveal complex temporal and spatial dynamics underlying a variety of cognitive processes. In the most recent fMRI effective connectivity measures, pair-wise GCM has commonly been applied based on single-voxel values or average values from special brain areas at the group level. Although a few novel conditional GCM methods have been proposed to quantify the connections between brain areas, our study is the first to propose a viable standardized approach for group analysis of fMRI data with GCM. To compare the effectiveness of our approach with traditional pair-wise GCM models, we applied a well-established conditional GCM to preselected time series of brain regions resulting from general linear model (GLM) and group spatial kernel independent component analysis of an fMRI data set in the temporal domain. Data sets consisting of one task-related and one resting-state fMRI were used to investigate connections among brain areas with the conditional GCM method. With the GLM-detected brain activation regions in the emotion-related cortex during the block design paradigm, the conditional GCM method was proposed to study the causality of the habituation between the left amygdala and pregenual cingulate cortex during emotion processing. For the resting-state data set, it is possible to calculate not only the effective connectivity between networks but also the heterogeneity within a single network. Our results have further shown a particular interacting pattern of default mode network that can be characterized as both afferent and efferent influences on the medial prefrontal cortex and posterior cingulate cortex. These results suggest that the conditional GCM approach based on a linear multivariate vector autoregressive model can achieve greater accuracy in detecting network connectivity than the widely used pair-wise GCM, and this group analysis methodology can be quite useful to extend the information obtainable in fMRI.     

Evaluation of preprocessing steps to compensate for magnetic field distortions due to body movements in BOLD fMRI

Blood oxygenation level-dependent (BOLD) functional magnetic resonance imaging (fMRI) is currently the dominant technique for non-invasive investigation of brain functions. One of the challenges with BOLD fMRI, particularly at high fields, is compensation for the effects of spatiotemporally varying magnetic field inhomogeneities (ΔB₀) caused by normal subject respiration and, in some studies, movement of the subject during the scan to perform tasks related to the functional paradigm. The presence of ΔB₀ during data acquisition distorts reconstructed images and introduces extraneous fluctuations in the fMRI time series that decrease the BOLD contrast-to-noise ratio. Optimization of the fMRI data-processing pipeline to compensate for geometric distortions is of paramount importance to ensure high quality of fMRI data. To investigate ΔB₀ caused by subject movement, echo-planar imaging scans were collected with and without concurrent motion of a phantom arm. The phantom arm was constructed and moved by the experimenter to emulate forearm motions while subjects remained still and observed a visual stimulation paradigm. These data were then subjected to eight different combinations of preprocessing steps. The best preprocessing pipeline included navigator correction, a complex phase regressor and spatial smoothing. The synergy between navigator correction and phase regression reduced geometric distortions better than either step in isolation and preconditioned the data to make them more amenable to the benefits of spatial smoothing. The combination of these steps provided a 10% increase in t-statistics compared to only navigator correction and spatial smoothing and reduced the noise and false activations in regions where no legitimate effects would occur.     

Cortex-based independent component analysis of fMRI time series

The cerebral cortex is the main target of analysis in many functional magnetic resonance imaging (fMRI) studies. Since only about 20% of the voxels of a typical fMRI data set lie within the cortex, statistical analysis can be restricted to the subset of the voxels obtained after cortex segmentation. While such restriction does not influence conventional univariate statistical tests, it may have a substantial effect on the performance of multivariate methods.

Here, we describe a novel approach for data-driven analysis of single-subject fMRI time series that combines techniques for the segmentation and reconstruction of the cortical surface of the brain and the spatial independent component analysis (sICA) of the functional time courses (TCs). We use the mesh of the white matter/gray matter boundary, automatically reconstructed from high-spatial-resolution anatomical MR images, to limit the sICA decomposition of a coregistered functional time series to those voxels which are within a specified region with respect to the cortical sheet (cortex-based ICA, or cbICA). We illustrate our analysis method in the context of fMRI blocked and event-related experimental designs and in an fMRI experiment with perceptually ambiguous stimulation, in which an a priori specification of the stimulation protocol is not possible.

A comparison between cbICA and conventional hypothesis-driven statistical methods shows that cortical surface maps and component TCs blindly obtained with cbICA reliably reflect task-related spatiotemporal activation patterns. Furthermore, the advantages of using cbICA when the specification of a temporal model of the expected hemodynamic response is not straightforward are illustrated and discussed. A comparison between cbICA and anatomically unconstrained ICA reveals that — beside reducing computational demand — the cortex-based approach improves the fitting of the ICA model in the gray matter voxels, the separation of cortical components and the estimation of their TCs, particularly in the case of fMRI data sets with a complex spatiotemporal statistical structure.     

Fixed-point algorithms for constrained ICA and their applications in fMRI data analysis

Constrained independent component analysis (CICA) eliminates the order ambiguity of standard ICA by incorporating prior information into the learning process to sort the components intrinsically. However, the original CICA (OCICA) and its variants depend on a learning rate, which is not easy to be tuned for various applications. To solve this problem, two learning-rate-free CICA algorithms were derived in this paper using the fixed-point learning concept. A complete stability analysis was provided for the proposed methods, which also made a correction to the stability analysis given to OCICA. Variations for adding constraints either to the components or to the associated time courses were derived too. Using synthetic data, the proposed methods yielded a better stability and a better source separation quality in terms of higher signal-to-noise-ratio and smaller performance index than OCICA. For the artificially generated brain activations, the new CICAs demonstrated a better sensitivity/specificity performance than standard univariate general linear model (GLM) and standard ICA. Original CICA showed a similar sensitivity/specificity gain but failed to converge for several times. Using functional magnetic resonance imaging (fMRI) data acquired with a well-characterized sensorimotor task, the proposed CICAs yielded better sensitivity than OCICA, standard ICA and GLM in all the target functional regions in terms of either higher t values or larger suprathreshold cluster extensions using the same significance threshold. In addition, they were more stable than OCICA and standard ICA for analyzing the sensorimotor fMRI data.     

Improved functional mapping of the human amygdala using a standard functional magnetic resonance imaging sequence with simple modifications

As the amygdala is involved in various aspects of emotional processing, its characterization using neuroimaging modalities, such as functional magnetic resonance imaging (fMRI), is of great interest. However, in fMRI, the amygdala region suffers from susceptibility artifacts that are composed of signal dropouts and image distortions. Various technically demanding approaches to reduce these artifacts have been proposed, and most require alterations beyond a mere change of the acquisition parameters and cannot be easily implemented by the user without changing the MR sequence code. In the present study, we therefore evaluated the impact of simple alterations of the acquisition parameters of a standard gradient-echo echo-planar imaging technique at 3 T composed of echo times (TEs) of 27 and 36 ms as well as section thicknesses of 2 and 4 mm while retaining a section orientation parallel to the intercommissural plane and an in-plane resolution of 2x2 mm(2). In contrast to previous studies, we based our evaluation on the resulting activation maps using an emotional stimulation paradigm rather than on MR raw image quality only. Furthermore, we tested the effects of spatial smoothing of the functional raw data in the course of postprocessing using spatial filters of 4 and 8 mm. Regarding MR raw image quality, a TE of 27 ms and 2-mm sections resulted in the least susceptibility artifacts in the anteromedial aspect of the temporal lobe. The emotional stimulation paradigm resulted in robust bilateral amygdala activation for the approaches with 2-mm sections only -- but with larger activation volumes for a TE of 36 ms as compared with that of 27 ms. Moderate smoothing with a 4-mm spatial filter represented a good compromise between increased sensitivity and preserved specificity. In summary, we showed that rather than applying advanced modifications of the MR sequence, a simple increase in spatial resolution (i.e., the reduction of section thickness) is sufficient to improve the detectability of amygdala activation.     

Semiautomated segmentation of the human spine based on echoplanar images

The number of functional magnetic resonance imaging (fMRI) studies performed on the human spinal cord (SC) has considerably increased in recent years. The lack of a validated processing pipeline is, however, a significant obstacle to the spread of SC fMRI. One component likely to be involved in any such pipeline is the process of SC masking, analogous to brain extraction in cerebral fMRI. In general, SC masking has been performed manually, with the incumbent costs of being very time consuming and operator dependent. To overcome these drawbacks, we have developed a tailored semiautomatic method for segmenting echoplanar images (EPI) of human spine that is able to identify the spinal canal and the SC. The method exploits both temporal and spatial features of the EPI series and was tested and optimized on EPI images of cervical spine acquired at 3 T. The dependence of algorithm performance on the degree of EPI image distortion was assessed by computing the displacement warping field that best matched the EPI to the corresponding high-resolution T(2) images. Segmentation accuracy was above 80%, a significant improvement over values obtained with similar approaches, but not exploiting temporal information. Geometric distortion was found to explain about 50% of the variance of algorithm classification efficiency.     

Performance evaluation of nonnegative matrix factorization algorithms to estimate task-related neuronal activities from fMRI data

Nonnegative matrix factorization (NMF) is a blind source separation (BSS) algorithm which is based on the distinct constraint of nonnegativity of the estimated parameters as well as on the measured data. In this study, according to the potential feasibility of NMF for fMRI data, the four most popular NMF algorithms, corresponding to the following two types of (1) least-squares based update [i.e., alternating least-squares NMF (ALSNMF) and projected gradient descent NMF] and (2) multiplicative update (i.e., NMF based on Euclidean distance and NMF based on divergence cost function), were investigated by using them to estimate task-related neuronal activities. These algorithms were applied firstly to individual data from a single subject and, subsequently, to group data sets from multiple subjects. On the single-subject level, although all four algorithms detected task-related activation from simulated data, the performance of multiplicative update NMFs was significantly deteriorated when evaluated using visuomotor task fMRI data, for which they failed in estimating any task-related neuronal activities. In group-level analysis on both simulated data and real fMRI data, ALSNMF outperformed the other three algorithms. The presented findings may suggest that ALSNMF appears to be the most promising option among the tested NMF algorithms to extract task-related neuronal activities from fMRI data.     

To smooth or not to smooth? ROC analysis of perfusion fMRI data

Blood oxygenation level dependent (BOLD) contrast has been widely used for visualizing regional neural activation. Temporal filtering and parameter estimation algorithms are generally used to account for the intrinsic temporal autocorrelation present in BOLD data. Arterial spin labeling perfusion imaging is an emerging methodology for visualizing regional brain function both at rest and during activation. Perfusion contrast manifests different noise properties compared with BOLD contrast, represented by the even distribution of noise power and spatial coherence across the frequency spectrum. Consequently, different strategies are expected to be employed in the statistical analysis of functional magnetic resonance imaging (fMRI) data based on perfusion contrast. In this study, the effect of different analysis methods upon signal detection efficacy, as assessed by receiver operator characteristic (ROC) measures, was examined for perfusion fMRI data. Simulated foci of neural activity of varying amplitude and spatial extent were added to resting perfusion data, and the accuracy of each analysis was evaluated by comparing the results with the known distribution of pseudo-activation. In contrast to the BOLD fMRI, temporal smoothing or filtering reduces the power of perfusion fMRI data analyses whereas spatial smoothing is beneficial to the efficacy of analyses.     

Optimizing the mapping of finger areas in primary somatosensory cortex using functional MRI

Functional magnetic resonance imaging mapping of the finger somatotopy in the primary somatosensory cortex requires a reproducible and precise stimulation. The highly detailed functional architecture in this region of the brain also requires careful consideration in choice of spatial resolution and postprocessing parameters. The purpose of this study is therefore to investigate the impact of spatial resolution and level of smoothing during tactile stimulation using a precise stimuli system. Twenty-one volunteers were scanned using 2³ mm³ and 3³ mm³ voxel volume and subsequently evaluated using three different smoothing kernel widths. The overall activation reproducibility was also evaluated. Using a high spatial resolution proved advantageous for all fingers. At 2³ mm³ voxel volume, activation of the thumb, middle finger and little finger areas was seen in 89%, 67% and 50% of the volunteers, compared to 78%, 61% and 33% at 3³ mm³, respectively. The sensitivity was comparable for nonsmoothed and slightly smoothed (4 mm kernel width) data; however, increasing the smoothing kernel width from 4 to 8 mm resulted in a critical decrease (50%) in sensitivity. In repeated measurements of the same subject at six different days, the localization reproducibility of all fingers was within 4 mm (1 S.D. of the mean). The precise computer-controlled stimulus, together with data acquisition at high spatial resolution and with only minor smoothing during evaluation, could be a very useful strategy in studies of brain plasticity and rehabilitation strategies in hand and finger disorders and injuries.     

Functional MRI of the motor cortex using a conventional gradient system: comparison of FLASH and EPI techniques

Gradient echo (GE) and echo planar imaging (EPI) techniques are two different approaches to functional MRI (fMRI). In contrast to GE sequences, the ultra short EPI technique facilitates fMRI experiments with high spatial and temporal resolution or mapping of the whole brain. Although it has become the method of choice for fMRI, EPI is generally restricted to modern scanners with a strong gradient system. The aim of our study was to evaluate the applicability of EPI for fMRI of the motor cortex using a 1.5 T scanner with a conventional gradient system of 10 mT/m (rise time: 1 ms). Therefore, EPI was compared with a well-established high resolution fast low angle shot (FLASH) technique (matrix size 128²). The FLASH technique was applied additionally with a 64² matrix size to exclude influences caused by different spatial resolution, because the EPI sequence was restricted to a 64² matrix size. A total of 35 healthy volunteers were included in this study. The task consisted of clenching and spreading of the right hand. FLASH and EPI techniques were compared regarding geometric distortions as well as qualitative and quantitative fMRI criteria: Mean signal increase between activation and rest and the area of activation were measured within the contralateral, ipsilateral, and supplementary motor cortex. The quality of subtraction images between activation and rest, as well as the quality of z-maps and time course within activated regions of interest, was evaluated visually. EPI revealed significant distortions of the anterior and postior brain margins; lateral distortions (relevant for the motor cortex) could be neglected in most cases. The mean signal increase was significantly higher using FLASH 128² compared to FLASH 64² and EPI 64², whereas the activated areas proved to be smaller in FLASH 128² functional images. Both results can be explained by well-documented partial volume effects, caused by different voxel size. Similar quality of the subtraction images and of the time courses in different regions of interest were found for all techniques under investigation, but slightly reduced quality of z-map in FLASH 128². Within the limits of reproducibility and measurement accuracy, the location of contralateral activation was similar using FLASH and EPI sequences. In conclusion, EPI proved to be a reliable technique for fMRI of the motor cortex, even on an MR scanner with a conventional gradient system.     

Effects of spatial smoothing on inter-subject correlation based analysis of FMRI

This study evaluates the effects of spatial smoothing on inter-subject correlation (ISC) analysis for FMRI data using the traditional model based analysis as a reference. So far within ISC analysis the effects of smoothing have not been studied systematically and linear Gaussian filters with varying kernel widths have been used without better knowledge about the effects of filtering. Instead, with the traditional general linear model (GLM) based analysis, the effects of smoothing have been studied extensively.     

In vivo mapping of functional domains and axonal connectivity in cat visual cortex using magnetic resonance imaging

Noninvasive cognitive neuroimaging studies based on functional magnetic resonance imaging (fMRI) are of ever-increasing importance for basic and clinical neurosciences. The explanatory power of fMRI could be greatly expanded, however, if the pattern of the neuronal circuitry underlying functional activation could be made visible in an equally noninvasive manner. In this study, blood oxygenation level-dependent (BOLD)-based fMRI and diffusion tensor imaging (DTI) were performed in the same cat visual cortex, and the foci of fMRI activation utilized as seeding points for 3D DTI fiber reconstruction algorithms, thus providing the map of the axonal circuitry underlying visual information processing. The methods developed in this study will lay the foundation for in vivo neuroanatomy and the ability for noninvasive longitudinal studies of brain development.     

Empirical optimization of ASL data analysis using an ASL data processing toolbox: ASLtbx

Arterial spin labeling (ASL) perfusion fMRI data differ in important respects from the more familiar blood oxygen level-dependent (BOLD) fMRI data and require specific processing strategies. In this paper, we examined several factors that may influence ASL data analysis, including data storage bit resolution, motion correction, preprocessing for cerebral blood flow (CBF) calculations and nuisance covariate modeling. Continuous ASL data were collected at 3 T from 10 subjects while they performed a simple sensorimotor task with an epoch length of 48 s. These data were then analyzed using systematic variations of the factors listed above to identify the approach that yielded optimal signal detection for task activation. Improvements in statistical power were found for use of at least 10 bits for data storage at 3 T. No significant difference was found in motor cortex regarding using simple subtraction or sinc subtraction, but the former presented minor but significantly (P<.024) larger peak t value in visual cortex. While artifactual head motion patterns were observed in synthetic data and background-suppressed ASL data when label/control images were realigned to a common target, independent realignment of label and control images did not yield significant improvements in activation in the sensorimotor data. It was also found that CBF calculations should be performed prior to spatial normalization and that modeling of global fluctuations yielded significantly increased peak t value in motor cortex. The implementation of all ASL data processing approaches is easily accomplished within an open-source toolbox, ASLtbx, and is advocated for most perfusion fMRI data sets.     

False positive control of activated voxels in single fMRI analysis using bootstrap resampling in comparison to spatial smoothing

Functional magnetic resonance imaging (fMRI) is an effective tool for the measurement of brain neuronal activities. To date, several statistical methods have been proposed for analyzing fMRI datasets to select true active voxels among all the voxels appear to be positively activated. Finding a reliable and valid activation map is very important and becomes more crucial in clinical and neurosurgical investigations of single fMRI data, especially when pre-surgical planning requires accurate lateralization index as well as a precise localization of activation map.Defining a proper threshold to determine true activated regions, using common statistical processes, is a challenging task. This is due to a number of variation sources such as noise, artifacts, and physiological fluctuations in time series of fMRI data which affect spatial distribution of noise in an expected uniform activated region. Spatial smoothing methods are frequently used as a preprocessing step to reduce the effect of noise and artifacts. The smoothing may lead to a shift and enlargement of activation regions, and in some extend, unification of distinct regions.In this article, we propose a bootstrap resampling technique for analyzing single fMRI dataset with the aim of finding more accurate and reliable activated regions. This method can remove false positive voxels and present high localization accuracy in activation map without any spatial smoothing and statistical threshold setting.     

Comprehensive mathematical simulation of functional magnetic resonance imaging time series including motion-related image distortion and spin saturation effect

There has been vast interest in determining the feasibility of functional magnetic resonance imaging (fMRI) as an accurate method of imaging brain function for patient evaluations. The assessment of fMRI as an accurate tool for activation localization largely depends on the software used to process the time series data. The performance evaluation of different analysis tools is not reliable unless truths in motion and activation are known. Lack of valid truths has been the limiting factor for comparisons of different algorithms. Until now, currently available phantom data do not include comprehensive accounts of head motion. While most fMRI studies assume no interslice motion during the time series acquisition in fMRI data acquired using a multislice and single-shot echo-planar imaging sequence, each slice is subject to a different set of motion parameters. In this study, in addition to known three-dimensional motion parameters applied to each slice, included in the time series computation are geometric distortion from field inhomogeneity and spin saturation effect as a result of out-of-plane head motion. We investigated the effect of these head motion-related artifacts and present a validation of the mapping slice-to-volume (MSV) algorithm for motion correction and activation detection against the known truths. MSV was evaluated, and showed better performance in comparison with other widely used fMRI data processing software, which corrects for head motion with a volume-to-volume realignment method. Furthermore, improvement in signal detection was observed with the implementation of the geometric distortion correction and spin saturation effect compensation features in MSV.     

Functional magnetic resonance imaging: Comparison between activation maps and computation pipelines in a clinical context

In this study new evaluation strategies for comparing different Statistical Parametric Maps computed from fMRI time-series analysis software tools are proposed. The aim of our work is to assess and quantitatively evaluate the statistical agreement of activation maps. Some pre-processing steps are necessary to compare SPMs (Statistical Parametric Maps), including segmentation and co-registration. The study of the statistical agreement is carried out following two ways. The first way considers SPMs as the result of two classification processes and extracts confusion matrix and Cohen's kappa index to assess agreement. Some considerations will be made on the statistical dependence of classes and a new formulation of kappa index will be used for overcoming this problem. The second way considers SPMs as two 3D images, and computes the similarity of SPMs images with a fuzzy formulation of the Jaccard Index. Several experiments were conducted both to assess the performance of the proposed evaluation tools and to compare activation maps computation pipelines from two widely used software tools in a clinical context.     

Improvement of spectral density-based activation detection of event-related fMRI data

For event-related data obtained from an experimental paradigm with a periodic design, spectral density at the fundamental frequency of the paradigm has been used as a template-free activation detection measure. In this article, we build and expand upon this detection measure to create an improved, integrated measure. Such an integrated measure linearly combines information contained in the spectral densities at the fundamental frequency as well as the harmonics of the paradigm and in a spatial correlation function characterizing the degree of co-activation among neighboring voxels. Several figures of merit are described and used to find appropriate values for the coefficients in the linear combination. Using receiver-operating characteristic analysis on simulated functional magnetic resonance imaging (fMRI) data sets, we quantify and validate the improved performance of the integrated measure over the spectral density measure based on the fundamental frequency as well as over some other popular template-free data analysis methods. We then demonstrate the application of the new method on an experimental fMRI data set. Finally, several extensions to this work are suggested.     

Effect of slice orientation on reproducibility of fMRI motor activation at 3 Tesla

The effect of slice orientation on reproducibility and sensitivity of 3T fMRI activation using a motor task has been investigated in six normal volunteers. Four slice orientations were used; axial, oblique axial, coronal and sagittal. We applied analysis of variance (ANOVA) to suprathreshold voxel statistics to quantify variability in activation between orientations and between subjects. We also assessed signal detection accuracy in voxels across the whole brain by using a finite mixture model to fit receiver operating characteristic (ROC) curves to the data. Preliminary findings suggest that suprathreshold cluster characteristics demonstrate high motor reproducibility across subjects and orientations, although a significant difference between slice orientations in number of activated voxels was demonstrated in left motor cortex but not cerebellum. Subtle inter-orientation differences are highlighted in the ROC analyses, which are not obvious by ANOVA; the oblique axial slice orientation offers the highest signal detection accuracy, whereas coronal slices give the lowest.     

Discussion on the choice of separated components in fMRI data analysis by spatial independent component analysis

By measuring the changes of magnetic resonance signals during a stimulation, the functional magnetic resonance imaging (fMRI) is able to localize the neural activation in the brain. In this report, we discuss the fMRI application of the spatial independent component analysis (spatial ICA), which maximizes statistical independence over spatial images. Included simulations show the possibility of the spatial ICA on discriminating asynchronous activations or different response patterns in an fMRI data set. An in vivo visual stimulation fMRI test was conducted, and the result shows a proper sum of the separated components as the final image is better than a single component, using fMRI data analysis by spatial ICA. Our result means that spatial ICA is a useful tool for the detection of different response activations and suggests that a proper sum of the separated independent components should be used for the imaging result of fMRI data processing.