Stephen C. Strother

YNICL Journal 2025 Journal Article

Multi-voxel pattern analysis for characterizing functional connectivity and neurocognitive function in major depression: A CAN-BIND-1 report

• Alice Rueda
• Ilya Demchenko
• Vanessa K. Tassone
• Fatemeh Gholamali Nezhad
• Vanessa Peters
• Nathan W. Churchill
• Benicio N. Frey
• Stefanie Hassel

BACKGROUND: Major depressive disorder (MDD) affects not only mood but also neurocognitive function. In this study, we used whole-brain functional connectivity multi-voxel pattern analysis (fc-MVPA) to examine the relationship between resting-state functional connectivity (rsFC) and neurocognitive function in individuals with MDD compared to healthy controls (HC). METHODS: Baseline functional magnetic resonance imaging (fMRI) scans from the CAN-BIND-1 dataset were analyzed using a data-driven whole-brain fc-MVPA approach in 147 individuals with MDD and 98 HC. All participants completed the Computerized Neurocognitive Assessment Vital Signs (CNS-VS) battery outside the scanner, and correlations between rsFC differences and CNS-VS domain scores were explored. RESULTS: The fc-MVPA reduced the dimensionality of fMRI data at both individual and group levels, identifying six clusters with altered rsFC in MDD relative to HC: left cerebellar crus I, right precuneus, left superior lateral occipital cortex, right ventral caudate, left superior parietal lobule, and left dorsal anterior cingulate cortex. Using these clusters as seeds, post-hoc analyses identified 24 patterns of altered rsFC in MDD involving the default mode, central executive, visual recognition, salience, and sensorimotor networks. Five of these patterns showed significant correlations with CNS-VS domain scores for composite memory, neurocognition index, processing speed, executive function, and simple attention in HC, but these associations were absent in individuals with MDD. CONCLUSIONS: Our findings highlight that MDD is associated with disrupted rsFC across networks relevant to neurocognitive function. The data-driven nature of the fc-MVPA identified the left cerebellar crus I as the most significant region of aberrant rsFC.

YNICL Journal 2022 Journal Article

Predicting escitalopram treatment response from pre-treatment and early response resting state fMRI in a multi-site sample: A CAN-BIND-1 report

• Jacqueline K. Harris
• Stefanie Hassel
• Andrew D. Davis
• Mojdeh Zamyadi
• Stephen R. Arnott
• Roumen Milev
• Raymond W. Lam
• Benicio N. Frey

Many previous intervention studies have used functional magnetic resonance imaging (fMRI) data to predict the antidepressant response of patients with major depressive disorder (MDD); however, practical constraints have limited many of those attempts to small, single centre studies which may not adequately reflect how these models will generalize when used in clinical practice. Not only does the act of collecting data at multiple sites generally increase sample sizes (a critical point in machine learning development) it also generates a more heterogeneous dataset due to systematic differences in scanners at different sites, and geographical differences in patient populations. As part of the Canadian Biomarker Integration Network in Depression (CAN-BIND-1) study, 144 MDD patients from six sites underwent resting state fMRI prior to starting escitalopram treatment, and again two weeks after the start. Here, we consider ways to use machine learning techniques to produce models that can predict response (measured at eight weeks after initiation), based on various parcellations, functional connectivity (FC) metrics, dimensionality reduction algorithms, and base learners, and also whether to use scans from one or both time points. Models that use only baseline (pre-treatment) or only week 2 (early-response) whole-brain FC features consistently failed to perform significantly better than default models. Utilizing the change in FC between these two time points, however, yielded significant results, with the best performing analytical pipeline achieving 69.6% (SD 10.8) accuracy. These results appear contrary to findings from many smaller single-site studies, which report substantially higher predictive accuracies from models trained on only baseline resting state FC features, suggesting these models may not generalize well beyond data used for development. Further, these results indicate the potential value of collecting data both before and shortly after treatment initiation.

YNICL Journal 2021 Journal Article

Accelerated brain aging in major depressive disorder and antidepressant treatment response: A CAN-BIND report

• Pedro L. Ballester
• Jee Su Suh
• Nikita Nogovitsyn
• Stefanie Hassel
• Stephen C. Strother
• Stephen R. Arnott
• Luciano Minuzzi
• Roberto B. Sassi

OBJECTIVES: Previous studies suggest that major depressive disorder (MDD) may be associated with volumetric indications of accelerated brain aging. This study investigated neuroanatomical signs of accelerated aging in MDD and evaluated whether a brain age gap is associated with antidepressant response. METHODS: Individuals in a major depressive episode received escitalopram treatment (10-20 mg/d) for 8 weeks. Depression severity was assessed at baseline and at weeks 8 and 16 using the Montgomery-Asberg Depression Rating Scale (MADRS). Response to treatment was characterized by a significant reduction in the MADRS (≥50%). Nonresponders received adjunctive aripiprazole treatment (2-10 mg/d) for a further 8 weeks. The brain-predicted age difference (brain-PAD) at baseline was determined using machine learning methods trained on 3377 healthy individuals from seven publicly available datasets. The model used features from all brain regions extracted from structural magnetic resonance imaging data. RESULTS: Brain-PAD was significantly higher in older MDD participants compared to younger MDD participants [t(147.35) = -2.35, p < 0.03]. BMI was significantly associated with brain-PAD in the MDD group [r(155) = 0.19, p < 0.03]. Response to treatment was not significantly associated with brain-PAD. CONCLUSION: We found an elevated brain age gap in older individuals with MDD. Brain-PAD was not associated with overall treatment response to escitalopram monotherapy or escitalopram plus adjunctive aripiprazole.

YNICL Journal 2020 Journal Article

An investigation of cortical thickness and antidepressant response in major depressive disorder: A CAN-BIND study report

• Jee Su Suh
• Luciano Minuzzi
• Pradeep Reddy Raamana
• Andrew Davis
• Geoffrey B. Hall
• Jacqueline Harris
• Stefanie Hassel
• Mojdeh Zamyadi

Major depressive disorder (MDD) is considered a highly heterogeneous clinical and neurobiological mental disorder. We employed a novel layered treatment design to investigate whether cortical thickness features at baseline differentiated treatment responders from non-responders after 8 and 16 weeks of a standardized sequential antidepressant treatment. Secondary analyses examined baseline differences between MDD and controls as a replication analysis and longitudinal changes in thickness after 8 weeks of escitalopram treatment. 181 MDD and 95 healthy comparison (HC) participants were studied. After 8 weeks of escitalopram treatment (10-20 mg/d, flexible dosage), responders (>50% decrease in Montgomery-Åsberg Depression Scale score) were continued on escitalopram; non-responders received adjunctive aripiprazole (2-10 mg/d, flexible dosage). MDD participants were classified into subgroups according to their response profiles at weeks 8 and 16. Baseline group differences in cortical thickness were analyzed with FreeSurfer between HC and MDD groups as well as between response groups. Two-stage longitudinal processing was used to investigate 8-week escitalopram treatment-related changes in cortical thickness. Compared to HC, the MDD group exhibited thinner cortex in the left rostral middle frontal cortex [MNI(X,Y,Z=-29,9,54.5,-7.7); CWP=0.0002]. No baseline differences in cortical thickness were observed between responders and non-responders based on week-8 or week-16 response profile. No changes in cortical thickness was observed after 8 weeks of escitalopram monotherapy. In a two-step 16-week sequential clinical trial we found that baseline cortical thickness does not appear to be associated to clinical response to pharmacotherapy at 8 or 16 weeks.

YNIMG Journal 2019 Journal Article

Optimization of preprocessing strategies in Positron Emission Tomography (PET) neuroimaging: A [11C]DASB PET study

• Martin Nørgaard
• Melanie Ganz
• Claus Svarer
• Vibe G. Frokjaer
• Douglas N. Greve
• Stephen C. Strother
• Gitte M. Knudsen

Positron Emission Tomography (PET) is an important neuroimaging tool to quantify the distribution of specific molecules in the brain. The quantification is based on a series of individually designed data preprocessing steps (pipeline) and an optimal preprocessing strategy is per definition associated with less noise and improved statistical power, potentially allowing for more valid neurobiological interpretations. In spite of this, it is currently unclear how to design the best preprocessing pipeline and to what extent the choice of each preprocessing step in the pipeline minimizes subject-specific errors. To evaluate the impact of various preprocessing strategies, we systematically examined 384 different pipeline strategies in data from 30 healthy participants scanned twice with the serotonin transporter (5-HTT) radioligand [11C]DASB. Five commonly used preprocessing steps with two to four options were investigated: (1) motion correction (MC) (2) co-registration (3) delineation of volumes of interest (VOI's) (4) partial volume correction (PVC), and (5) kinetic modeling. To quantitatively compare and evaluate the impact of various preprocessing strategies, we used the performance metrics: test-retest bias, within- and between-subject variability, the intraclass-correlation coefficient, and global signal-to-noise ratio. We also performed a power analysis to estimate the required sample size to detect either a 5% or 10% difference in 5-HTT binding as a function of preprocessing pipeline. The results showed a complex downstream dependency between the various preprocessing steps on the performance metrics. The choice of MC had the most profound effect on 5-HTT binding, prior to the effects caused by PVC and kinetic modeling, and the effects differed across VOI's. Notably, we observed a negative bias in 5-HTT binding across test and retest in 98% of pipelines, ranging from 0 to 6% depending on the pipeline. Optimization of the performance metrics revealed a trade-off in within- and between-subject variability at the group-level with opposite effects (i. e. minimization of within-subject variability increased between-subject variability and vice versa). The sample size required to detect a given effect size was also compromised by the preprocessing strategy, resulting in up to 80% increases in sample size needed to detect a 5% difference in 5-HTT binding. This is the first study to systematically investigate and demonstrate the effect of choosing different preprocessing strategies on the outcome of dynamic PET studies. We provide a framework to show how optimal and maximally powered neuroimaging results can be obtained by choosing appropriate preprocessing strategies and we provide recommendations depending on the study design. In addition, the results contribute to a better understanding of methodological uncertainty and variability in preprocessing decisions for future group- and/or longitudinal PET studies.

YNIMG Journal 2019 Journal Article

Testing a deep convolutional neural network for automated hippocampus segmentation in a longitudinal sample of healthy participants

• Nikita Nogovitsyn
• Roberto Souza
• Meghan Muller
• Amelia Srajer
• Stefanie Hassel
• Stephen R. Arnott
• Andrew D. Davis
• Geoffrey B. Hall

Subtle changes in hippocampal volumes may occur during both physiological and pathophysiological processes in the human brain. Assessing hippocampal volumes manually is a time-consuming procedure, however, creating a need for automated segmentation methods that are both fast and reliable over time. Segmentation algorithms that employ deep convolutional neural networks (CNN) have emerged as a promising solution for large longitudinal neuroimaging studies. However, for these novel algorithms to be useful in clinical studies, the accuracy and reproducibility should be established on independent datasets. Here, we evaluate the performance of a CNN-based hippocampal segmentation algorithm that was developed by Thyreau and colleagues – Hippodeep. We compared its segmentation outputs to manual segmentation and FreeSurfer 6. 0 in a sample of 200 healthy participants scanned repeatedly at seven sites across Canada, as part of the Canadian Biomarker Integration Network in Depression consortium. The algorithm demonstrated high levels of stability and reproducibility of volumetric measures across all time points compared to the other two techniques. Although more rigorous testing in clinical populations is necessary, this approach holds promise as a viable option for tracking volumetric changes in longitudinal neuroimaging studies.

YNICL Journal 2019 Journal Article

The Canadian Dementia Imaging Protocol: Harmonization validity for morphometry measurements

• Olivier Potvin
• Isabelle Chouinard
• Louis Dieumegarde
• Robert Bartha
• Pierre Bellec
• D. Louis Collins
• Maxime Descoteaux
• Rick Hoge

The harmonized Canadian Dementia Imaging Protocol (CDIP) has been developed to suit the needs of a number of co-occurring Canadian studies collecting data on brain changes across adulthood and neurodegeneration. In this study, we verify the impact of CDIP parameters compliance on total brain volume variance using 86 scans of the same individual acquired on various scanners. Data included planned data collection acquired within the Consortium pour l'identification précoce de la maladie Alzheimer - Québec (CIMA-Q) and Canadian Consortium on Neurodegeneration in Aging (CCNA) studies, as well as opportunistic data collection from various protocols. For images acquired from Philips scanners, lower variance in brain volumes were observed when the stated CDIP resolution was set. For images acquired from GE scanners, lower variance in brain volumes were noticed when TE/TR values were within 5% of the CDIP protocol, compared to values farther from that criteria. Together, these results suggest that a harmonized protocol like the CDIP may help to reduce neuromorphometric measurement variability in multi-centric studies.

YNICL Journal 2018 Journal Article

FDG PET Parkinson’s disease-related pattern as a biomarker for clinical trials in early stage disease

• Dawn C. Matthews
• Hedva Lerman
• Ana Lukic
• Randolph D. Andrews
• Anat Mirelman
• Miles N. Wernick
• Nir Giladi
• Stephen C. Strother

Background: The development of therapeutic interventions for Parkinson disease (PD) is challenged by disease complexity and subjectivity of symptom evaluation. A Parkinson's Disease Related Pattern (PDRP) of glucose metabolism via fluorodeoxyglucose positron emission tomography (FDG-PET) has been reported to correlate with motor symptom scores and may aid the detection of disease-modifying therapeutic effects. Objectives: We sought to independently evaluate the potential utility of the PDRP as a biomarker for clinical trials of early-stage PD. Methods: Two machine learning approaches (Scaled Subprofile Model (SSM) and NPAIRS with Canonical Variates Analysis) were performed on FDG-PET scans from 17 healthy controls (HC) and 23 PD patients. The approaches were compared regarding discrimination of HC from PD and relationship to motor symptoms. Results: = 0.16, p < 0.06). Conclusions: Two independent analyses performed in a cohort of mild PD patients replicated key features of the PDRP, confirming that FDG-PET and multivariate classification can provide an objective, sensitive biomarker of disease stage with the potential to detect treatment effects on PD progression.

YNIMG Journal 2017 Journal Article

Optimizing fMRI preprocessing pipelines for block-design tasks as a function of age

• Nathan W. Churchill
• Pradeep Raamana
• Robyn Spring
• Stephen C. Strother

Functional Magnetic Resonance Imaging (fMRI) is a powerful neuroimaging tool, which is often hampered by significant noise confounds. There is evidence that our ability to detect activations in task fMRI is highly dependent on the preprocessing steps used to control noise and artifact. However, the vast majority of studies examining preprocessing pipelines in fMRI have focused on young adults. Given the widespread use of fMRI for characterizing the neurobiology of aging, it is critical to examine how the impact of preprocessing choices varies as a function of age. In this study, we employ the NPAIRS cross-validation framework, which optimizes pipelines based on metrics of prediction accuracy (P) and spatial reproducibility (R), to compare the effects of pipeline optimization between young (21–33 years) and older (61–82 years) cohorts, for three different block-design contrasts. Motion is shown to be a greater issue in the older cohort, and we introduce new statistical approaches to control for potential biases due to head motion during pipeline optimization. In comparison, data-driven methods of physiological noise correction show comparable benefits for both young and old cohorts. Using our optimization framework, we demonstrate that the optimal pipelines tend to be highly similar across age cohorts. In addition, there is a comparable, significant benefit of pipeline optimization across age cohorts, for (P, R) metrics and independent validation measures of activation overlap (both between-subject, within-session and within-subject, between-session). The choice of task contrast consistently shows a greater impact than the age cohort, for (P, R) metrics and activation overlap. Finally, adaptive pipeline optimization per task run shows improved sensitivity to age-related changes in brain activity, particularly for weaker, more complex cognitive contrasts. The current study provides the first detailed examination of preprocessing pipelines across age cohorts, demonstrating a significant benefit of adaptive pipeline optimization across age groups.

YNIMG Journal 2016 Journal Article

The association between cerebrovascular reactivity and resting-state fMRI functional connectivity in healthy adults: The influence of basal carbon dioxide

• Ali M. Golestani
• Jonathan B. Kwinta
• Stephen C. Strother
• Yasha B. Khatamian
• J. Jean Chen

Although widely used in resting-state fMRI (fMRI) functional connectivity measurement (fcMRI), the BOLD signal is only an indirect measure of neuronal activity, and is inherently modulated by both neuronal activity and vascular physiology. For instance, cerebrovascular reactivity (CVR) varies widely across individuals irrespective of neuronal function, but the implications for fcMRI are currently unknown. This knowledge gap compromises our ability to correctly interpret fcMRI measurements. In this work, we investigate the relationship between CVR and resting fcMRI measurements in healthy young adults, in both the motor and the executive-control networks. We modulate CVR within each individual by subtly increasing and decreasing resting vascular tension through baseline end-tidal CO2 (PETCO2), and measure fcMRI during these hypercapnic, hypocapnic and normocapnic states. Furthermore, we assess the association between CVR and fcMRI within and across individuals. Within individuals, resting PETCO2 is found to significantly influence both CVR and resting fcMRI values. In addition, we find resting fcMRI to be significantly and positively associated with CVR across the group in both networks. This relationship is potentially mediated by concomitant alterations in BOLD signal fluctuation amplitude. This work clearly demonstrates and quantifies a major vascular modulator of resting fcMRI, one that is also subject and regional dependent. We suggest that individualized correction for CVR effects in fcMRI measurements is essential for fcMRI studies of healthy brains, and can be even more important in studying diseased brains.

YNIMG Journal 2016 Journal Article

The Center for Integrated Molecular Brain Imaging (Cimbi) database

• Gitte M. Knudsen
• Peter S. Jensen
• David Erritzoe
• William F.C. Baaré
• Anders Ettrup
• Patrick M. Fisher
• Nic Gillings
• Hanne D. Hansen

We here describe a multimodality neuroimaging containing data from healthy volunteers and patients, acquired within the Lundbeck Foundation Center for Integrated Molecular Brain Imaging (Cimbi) in Copenhagen, Denmark. The data is of particular relevance for neurobiological research questions related to the serotonergic transmitter system with its normative data on the serotonergic subtype receptors 5-HT1A, 5-HT1B, 5-HT2A, and 5-HT4 and the 5-HT transporter (5-HTT), but can easily serve other purposes. The Cimbi database and Cimbi biobank were formally established in 2008 with the purpose to store the wealth of Cimbi-acquired data in a highly structured and standardized manner in accordance with the regulations issued by the Danish Data Protection Agency as well as to provide a quality-controlled resource for future hypothesis-generating and hypothesis-driven studies. The Cimbi database currently comprises a total of 1100 PET and 1000 structural and functional MRI scans and it holds a multitude of additional data, such as genetic and biochemical data, and scores from 17 self-reported questionnaires and from 11 neuropsychological paper/computer tests. The database associated Cimbi biobank currently contains blood and in some instances saliva samples from about 500 healthy volunteers and 300 patients with e. g. , major depression, dementia, substance abuse, obesity, and impulsive aggression. Data continue to be added to the Cimbi database and biobank.

YNIMG Journal 2014 Journal Article

Pattern classification of fMRI data: Applications for analysis of spatially distributed cortical networks

• Grigori Yourganov
• Tanya Schmah
• Nathan W. Churchill
• Marc G. Berman
• Cheryl L. Grady
• Stephen C. Strother

The field of fMRI data analysis is rapidly growing in sophistication, particularly in the domain of multivariate pattern classification. However, the interaction between the properties of the analytical model and the parameters of the BOLD signal (e. g. signal magnitude, temporal variance and functional connectivity) is still an open problem. We addressed this problem by evaluating a set of pattern classification algorithms on simulated and experimental block-design fMRI data. The set of classifiers consisted of linear and quadratic discriminants, linear support vector machine, and linear and nonlinear Gaussian naive Bayes classifiers. For linear discriminant, we used two methods of regularization: principal component analysis, and ridge regularization. The classifiers were used (1) to classify the volumes according to the behavioral task that was performed by the subject, and (2) to construct spatial maps that indicated the relative contribution of each voxel to classification. Our evaluation metrics were: (1) accuracy of out-of-sample classification and (2) reproducibility of spatial maps. In simulated data sets, we performed an additional evaluation of spatial maps with ROC analysis. We varied the magnitude, temporal variance and connectivity of simulated fMRI signal and identified the optimal classifier for each simulated environment. Overall, the best performers were linear and quadratic discriminants (operating on principal components of the data matrix) and, in some rare situations, a nonlinear Gaussian naïve Bayes classifier. The results from the simulated data were supported by within-subject analysis of experimental fMRI data, collected in a study of aging. This is the first study that systematically characterizes interactions between analysis model and signal parameters (such as magnitude, variance and correlation) on the performance of pattern classifiers for fMRI.

YNIMG Journal 2013 Journal Article

PHYCAA+: An optimized, adaptive procedure for measuring and controlling physiological noise in BOLD fMRI

• Nathan W. Churchill
• Stephen C. Strother

The presence of physiological noise in functional MRI can greatly limit the sensitivity and accuracy of BOLD signal measurements, and produce significant false positives. There are two main types of physiological confounds: (1) high-variance signal in non-neuronal tissues of the brain including vascular tracts, sinuses and ventricles, and (2) physiological noise components which extend into gray matter tissue. These physiological effects may also be partially coupled with stimuli (and thus the BOLD response). To address these issues, we have developed PHYCAA+, a significantly improved version of the PHYCAA algorithm (Churchill et al. , 2011) that (1) down-weights the variance of voxels in probable non-neuronal tissue, and (2) identifies the multivariate physiological noise subspace in gray matter that is linked to non-neuronal tissue. This model estimates physiological noise directly from EPI data, without requiring external measures of heartbeat and respiration, or manual selection of physiological components. The PHYCAA+ model significantly improves the prediction accuracy and reproducibility of single-subject analyses, compared to PHYCAA and a number of commonly-used physiological correction algorithms. Individual subject denoising with PHYCAA+ is independently validated by showing that it consistently increased between-subject activation overlap, and minimized false-positive signal in non gray-matter loci. The results are demonstrated for both block and fast single-event task designs, applied to standard univariate and adaptive multivariate analysis models.

YNIMG Journal 2012 Journal Article

Enhancing reproducibility of fMRI statistical maps using generalized canonical correlation analysis in NPAIRS framework

• Babak Afshin-Pour
• Gholam-Ali Hossein-Zadeh
• Stephen C. Strother
• Hamid Soltanian-Zadeh

Common fMRI data processing techniques usually minimize a temporal cost function or fit a temporal model to extract an activity map. Here, we focus on extracting a highly, spatially reproducible statistical parametric map (SPM) from fMRI data using a cost function that does not depend on a model of the subjects' temporal response. Based on a generalized version of canonical correlation analysis (gCCA), we propose a method to extract a highly reproducible map by maximizing the sum of pair-wise correlations between some maps. In a group analysis, each map is calculated from a linear combination of fMRI scans of a subset of subjects under study. The proposed method is applied to BOLD fMRI datasets without any spatial smoothing from 10 subjects performing a simple reaction time (RT) task. Using the NPAIRS split-half resampling framework with a reproducibility measure based on SPM correlations, we compare the proposed approach with canonical variate analysis (CVA) and a simple general linear model (GLM). gCCA provides statistical parametric maps with higher reproducibility than CVA and GLM with correlation reproducibilities across independent split-half SPMs of 0. 78, 0. 46, and 0. 41, respectively. Our results show that gCCA is an efficient approach for extracting the default mode network, assessing brain connectivity, and processing event-related and resting-state datasets in which the temporal BOLD signal varies from subject to subject.

YNIMG Journal 2012 Journal Article

PHYCAA: Data-driven measurement and removal of physiological noise in BOLD fMRI

• Nathan W. Churchill
• Grigori Yourganov
• Robyn Spring
• Peter M. Rasmussen
• Wayne Lee
• Jon E. Ween
• Stephen C. Strother

The effects of physiological noise may significantly limit the reproducibility and accuracy of BOLD fMRI. However, physiological noise evidences a complex, undersampled temporal structure and is often non-orthogonal relative to the neuronally-linked BOLD response, which presents a significant challenge for identifying and removing such artifact. This paper presents a multivariate, data-driven method for the characterization and removal of physiological noise in fMRI data, termed PHYCAA (PHYsiological correction using Canonical Autocorrelation Analysis). The method identifies high frequency, autocorrelated physiological noise sources with reproducible spatial structure, using an adaptation of Canonical Correlation Analysis performed in a split-half resampling framework. The technique is able to identify physiological effects with vascular-linked spatial structure, and an intrinsic dimensionality that is task- and subject-dependent. We also demonstrate that increasing dimensionality of such physiological noise is correlated with increasing variability in externally-measured respiratory and cardiac processes. Using PHYCAA as a denoising technique significantly improves simulated signal detection with physiological noise, and real data-driven model prediction and reproducibility, for both block and event-related task designs. This is demonstrated compared to no physiological noise correction, and to the widely used RETROICOR (Glover et al. , 2000) physiological denoising algorithm, which uses externally measured cardiac and respiration signals.

YNIMG Journal 2011 Journal Article

Data-driven optimization and evaluation of 2D EPI and 3D PRESTO for BOLD fMRI at 7 Tesla: I. Focal coverage

• Robert L. Barry
• Stephen C. Strother
• J. Christopher Gatenby
• John C. Gore

Blood oxygenation level dependent (BOLD) functional magnetic resonance imaging (fMRI) is commonly performed using 2D single-shot echo-planar imaging (EPI). However, single-shot EPI at 7 Tesla (T) often suffers from significant geometric distortions (due to low bandwidth (BW) in the phase-encode (PE) direction) and amplified physiological noise. Recent studies have suggested that 3D multi-shot sequences such as PRESTO may offer comparable BOLD contrast-to-noise ratio with increased volume coverage and decreased geometric distortions. Thus, a four-way group-level comparison was performed between 2D and 3D acquisition sequences at two in-plane resolutions. The quality of fMRI data was evaluated via metrics of prediction and reproducibility using NPAIRS (Non-parametric Prediction, Activation, Influence and Reproducibility re-Sampling). Group activation maps were optimized for each acquisition strategy by selecting the number of principal components that jointly maximized prediction and reproducibility, and showed good agreement in sensitivity and specificity for positive BOLD changes. High-resolution EPI exhibited the highest z-scores of the four acquisition sequences; however, it suffered from the lowest BW in the PE direction (resulting in the worst geometric distortions) and limited spatial coverage, and also caused some subject discomfort through peripheral nerve stimulation (PNS). In comparison, PRESTO also had high z-scores (higher than EPI for a matched in-plane resolution), the highest BW in the PE direction (producing images with superior geometric fidelity), the potential for whole-brain coverage, and no reported PNS. This study provides evidence to support the use of 3D multi-shot acquisition sequences in lieu of single-shot EPI for ultra high field BOLD fMRI at 7T.

YNIMG Journal 2011 Journal Article

Dimensionality estimation for optimal detection of functional networks in BOLD fMRI data

• Grigori Yourganov
• Xu Chen
• Ana S. Lukic
• Cheryl L. Grady
• Steven L. Small
• Miles N. Wernick
• Stephen C. Strother

Estimation of the intrinsic dimensionality of fMRI data is an important part of data analysis that helps to separate the signal of interest from noise. We have studied multiple methods of dimensionality estimation proposed in the literature and used these estimates to select a subset of principal components that was subsequently processed by linear discriminant analysis (LDA). Using simulated multivariate Gaussian data, we show that the dimensionality that optimizes signal detection (in terms of the receiver operating characteristic (ROC) metric) goes through a transition from many dimensions to a single dimension as a function of the signal-to-noise ratio. This transition happens when the loci of activation are organized into a spatial network and the variance of the networked, task-related signals is high enough for the signal to be easily detected in the data. We show that reproducibility of activation maps is a metric that captures this switch in intrinsic dimensionality. Except for reproducibility, all of the methods of dimensionality estimation we considered failed to capture this transition: optimization of Bayesian evidence, minimum description length, supervised and unsupervised LDA prediction, and Stein's unbiased risk estimator. This failure results in sub-optimal ROC performance of LDA in the presence of a spatially distributed network, and may have caused LDA to underperform in many of the reported comparisons in the literature. Using real fMRI data sets, including multi-subject group and within-subject longitudinal analysis we demonstrate the existence of these dimensionality transitions in real data.

YNIMG Journal 2008 Journal Article

Evaluation and comparison of GLM- and CVA-based fMRI processing pipelines with Java-based fMRI processing pipeline evaluation system

• Jing Zhang
• Lichen Liang
• Jon R. Anderson
• Lael Gatewood
• David A. Rottenberg
• Stephen C. Strother

Activation patterns identified by fMRI processing pipelines or fMRI software packages are usually determined by the preprocessing options, parameters, and statistical models used. Previous studies that evaluated options of GLM (general linear model)-based fMRI processing pipelines are mainly based on simulated data with receiver operating characteristics (ROC) analysis, but evaluation of such fMRI processing pipelines on real fMRI data is rare. To understand the effect of processing options on performance of GLM-based fMRI processing pipelines with real fMRI data, we investigated the impact of commonly-used fMRI preprocessing steps; optimized the associated GLM-based single-subject processing pipelines; and quantitatively compared univariate GLM (in FSL. FEAT and NPAIRS. GLM) and multivariate CVA (canonical variates analysis) (in NPAIRS. CVA)-based analytic models in single-subject analysis with a recently developed fMRI processing pipeline evaluation system based on prediction accuracy (classification accuracy) and reproducibility performance metrics. For block-design data, we found that with GLM analysis (1) slice timing correction and global intensity normalization have little consistent impact on fMRI processing pipelines, spatial smoothing and high-pass filtering or temporal detrending significantly increases pipeline performance and thus are essential for robust fMRI statistical analysis; (2) combined optimization of spatial smoothing and temporal detrending improves pipeline performance; and (3) in general, the prediction performance of multivariate CVA is higher than that of the univariate GLM, while univariate GLM is more reproducible than multivariate CVA. Because of the different bias–variance trade-offs of univariate and multivariate models, it may be necessary to consider a consensus approach to obtain more accurate activation patterns in fMRI data.

YNIMG Journal 2006 Journal Article

Mapping cerebral blood flow during speech production in hereditary ataxia

• John J. Sidtis
• Christopher Gomez
• Ansam Groshong
• Stephen C. Strother
• David A. Rottenberg

Dysarthria is a significant feature of the dominantly inherited spinocerebellar ataxias (SCA), but little is known about the patterns of brain activity associated with this disorder of motor speech control. Positron emission tomography (PET) was used to study regional cerebral blood flow during speech and rest in a group of 24 subjects with hereditary ataxia with mild-to-moderate dysarthria. These data were compared to the results obtained from a group of 13 age-matched, normal speakers. In the ataxic subjects, speech rates during scanning were significantly slowed compared to normal speakers. Significant reductions in mean regional blood flow were found in the cerebellum but not in supratentorial regions in the ataxic subjects. Multiple linear regression was used to model speech rate from regional blood flow. Four regions were identified as having significant relationships with speech rate in the model: the left inferior frontal and transverse temporal regions, and the right inferior cerebellar region and caudate nucleus. The relationship between flow and rate was positive in the inferior frontal and cerebellar regions and negative in the caudate and the transverse temporal region. The ataxic model represents an elaboration of the relationship previously reported for normal speakers, likely reflecting both the effects of, and compensation for, cerebellar degeneration in motor speech control. Although the mean regional blood flow values presented a pattern of functional organization for motor speech control at odds with lesion data, the performance-based model was in agreement with clinical experience. Incorporating performance data in functional image analysis may be more revealing of system characteristics than simply examining mean blood flow values.

YNIMG Journal 2005 Journal Article

Functional connectivity during Stroop task performance

• Ben J. Harrison
• Marnie Shaw
• Murat Yücel
• Rosemary Purcell
• Warrick J. Brewer
• Stephen C. Strother
• Gary F. Egan
• James S. Olver

Using covariance-based multivariate analysis, we examined patterns of functional connectivity in rCBF on a practice-extended version of the Stroop color-word paradigm. Color-word congruent and incongruent conditions were presented in six AB trials to healthy subjects during 12 H2 15O PET scans. Analyses identified two reproducible canonical eigenimages (CE) from the PET data, which were converted to a standard Z score scale after cross-validation resampling and correction for random subject effects. The first CE corresponded to practice-dependent changes in covarying rCBF that occurred over early task repetitions and correlated with improved behavioral performance. This included many regions previously implicated by PET and fMRI studies of this task, which we suggest may represent two “parallel” networks: (i) a cingulo-frontal system that was initially engaged in selecting and mapping a task-relevant response (color naming) when the attentional demands of the task were greatest; and (ii) a ventral visual processing stream whose concurrent decrease in activity represented the task-irrelevant inhibition of word reading. The second CE corresponded to a consistent paradigmatic effect of Stroop interference on covarying rCBF. Coactivations were located in dorsal and ventral prefrontal regions as well as frontopolar cortex. This pattern supports existing evidence that prefrontal regions are involved in maintaining attentional control over conflicting response systems. Taken together, these findings may be more in line with theoretical models that emphasize a role for practice in the emergence of Stroop phenomena. These findings may also provide some additional insight into the nature of anterior cingulate- and prefrontal cortical contributions to implementing cognitive control in the brain.

YNIMG Journal 2003 Journal Article

Evaluating subject specific preprocessing choices in multisubject fMRI data sets using data-driven performance metrics

• Marnie E. Shaw
• Stephen C. Strother
• Maria Gavrilescu
• Katherine Podzebenko
• Anthony Waites
• John Watson
• Jon Anderson
• Graeme Jackson

This study investigated the possible benefit of subject specific optimization of preprocessing strategies in functional magnetic resonance imaging (fMRI) experiments. The optimization was performed using the data-driven performance metrics developed recently [Neuroimage 15 (2002), 747]. We applied numerous preprocessing strategies and a multivariate statistical analysis to each of the 20 subjects in our two example fMRI data sets. We found that the optimal preprocessing strategy varied, in general, from subject to subject. For example, in one data set, optimum smoothing levels varied from 16 mm (4 subjects), 10 mm (5 subjects), to no smoothing at all (1 subject). This strongly suggests that group-specific preprocessing schemes may not give optimum results. For both studies, optimizing the preprocessing for each subject resulted in an increased number of suprathresholded voxels in within-subject analyses. Furthermore, we demonstrated that we were able to aggregate the optimized data with a random effects group analysis, resulting in improved sensitivity in one study and the detection of interesting, previously undetected results in the other.

YNIMG Journal 2002 Journal Article

Abnormal Functional Connectivity in Posttraumatic Stress Disorder

• Marnie E. Shaw
• Stephen C. Strother
• Alexander C. McFarlane
• Philip Morris
• Jon Anderson
• C.Richard Clark
• Gary F. Egan

This study investigated the efficacy of a combined multivariate/resampling procedure for the analysis of PET activation studies. The covariance-based multivariate analysis was used to investigate distributed brain systems in posttraumatic stress disorder (PTSD) patients and matched controls during performance of a working memory task. The results were compared to univariate results obtained in an earlier study. We also examined whether the PTSD patients demonstrated a breakdown in functional connectivity that may be associated with working memory difficulties often experienced by these patients. A resampling procedure was used specifically to test the reliability of measured between-group effects, to avoid mistaken inference on the basis of random intersubject differences. Significant and reproducible differences in network connectivity were obtained for the two groups. The functional connectivity pattern of the patient group was characterized by relatively more activation in the bilateral inferior parietal lobes and the left precentral gyrus than the control group, and less activation in the inferior medial frontal lobe, bilateral middle frontal gyri and right inferior temporal gyrus. The resampling procedure provided direct evidence that working memory updating was abnormal in PTSD patients relative to matched controls. This work focuses on the need to identify extended brain networks (in addition to regionally specific changes) for the full characterization of brain responses in neuroimaging experiments. Our multivariate analysis explicitly measures the reliability of the patterns of functional connectivity we obtain and demonstrates the potential of such analyses for the study of brain network dysfunction in psychopathology.

AIIM Journal 2002 Journal Article

An evaluation of methods for detecting brain activations from functional neuroimages

• Ana S. Lukic
• Miles N. Wernick
• Stephen C. Strother

Brain activation studies based on PET or fMRI seek to explore neuroscience questions by statistically analyzing the acquired images to produce statistical parametric images (SPIs). An increasingly wide range of univariate and multivariate analysis techniques are used to generate SPIs in order to detect mean-signal activations and/or long-range spatial interactions. However, little is known about the comparative detection performance of even simple techniques in finite data sets. Our aims are (1) to empirically compare the detection performance of a range of techniques using simulations of a simple image phantom and receiver operating characteristics (ROC) analysis, and (2) to construct two near-optimal detectors, both generalized likelihood ratio tests as upper performance bounds. We found that for finite samples of (10–100) images, even when the t-test with single-voxel variance estimates (single-voxel t-test) is the “correct” (i. e. unbiased) model for simple local additive signals, better detection performance is obtained using pooled variance estimates or adaptive, multivariate covariance-based detectors. Normalization by voxel-based variance estimates causes significantly decreased detection performance using either single-voxel t-tests or correlation-coefficient thresholding compared to pooled-variance t-tests or covariance thresholding, respectively. Moreover, we found that SVD by itself, or followed by an adaptive Fisher linear discriminant, provides a detector that is (1) more sensitive to mean differences than a single-voxel t-test, (2) insensitive to the large local signal variances detected by covariance thresholding, and (3) much more sensitive to signal correlations than correlation-coefficient thresholding. Adaptive, multivariate covariance-based approaches and pooled-variance t-tests represent promising directions for obtaining optimal signal detection in functional neuroimaging studies.

YNIMG Journal 2002 Journal Article

The Quantitative Evaluation of Functional Neuroimaging Experiments: The NPAIRS Data Analysis Framework

• Stephen C. Strother
• Jon Anderson
• Lars Kai Hansen
• Ulrik Kjems
• Rafal Kustra
• John Sidtis
• Sally Frutiger
• Suraj Muley

We introduce a data-analysis framework and performance metrics for evaluating and optimizing the interaction between activation tasks, experimental designs, and the methodological choices and tools for data acquisition, preprocessing, data analysis, and extraction of statistical parametric maps (SPMs). Our NPAIRS (nonparametric prediction, activation, influence, and reproducibility resampling) framework provides an alternative to simulations and ROC curves by using real PET and fMRI data sets to examine the relationship between prediction accuracy and the signal-to-noise ratios (SNRs) associated with reproducible SPMs. Using cross-validation resampling we plot training–test set predictions of the experimental design variables (e. g. , brain-state labels) versus reproducibility SNR metrics for the associated SPMs. We demonstrate the utility of this framework across the wide range of performance metrics obtained from [15O]water PET studies of 12 age- and sex-matched data sets performing different motor tasks (8 subjects/set). For the 12 data sets we apply NPAIRS with both univariate and multivariate data-analysis approaches to: (1) demonstrate that this framework may be used to obtain reproducible SPMs from any data-analysis approach on a common Z-score scale (rSPM{Z}); (2) demonstrate that the histogram of a rSPM{Z} image may be modeled as the sum of a data-analysis-dependent noise distribution and a task-dependent, Gaussian signal distribution that scales monotonically with our reproducibility performance metric; (3) explore the relation between prediction and reproducibility performance metrics with an emphasis on bias-variance tradeoffs for flexible, multivariate models; and (4) measure the broad range of reproducibility SNRs and the significant influence of individual subjects. A companion paper describes learning curves for four of these 12 data sets, which describe an alternative mutual-information prediction metric and NPAIRS reproducibility as a function of training-set sizes from 2 to 18 subjects. We propose the NPAIRS framework as a validation tool for testing and optimizing methodological choices and tools in functional neuroimaging.

YNIMG Journal 2001 Journal Article

Consensus Inference in Neuroimaging

• Lars Kai Hansen
• Finn Årup Nielsen
• Stephen C. Strother
• Nicholas Lange

We introduce model averaging in neuroimaging. We show that model summary images can be directly compared and averaged after histogram equalization. We demonstrate that averaging enhances the ROC curve in a simulation study. The averaging procedure is applied to a fMRI study of motor cortex.

YNIMG Journal 2001 Journal Article

Effects of Changes in Experimental Design on PET Studies of Isometric Force

• Suraj A. Muley
• Stephen C. Strother
• James Ashe
• Sally A. Frutiger
• John R. Anderson
• John J. Sidtis
• David A. Rottenberg

Based on single-cell recordings in primates, the relationship between neuronal activity and force magnitude is thought to be monotonic, at least for a subset of pyramidal cells in the motor cortex. Functional neuroimaging studies have also suggested a monotonic relationship between cerebral activation and force magnitude. In order to more precisely define this relationship and to characterize the activation pattern(s) associated with the modulation of static force, we studied 40 normal subjects using [15O]water PET and a simple visuomotor task—application of static force on a micro force sensor with the thumb and index finger of the right hand. When our experimental design did not produce the expected result (evidence of a relationship between cerebral activation and force magnitude in ten subjects), we made serial changes in the experimental protocol, including the addition of control (baseline) trials, and increased the number of subjects in an effort to increase our sensitivity to variations in force magnitude. We compared univariate and multivariate data-analytic strategies, but we relied on our multivariate results to elucidate the interaction of attentional and motor networks. We found that increasing the number of subjects from 10 to 20 resulted in an increase in statistical power and a more stable (i. e. , more replicable) but qualitatively similar result, and that the inclusion of control trials in a 10-subject group did not enhance our ability to discern significant brain-behavior relationships. Our results suggest that sample sizes greater than 20 may be required to detect parametric variation in some instances and that failure to detect such variation may result from unanticipated neurobehavioral effects.

YNIMG Journal 2000 Journal Article

Learning curves for evaluation of neuroimaging experiments

• Ulrik Kjems
• Lars Kai Hansen
• Suraj Muley
• Jon Anderson
• David Rottenburg
• Sally Frutiger
• Stephen C. Strother

YNIMG Journal 2000 Journal Article

Multivariate Predictive Relationship between Kinematic and Functional Activation Patterns in a PET Study of Visuomotor Learning

• Sally A. Frutiger
• Stephen C. Strother
• Jon R. Anderson
• John J. Sidtis
• James B. Arnold
• David A. Rottenberg

Imaging studies of visuomotor learning have reported practice-related activation in brain regions mediating sensorimotor functions. However, development and testing of functional motor learning models, based on the relationship between imaging and behavioral measures, is complicated by the multidimensional nature of motoric control. In the present study, multivariate techniques were used to analyze [15O]water PET and kinematic correlates of learning in a visuomotor tracing task. Fourteen subjects traced a geometric form over a series of eight tracing trials, preceded and followed by baseline trials in which they passively viewed the geometric form. Simultaneous evaluation of multiple behavioral measures indicated that performance improvement was most strongly associated with a global performance measure and least strongly associated with measures of fine motor control. Results of three independent analytic techniques (i. e. , intertrial correlation matrices, power function modeling, iterative canonical variate analysis) indicated that imaging and behavioral measures were most closely related on early learning trials. Performance improvement was associated with covarying increases in normalized activity among superior parietal, postcentral gyrus, and premotor regions and covarying decreases in normalized activity among cerebellar, inferior parietal, pallidal, and medial occipital regions. These findings suggest that performance improvement may be associated with increased activation in neural systems previously implicated in visually guided reaching and decreased activation in neural systems previously implicated in attentive visuospatial processing.

YNIMG Journal 1999 Journal Article

Are Brain Functions Really Additive?

• John J. Sidtis
• Stephen C. Strother
• Jon R. Anderson
• David A. Rottenberg

Although Positron Emission Tomography (PET) and functional magnetic resonance imaging (fMRI) studies commonly subtract data obtained during two or more experimental conditions to decompose a complex task, there have been few opportunities to evaluate this approach directly. In the present study, PET was used to study three motor speech tasks selected such that two were constituent components of the third, making possible a direct examination of decomposition by subtraction. In Experiment 1, a group of 13 right-handed normal volunteers participated in three activation studies: syllable repetition; phonation; and repetitive lip closure. A scanning session was devoted to a single task, repeated four times. In Experiment 2, six of the original subjects performed the same three activation studies during a single scanning session. Whether tasks were studied in separate scanning sessions or combined within a single session, the results of decomposition by compound subtraction differed significantly from the results obtained when individual tasks were compared to a simple baseline condition. These data failed to demonstrate task additivity, a necessary property if decomposition by subtraction is to provide an accurate characterization of the brain activity accompanying complex behavior.

YNIMG Journal 1999 Journal Article

Generalizable Patterns in Neuroimaging: How Many Principal Components?

• Lars Kai Hansen
• Jan Larsen
• Finn Årup Nielsen
• Stephen C. Strother
• Egill Rostrup
• Robert Savoy
• Nicholas Lange
• John Sidtis

Generalization can be defined quantitatively and can be used to assess the performance of principal component analysis (PCA). The generalizability of PCA depends on the number of principal components retained in the analysis. We provide analytic and test set estimates of generalization. We show how the generalization error can be used to select the number of principal components in two analyses of functional magnetic resonance imaging activation sets.

YNIMG Journal 1999 Journal Article

Plurality and Resemblance in fMRI Data Analysis

• Nicholas Lange
• Stephen C. Strother
• Jon R. Anderson
• Finn Å. Nielsen
• Andrew P. Holmes
• Thomas Kolenda
• Robert Savoy
• Lars Kai Hansen

We apply nine analytic methods employed currently in imaging neuroscience to simulated and actual BOLD fMRI signals and compare their performances under each signal type. Starting with baseline time series generated by a resting subject during a null hypothesis study, we compare method performance with embedded focal activity in these series of three different types whose magnitudes and time courses are simple, convolved with spatially varying hemodynamic responses, and highly spatially interactive. We then apply these same nine methods to BOLD fMRI time series from contralateral primary motor cortex and ipsilateral cerebellum collected during a sequential finger opposition study. Paired comparisons of results across methods include a voxel-specific concordance correlation coefficient for reproducibility and a resemblance measure that accommodates spatial autocorrelation of differences in activity surfaces. Receiver-operating characteristic curves show considerable model differences in ranges less than 10% significance level (false positives) and greater than 80% power (true positives). Concordance and resemblance measures reveal significant differences between activity surfaces in both data sets. These measures can assist researchers by identifying groups of models producing similar and dissimilar results, and thereby help to validate, consolidate, and simplify reports of statistical findings. A pluralistic strategy for fMRI data analysis can uncover invariant and highly interactive relationships between local activity foci and serve as a basis for further discovery of organizational principles of the brain. Results also suggest that a pluralistic empirical strategy coupled formally with substantive prior knowledge can help to uncover new brain–behavior relationships that may remain hidden if only a single method is employed.