Power spectral based detection of brain activation from fMR images

An efficient method for detecting activation on single and multiple epoch functional MRI (fMRI) data based on power spectral density of time-series and hidden Markov model is presented. Conventional methods of analysis of fMRI data are generally based on time-domain correlation analysis concentrating mainly on the multiple epoch data and generally do not provide good results for single epoch data. The main focus of this study is the analysis of single epoch data, constrained by certain experiments such as pain response, sleep, administration of pharmacological agents, which can only have a single or very few stimulus cycles. Further, our method obviates the need to exclusively model the hemodynamic response function and correctly identifies the voxels with delayed activation. We demonstrate the efficacy of our method in detecting brain activation by using both synthetic and real fMRI data.     

基于栈式循环神经网络的血液动力学状态估计方法

利用fMRI数据准确地估计血液动力学状态, 能得到一种更接近神经元层面的大脑活动的客观表示, 这将促进人们对大脑运行机理的深刻理解, 推动脑认知的进一步发展.迄今为止, 人们已经提出了许多血液动力学状态估计方法.然而, 这些方法大都只考虑了相邻时刻血液动力学状态之间的关系, 忽视了更深层次的时序特征.而对模型参数先验信息的需求也使一些方法在实际应用中受到了限制.为此, 本文提出了一种基于循环神经网络的血液动力学状态估计新方法.首先, 利用血液动力学模型中非线性函数的反函数建立BOLD信号与血液动力学状态之间的映射关系, 并构建模型的反演过程.然后, 采用一种堆叠三个RNN模块的栈式神经网络结构来拟合这种映射关系, 使其能够以BOLD信号作为输入, 得到血液动力学状态的估计值.最后, 在仿真数据上验证新方法的性能.实验结果表明:与一些代表算法相比, 新方法能够更合理地提取fMRI数据中的时间特性, 有效地拟合BOLD信号与血液动力学状态之间的动态非线性关系.     

Statistical Methods in the Fourier Domain to Enhance and Classify Images

A mathematical model, for which rigorous methods of statistical inference are available, is described and techniques for image enhancement and linear discriminant analysis of groups are developed. Since the gray values of neighboring pixels in tomographically produced medical images are spatially correlated, the calculations are carried out in the Fourier domain to insure statistical independence of the variables. Furthermore, to increase the power of statistical tests the known spatial covariance was used to specify constraints in the spectral domain. These methods were compared to statistical procedures carried out in the spatial domain. Positron emission tomography (PET) images of alcoholics with organic brain disorders were compared by these techniques to age-matched normal volunteers. Although these techniques are employed to analyze group characteristics of functional images, they provide a comprehensive set of mathematical and statistical procedures in the spectral domain that can also be applied to images of other modalities, such as computed tomography (CT) or magnetic resonance imaging (MRI).     

自适应阈值化功能磁共振统计参数映射图

为了准确检测及定位功能激发信息,需选择一个客观、有效的阈值来阈值化功能磁共振统计参数映射图.为此提出了一种组合控制错误发现率及分析皮层血流动力学响应的空间相关性阈值化功能磁共振统计参数映射图的方法.该方法首先采用基于控制错误发现率的方法确定阈值,进行激发体素判别,然后分析已判别为激发的体素与其三维空间26-邻域体素的血流动力学响应的相关性,并进行空间相关检验.该方法不仅能够自适应地选取阈值,而且能够识别由于随机因素而导致的伪激发体素,具有更好脑功能激发信息检测及空间定位能力.     

Identifying distributed and overlapping clusters of hemodynamic synchrony in fMRI data sets

Natural sensory stimuli elicit complex brain responses that manifest in fMRI as widely distributed and overlapping clusters of hemodynamic responses. We propose a statistical signal processing method for finding synchronous hemodynamic activity that directly or transiently reflects information about the experimental condition. When applied to fMRI data, the method searches for voxels with activation patterns exhibiting high coherence and simultaneously high variance across brain scans. The crux of the method is functional principal component analysis (fPCA) of activation patterns stored in a two-dimensional data matrix, with rows and columns representing voxels and scans, respectively. Without external information, fPCA is performed directly on this data matrix. Otherwise, the data matrix is first transformed to highlight a specific source of variation, enabling fully or partially supervised fPCA with a single parameter determining the degree of supervision. We evaluated our method on a public benchmark of fMRI scans of subjects viewing natural movies. Our method turns out to be very suitable for flexibly uncovering distributed and overlapping hemodynamic patterns that distinguish well between experimental conditions or cognitive states.     

fMRI analysis on the GPU-possibilities and challenges

Functional magnetic resonance imaging (fMRI) makes it possible to non-invasively measure brain activity with high spatial resolution. There are however a number of issues that have to be addressed. One is the large amount of spatio-temporal data that needs to be processed. In addition to the statistical analysis itself, several preprocessing steps, such as slice timing correction and motion compensation, are normally applied. The high computational power of modern graphic cards has already successfully been used for MRI and fMRI. Going beyond the first published demonstration of GPU-based analysis of fMRI data, all the preprocessing steps and two statistical approaches, the general linear model (GLM) and canonical correlation analysis (CCA), have been implemented on a GPU. For an fMRI dataset of typical size (80 volumes with 64 × 64 × 22 voxels), all the preprocessing takes about 0.5 s on the GPU, compared to 5 s with an optimized CPU implementation and 120 s with the commonly used statistical parametric mapping (SPM) software. A random permutation test with 10,000 permutations, with smoothing in each permutation, takes about 50 s if three GPUs are used, compared to 0.5-2.5 h with an optimized CPU implementation. The presented work will save time for researchers and clinicians in their daily work and enables the use of more advanced analysis, such as non-parametric statistics, both for conventional fMRI and for real-time fMRI.     

基于PEM的三自由度直升机模型辨识

针对三自由度(3-DOF)直升机平台的特点,提出了一种基于预测误差法(PEM)的模型频域辨识方法,建立了机理模型,运用扫频技术得到巡航飞行状态直升机3个通道的输入-输出数据;分析了偏相干函数和复合窗函数,通过PEM进行了模型的频域辨识,得到了状态空间方程的待辨识参数和直升机的参数化模型.通过时域飞行和模型预测响应的对比,验证了该模型的准确性和该辨识方法的有效性.     

基于fMRI瞬时功率的独立成分分析

独立成分分析(independent component analysis,ICA)采用一种统计隐变量模型,假设信号是由各信源线性叠加构成.为了解决功能磁共振数据(functional magnetic resonance imaging,fMRI)中由于信源非线性叠加造成的ICA检测误差,提出了基于瞬时功率的ICA方法.首先,由电流能量形式将fMRI数据推广为fMRI能量信号;然后,由血氧水平依赖(blood oxygenation level dependent,BOLD)信号与T2*信号的关系,给出了两种反映BOLD能量变化的瞬时功率fMRI信号;最后,采用空间ICA分析fMRI瞬时功率信号,得到与各脑部活跃区域能量相关的独立成分.从理论和仿真试验两个方面阐明了新方法的合理性和优越性,同时应用于实际癫痫fMRI数据,经与传统ICA方法比较,该方法能够在静息态下鲁棒地检测脑部能量异常区域.     

Analytical modeling and simulation of a 2-DOF drive and 1-DOF sense gyro-accelerometer

In this paper, we present a new analytical model for frequency as well as transient analysis of a 3-DOF gyro-accelerometer system having 2-DOF in drive and 1-DOF in sense direction respectively. The model constructs lumped parameter differential equations by vector analysis associated with each degree of freedom that comprises Coriolis action, Euler’s action and action due to external acceleration along with biasing counterparts. These coupled differential equations are then solved explicitly in the frequency domain by taking their Laplace transforms. Based on these formulations, a thorough system analysis has been carried out keeping in view the various parameters and issues related to the device design. Furthermore, a discriminating scheme for time varying angular rate and linear acceleration by combining the structural model of a gyro-accelerometer with the demodulation and filtering processes to investigate frequency response of a micro gyro-accelerometer has also been presented by taking into account the presently derived settled transient solution of sense mode response. Finally we have verified the present model with MATLAB Simulink, showing their excellent agreement with each other.     

A Bayesian multilevel model for fMRI data analysis

Bayesian approaches have been proposed by several functional magnetic resonance imaging (fMRI) researchers in order to overcome the fundamental limitations of the popular statistical parametric mapping method. However, the difficulties associated with subjective prior elicitation have prevented the widespread adoption of the Bayesian methodology by the neuroimaging community. In this paper, we present a Bayesian multilevel model for the analysis of brain fMRI data. The main idea is to consider that all the estimated group effects (fMRI activation patterns) are exchangeable. This means that all the collected voxel time series are considered manifestations of a few common underlying phenomena. In contradistinction to other Bayesian approaches, we think of the estimated activations as multivariate random draws from the same distribution without imposing specific prior spatial and/or temporal information for the interaction between voxels. Instead, a two-stage empirical Bayes prior approach is used to relate voxel regression equations through correlations between the regression coefficient vectors. The adaptive shrinkage properties of the Bayesian multilevel methodology are exploited to deal with spatial variations, and noise outliers. The characteristics of the proposed model are evaluated by considering its application to two real data sets.     

一种基于自然图像统计模型的图像合成检测方法

Nowadays, digital images can be easily tampered due to the availability of powerful image processing software. As digital cameras continue to replace their analog counterparts, the importance of authenticating digital images, identifying their sources, and detecting forgeries is increasing. Blind image forensics is used to analyze an image in the complete absence of any digital watermark or signature. Image compositing is the most common form of digital tampering. Assuming that image compositing operations affect the inherent statistics of the image, we propose an image compositing detection method on based on a statistical model for natural image in the wavelet transform domain. The generalized Gaussian model (GGD) is employed to describe the marginal distribution of wavelet coefficients of images, and the parameters of GGD are obtained using maximum-likelihood estimator. The statistical features include GGD parameters, prediction error, mean, variance, skewness, and kurtosis at each wavelet detail subband. Then, these feature vectors are used to discriminate between natural images and composite images using support vector machine (SVM). To evaluate the performance of our proposed method, we carried out tests on the Columbia Uncompressed Image Splicing Detection Dataset and another advanced dataset, and achieved a detection accuracy of 92% and 79%, respectively. The detection performance of our method is better than that of the method using camera response function on the same dataset.     

Lattice estimation from images of patterns that exhibit translational symmetry

The analysis of regular texture images is cast in a model comparison framework. Texel lattice hypotheses are used to define statistical models which are compared in terms of their ability to explain the images. This approach is used to estimate lattice geometry from patterns that exhibit translational symmetry (regular textures). It is also used to determine whether images consist of such regular textures. A method based on this approach is described in which lattice hypotheses are generated using analysis of peaks in the image autocorrelation function, statistical models are based on Gaussian or Gaussian mixture clusters, and model comparison is performed using the marginal likelihood as approximated by the Bayes Information Criterion (BIC). Experiments on public domain images and a commercial textile image archive demonstrate substantially improved accuracy compared to several alternative methods.     

Classified image compression using optimally structured auto-association networks

Here, an application of a set of auto-association networks with linear output neurons and sigmoidal hidden neurons for classified image compression is carried out. Simulations and statistical analysis of this type of network have shown that, at convergence, the hidden neurons operate mainly in their linear region. The nearly linear behaviour of the hidden neurons is exploited in finding out the minimum number of hidden neurons needed to reconstruct image data within a certain error threshold. Four optimally structured auto-association networks are set up so that each network is trained to encode a certain variance-based class of image blocks. Results have shown excellent performance of the proposed architecture in reproducing high-quality images at a low bit rate.     

Motion blur identification in noisy images using mathematical models and statistical measures

Motion blur is one of the most common blurs that degrades images. Restoration of such images is highly dependent on estimation of motion blur parameters. Since 1976, many researchers have developed algorithms to estimate linear motion blur parameters. These algorithms are different in their performance, time complexity, precision and robustness in noisy environments. In this paper, we have presented a novel algorithm to estimate linear motion blur parameters such as direction and length. We used Radon transform to find direction and bispectrum modeling to find the length of motion. Our algorithm is based on the combination of spatial and frequency domain analysis. The great benefit of our algorithm is its robustness and precision in noisy images. We used statistical measures to prove goodness of our model. Our method was tested on 80 standard images that were degraded with different directions and motion lengths, with additive Gaussian noise. The error tolerance average of the estimated parameters was 0.9° in direction and 0.95 pixel in length and the standard deviations were 0.69 and 0.85, respectively.     

Linear Feature Transform and Enhancement of Classification on Deep Neural Network

A weighted and convex regularized nuclear norm model is introduced to construct a rank constrained linear transform on feature vectors of deep neural networks. The feature vectors of each class are modeled by a subspace, and the linear transform aims to enlarge the pairwise angles of the subspaces. The weight and convex regularization resolve the rank degeneracy of the linear transform. The model is computed by a difference of convex function algorithm whose descent and convergence properties are analyzed. Numerical experiments are carried out in convolutional neural networks on CAFFE platform for 10 class handwritten digit images (MNIST) and small object color images (CIFAR-10) in the public domain. The transformed feature vectors improve the accuracy of the network in the regime of low dimensional features subsequent to dimensional reduction via principal component analysis. The feature transform is independent of the network structure, and can be applied to reduce complexity of the final fully-connected layer without retraining the feature extraction layers of the network.     

基于马尔科夫模型的通用隐秘检测技术

对Sullivan等人提出的马尔科夫链(Markov Chain,MC)模型进行深入研究,并分析了其不能同时对空间域和频率域进行有效检测的缺陷。针对该缺陷,利用图像小波域的特性,提出了一个能够同时对空间域和频率域进行检测的改进方案,在基于模型的检测理论下,证明了该方案的有效性。实验表明,该方案具有更高的通用性和检测率。     

脑图像数据中的独立分量分析方法

针对脑功能磁共振成像在处理数据时空间维数较大的问题,提出一种空间独立分量分析(ICA)方法。研究空间ICA方法的基本模型结构和空间ICA的3种常见算法,即Infomax算法、Fixed-Point算法和Orth-Infomax算法。设计中文词义辨别实验,并使用线性相关方法进行算法比较。实验结果表明,与Infomax算法、Fixed-Point算法相比,Orth-Infomax算法任务相关分量的时间序列与参考函数的平均相关系数最大,具有较高的求解质量和求解效率,能够有效处理脑功能磁共振成像系统中存在的大量数据。     

小波域功能磁共振图像的处理和功能区的检测

传统的Gaussian变换不能够很好地去除功能磁共振图像当中的相关性噪声,影响对功能区的检测结果.为了更准确地检测和定位功能区,提出了基于小波变换的方法对功能磁共振图像进行处理.首先采用非线性小波变换阈值法在小波域对功能磁共振图像进行降噪处理;然后结合小波域的错误发现率算法对脑激活区进行检验.多套数据的统计结果表明,与传统的Gaussian变换相比,文中方法在保持检出敏感性的同时减少检测结果中假阳性点的数量,具有较高的检出特异性和定位可靠性.     

基于同步EEG-fMRI采集的情绪认知重评数据特征融合分析研究

脑电(Electroencephalography, EEG)与功能磁共振成像(Functional magnetic resonance imaging, fMRI)为脑科学研究提供了互补的时空信息. 为研究大脑在对情绪图片采取认知重评策略时的神经活动, 基于同步采集的EEG-fMRI数据, 应用典型相关分析、经验模态分解及k-均值聚类等算法对融合情绪数据进行交叉关联和盲源分离, 得到空间上的fMRI图像和与之对应的EEG时间演变信号. 结果表明: 时域上, CCA分离出的脑电成分在认知重评状态下有明显的晚期正电位(Late positive potential, LPP) (潜伏期200ms～900ms)出现, 而且认知重评策略诱发下的LPP 波幅明显小于观看负性诱发的LPP波幅(F(1, 224)= 28.72, P<0.01), 而大于观看中性诱发的LPP波幅(F(1, 224)= 63.32, P<0.01); 与之对应的空域上, 可以明显地看出和情绪调节相关的扣带回, 额叶、颞叶等区域有明显激活区, 采用情绪认知重评策略时的脑区激活强度明显小于观看负性状态, 而大于观看中性, 且观看中性状态下被激活的与情绪相关的区域相对较少. 研究表明, 这种融合数据分析技术通过计算两种模态数据之间潜在的线性相关性, 可以有效地分离出大脑在时空上神经活动情况, 达到了同时描绘出大脑神经活动的时间信息与空间信息的效果.     

The use of machine learning and deep learning algorithms in functional magnetic resonance imaging—A systematic review

Functional Magnetic Resonance Imaging (fMRI) is presently one of the most popular techniques for analysing the dynamic states in brain images using various kinds of algorithms. From the last decade, there is an exponential rise in the use of the machine and deep learning algorithms of artificial intelligence for analysing fMRI data. However, it is a big challenge for every researcher to choose a suitable machine or deep learning algorithm for analysing fMRI data due to the availability of a large number of algorithms in the literature. It takes much time for each researcher to know about the various approaches and algorithms which are in use for fMRI data. This paper provides a review in a systematic manner for the present literature of fMRI data that makes use of the machine and deep learning algorithms. The major goals of this review paper are to (a) identify machine learning and deep learning research trends for the implementation of fMRI; (b) identify usage of Machine Learning Algorithms and deep learning in fMRI, and (c) help new researchers based on fMRI to put their new findings appropriately in existing domain of fMRI research. The results of this systematic review identified various fMRI studies and classified them based on fMRI types, mental diseases, use of machine learning and deep learning algorithms. The authors have provided the studies with the best performance of machine learning and deep learning algorithms used in fMRI. The authors believe that this systematic review will help incoming researchers on fMRI in their future works.