Contextual clustering for analysis of functional MRI data

Sensitivity of Functional Arterial Spin Labelling in Detecting Cerebral Blood Flow Changes

Aims/Background Arterial spin labelling (ASL) is a non-invasive magnetic resonance imaging (MRI) method. ASL techniques can quantitatively measure cerebral perfusion by fitting a kinetic model to the difference between labelled images (tag images) and ones which are acquired without labelling (control images). ASL functional MRI (fMRI) provides quantitative perfusion maps by using arterial water as an endogenous tracer instead of depending on vascular blood oxygenation level.This study aimed to assess the number of pulsed ASL blocks that were needed to provide accurate and reliable regional estimates of cerebral blood flow (CBF) changes when participants engaged in visually guided saccade and fixation task; evaluate the localization to cortical control saccade versus fixation; investigate the relationship between the sensitivity of ASL fMRI and the number of blocks; and compare the sensitivity of blood oxygen level-dependent (BOLD) fMRI and ASL fMRI. Methods The experiment was a block-design paradigm consisting of two conditions: fixation and saccade. No response other than the eye movements of the participants was recorded during the scans. ASL and BOLD fMRI scans were conducted on all participants during the same session. The fMRI study consisted of two functional experiments: a CBF contrast was provided using the ASL sequence, and an optimized BOLD contrast was provided using the BOLD sequence. Results From group analysis in all divided blocks of ASL sessions (4, 6, 8...... 14, 16, 18......26, 28, 30), ASL yielded significant activation clusters in the visual cortex of the bilateral hemisphere from block 4. There was no false activation from block 4. No activation cluster was found by reversing analysis of block 2. Robust and consistent activation in the visual cortex was observed in each of the 14 divided blocks group analysis, and no activation was found in the eye field of the brain. The sensitivity of 4-block was found to be better than that of 8-block. More significant activation clusters of the visual cortex were found in BOLD than in ASL. No activation cluster of parietal eye field (PEF), frontal eye field (FEF) and supplementary eye field (SEF) was detected in ASL. The voxel size of the activation cluster increased with the increasing number of blocks, and the percent signal change in the activation cluster decreased with the escalating block number. The voxel size was positively correlated with the number of blocks (correlation coefficient = 0.98, p < 0.0001), and the percent signal change negatively correlated with the number of blocks (correlation coefficient = -0.90, p < 0.0001). Conclusion The 4-block pulsed functional ASL (fASL) presents accurate and reliable activation, with minimal time-on-task effect and little adverse impact of time, in participants engaging in visually guided saccade and fixation tasks. Despite having lower sensitivity than BOLD fMRI, ASL can determine accurate activation location. Although the time-on-task effects affect the observation for the sensitivity of ASL over task time, it is suggested that ASL fMRI may provide a powerful method for pinpointing the time-on-task effect over a long period of time.

ISLES 2015 - A public evaluation benchmark for ischemic stroke lesion segmentation from multispectral MRI

Ischemic stroke is the most common cerebrovascular disease, and its diagnosis, treatment, and study relies on non-invasive imaging. Algorithms for stroke lesion segmentation from magnetic resonance imaging (MRI) volumes are intensely researched, but the reported results are largely incomparable due to different datasets and evaluation schemes. We approached this urgent problem of comparability with the Ischemic Stroke Lesion Segmentation (ISLES) challenge organized in conjunction with the MICCAI 2015 conference. In this paper we propose a common evaluation framework, describe the publicly available datasets, and present the results of the two sub-challenges: Sub-Acute Stroke Lesion Segmentation (SISS) and Stroke Perfusion Estimation (SPES). A total of 16 research groups participated with a wide range of state-of-the-art automatic segmentation algorithms. A thorough analysis of the obtained data enables a critical evaluation of the current state-of-the-art, recommendations for further developments, and the identification of remaining challenges. The segmentation of acute perfusion lesions addressed in SPES was found to be feasible. However, algorithms applied to sub-acute lesion segmentation in SISS still lack accuracy. Overall, no algorithmic characteristic of any method was found to perform superior to the others. Instead, the characteristics of stroke lesion appearances, their evolution, and the observed challenges should be studied in detail. The annotated ISLES image datasets continue to be publicly available through an online evaluation system to serve as an ongoing benchmarking resource (www.isles-challenge.org).

Automatic segmentation for detecting uterine fibroid regions treated with MR-guided high intensity focused ultrasound (MR-HIFU)

PURPOSE

Up to 25% of women suffer from uterine fibroids (UF) that cause infertility, pain, and discomfort. MR-guided high intensity focused ultrasound (MR-HIFU) is an emerging technique for noninvasive, computer-guided thermal ablation of UFs. The volume of induced necrosis is a predictor of the success of the treatment. However, accurate volume assessment by hand can be time consuming, and quick tools produce biased results. Therefore, fast and reliable tools are required in order to estimate the technical treatment outcome during the therapy event so as to predict symptom relief.

METHODS

A novel technique has been developed for the segmentation and volume assessment of the treated region. Conventional algorithms typically require user interaction ora priori knowledge of the target. The developed algorithm exploits the treatment plan, the coordinates of the intended ablation, for fully automatic segmentation with no user input.

RESULTS

A good similarity to an expert-segmented manual reference was achieved (Dice similarity coefficient = 0.880 ± 0.074). The average automatic segmentation time was 1.6 ± 0.7 min per patient against an order of tens of minutes when done manually.

CONCLUSIONS

The results suggest that the segmentation algorithm developed, requiring no user-input, provides a feasible and practical approach for the automatic evaluation of the boundary and volume of the HIFU-treated region.

Oriented Markov random field based dendritic spine segmentation for fluorescence microscopy images

Dendritic spines have been shown to be closely related to various functional properties of the neuron. Usually dendritic spines are manually labeled to analyze their morphological changes, which is very time-consuming and susceptible to operator bias, even with the assistance of computers. To deal with these issues, several methods have been recently proposed to automatically detect and measure the dendritic spines with little human interaction. However, problems such as degraded detection performance for images with larger pixel size (e.g. 0.125 μm/pixel instead of 0.08 μm/pixel) still exist in these methods. Moreover, the shapes of detected spines are also distorted. For example, the "necks" of some spines are missed. Here we present an oriented Markov random field (OMRF) based algorithm which improves spine detection as well as their geometric characterization. We begin with the identification of a region of interest (ROI) containing all the dendrites and spines to be analyzed. For this purpose, we introduce an adaptive procedure for identifying the image background. Next, the OMRF model is discussed within a statistical framework and the segmentation is solved as a maximum a posteriori estimation (MAP) problem, whose optimal solution is found by a knowledge-guided iterative conditional mode (KICM) algorithm. Compared with the existing algorithms, the proposed algorithm not only provides a more accurate representation of the spine shape, but also improves the detection performance by more than 50% with regard to reducing both the misses and false detection.

Combining spatial priors and anatomical information for fMRI detection

In this paper, we analyze Markov Random Field (MRF) as a spatial regularizer in fMRI detection. The low signal-to-noise ratio (SNR) in fMRI images presents a serious challenge for detection algorithms, making regularization necessary to achieve good detection accuracy. Gaussian smoothing, traditionally employed to boost SNR, often produces over-smoothed activation maps. Recently, the use of MRF priors has been suggested as an alternative regularization approach. However, solving for an optimal configuration of the MRF is NP-hard in general. In this work, we investigate fast inference algorithms based on the Mean Field approximation in application to MRF priors for fMRI detection. Furthermore, we propose a novel way to incorporate anatomical information into the MRF-based detection framework and into the traditional smoothing methods. Intuitively speaking, the anatomical evidence increases the likelihood of activation in the gray matter and improves spatial coherency of the resulting activation maps within each tissue type. Validation using the receiver operating characteristic (ROC) analysis and the confusion matrix analysis on simulated data illustrates substantial improvement in detection accuracy using the anatomically guided MRF spatial regularizer. We further demonstrate the potential benefits of the proposed method in real fMRI signals of reduced length. The anatomically guided MRF regularizer enables significant reduction of the scan length while maintaining the quality of the resulting activation maps.

Probabilistic framework for brain connectivity from functional MR images

This paper unifies our earlier work on detection of brain activation (Rajapakse and Piyaratna, 2001) and connectivity (Rajapakse and Zhou, 2007) in a probabilistic framework for analyzing effective connectivity among activated brain regions from functional magnetic resonance imaging (fMRI) data. Interactions among brain regions are expressed by a dynamic Bayesian network (DBN) while contextual dependencies within functional images are formulated by a Markov random field. The approach simultaneously considers both the detection of brain activation and the estimation of effective connectivity and does not require a priori model of connectivity. Experimental results show that the present approach outperforms earlier fMRI analysis techniques on synthetic functional images and robustly derives brain connectivity from real fMRI data.

Experimental design and analysis in functional MRI

As the experimental diversity of functional magnetic resonance imaging (fMRI) has grown since introduction of the technique in 1991, the appropriate application of statistical evaluation has become a significant research need. Advances in our understanding of both the dynamics of the underlying neurophysiology and the measurement process make fMRI a fertile ground for novel signal and image processing research. Recent analysis procedures seek to incorporate knowledge of the temporal and spatial characteristics of the blood oxygenation level dependent (BOLD) response measured by fMRI. Efforts that incorporate the nonstationary aspects of the BOLD response will become more important as future applications of fMRI are likely to be in conjunction with additional imaging modalities (e.g., fMRI combined with EEG) to acquire more complete physiologic data and permit greater refinement of our characterization of the central nervous system responses arising due to an applied stimulus.

LEVEL SET BASED CLUSTERING FOR ANALYSIS OF FUNCTIONAL MRI DATA

We present a level set based clustering technique to detect activation regions from functional brain images using contextual information. Earlier similar approaches have been primarily concerned with local spatial context. Our approach relies on the idea that voxels within a functional region have similar temporal behavior. Using a level set formulation, a two-dimensional curve is evolved with a speed proportional to a similarity measure between the fMRI signals of voxels lying on the curve and their neighbors in the direction of propagation. The correlation coefficient is used to quantify similarity in time series of adjacent voxels. Simulation results from synthetic images demonstrate that using spatio-temporal contextual information provides better segmentation than a context-free, voxel-wise technique. Results from a real fMRI experiment using auditory stimulation are also presented.

Sensorimotor Cortex Localization: Comparison of Magnetoencephalography, Functional MR Imaging, and Intraoperative Cortical Mapping

PURPOSE

To prospectively evaluate magnetoencephalography (MEG) and functional magnetic resonance (MR) imaging, as compared with intraoperative cortical mapping, for identification of the central sulcus.

MATERIALS AND METHODS

Fifteen patients (six men, nine women; age range, 25-58 years) with a lesion near the primary sensorimotor cortex (13 gliomas, one cavernous hemangioma, and one meningioma) were examined after institutional review board approval and written informed consent from each patient were obtained. At MEG, evoked magnetic fields to median nerve stimulation were recorded; at functional MR imaging, hemodynamic responses to self-paced palmar flexion of the wrist were imaged. General linear model analysis with contextual clustering (P < .01) was used to analyze functional MR imaging data, and dipole modeling was used to analyze MEG data. MEG and functional MR localizations were compared with intraoperative cortical mappings. The distance from the area of functional MR imaging activation to the tumor margin was compared between the patients with discordant and those with concordant intraoperative mapping findings by using unpaired t testing.

RESULTS

MEG depicted the central sulcus correctly in all 15 patients, as verified at intraoperative mapping. The functional MR imaging localization results agreed with the intraoperative mappings in 11 patients. In all four patients with a false localization, the primary activation was in the postcentral sulcus region, but it did not differ significantly from the primary activation in the patients with correct localization with respect to proximity to the tumor (P = .38). Furthermore, at functional MR imaging, multiple nonprimary areas were activated, with considerable interindividual variation.

CONCLUSION

Although both MEG and functional MR imaging can provide useful information for neurosurgical planning, in the present study, MEG proved to be superior for locating the central sulcus. Activation of multiple nonprimary cerebral areas may confound the interpretation of functional MR imaging results.

A cluster mass permutation test with contextual enhancement for fMRI activation detection

Gaussian random field (GRF)-based methods are commonly used for statistical inference and to control the family-wise error rate (FWE) in neuroimaging. They require that the error fields are reasonable lattice approximations to an underlying continuous multivariate Gaussian random field and have differentiable and invertible spatial autocorrelation function. Permutation test estimates the distribution of the test statistic from the data and adjusts automatically for the FWE. Here we present a new analysis procedure, the cluster mass permutation test with contextual enhancement (CMPCE), and compare it to GRF. In CMPCE, the data are first pre-whitened to remove temporal autocorrelations. The FWE rates, the cluster detection probability and delineation accuracy of CMPCE and GRF were compared using measured null data and null data containing simulated activations. We also applied both methods to an fMRI experiment where tactile somatosensory stimulation into the right hand was used. When analyzing the FWE using null data, both CMPCE and GRF gave significantly higher FWEs (CMPCE up to 0.12, GRF up to 0.18) than the nominal significance level 0.05, indicating that the pre-whitening, motion correction or high-pass filtering partially failed. In the simulated activation data, CMPCE gave less falsely classified voxels for the same cluster detection probability level than GRF. The maximal cluster detection probability was on the other hand higher in the GRF-based method. Both methods gave qualitatively similar results in the tactile fMRI data. CMPCE seems to be a promising fMRI analysis method, especially if high delineation accuracy is required.

Contextual modeling of functional MR images with conditional random fields

This paper presents a conditional random field (CRF) approach to fuse contextual dependencies in functional magnetic resonance imaging (fMRI) data for the detection of brain activation. The interactions among both activation (activated/inactive) labels and observed data of brain voxels are unified in a probabilistic framework based on the CRF, where the interaction strength can be adaptively adjusted in terms of the data similarity of neighboring sites. Compared to earlier detection methods, including statistical parametric mapping and Markov random field, the proposed method avoids the suppression of high frequency information and relaxes the strong assumption of conditional independence of observed data. Experimental results show that the proposed approach effectively integrates contextual constraints within the detection process and robustly detects brain activities from fMRI data.

Model-independent method for fMRI analysis

This paper presents a fast method for delineation of activated areas of the brain from functional magnetic resonance imaging (fMRI) time series data. The steps of the work accomplished are as follows. 1) It is shown that the detection performance evaluated by the area under the receiver operating characteristic curve is directly related to the signal-to-noise ratio (SNR) of the composite image generated in the detection process. 2) Detection and segmentation of activated areas are formulated in a vector space framework. In this formulation, a linear transformation (image combination method) is shown to be desirable to maximize the SNR of the activated areas subject to the constraint of removing inactive areas. 3) An analytical solution for the problem is found. 4) Image pixel vectors and expected time series pattern (signature) for inactive pixels are used to calculate weighting vector and identify activated regions. 5) Signatures of the activated regions are used to segment different activities. 6) Segmented images by the proposed method are compared with those generated by the conventional methods (correlation, t-statistic, and z statistic). Detection performance and SNRs of the images are compared. The proposed approach outperforms the conventional methods of fMRI analysis. In addition, it is model-independent and does not require a priori knowledge of the fMRI response to the paradigm. Since the method is linear and most of the work is done analytically, numerical implementation and execution of the method are much faster than the conventional methods.