Diffusion tensor brain imaging at 0.55T: A feasibility study.

PURPOSE: To investigate the feasibility of diffusion tensor brain imaging at 0.55T with comparisons against 3T. METHODS: Diffusion tensor imaging data with 2 mm isotropic resolution was acquired on a cohort of five healthy subjects using both 0.55T and 3T scanners. The signal-to-noise ratio (SNR) of the 0.55T data was improved using a previous SNR-enhancing joint reconstruction method that jointly reconstructs the entire set of diffusion weighted images from k-space using shared-edge constraints. Quantitative diffusion tensor parameters were estimated and compared across field strengths. We also performed a test-retest assessment of repeatability at each field strength. RESULTS: After applying SNR-enhancing joint reconstruction, the diffusion tensor parameters obtained from 0.55T data were strongly correlated ( R 2 ≥ 0 . 70 $$ {R}^2\ge 0.70 $$ ) with those obtained from 3T data. Test-retest analysis showed that SNR-enhancing reconstruction improved the repeatability of the 0.55T diffusion tensor parameters. CONCLUSION: High-resolution in vivo diffusion MRI of the human brain is feasible at 0.55T when appropriate noise-mitigation strategies are applied.

BundleCleaner: Unsupervised Denoising and Subsampling of Diffusion MRI-Derived Tractography Data.

We present BundleCleaner, an unsupervised multi-step framework that can filter, denoise and subsample bundles derived from diffusion MRI-based whole-brain tractography. Our approach considers both the global bundle structure and local streamline-wise features. We apply BundleCleaner to bundles generated from single-shell diffusion MRI data in an independent clinical sample of older adults from India using probabilistic tractography and the resulting 'cleaned' bundles can better align with the atlas bundles with reduced overreach. In a downstream tractometry analysis, we show that the cleaned bundles, represented with less than 20% of the original set of points, can robustly localize along-tract microstructural differences between 32 healthy controls and 34 participants with Alzheimer's disease ranging in age from 55 to 84 years old. Our approach can help reduce memory burden and improving computational efficiency when working with tractography data, and shows promise for large-scale multi-site tractometry.

BrainSuite BIDS App: Containerized Workflows for MRI Analysis.

There has been a concerted effort by the neuroimaging community to establish standards for computational methods for data analysis that promote reproducibility and portability. In particular, the Brain Imaging Data Structure (BIDS) specifies a standard for storing imaging data, and the related BIDS App methodology provides a standard for implementing containerized processing environments that include all necessary dependencies to process BIDS datasets using image processing workflows. We present the BrainSuite BIDS App, which encapsulates the core MRI processing functionality of BrainSuite within the BIDS App framework. Specifically, the BrainSuite BIDS App implements a participant-level workflow comprising three pipelines and a corresponding set of group-level analysis workflows for processing the participant-level outputs. The BrainSuite Anatomical Pipeline (BAP) extracts cortical surface models from a T1-weighted (T1w) MRI. It then performs surface-constrained volumetric registration to align the T1w MRI to a labeled anatomical atlas, which is used to delineate anatomical regions of interest in the MRI brain volume and on the cortical surface models. The BrainSuite Diffusion Pipeline (BDP) processes diffusion-weighted imaging (DWI) data, with steps that include coregistering the DWI data to the T1w scan, correcting for geometric image distortion, and fitting diffusion models to the DWI data. The BrainSuite Functional Pipeline (BFP) performs fMRI processing using a combination of FSL, AFNI, and BrainSuite tools. BFP coregisters the fMRI data to the T1w image, then transforms the data to the anatomical atlas space and to the Human Connectome Project's grayordinate space. Each of these outputs can then be processed during group-level analysis. The outputs of BAP and BDP are analyzed using the BrainSuite Statistics in R (bssr) toolbox, which provides functionality for hypothesis testing and statistical modeling. The outputs of BFP can be analyzed using atlas-based or atlas-free statistical methods during group-level processing. These analyses include the application of BrainSync, which synchronizes the time-series data temporally and enables comparison of resting-state or task-based fMRI data across scans. We also present the BrainSuite Dashboard quality control system, which provides a browser-based interface for reviewing the outputs of individual modules of the participant-level pipelines across a study in real-time as they are generated. BrainSuite Dashboard facilitates rapid review of intermediate results, enabling users to identify processing errors and make adjustments to processing parameters if necessary. The comprehensive functionality included in the BrainSuite BIDS App provides a mechanism for rapidly deploying the BrainSuite workflows into new environments to perform large-scale studies. We demonstrate the capabilities of the BrainSuite BIDS App using structural, diffusion, and functional MRI data from the Amsterdam Open MRI Collection's Population Imaging of Psychology dataset.

Meta Transfer of Self-Supervised Knowledge: Foundation Model in Action for Post-Traumatic Epilepsy Prediction.

Despite the impressive advancements achieved using deep-learning for functional brain activity analysis, the heterogeneity of functional patterns and scarcity of imaging data still pose challenges in tasks such as prediction of future onset of Post-Traumatic Epilepsy (PTE) from data acquired shortly after traumatic brain injury (TBI). Foundation models pre-trained on separate large-scale datasets can improve the performance from scarce and heterogeneous datasets. For functional Magnetic Resonance Imaging (fMRI), while data may be abundantly available from healthy controls, clinical data is often scarce, limiting the ability of foundation models to identify clinically-relevant features. We overcome this limitation by introducing a novel training strategy for our foundation model by integrating meta-learning with self-supervised learning to improve the generalization from normal to clinical features. In this way we enable generalization to other downstream clinical tasks, in our case prediction of PTE. To achieve this, we perform self-supervised training on the control dataset to focus on inherent features that are not limited to a particular supervised task while applying meta-learning, which strongly improves the model's generalizability using bi-level optimization. Through experiments on neurological disorder classification tasks, we demonstrate that the proposed strategy significantly improves task performance on small-scale clinical datasets. To explore the generalizability of the foundation model in downstream applications, we then apply the model to an unseen TBI dataset for prediction of PTE using zero-shot learning. Results further demonstrated the enhanced generalizability of our foundation model.

Chronic anemia: The effects on the connectivity of white matter.

Chronic anemia is commonly observed in patients with hemoglobinopathies, mainly represented by disorders of altered hemoglobin (Hb) structure (sickle cell disease, SCD) and impaired Hb synthesis (e.g. thalassemia syndromes, non-SCD anemia). Both hemoglobinopathies have been associated with white matter (WM) alterations. Novel structural MRI research in our laboratory demonstrated that WM volume was diffusely lower in deep, watershed areas proportional to anemia severity. Furthermore, diffusion tensor imaging analysis has provided evidence that WM microstructure is disrupted proportionally to Hb level and oxygen saturation. SCD patients have been widely studied and demonstrate lower fractional anisotropy (FA) in the corticospinal tract and cerebellum across the internal capsule and corpus callosum. In the present study, we compared 19 SCD and 15 non-SCD anemia patients with a wide range of Hb values allowing the characterization of the effects of chronic anemia in isolation of sickle Hb. We performed a tensor analysis to quantify FA changes in WM connectivity in chronic anemic patients. We calculated the volumetric mean of FA along the pathway of tracks connecting two regions of interest defined by BrainSuite's BCI-DNI atlas. In general, we found lower FA values in anemic patients; indicating the loss of coherence in the main diffusion direction that potentially indicates WM injury. We saw a positive correlation between FA and hemoglobin in these same regions, suggesting that decreased WM microstructural integrity FA is highly driven by chronic hypoxia. The only connection that did not follow this pattern was the connectivity within the left middle-inferior temporal gyrus. Interestingly, more reductions in FA were observed in non-SCD patients (mainly along with intrahemispheric WM bundles and watershed areas) than the SCD patients (mainly interhemispheric).

A hybrid high-resolution anatomical MRI atlas with sub-parcellation of cortical gyri using resting fMRI.

We present a new high-quality, single-subject atlas with sub-millimeter voxel resolution, high SNR, and excellent gray-white tissue contrast to resolve fine anatomical details. The atlas is labeled into two parcellation schemes: 1) the anatomical BCI-DNI atlas, which is manually labeled based on known morphological and anatomical features, and 2) the hybrid USCBrain atlas, which incorporates functional information to guide the sub-parcellation of cerebral cortex. In both cases, we provide consistent volumetric and cortical surface-based parcellation and labeling. The intended use of the atlas is as a reference template for structural coregistration and labeling of individual brains. A single-subject T1-weighted image was acquired five times at a resolution of 0.547 mm × 0.547 mm × 0.800 mm and averaged. Images were processed by an expert neuroanatomist using semi-automated methods in BrainSuite to extract the brain, classify tissue-types, and render anatomical surfaces. Sixty-six cortical and 29 noncortical regions were manually labeled to generate the BCI-DNI atlas. The cortical regions were further sub-parcellated into 130 cortical regions based on multi-subject connectivity analysis using resting fMRI (rfMRI) data from the Human Connectome Project (HCP) database to produce the USCBrain atlas. In addition, we provide a delineation between sulcal valleys and gyral crowns, which offer an additional set of 26 sulcal subregions per hemisphere. Lastly, a probabilistic map is provided to give users a quantitative measure of reliability for each gyral subdivision. Utility of the atlas was assessed by computing Adjusted Rand Indices (ARIs) between individual sub-parcellations obtained through structural-only coregistration to the USCBrain atlas and sub-parcellations obtained directly from each subject's resting fMRI data. Both atlas parcellations can be used with the BrainSuite, FreeSurfer, and FSL software packages.

Deep Quantile Regression for Uncertainty Estimation in Unsupervised and Supervised Lesion Detection.

Despite impressive state-of-the-art performance on a wide variety of machine learning tasks in multiple applications, deep learning methods can produce over-confident predictions, particularly with limited training data. Therefore, quantifying uncertainty is particularly important in critical applications such as lesion detection and clinical diagnosis, where a realistic assessment of uncertainty is essential in determining surgical margins, disease status and appropriate treatment. In this work, we propose a novel approach that uses quantile regression for quantifying aleatoric uncertainty in both supervised and unsupervised lesion detection problems. The resulting confidence intervals can be used for lesion detection and segmentation. In the unsupervised setting, we combine quantile regression with the Variational AutoEncoder (VAE). The VAE is trained on lesion-free data, so when presented with an image with a lesion, it tends to reconstruct a lesion-free version of the image. To detect the lesion, we then compare the input (lesion) and output (lesion-free) images. Here we address the problem of quantifying uncertainty in the images that are reconstructed by the VAE as the basis for principled outlier or lesion detection. The VAE models the output as a conditionally independent Gaussian characterized by its mean and variance. Unfortunately, joint optimization of both mean and variance in the VAE leads to the well-known problem of shrinkage or underestimation of variance. Here we describe an alternative Quantile-Regression VAE (QR-VAE) that avoids this variance shrinkage problem by directly estimating conditional quantiles for the input image. Using the estimated quantiles, we compute the conditional mean and variance for the input image from which we then detect outliers by thresholding at a false-discovery-rate corrected p-value. In the supervised setting, we develop binary quantile regression (BQR) for the supervised lesion segmentation task. We show how BQR can be used to capture uncertainty in lesion boundaries in a manner that characterizes expert disagreement.

A robust variational autoencoder using beta divergence.

The presence of outliers can severely degrade learned representations and performance of deep learning methods and hence disproportionately affect the training process, leading to incorrect conclusions about the data. For example, anomaly detection using deep generative models is typically only possible when similar anomalies (or outliers) are not present in the training data. Here we focus on variational autoencoders (VAEs). While the VAE is a popular framework for anomaly detection tasks, we observe that the VAE is unable to detect outliers when the training data contains anomalies that have the same distribution as those in test data. In this paper we focus on robustness to outliers in training data in VAE settings using concepts from robust statistics. We propose a variational lower bound that leads to a robust VAE model that has the same computational complexity as the standard VAE and contains a single automatically-adjusted tuning parameter to control the degree of robustness. We present mathematical formulations for robust variational autoencoders (RVAEs) for Bernoulli, Gaussian and categorical variables. The RVAE model is based on beta-divergence rather than the standard Kullback-Leibler (KL) divergence. We demonstrate the performance of our proposed ß-divergence-based autoencoder for a variety of image and categorical datasets showing improved robustness to outliers both qualitatively and quantitatively. We also illustrate the use of our robust VAE for detection of lesions in brain images, formulated as an anomaly detection task. Finally, we suggest a method to tune the hyperparameter of RVAE which makes our model completely unsupervised.

Quantification of kappa opioid receptor ligand potency, efficacy and desensitization using a real-time membrane potential assay.

We explored the utility of the real-time FLIPR Membrane Potential (FMP) assay as a method to assess kappa opioid receptor (KOR)-induced hyperpolarization. The FMP Blue dye was used to measure fluorescent signals reflecting changes in membrane potential in KOR expressing CHO (CHO-KOR) cells. Treatment of CHO-KOR cells with kappa agonists U50,488 or dynorphin [Dyn (1-13)NH2] produced rapid and concentration-dependent decreases in FMP Blue fluorescence reflecting membrane hyperpolarization. Both the nonselective opioid antagonist naloxone and the κ-selective antagonists nor-binaltorphimine (nor-BNI) and zyklophin produced rightward shifts in the U50,488 concentration-response curves, consistent with competitive antagonism of the KOR mediated response. The decrease in fluorescent emission produced by U50,488 was blocked by overnight pertussis toxin pretreatment, indicating the requirement for PTX-sensitive G proteins in the KOR mediated response. We directly compared the potency of U50,488 and Dyn (1-13)NH2 in the FMP and [35S]GTPγS binding assays, and found that both were approximately 10 times more potent in the cellular fluorescence assay. The maximum responses of both U50,488 and Dyn (1-13)NH2 declined following repeated additions, reflecting receptor desensitization. We assessed the efficacy and potency of structurally distinct KOR small molecule and peptide ligands. The FMP assay reliably detected both partial agonists and stereoselectivity. Using KOR-selective peptides with varying efficacies, we found that the FMP assay allowed high throughput quantification of peptide efficacy. These data demonstrate that the FMP assay is a sensitive method for assessing κ-opioid receptor induced hyperpolarization, and represents a useful approach for quantification of potency, efficacy and desensitization of KOR ligands.

Validation of semi-automated anatomically labeled SEEG contacts in a brain atlas for mapping connectivity in focal epilepsy.

OBJECTIVE: Stereotactic electroencephalography (SEEG) has been widely used to explore the epileptic network and localize the epileptic zone in patients with medically intractable epilepsy. Accurate anatomical labeling of SEEG electrode contacts is critically important for correctly interpreting epileptic activity. We present a method for automatically assigning anatomical labels to SEEG electrode contacts using a 3D-segmented cortex and coregistered postoperative CT images. METHOD: Stereotactic electroencephalography electrode contacts were spatially localized relative to the brain volume using a standard clinical procedure. Each contact was then assigned an anatomical label by clinical epilepsy fellows. Separately, each contact was automatically labeled by coregistering the subject's MRI to the USCBrain atlas using the BrainSuite software and assigning labels from the atlas based on contact locations. The results of both labeling methods were then compared, and a subsequent vetting of the anatomical labels was performed by expert review. RESULTS: Anatomical labeling agreement between the two methods for over 17 000 SEEG contacts was 82%. This agreement was consistent in patients with and without previous surgery (P = .852). Expert review of contacts in disagreement between the two methods resulted in agreement with the atlas based over manual labels in 48% of cases, agreement with manual over atlas-based labels in 36% of cases, and disagreement with both methods in 16% of cases. Labels deemed incorrect by the expert review were then categorized as either in a region directly adjacent to the correct label or as a gross error, revealing a lower likelihood of gross error from the automated method. SIGNIFICANCE: The method for semi-automated atlas-based anatomical labeling we describe here demonstrates potential to assist clinical workflow by reducing both analysis time and the likelihood of gross anatomical error. Additionally, it provides a convenient means of intersubject analysis by standardizing the anatomical labels applied to SEEG contact locations across subjects.

Robust brain network identification from multi-subject asynchronous fMRI data.

We describe a novel method for robust identification of common brain networks and their corresponding temporal dynamics across subjects from asynchronous functional MRI (fMRI) using tensor decomposition. We first temporally align asynchronous fMRI data using the orthogonal BrainSync transform, allowing us to study common brain networks across sessions and subjects. We then map the synchronized fMRI data into a 3D tensor (vertices × time × subject/session). Finally, we apply Nesterov-accelerated adaptive moment estimation (Nadam) within a scalable and robust sequential Canonical Polyadic (CP) decomposition framework to identify a low rank tensor approximation to the data. As a result of CP tensor decomposition, we successfully identified twelve known brain networks with their corresponding temporal dynamics from 40 subjects using the Human Connectome Project's language task fMRI data without any prior information regarding the specific task designs. Seven of these networks show distinct subjects' responses to the language task with differing temporal dynamics; two show sub-components of the default mode network that exhibit deactivation during the tasks; the remaining three components reflect non-task-related activities. We compare results to those found using group independent component analysis (ICA) and canonical ICA. Bootstrap analysis demonstrates increased robustness of networks found using the CP tensor approach relative to ICA-based methods.

The FAST graph: A novel framework for the anatomically-guided visualization and analysis of cortico-cortical evoked potentials.

BACKGROUND: Intracerebral electroencephalography (iEEG) using stereoelectroencephalography (SEEG) methodology for epilepsy surgery gives rise to complex data sets. The neurophysiological data obtained during the in-patient period includes categorization of the evoked potentials resulting from direct electrical cortical stimulation such as cortico-cortical evoked potentials (CCEPs). These potentials are recorded by hundreds of contacts, making these waveforms difficult to quickly interpret over such high-density arrays that are organized in three dimensional fashion. NEW METHOD: The challenge in analyzing CCEPs data arises not just from the density of the array, but also from the stimulation of a number of different intracerebral sites. A systematic methodology for visualization and analysis of these evoked data is lacking. We describe the process of incorporating anatomical information into the visualizations, which are then compared to more traditional plotting techniques to highlight the usefulness of the new framework. RESULTS: We describe here an innovative framework for sorting, registering, labeling, ordering, and quantifying the functional CCEPs data, using the anatomical labelling of the brain, to provide an informative visualization and summary statistics which we call the "FAST graph" (Functional-Anatomical STacked area graphs). The FAST graph analysis is used to depict the significant CCEPs responses in patient with focal epilepsy. CONCLUSIONS: The novel plotting approach shown here allows us to visualize high-density stimulation data in a single summary plot for subsequent detailed analyses. Improving the visual presentation of complex data sets aides in enhancing the clinical utility of the data.

Temporal non-local means filtering for studies of intrinsic brain connectivity from individual resting fMRI.

Characterizing functional brain connectivity using resting functional magnetic resonance imaging (fMRI) is challenging due to the relatively small Blood-Oxygen-Level Dependent contrast and low signal-to-noise ratio. Denoising using surface-based Laplace-Beltrami (LB) or volumetric Gaussian filtering tends to blur boundaries between different functional areas. To overcome this issue, a time-based Non-Local Means (tNLM) filtering method was previously developed to denoise fMRI data while preserving spatial structure. The kernel and parameters that define the tNLM filter need to be optimized for each application. Here we present a novel Global PDF-based tNLM filtering (GPDF) algorithm that uses a data-driven kernel function based on a Bayes factor to optimize filtering for spatial delineation of functional connectivity in resting fMRI data. We demonstrate its performance relative to Gaussian spatial filtering and the original tNLM filtering via simulations. We also compare the effects of GPDF filtering against LB filtering using individual in-vivo resting fMRI datasets. Our results show that LB filtering tends to blur signals across boundaries between adjacent functional regions. In contrast, GPDF filtering enables improved noise reduction without blurring adjacent functional regions. These results indicate that GPDF may be a useful preprocessing tool for analyses of brain connectivity and network topology in individual fMRI recordings.

BRAIN LESION DETECTION USING A ROBUST VARIATIONAL AUTOENCODER AND TRANSFER LEARNING.

Automated brain lesion detection from multi-spectral MR images can assist clinicians by improving sensitivity as well as specificity. Supervised machine learning methods have been successful in lesion detection. However, these methods usually rely on a large number of manually delineated images for specific imaging protocols and parameters and often do not generalize well to other imaging parameters and demographics. Most recently, unsupervised models such as autoencoders have become attractive for lesion detection since they do not need access to manually delineated lesions. Despite the success of unsupervised models, using pre-trained models on an unseen dataset is still a challenge. This difficulty is because the new dataset may use different imaging parameters, demographics, and different pre-processing techniques. Additionally, using a clinical dataset that has anomalies and outliers can make unsupervised learning challenging since the outliers can unduly affect the performance of the learned models. These two difficulties make unsupervised lesion detection a particularly challenging task. The method proposed in this work addresses these issues using a two-prong strategy: (1) we use a robust variational autoencoder model that is based on robust statistics, specifically the ß-divergence that can be trained with data that has outliers; (2) we use a transfer-learning method for learning models across datasets with different characteristics. Our results on MRI datasets demonstrate that we can improve the accuracy of lesion detection by adapting robust statistical models and transfer learning for a variational autoencoder model.

A NETWORK-BASED APPROACH TO STUDY OF ADHD USING TENSOR DECOMPOSITION OF RESTING STATE FMRI DATA.

Identifying changes in functional connectivity in Attention Deficit Hyperactivity Disorder (ADHD) using functional magnetic resonance imaging (fMRI) can help us understand the neural substrates of this brain disorder. Many studies of ADHD using resting state fMRI (rs-fMRI) data have been conducted in the past decade with either manually crafted features that do not yield satisfactory performance, or automatically learned features that often lack interpretability. In this work, we present a tensor-based approach to identify brain networks and extract features from rs-fMRI data. Results show the identified networks are interpretable and consistent with our current understanding of ADHD conditions. The extracted features are not only predictive of ADHD score but also discriminative for classification of ADHD subjects from typically developed children.

Anemia predicts lower white matter volume and cognitive performance in sickle and non-sickle cell anemia syndrome.

Severe chronic anemia is an independent predictor of overt stroke, white matter damage, and cognitive dysfunction in the elderly. Severe anemia also predisposes to white matter strokes in young children, independent of the anemia subtype. We previously demonstrated symmetrically decreased white matter (WM) volumes in patients with sickle cell disease (SCD). In the current study, we investigated whether patients with non-sickle anemia also have lower WM volumes and cognitive dysfunction. Magnetic Resonance Imaging was performed on 52 clinically asymptomatic SCD patients (age = 21.4 ± 7.7; F = 27, M = 25; hemoglobin = 9.6 ± 1.6 g/dL), 26 non-sickle anemic patients (age = 23.9 ± 7.9; F = 14, M = 12; hemoglobin = 10.8 ± 2.5 g/dL) and 40 control subjects (age = 27.7 ± 11.3; F = 28, M = 12; hemoglobin = 13.4 ± 1.3 g/dL). Voxel-wise changes in WM brain volumes were compared to hemoglobin levels to identify brain regions that are vulnerable to anemia. White matter volume was diffusely lower in deep, watershed areas proportionally to anemia severity. After controlling for age, sex, and hemoglobin level, brain volumes were independent of disease. WM volume loss was associated with lower Full Scale Intelligence Quotient (FSIQ; P = .0048; r2 = .18) and an abnormal burden of silent cerebral infarctions (P = .029) in males, but not in females. Hemoglobin count and cognitive measures were similar between subjects with and without white-matter hyperintensities. The spatial distribution of volume loss suggests chronic hypoxic cerebrovascular injury, despite compensatory hyperemia. Neurocognitive consequences of WM volume changes and silent cerebral infarction were strongly sexually dimorphic. Understanding the possible neurological consequences of chronic anemia may help inform our current clinical practices.

A principal component analysis-based framework for statistical modeling of bone displacement during wrist maneuvers.

We present a method for the statistical modeling of the displacements of wrist bones during the performance of coordinated maneuvers, such as radial-ulnar deviation (RUD). In our approach, we decompose bone displacement via a set of basis functions, identified via principal component analysis (PCA). We utilized MRI wrist scans acquired at multiple static positions for deriving these basis functions. We then utilized these basis functions to compare the displacements undergone by the bones of the left versus right wrist in the same individual, and between bones of the wrists of men and women, during the performance of the coordinated RUD maneuver. Our results show that the complex displacements of the wrist bones during RUD can be modeled with high reliability with just 5 basis functions, that captured over 91% of variation across individuals. The basis functions were able to predict intermediate wrist bone poses with an overall high accuracy (mean error of 0.26 mm). Our proposed approach found statistically significant differences between bone displacement trajectories in women versus men, however, did not find significant differences in those of the left versus right wrist in the same individual. Our proposed method has the potential to enable detailed analysis of wrist kinematics for each sex, and provide a robust framework for characterizing the normal and pathologic displacement of the wrist bones, such as in the context of wrist instability.

Predicting Cognitive Scores from Resting fMRI Data and Geometric Features of the Brain.

Anatomical T1 weighted Magnetic Resonance Imaging (MRI) and functional magnetic resonance imaging collected during resting (rfMRI) are promising markers that offer insight into the structure and function of the human brain. The objective of this work is to explore the use of a deep learning neural network to predict cognitive performance scores for a population of normal controls and subjects with Attention Deficit Hyperactivity Disorder (ADHD). Specifically, we predict verbal and performance IQs and ADHD index from features derived from T1 and rfMRI imaging data. First, we processed the rfMRI and MRI data of subjects using the BrainSuite fMRI Processing (BFP) pipeline to perform anatomical and functional preprocessing. This produces for each subject fMRI and geometric (anatomical) features represented in a standardized grayordinate system. The geometric and functional cortical data corresponding to the two hemispheres were then transformed to 128×128 multichannel images and input to a convolutional component of the neural network. Subcortical data were presented in a standard vector form and inputted to a input layer of the network. The neural network was implemented in Python using the Keras library with a TensorFlow backend. Training was performed on 168 images with 90 images used for testing. We observed a high correlation between predicted and actual values of the indices tested: Performance IQ: 0.47; Verbal IQ: 0.41, ADHD: 0.57. Comparing these values to those from network trained on functional-only and structural-only data, we saw that rfMRI is more informative than MRI, but the two modalities are highly complementary in terms of predicting these indices.

A Matched Filter Decomposition of fMRI into Resting and Task Components.

The human brain exhibits dynamic interactions among brain regions when responding to stimuli and executing tasks, which can be recorded using functional magnetic resonance imaging (fMRI). Functional MRI signals collected in response to specific tasks consist of a combination of task-related and spontaneous (task-independent) activity. By exploiting the highly structured spatiotemporal patterns of resting state networks, this paper presents a matched-filter approach to decomposing fMRI signals into task and resting-state components. To perform the decomposition, we first use a temporal alignment procedure that is a windowed version of the brainsync transform to synchronize a resting template to the brain's response to tasks. The resulting 'matched filter' removes the components of the fMRI signal that can be described by resting connectivity, leaving the portion of brain activity directly related to tasks. We present a closed-form expression for the windowed synchronization transform that is used by the matched filter. We demonstrate performance of this procedure in application to motor task and language task fMRI data. We show qualitatively and quantitatively that by removing the resting activity, we are able to identify task activated regions in the brain more clearly. Additionally, we show improved prediction accuracy in multivariate pattern analysis when using the matched filtered fMRI data.

Are you thinking what I'm thinking? Synchronization of resting fMRI time-series across subjects.

We describe BrainSync, an orthogonal transform that allows direct comparison of resting fMRI (rfMRI) time-series across subjects. For this purpose, we exploit the geometry of the rfMRI signal space to propose a novel orthogonal transformation that synchronizes rfMRI time-series across sessions and subjects. When synchronized, rfMRI signals become approximately equal at homologous locations across subjects. The method is based on the observation that rfMRI data exhibit similar connectivity patterns across subjects, as reflected in the pairwise correlations between different brain regions. We show that if the data for two subjects have similar correlation patterns then their time courses can be approximately synchronized by an orthogonal transformation. This transform is unique, invertible, efficient to compute, and preserves the connectivity structure of the original data for all subjects. Analogously to image registration, where we spatially align structural brain images, this temporal synchronization of brain signals across a population, or within-subject across sessions, facilitates cross-sectional and longitudinal studies of rfMRI data. The utility of the BrainSync transform is illustrated through demonstrative simulations and applications including quantification of rfMRI variability across subjects and sessions, cortical functional parcellation across a population, timing recovery in task fMRI data, comparison of task and resting state data, and an application to complex naturalistic stimuli for annotation prediction.