IMPROVED DETECTION OF TIME WINDOWS OF BRAIN RESPONSES IN fMRI USING MODIFIED TEMPORAL CLUSTERING ANALYSIS

Temporal clustering analysis (TCA) has been proposed recently as a method to detect time windows of brain responses in functional MRI (fMRI) studies when the timing and location of the activation are completely unknown. Modifications to the TCA technique are introduced in this report to further improve the sensitivity in detecting brain activation.     

An improved temporal clustering analysis method applied to whole-brain data in fMRI study

Temporal clustering analysis (TCA) has been proposed as a method to detect the brain responses of an fMRI time series when the time and location of the activation are completely unknown. But TCA is still incompetent in dealing with the time series of the whole brain due to the existence of many inactive pixels. If only active pixels are considered, the sensitivity of TCA will be improved greatly and it could be applied to the whole brain. In this study, some modifications were made to TCA to remove inactive pixels, and the applicability of the modified TCA to the whole brain was validated with a set of visual fMRI data. Based on the time series of the modified TCA, activations of the whole brain corresponding to the visual stimulation were detected. Compared with the previous TCA, the modified TCA method shows a significant improvement in the sensitivity to detect activation peaks of the whole brain.     

Iterative temporal clustering analysis for the detection of multiple response peaks in fMRI

The temporal clustering analysis (TCA) is a novel and effective technique for obtaining brain activation maps when the timing and location of the activation are completely unknown. Performing the TCA method once can only detect the largest peak of the activation time windows well, if multiple response peaks at the same location of the brain occur. However, this limitation can be removed by using a TCA method in an iterative way in order for the smaller peaks to be detected. Our in vivo fMRI experiments with event-related visual tasks have demonstrated this ability.     

Improved temporal clustering analysis method applied to whole-brain data in acupuncture fMRI study

Temporal clustering analysis (TCA) has been proposed as a method for detecting the brain responses of a functional magnetic resonance imaging (fMRI) time series when the time and location of activation are completely unknown. But TCA is not suitable for treating the time series of the whole brain due to the existence of many inactive pixels. In theory, active pixels are located only in gray matter (GM). In this study, SPM2 was used to segment functional images into GM, white matter and cerebrospinal fluid, and only the pixels in GM were considered. Thus, most of inactive pixels are deleted, so that the sensitivity of TCA is greatly improved in the analysis of the whole brain. The same set of acupuncture fMRI data was treated using both conventional TCA and modified TCA (MTCA) for comparing their analytical ability. The results clearly show a significant improvement in the sensitivity achieved by MTCA.     

Derivative temporal clustering analysis: detecting prolonged neuronal activity

Temporal clustering analysis (TCA) and independent component analysis (ICA) are promising data-driven techniques in functional magnetic resonance imaging (fMRI) experiments to obtain brain activation maps in conditions with unknown temporal information regarding the neuronal activity. Although comparable to ICA in detecting transient neuronal activities, TCA fails to detect prolonged plateau brain activations. To eliminate this pitfall, a novel derivative TCA (DTCA) method was introduced and its algorithms with different subtraction intervals were tested on simulated data with a pattern of prolonged plateau brain activation. It was found that the best performance of DTCA method in generating functional maps could be obtained if the subtraction interval is equal to or larger than the length of the rising time of the fMRI response. The DTCA method and its theoretical predication were further investigated and validated using in vivo fMRI data sets. By removing the limitations in the previous TCA, DTCA has shown its powerful capability in detecting prolonged plateau neuronal activities.     

Neonatal auditory activation detected by functional magnetic resonance imaging

The objective of this study was to detect auditory cortical activation in non-sedated neonates employing functional magnetic resonance imaging (fMRI). Using echo-planar functional brain imaging, subjects were presented with a frequency-modulated pure tone; the BOLD signal response was mapped in 5 mm-thick slices running parallel to the superior temporal gyrus. Twenty healthy neonates (13 term, 7 preterm) at term and 4 adult control subjects. Blood oxygen level-dependent (BOLD) signal in response to auditory stimulus was detected in all 4 adults and in 14 of the 20 neonates. FMRI studies of adult subjects demonstrated increased signal in the superior temporal regions during auditory stimulation. In contrast, signal decreases were detected during auditory stimulation in 9 of 14 newborns with BOLD response. fMRI can be used to detect brain activation with auditory stimulation in human infants.     

Sensitivity limitations of high-resolution perfusion-based human fMRI at 7 Tesla

The study of the brain's functional organization at laminar and columnar level of the cortex with blood oxygenation-level dependent (BOLD) functional MRI (fMRI) is affected by the contribution of large veins downstream from the microvascular response to brain activity. Blood volume- and especially perfusion-based techniques may reduce this problem because of their reduced sensitivity to venous effects, but may not allow the same spatial resolution because of smaller signal changes associated with brain activity. Here we investigated the practical resolution limits of perfusion-weighted fMRI in human visual stimulation experiments. For this purpose, we used a highly sensitive, single-shot perfusion labeling (SSPL) technique at 7 T and compared sensitivity to detect visual activation at low (2 mm, n = 10) and high (1 mm, n = 8) nominal isotropic spatial, and 3 s temporal, resolution with BOLD in 5½-minute-long experiments. Despite the smaller absolute signal change with activation, 2 mm resolution SSPL yielded comparable sensitivity to BOLD. This was attributed to a superior suppression of physiological noise with SSPL. However, at 1 mm nominal resolution, SSPL sensitivity fell on average at least 42% below that of BOLD, and detection of visual activation was compromised. This is explained by the fact that at high resolution, with both techniques, typically thermal noise rather than physiological noise dominates sensitivity. The observed sensitivity loss implies that to perform 1-mm resolution, perfusion weighted fMRI with a robustness similar to BOLD, scan times that are almost 3 times longer than the comparable BOLD experiment are required. This is in line with or slightly better than previous comparisons between perfusion-weighted fMRI and BOLD. The lower sensitivity has to be weighed against the spatial fidelity advantages of high-resolution perfusion-weighted fMRI.     

Activation mapping as a percentage of local excitation: fMRI stability within scans, between scans and across field strengths

Functional magnetic resonance imaging (fMRI) does not typically yield highly reproducible maps of brain activation. Maps can vary significantly even with constant scanning parameters and consistent task performance conditions (Liu et al., Magn. Reson. Med., 2004, 52:751-760). Reproducibility is even more of a problem when comparing fMRI signal magnitude and spatial extent of activation across scans involving different task performance levels, scan durations, pulse sequences or magnetic field strengths. In this report, the consistency of fMRI was reexamined by considering the relative spatial and temporal distribution of fMRI blood oxygen level dependent (BOLD) activation signals separately from the absolute magnitude of the activation signal in each brain area. Subjects repeatedly performed the same simple motor task but under a variety of imaging conditions, using both spiral and standard echo-planar pulse sequences and at 1.5- and 4.0-T magnetic field strengths. The results demonstrate that the absolute amplitude of BOLD statistical activation signals varied significantly across time and scanning conditions, but the relative spatial pattern of BOLD activation was highly reproducible across all conditions. Analysis of realistic simulated fMRI data sets indicates that stability of relative activation patterns could provide a useful tool for assessing the accuracy of fMRI maps.     

Characterization of BOLD-fMRI signal during a verbal fluency paradigm in patients with intracerebral tumors affecting the frontal lobe

Previous studies have indicated that the BOLD-fMRI signal can be modified by tumor processes in close vicinity to functional brain areas. This effect has been investigated primarily for the perirolandic area but there is only a limited number of studies concerning frontal cortical regions. Therefore, the aim of the current study was to characterize BOLD-fMRI signal and activation patterns in patients with frontal brain tumors while performing a verbal fluency task. Six patients (ages 31-56 years) suffering from frontal (5 left sided and 1 right sided) intracerebral tumors were examined with fMRI while performing a verbal fluency task in a blocked paradigm design. Eight healthy volunteers served as the control group. The patients (5 right and 1 left handed) demonstrated left frontal activation which could be clearly located outside the tumor area and adjacent edema with varying degrees of additional right frontal activation. In the predominant left frontal activation cluster, the mean voxel based z-score and cluster size were not statistically different between patients and controls. The present fMRI study is indicating that language related BOLD signal changes in the frontal cortex of patients with tumors close to functional areas were comparable to the signal in normal controls. Additionally, the temporal hemodynamic response characteristic was comparable in both groups. This is an important finding consistent with PET results and corroborates the feasibility of functional mapping approaches in patients with tumors affecting the frontal lobe. Additional studies investigating alterations of the hemodynamic response depending on tumor location and histology are required in order to further elucidate the association between pathophysiology and BOLD fMRI signal.     

Dissecting cognitive stages with time-resolved fMRI data: a comparison of fuzzy clustering and independent component analysis

In combination with cognitive tasks entailing sequences of sensory and cognitive processes, event-related acquisition schemes allow using functional MRI to examine not only the topography but also the temporal sequence of cortical activation across brain regions (time-resolved fMRI). In this study, we compared two data-driven methods--fuzzy clustering method (FCM) and independent component analysis (ICA)--in the context of time-resolved fMRI data collected during the performance of a newly devised visual imagery task. We analyzed a multisubject fMRI data set using both methods and compared their results in terms of within- and between-subject consistency and spatial and temporal correspondence of obtained maps and time courses. Both FCM and spatial ICA allowed discriminating the contribution of distinct networks of brain regions to the main cognitive stages of the task (auditory perception, mental imagery and behavioural response), with good agreement across methods. Whereas ICA worked optimally on the original time series, averaging with respect to the task onset (and thus introducing some a priori information on the stimulation protocol) was found to be indispensable in the case of FCM. On averaged time series, FCM led to a richer decomposition of the spatio-temporal patterns of activation and allowed a finer separation of the neurocognitive processes subserving the mental imagery task. This study confirms the efficacy of the two examined methods in the data-driven estimation of hemodynamic responses in time-resolved fMRI studies and provides empirical guidelines to their use.     

Synchronized detection of minute electrical currents with MRI using Lorentz effect imaging

The blood oxygenation level-dependent (BOLD) effect is the most commonly used contrast mechanism in functional magnetic resonance imaging (fMRI), due to its relatively high spatial resolution and sensitivity. However, the ability of BOLD fMRI to accurately localize neuronal activation in space and time is limited by the inherent hemodynamic modulation. There is hence a need to develop alternative MRI methods that can directly image neuroelectric activity, thereby achieving both a high temporal resolution and spatial specificity as compared to conventional BOLD fMRI. In this paper, we extend the Lorentz effect imaging technique, which can detect spatially incoherent yet temporally synchronized minute electrical activity in a strong magnetic field, and demonstrate its feasibility for imaging randomly oriented electrical currents on the order of microamperes with a temporal resolution on the order of milliseconds in gel phantoms. This constitutes a promising step towards its application to direct imaging of neuroelectric activity in vivo, which has the same order of current density and temporal synchrony.     

Fast functional brain signal changes detected by diffusion weighted fMRI

Functional magnetic resonance imaging (fMRI) has become the method of choice in the study of system neuroscience, as evidenced by an explosion of such literature in the past decade. Contrast mechanisms based on the blood oxygenation level, volume, and flow changes have been used to non-invasively detect brain activation secondary to the neuronal activity. However, because of the hemodynamic modulations inherent in these signals, their spatial and temporal characteristics are influenced by the complex geometry and varying delivery speed of the brain vasculature. Consequently, spatial dispersions and temporal delays are commonly seen in the brain activity using fMRI. It is thus of critical importance to investigate alternative contrast mechanisms that may offer shorter temporal delays and more direct spatial localization. In light of a recent phantom study which demonstrated the possibility to detect the destructive phase addition from the spatially incoherent, yet temporally synchronized, displacements caused by the Lorentz force experienced during electrical conduction within a strong magnetic field, we seek to apply similar imaging technique to investigate the functional signal changes that may provide alternative temporal and spatial characteristics. It is found that by using heavy diffusion weighting, which is one form of displacement encoding strategies, to remove the vascular signal and sensitize the minute and incoherent displacement, one can detect fast dynamic signal changes synchronized to the task. This finding may help take an initial step toward direct non-invasive MRI detection of the neuronal activity with improved temporal accuracy.     

On applicability of PCA,voxel-wise variance normalization and dimensionality assumptions for sliding temporal window sICA in resting-state fMRI

Subject-level resting-state fMRI (RS-fMRI) spatial independent component analysis (sICA) may provide new ways to analyze the data when performed in the sliding time window. However, whether principal component analysis (PCA) and voxel-wise variance normalization (VN) are applicable pre-processing procedures in the sliding-window context, as they are for regular sICA, has not been addressed so far. Also model order selection requires further studies concerning sliding-window sICA. In this paper we have addressed these concerns. First, we compared PCA-retained subspaces concerning overlapping parts of consecutive temporal windows to answer whether in-window PCA and VN can confound comparisons between sICA analyses in consecutive windows. Second, we compared the PCA subspaces between windowed and full data to assess expected comparability between windowed and full-data sICA results. Third, temporal evolution of dimensionality estimates in RS-fMRI data sets was monitored to identify potential challenges in model order selection in a sliding-window sICA context. Our results illustrate that in-window VN can be safely used, in-window PCA is applicable with most window widths and that comparisons between windowed and full data should not be performed from a subspace similarity point of view. In addition, our studies on dimensionality estimates demonstrated that there are sustained, periodic and very case-specific changes in signal-to-noise ratio within RS-fMRI data sets. Consequently, dimensionality estimation is needed for well-founded model order determination in the sliding-window case. The observed periodic changes correspond to a frequency band of ≤ 0.1 Hz, which is commonly associated with brain activity in RS-fMRI and become on average most pronounced at window widths of 80 and 60 time points (144 and 108 s, respectively). Wider windows provided only slightly better comparability between consecutive windows, and 60 time point or shorter windows also provided the best comparability with full-data results. Further studies are needed to determine the cause for dimensionality variations.     

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.     

Evaluation and optimization of fMRI single-subject processing pipelines with NPAIRS and second-level CVA

In functional magnetic resonance imaging (fMRI) analysis, although the univariate general linear model (GLM) is currently the dominant approach to brain activation detection, there is growing interest in multivariate approaches such as principal component analysis, canonical variate analysis (CVA), independent component analysis and cluster analysis, which have the potential to reveal neural networks and functional connectivity in the brain. To understand the effect of processing options on performance of multivariate model-based fMRI processing pipelines with real fMRI data, we investigated the impact of commonly used fMRI preprocessing steps and optimized the associated multivariate CVA-based, single-subject processing pipelines with the NPAIRS (nonparametric prediction, activation, influence and reproducibility resampling) performance metrics [prediction accuracy and statistical parametric image (SPI) reproducibility] on the Fiswidgets platform. We also compared the single-subject SPIs of univariate GLM with multivariate CVA-based processing pipelines from SPM, FSL.FEAT, NPAIRS.GLM and NPAIRS.CVA software packages (or modules) using a novel second-level CVA. We found that for the block-design data, (a) slice timing correction and global intensity normalization have little consistent impact on the fMRI processing pipeline, but spatial smoothing, temporal detrending or high-pass filtering, and motion correction significantly improved pipeline performance across all subjects; (b) the combined optimization of spatial smoothing, temporal detrending and CVA model parameters on average improved between-subject reproducibility; and (c) the most important pipeline choices include univariate or multivariate statistical models and spatial smoothing. This study suggests that considering options other than simply using GLM with a fixed spatial filter may be of critical importance in determining activation patterns in BOLD fMRI studies.     

Integration of EEG source imaging and fMRI during continuous viewing of natural movies

Electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) are noninvasive neuroimaging tools which can be used to measure brain activity with excellent temporal and spatial resolution, respectively. By combining the neural and hemodynamic recordings from these modalities, we can gain better insight into how and where the brain processes complex stimuli, which may be especially useful in patients with different neural diseases. However, due to their vastly different spatial and temporal resolutions, the integration of EEG and fMRI recordings is not always straightforward. One fundamental obstacle has been that paradigms used for EEG experiments usually rely on event-related paradigms, while fMRI is not limited in this regard. Therefore, here we ask whether one can reliably localize stimulus-driven EEG activity using the continuously varying feature intensities occurring in natural movie stimuli presented over relatively long periods of time. Specifically, we asked whether stimulus-driven aspects in the EEG signal would be co-localized with the corresponding stimulus-driven BOLD signal during free viewing of a movie. Secondly, we wanted to integrate the EEG signal directly with the BOLD signal, by estimating the underlying impulse response function (IRF) that relates the BOLD signal to the underlying current density in the primary visual area (V1). We made sequential fMRI and 64-channel EEG recordings in seven subjects who passively watched 2-min-long segments of a James Bond movie. To analyze EEG data in this natural setting, we developed a method based on independent component analysis (ICA) to reject EEG artifacts due to blinks, subject movement, etc., in a way unbiased by human judgment. We then calculated the EEG source strength of this artifact-free data at each time point of the movie within the entire brain volume using low-resolution electromagnetic tomography (LORETA). This provided for every voxel in the brain (i.e., in 3D space) an estimate of the current density at every time point. We then carried out a correlation between the time series of visual contrast changes in the movie with that of EEG voxels. We found the most significant correlations in visual area V1, just as seen in previous fMRI studies (Bartels A, Zeki, S, Logothetis NK. Natural vision reveals regional specialization to local motion and to contrast-invariant, global flow in the human brain. Cereb Cortex 2008;18(3):705–717), but on the time scale of milliseconds rather than of seconds. To obtain an estimate of how the EEG signal relates to the BOLD signal, we calculated the IRF between the BOLD signal and the estimated current density in area V1. We found that this IRF was very similar to that observed using combined intracortical recordings and fMRI experiments in nonhuman primates. Taken together, these findings open a new approach to noninvasive mapping of the brain. It allows, firstly, the localization of feature-selective brain areas during natural viewing conditions with the temporal resolution of EEG. Secondly, it provides a tool to assess EEG/BOLD transfer functions during processing of more natural stimuli. This is especially useful in combined EEG/fMRI experiments, where one can now potentially study neural-hemodynamic relationships across the whole brain volume in a noninvasive manner.     

Increased BOLD sensitivity in the orbitofrontal cortex using slice-dependent echo times at 3 T

Functional magnetic resonance imaging (fMRI) exploits the blood oxygenation level dependent (BOLD) effect to detect neuronal activation related to various experimental paradigms. Some of these, such as reversal learning, involve the orbitofrontal cortex and its interaction with other brain regions like the amygdala, striatum or dorsolateral prefrontal cortex. These paradigms are commonly investigated with event-related methods and gradient echo-planar imaging (EPI) with short echo time of 27 ms. However, susceptibility-induced signal losses and image distortions in the orbitofrontal cortex are still a problem for this optimized sequence as this brain region consists of several slices with different optimal echo times. An EPI sequence with slice-dependent echo times is suitable to maximize BOLD sensitivity in all slices and might thus improve signal detection in the orbitofrontal cortex. To test this hypothesis, we first optimized echo times via BOLD sensitivity simulation. Second, we measured 12 healthy volunteers using a standard EPI sequence with an echo time of 27 ms and a modified EPI sequence with echo times ranging from 22 ms to 47 ms. In the orbitofrontal cortex, the number of activated voxels increased from 87±44 to 549±83 and the maximal t-value increased from 4.4±0.3 to 5.4±0.3 when the modified EPI was used. We conclude that an EPI with slice-dependent echo times may be a valuable tool to mitigate susceptibility artifacts in event-related whole-brain fMRI studies with a focus on the orbitofrontal cortex.     

High-resolution, multiple gradient-echo functional MRI at 1.5 T

A multiple gradient echo, high resolution imaging method is proposed to better visualize different sources of activation in functional magnetic resonance imaging (fMRI) experiments. Eight echoes are collected from 30 ms to 205 ms with an echo spacing of 25 ms. All echoes show significant activation, but each echo reveals its own pattern of activation. From this variability, it appears that large vessel contributions can be separated from small vessel contributions using a fuzzy cluster analysis across echo times. The results demonstrate the importance of a multiple gradient echo data acquisition approach in localizing various vascular contributions to brain activation in fMRI.     

Failure to direct detect magnetic field dephasing corresponding to ERP generation

Functional magnetic resonance imaging (fMRI) has become the method of choice for mapping brain activity in human subjects and detects changes in regional blood oxygenation and volume associated with local changes in neuronal activity. While imaging based on blood oxygenation level dependent (BOLD) contrast has good spatial resolution and sensitivity, the hemodynamic signal develops relatively slowly and is only indirectly related to neuronal activity. An alternative approach termed magnetic source magnetic resonance imaging (msMRI) is based on the premise that neural activity may be mapped by magnetic resonance imaging (MRI) with greater temporal resolution by detecting the local magnetic field perturbations associated with local neuronal electric currents. We used a hybrid ms/BOLD MRI method to investigate whether msMRI could detect signal changes that occur simultaneously at the time of the production of well-defined event-related potentials, the P300 and N170, in regions that previously have been identified as generators of these electrical signals. Robust BOLD activations occurred after some seconds, but we were unable to detect any significant changes in the T2*-weighted signal in these locations that correlated temporally with the timings of the evoked response potentials (ERPs).