scholarly journals Statistical significance in DTI group analyses: How the choice of the estimator can inflate effect sizes

2019 ◽  
Author(s):  
Szabolcs David ◽  
Hamed Y. Mesri ◽  
Max A. Viergever ◽  
Alexander Leemans
Keyword(s):  
Least Squares ◽  
Diffusion Tensor ◽  
Statistical Significance ◽  
Group Analysis ◽  
Final Outcome ◽  
Effect Sizes ◽  
Estimation Methods ◽  
Linear Least Squares ◽  
Human Connectome Project

AbstractDiffusion magnetic resonance imaging (dMRI) is one of the most prevalent methods to investigate the micro- and macrostructure of the human brain in vivo. Prior to any group analysis, dMRI data are generally processed to alleviate adverse effects of known artefacts such as signal drift, data noise and outliers, subject motion, and geometric distortions. These dMRI data processing steps are often combined in automated pipelines, such as the one of the Human Connectome Project (HCP). While improving the performance of processing tools has clearly shown its benefits at each individual step along the pipeline, it remains unclear whether – and to what degree – choices for specific user-defined parameter settings can affect the final outcome of group analyses. In this work, we demonstrate how making such a choice for a particular processing step of the pipeline drives the final outcome of a group study. More specifically, we performed a dMRI group analysis on gender using HCP data sets and compared the results obtained with two diffusion tensor imaging estimation methods: the widely used ordinary linear least squares (OLLS) and the more reliable iterative weighted linear least squares (IWLLS). Our results show that the effect sizes for group analyses are significantly smaller with IWLLS than with OLLS. While previous literature has demonstrated higher estimation reliability with IWLLS than with OLLS using simulations, this work now also shows how OLLS can produce a larger number of false positives than IWLLS in a typical group study. We therefore highly recommend using the IWLLS method. By raising awareness of how the choice of estimator can artificially inflate effect size and thus alter the final outcome, this work may contribute to improvement of the reliability and validity of dMRI group studies.

scholarly journals Influence of activation pattern estimates and statistical significance tests in fMRI decoding analysis

2018 ◽  
Author(s):  
Juan E. Arco ◽  
Carlos González-García ◽  
Paloma Díaz-Gutiérrez ◽  
Javier Ramírez ◽  
María Ruz
Keyword(s):  
Least Squares ◽  
Pattern Analysis ◽  
Block Design ◽  
Statistical Significance ◽  
Estimation Methods ◽  
Permutation Testing ◽  
Functional Magnetic Resonance ◽  
Resonance Imaging ◽  
Imaging Studies ◽  
Sustained Activity

AbstractThe use of Multi-Voxel Pattern Analysis (MVPA) has increased considerably in recent functional magnetic resonance imaging studies. A crucial step consists in the choice of methods for the estimation of responses and their statistical significance. However, a systematic comparison of these and their adequacy to predominant experimental design is missing.In the current study, we compared three pattern estimation methods: Least-Squares Unitary (LSU), based on run-wise estimation, Least-Squares All (LSA) and Least-Squares Separate (LSS), which rely on trial-wise estimation. We compared the efficiency of these methods in an experiment where sustained activity had to be isolated from zero-duration events as well as in a block-design approach and in an event-related design. We evaluated the sensitivity of the t-test in comparison with two non-parametric methods based on permutation testing: one proposed in Stelzer et al. (2013), equivalent to performing a permutation in each voxel separately and the Threshold-Free Cluster Enhancement (Smith and Nichols, 2009).LSS resulted the most accurate approach to address the large overlap of signal among close events in the event-related designs. We found a larger sensitivity of Stelzer’s method in all settings, especially in the event-related designs, where voxels close to surpass the statistical threshold with the other approaches were now marked as informative regions.Our results provide evidence that LSS is the most accurate approach for unmixing events with different duration and large overlap of signal, consistent with previous studies showing better handling of collinearity in LSS. Moreover, Stelzer’s potentiates this better estimation with its larger sensitivity.

scholarly journals Effect Size and Power in fMRI Group Analysis

2018 ◽  
Author(s):  
Stephan Geuter ◽  
Guanghao Qi ◽  
Robert C. Welsh ◽  
Tor D. Wager ◽  
Martin A. Lindquist
Keyword(s):  
Sample Size ◽  
Group Analysis ◽  
Population Level ◽  
Small Sample ◽  
Effect Sizes ◽  
Medium Effect ◽  
Single Subject ◽  
Parameter Estimates ◽  
Sample Sizes ◽  
Human Connectome Project

AbstractMulti-subject functional magnetic resonance imaging (fMRI) analysis is often concerned with determining whether there exists a significant population-wide ‘activation’ in a comparison between two or more conditions. Typically this is assessed by testing the average value of a contrast of parameter estimates (COPE) against zero in a general linear model (GLM) analysis. In this work we investigate several aspects of this type of analysis. First, we study the effects of sample size on the sensitivity and reliability of the group analysis, allowing us to evaluate the ability of small sampled studies to effectively capture population-level effects of interest. Second, we assess the difference in sensitivity and reliability when using volumetric or surface based data. Third, we investigate potential biases in estimating effect sizes as a function of sample size. To perform this analysis we utilize the task-based fMRI data from the 500-subject release from the Human Connectome Project (HCP). We treat the complete collection of subjects (N = 491) as our population of interest, and perform a single-subject analysis on each subject in the population. We investigate the ability to recover population level effects using a subset of the population and standard analytical techniques. Our study shows that sample sizes of 40 are generally able to detect regions with high effect sizes (Cohen’s d > 0.8), while sample sizes closer to 80 are required to reliably recover regions with medium effect sizes (0.5 < d < 0.8). We find little difference in results when using volumetric or surface based data with respect to standard mass-univariate group analysis. Finally, we conclude that special care is needed when estimating effect sizes, particularly for small sample sizes.

Water-Fat Decomposition by IDEAL-MRI With Phase Estimation: A Method to Determine Chemical Contents In Vivo

2010 ◽  
Author(s):  
Jing Xu ◽  
Xiaofei Hu ◽  
Haiying Tang ◽  
Richard Kennan ◽  
Karim Azer
Keyword(s):  
Least Squares ◽  
Small Animal ◽  
Region Growing ◽  
Least Squares Estimation ◽  
Estimation Methods ◽  
Chemical Components ◽  
Proton Density ◽  
Phase Estimation ◽  
Field Inhomogeneity

High-resolution Magnetic Resonance Imaging (MRI) of humans and animals in vivo is routine and non-invasive. Identifying and quantifying chemical composition of tissue from acquired images is a challenge. MR spectroscopy (MRS) may be used to identify chemical components accurately over a finite volume in the tissue. However, the temporal and spatial resolutions are limited. Multi-spectral MRI exploits the multiple modes of MR such as T1, T2 and proton density maps and classifies voxels into different tissue types, but the chemical identity of the tissue remains unknown. Many fat suppression methods were developed because the unwanted fat signal often compromises image interpretability in clinical MRI, but these techniques are sensitive to MR field inhomogeneity. Multi-point Dixon methods separate MR images into water and fat images and are less sensitive to field inhomogeneity [1] and IDEAL-MRI (iterative decomposition of water and fat with echo asymmetry and least-squares estimation) improved upon the Dixon methods by avoiding the problem of phase unwrapping [2]. However, special care has to be taken when estimating the field map to avoid erroneous solutions to the least-squares estimation problem which lead to artifacts such as swapping of water and fat. The use of region growing schemes (with a reliable seed) mitigates this problem as demonstrated in previous studies [3][4]. However, the seed is not always reliable and growing schemes can be sensitive to phase discontinuities. Moreover, although the technology was successfully demonstrated on many clinical scanners, only limited applications were found in preclinical scanners with high MR field where the field inhomogeneity can be far worse [5]. We developed a robust and accurate algorithm to compute water and fat content on an 11.7T small animal scanner by improving upon existing phase estimation methods through multiple starting pixels and consensus-based region growing. The method, after further validation, has the potential of providing a translatable assay to study disease progression and regression related to fat and water contents in various animal models, such as studying atherosclerotic plaque composition.

Statistical significance of linear least-squares parameters

1979 ◽  
Vol 56 (4) ◽  
pp. 244 ◽  
Author(s):  
Merle D. Pattengill ◽  
Donald E. Sands
Keyword(s):  
Least Squares ◽  
Statistical Significance ◽  
Linear Least Squares

scholarly journals The Residual Plot for a Non–Linear Regression Model with the Presence of Outliers and Heteroscedastic Errors

2012 ◽  
Author(s):  
Habshah Midi ◽  
Azmi Jaafar
Keyword(s):  
Least Squares ◽  
Robust Regression ◽  
Estimation Methods ◽  
Linear Least Squares ◽  
Data Set ◽  
Leverage Points ◽  
Non Linear ◽  
Residual Plot ◽  
Heteroscedastic Errors ◽  
Residual Plots

Regresi teguh adalah sangat berguna bagi menilai kecukupan satu penyesuaian dan mencadangkan penjelmaan yang sesuai. Ini boleh dicapai dalam hanya satu pelaksanaan penganggaran teguh dan bukannya membina satu diagnostik titik terpencil. Dalam kertas ini, prestasi plot reja bagi Penganggar Teguh MM Berpemberat (WMM) dibandingkan dengan plot reja Kuasadua Terkecil Tak Linear (NLLS), Kuasadua Terkecil Tak Linear Teritlak (GNLLS) dan Penganggar Teguh MM. Dari keputusan berangka yang diperoleh, menyatakan bahawa plot reja NLLS dan GNLLS sukar mengenal pasti titik terpencil dan titik pelarasan tinggi. Lagipun, ianya tidak menunjukkan corak impian dalam selang reja [–2.5, 2.5]. Plot reja GNLLS dan WMM menghasilkan corak impian apabila varian ralat adalah heterosedastik dan tiada kekotoran berlaku dalam data set. Plot reja dari MM boleh mengenal pasti titik terpencil tetapi reja dalam selang [–2.5, 2.5] menyatakan bahawa ianya menokok dengan menokoknya tindakbalas penganggar, yang mungkin memerlukan penjelmaan yang sesuai. Plot reja WMM mempamerkan corak impian unggul dengan rejanya yang bertaburan secara rawak di dalam selang [–2.5, 2.5]. Kata kunci: Titik terpencil, regresi teguh, heterosedastik, penganggar MM berpemberat Robust regression is extremely useful in assessing the adequacy of a fit and suggesting appropriate transformations. This can be achieved in a single run by using robust estimation methods instead of constructing outlier diagnostics. In this paper, the performance of the residual plot of the robust Weighted MM estimators (WMM) was compared with the Non–Linear Least Squares (NLLS), Generalized Non–Linear Least Squares (GNLLS), and robust MM residual plots. The results obtained from numerical examples signified that the residual plots from the NLLS and the GNLLS fit can hardly identify outliers and high leverage points. Furthermore, it did not show an ideal pattern in the residuals within [–2.5, 2.5] interval. The GNLLS and the WMM residual plots revealed an ideal pattern when the error variances were heteroscedastic and no contamination occured in the data set. The residual plot from the MM fit can identify outliers but the residuals within the [–2.5, 2.5] interval indicated that the residuals increased with increasing estimate response, which may suggest an appropriate transformation. The WMM residual plot exhibited a pronounced ideal pattern denoted by its residuals, which were randomly distributed within [–2.5, 2.5] interval. Key words: Outliers, robust regression, heteroscedastic, weighted MM estimator

In vivo diffusion tensor imaging of the neonatal rat brain development

2013 ◽  
Vol 44 (S 01) ◽  
Author(s):  
M Breu ◽  
D Reisinger ◽  
D Wu ◽  
Y Zhang ◽  
A Fatemi ◽  
...  
Keyword(s):  
Diffusion Tensor Imaging ◽  
Brain Development ◽  
Rat Brain ◽  
Diffusion Tensor ◽  
Neonatal Rat ◽  
Neonatal Rat Brain

Diffusion tensor imaging reveals microstructural differences between subtypes of trigeminal neuralgia

2020 ◽  
Vol 133 (2) ◽  
pp. 573-579 ◽  
Author(s):  
Matthew S. Willsey ◽  
Kelly L. Collins ◽  
Erin C. Conrad ◽  
Heather A. Chubb ◽  
Parag G. Patil
Keyword(s):  
Diffusion Coefficient ◽  
Diffusion Tensor Imaging ◽  
Trigeminal Neuralgia ◽  
Apparent Diffusion Coefficient ◽  
Diffusion Tensor ◽  
Statistical Significance ◽  
Region Of Interest ◽  
Pain Syndrome ◽  
Radial Diffusivity ◽  
Apparent Diffusion

OBJECTIVETrigeminal neuralgia (TN) is an uncommon idiopathic facial pain syndrome. To assist in diagnosis, treatment, and research, TN is often classified as type 1 (TN1) when pain is primarily paroxysmal and episodic or type 2 (TN2) when pain is primarily constant in character. Recently, diffusion tensor imaging (DTI) has revealed microstructural changes in the symptomatic trigeminal root and root entry zone of patients with unilateral TN. In this study, the authors explored the differences in DTI parameters between subcategories of TN, specifically TN1 and TN2, in the pontine segment of the trigeminal tract.METHODSThe authors enrolled 8 patients with unilateral TN1, 7 patients with unilateral TN2, and 23 asymptomatic controls. Patients underwent DTI with parameter measurements in a region of interest within the pontine segment of the trigeminal tract. DTI parameters were compared between groups.RESULTSIn the pontine segment, the radial diffusivity (p = 0.0049) and apparent diffusion coefficient (p = 0.023) values in TN1 patients were increased compared to the values in TN2 patients and controls. The DTI measures in TN2 were not statistically significant from those in controls. When comparing the symptomatic to asymptomatic sides in TN1 patients, radial diffusivity was increased (p = 0.025) and fractional anisotropy was decreased (p = 0.044) in the symptomatic sides. The apparent diffusion coefficient was increased, with a trend toward statistical significance (p = 0.066).CONCLUSIONSNoninvasive DTI analysis of patients with TN may lead to improved diagnosis of TN subtypes (e.g., TN1 and TN2) and improve patient selection for surgical intervention. DTI measurements may also provide insights into prognosis after intervention, as TN1 patients are known to have better surgical outcomes than TN2 patients.

Evaluation for estimating of the PDF and the CDF of Generalized Inverted Exponential Distribution with Application in Industry

2020 ◽  
pp. 507-522
Author(s):  
Parisa Torkaman
Keyword(s):  
Least Squares ◽  
Exponential Distribution ◽  
Mean Squared Error ◽  
Weighted Least Squares ◽  
Real Data ◽  
Minimum Variance ◽  
Cumulative Distribution ◽  
Estimation Methods ◽  
Data Set ◽  
Better Than

The generalized inverted exponential distribution is introduced as a lifetime model with good statistical properties. This paper, the estimation of the probability density function and the cumulative distribution function of with five different estimation methods: uniformly minimum variance unbiased(UMVU), maximum likelihood(ML), least squares(LS), weighted least squares (WLS) and percentile(PC) estimators are considered. The performance of these estimation procedures, based on the mean squared error (MSE) by numerical simulations are compared. Simulation studies express that the UMVU estimator performs better than others and when the sample size is large enough the ML and UMVU estimators are almost equivalent and efficient than LS, WLS and PC. Finally, the result using a real data set are analyzed.

scholarly journals Whole-brain 3D MR fingerprinting brain imaging: clinical validation and feasibility to patients with meningioma

2021 ◽  
Author(s):  
Thomaz R. Mostardeiro ◽  
Ananya Panda ◽  
Robert J. Witte ◽  
Norbert G. Campeau ◽  
Kiaran P. McGee ◽  
...  
Keyword(s):  
White Matter ◽  
Statistical Significance ◽  
Relaxation Times ◽  
Brain Structures ◽  
Centrum Semiovale ◽  
In Vivo Evaluation ◽  
Whole Brain ◽  
Mean Values ◽  
Caudate Head

Abstract Purpose MR fingerprinting (MRF) is a MR technique that allows assessment of tissue relaxation times. The purpose of this study is to evaluate the clinical application of this technique in patients with meningioma. Materials and methods A whole-brain 3D isotropic 1mm3 acquisition under a 3.0T field strength was used to obtain MRF T1 and T2-based relaxometry values in 4:38 s. The accuracy of values was quantified by scanning a quantitative MR relaxometry phantom. In vivo evaluation was performed by applying the sequence to 20 subjects with 25 meningiomas. Regions of interest included the meningioma, caudate head, centrum semiovale, contralateral white matter and thalamus. For both phantom and subjects, mean values of both T1 and T2 estimates were obtained. Statistical significance of differences in mean values between the meningioma and other brain structures was tested using a Friedman’s ANOVA test. Results MR fingerprinting phantom data demonstrated a linear relationship between measured and reference relaxometry estimates for both T1 (r2 = 0.99) and T2 (r2 = 0.97). MRF T1 relaxation times were longer in meningioma (mean ± SD 1429 ± 202 ms) compared to thalamus (mean ± SD 1054 ± 58 ms; p = 0.004), centrum semiovale (mean ± SD 825 ± 42 ms; p < 0.001) and contralateral white matter (mean ± SD 799 ± 40 ms; p < 0.001). MRF T2 relaxation times were longer for meningioma (mean ± SD 69 ± 27 ms) as compared to thalamus (mean ± SD 27 ± 3 ms; p < 0.001), caudate head (mean ± SD 39 ± 5 ms; p < 0.001) and contralateral white matter (mean ± SD 35 ± 4 ms; p < 0.001) Conclusions Phantom measurements indicate that the proposed 3D-MRF sequence relaxometry estimations are valid and reproducible. For in vivo, entire brain coverage was obtained in clinically feasible time and allows quantitative assessment of meningioma in clinical practice.

Sign in / Sign up

Export Citation Format

Share Document