Dynamic network model with continuous valued nodes for longitudinal brain morphometry.

Longitudinal brain morphometry probes time-related brain morphometric patterns. We propose a method called dynamic network modeling with continuous valued nodes to generate a dynamic brain network from continuous valued longitudinal morphometric data. The mathematical framework of this method is based on state-space modeling. We use a bootstrap-enhanced least absolute shrinkage operator to solve the network-structure generation problem. In contrast to discrete dynamic Bayesian network modeling, the proposed method enables network generation directly from continuous valued high-dimensional short sequence data, being free from any discretization process. We applied the proposed method to a study of normal brain development.

Association of dynamic susceptibility contrast enhanced MR Perfusion parameters with prognosis in elderly patients with glioblastomas.

OBJECTIVES: We aimed to evaluate the prognostic value of dynamic susceptibility contrast (DSC) MR perfusion in elderly patients with glioblastomas (GBM). METHODS: Thirty five patients aged ≥65 and 35 aged <65 years old, (referred to as elderly and younger, respectively) were included in this retrospective study. The median relative cerebral volume (rCBV) from the enhancing region (rCBVER-Med) and immediate peritumoral region (rCBVIPR-Med) and maximum rCBV from the enhancing region of the tumor (rCBVER-Max) were compared and correlated with survival data. Analysis was repeated after rCBVs were dichotomized into high and low values and after excluding elderly patients who did not receive postoperative chemoradiation (34.3%). Kaplan-Meyer survival curves and parametric and semi-parametric regression tests were used for analysis. RESULTS: All rCBV parameters were higher in elderly compared to younger patients (p < 0.05). After adjustment for age, none were independently associated with shorter survival (p > 0.05). After rCBV dichotomization into high and low values, high rCBV in elderly was independently associated with shorter survival compared to low rCBV in elderly, or any rCBV in younger patients (p < 0.05). CONCLUSION: rCBV can be an imaging biomarker to identify a subgroup of GBM patients in the elderly with worse prognosis compared to others. KEY POINTS: • GBM perfusion parameters are higher in elderly compared to younger patients. • rCBV can identify a subgroup of elderly patients with worse prognosis. • rCBV can be an imaging biomarker for prognostication in GBM. • The identified elderly patients may benefit from anti-angiogenic treatment.

Changes in Brain Activation Patterns During Working Memory Tasks in People With Post-COVID Condition and Persistent Neuropsychiatric Symptoms.

BACKGROUND AND OBJECTIVES: Post-COVID condition (PCC) is common and often involves neuropsychiatric symptoms. This study aimed to use blood oxygenation level-dependent fMRI (BOLD-fMRI) to assess whether participants with PCC had abnormal brain activation during working memory (WM) and whether the abnormal brain activation could predict cognitive performance, motor function, or psychiatric symptoms. METHODS: The participants with PCC had documented coronavirus disease 2019 (COVID-19) at least 6 weeks before enrollment. Healthy control participants had no prior history of COVID-19 and negative tests for severe acute respiratory syndrome coronavirus 2. Participants were assessed using 3 NIH Toolbox (NIHTB) batteries for Cognition (NIHTB-CB), Emotion (NIHTB-EB), and Motor function (NIHTB-MB) and selected tests from the Patient-Reported Outcomes Measurement Information System (PROMIS). Each had BOLD-fMRI at 3T, during WM (N-back) tasks with increasing attentional/WM load. RESULTS: One hundred sixty-nine participants were screened; 50 fulfilled the study criteria and had complete and usable data sets for this cross-sectional cohort study. Twenty-nine participants with PCC were diagnosed with COVID-19 242 ± 156 days earlier; they had similar ages (42 ± 12 vs 41 ± 12 years), gender proportion (65% vs 57%), racial/ethnic distribution, handedness, education, and socioeconomic status, as the 21 uninfected healthy controls. Despite the high prevalence of memory (79%) and concentration (93%) complaints, the PCC group had similar performance on the NIHTB-CB as the controls. However, participants with PCC had greater brain activation than the controls across the network (false discovery rate-corrected p = 0.003, Tmax = 4.17), with greater activation in the right superior frontal gyrus (p = 0.009, Cohen d = 0.81, 95% CI 0.15-1.46) but lesser deactivation in the default mode regions (p = 0.001, d = 1.03, 95% CI 0.61-1.99). Compared with controls, participants with PCC also had poorer dexterity and endurance on the NIHTB-MB, higher T scores for negative affect and perceived stress, but lower T scores for psychological well-being on the NIHTB-EB, as well as more pain symptoms and poorer mental and physical health on measures from the PROMIS. Greater brain activation predicted poorer scores on measures that were abnormal on the NIHTB-EB. DISCUSSION: Participants with PCC and neuropsychiatric symptoms demonstrated compensatory neural processes with greater usage of alternate brain regions, and reorganized networks, to maintain normal performance during WM tasks. BOLD-fMRI was sensitive for detecting brain abnormalities that correlated with various quantitative neuropsychiatric symptoms.

Dynamic Bayesian network modeling for longitudinal brain morphometry.

Identifying interactions among brain regions from structural magnetic-resonance images presents one of the major challenges in computational neuroanatomy. We propose a Bayesian data-mining approach to the detection of longitudinal morphological changes in the human brain. Our method uses a dynamic Bayesian network to represent evolving inter-regional dependencies. The major advantage of dynamic Bayesian network modeling is that it can represent complicated interactions among temporal processes. We validated our approach by analyzing a simulated atrophy study, and found that this approach requires only a small number of samples to detect the ground-truth temporal model. We further applied dynamic Bayesian network modeling to a longitudinal study of normal aging and mild cognitive impairment--the Baltimore Longitudinal Study of Aging. We found that interactions among regional volume-change rates for the mild cognitive impairment group are different from those for the normal-aging group.

Structural MRI in autism spectrum disorder.

Magnetic resonance (MR) examination provides a powerful tool for investigating brain structural changes in children with autism spectrum disorder (ASD). We review recent advances in the understanding of structural MR correlates of ASD. We summarize findings from studies based on voxel-based morphometry, surface-based morphometry, tensor-based morphometry, and diffusion-tensor imaging. Finally, we discuss diagnostic models of ASD based on MR-derived features.

Artificial intelligence in molecular imaging.

AI has, to varying degrees, affected all aspects of molecular imaging, from image acquisition to diagnosis. During the last decade, the advent of deep learning in particular has transformed medical image analysis. Although the majority of recent advances have resulted from neural-network models applied to image segmentation, a broad range of techniques has shown promise for image reconstruction, image synthesis, differential-diagnosis generation, and treatment guidance. Applications of AI for drug design indicate the way forward for using AI to facilitate molecular-probe design, which is still in its early stages. Deep-learning models have demonstrated increased efficiency and image quality for PET reconstruction from sinogram data. Generative adversarial networks (GANs), which are paired neural networks that are jointly trained to generate and classify images, have found applications in modality transformation, artifact reduction, and synthetic-PET-image generation. Some AI applications, based either partly or completely on neural-network approaches, have demonstrated superior differential-diagnosis generation relative to radiologists. However, AI models have a history of brittleness, and physicians and patients may not trust AI applications that cannot explain their reasoning. To date, the majority of molecular-imaging applications of AI have been confined to research projects, and are only beginning to find their ways into routine clinical workflows via commercialization and, in some cases, integration into scanner hardware. Evaluation of actual clinical products will yield more realistic assessments of AI's utility in molecular imaging.

Voxel-based Bayesian lesion-symptom mapping.

Most existing voxel-based lesion-symptom mapping methods are based on the same statistical foundation: null hypothesis significance testing (NHST). The two major limitations of these methods are the inability to infer that there is no difference in lesion proportions, and a requirement for multiple-comparison correction. We propose a Bayesian approach that directly models the posterior distribution of lesion-proportion difference, and makes decisions based on inference on this posterior distribution. Compared to NHST-based approaches, our Bayesian approach yields inference results with clearer semantics, and does not require multiple-comparison correction. We evaluated our Bayesian method using simulated data, and data from a study of acute ischemic left-hemispheric stroke. Results of both experiments indicate that the Bayesian approach is sensitive in detecting regions that characterize group differences.

Machine-learning techniques for building a diagnostic model for very mild dementia.

Many researchers have sought to construct diagnostic models to differentiate individuals with very mild dementia (VMD) from healthy elderly people, based on structural magnetic-resonance (MR) images. These models have, for the most part, been based on discriminant analysis or logistic regression, with few reports of alternative approaches. To determine the relative strengths of different approaches to analyzing structural MR data to distinguish people with VMD from normal elderly control subjects, we evaluated seven different classification approaches, each of which we used to generate a diagnostic model from a training data set acquired from 83 subjects (33 VMD and 50 control). We then evaluated each diagnostic model using an independent data set acquired from 30 subjects (13 VMD and 17 controls). We found that there were significant performance differences across these seven diagnostic models. Relative to the diagnostic models generated by discriminant analysis and logistic regression, the diagnostic models generated by other high-performance diagnostic-model-generation algorithms manifested increased generalizability when diagnostic models were generated from all atlas structures.

Predictive models of autism spectrum disorder based on brain regional cortical thickness.

Autism spectrum disorder (ASD) is a neurodevelopmental disorder with a wide phenotypic range, often affecting personality and communication. Previous voxel-based morphometry (VBM) studies of ASD have identified both gray- and white-matter volume changes. However, the cerebral cortex is a 2-D sheet with a highly folded and curved geometry, which VBM cannot directly measure. Surface-based morphometry (SBM) has the advantage of being able to measure cortical surface features, such as thickness. The goals of this study were twofold: to construct diagnostic models for ASD, based on regional thickness measurements extracted from SBM, and to compare these models to diagnostic models based on volumetric morphometry. Our study included 22 subjects with ASD (mean age 9.2+/-2.1 years) and 16 volunteer controls (mean age 10.0+/-1.9 years). Using SBM, we obtained regional cortical thicknesses for 66 brain structures for each subject. In addition, we obtained volumes for the same 66 structures for these subjects. To generate diagnostic models, we employed four machine-learning techniques: support vector machines (SVMs), multilayer perceptrons (MLPs), functional trees (FTs), and logistic model trees (LMTs). We found that thickness-based diagnostic models were superior to those based on regional volumes. For thickness-based classification, LMT achieved the best classification performance, with accuracy=87%, area under the receiver operating characteristic (ROC) curve (AUC)=0.93, sensitivity=95%, and specificity=75%. For volume-based classification, LMT achieved the highest accuracy, with accuracy=74%, AUC=0.77, sensitivity=77%, and specificity=69%. The thickness-based diagnostic model generated by LMT included 7 structures. Relative to controls, children with ASD had decreased cortical thickness in the left and right pars triangularis, left medial orbitofrontal gyrus, left parahippocampal gyrus, and left frontal pole, and increased cortical thickness in the left caudal anterior cingulate and left precuneus. Overall, thickness-based classification outperformed volume-based classification across a variety of classification methods.

Computational framework for detection of subtypes of neuropsychiatric disorders based on DTI-derived anatomical connectivity.

Many brain disorders - such as Alzheimer's disease, Parkinson's disease, schizophrenia and autism - are heterogeneous, that is, they may have several subtypes. Traditionally, clinicians have identified subtypes, such as subtypes of psychosis, using clinical criteria. Neuroimaging has the potential to detect subtypes based on objective biomarker-based criteria; however, there are no studies that evaluate the application of combining unsupervised machine learning and anatomical connectivity analysis to accomplish this goal. We propose a computational framework to detect subtypes based on anatomical connectivity computed from diffusion tensor imaging data, in a data-driven and fully automated way. The proposed method exhibits excellent performance on simulated data. We also applied this approach to a real-world dataset: the Nathan Kline Institute data set. The Nathan Kline Institute study consists of 137 normal adult subjects (mean age 41 years (standard deviation 18), male/female 85/52). We examined the association between detected subtypes and the impulsive behavior scale. We found that a subtype characterized by lower connectivity scores was associated with a higher positive urgency score; positive urgency is a vulnerability marker for drug addiction. The top-ranked connections characterizing subtypes involve several brain regions, including the anterior cingulate gyrus, median cingulate gyrus, thalamus, superior frontal gyrus (medial), middle frontal gyrus (orbital part), inferior frontal gyrus (triangular part), superior frontal gyrus, precuneus and putamen. The proposed framework is extendable, and can be used to detect subtypes from other features, including clinical and genomic biomarkers.

Where (in the brain) do semantic errors come from?

BACKGROUND: Semantic errors result from the disruption of access either to semantics or to lexical representations. One way to determine the origins of these errors is to evaluate comprehension of words that elicit semantic errors in naming. We hypothesized that in acute stroke there are different brain regions where dysfunction results in semantic errors in both naming and comprehension versus those with semantic errors in oral naming alone. METHODS: A consecutive series of 196 patients with acute left hemispheric stroke who met inclusion criteria were evaluated with oral naming and spoken word/picture verification tasks and magnetic resonance imaging within 48 h of stroke onset. We evaluated the relationship between tissue dysfunction in 10 pre-specified Brodmann's areas (BA) and the production of coordinate semantic errors resulting from (1) semantic deficits or (2) lexical access deficits. RESULTS: Semantic errors arising from semantic deficits were most associated with tissue dysfunction/infarct of left BA 22. Semantic errors resulting from lexical access deficits were associated with hypoperfusion/infarct of left BA 37. CONCLUSION: Our study shows that semantic errors arising from damage to distinct cognitive processes reflect dysfunction of different brain regions.

Neural regions essential for reading and spelling of words and pseudowords.

OBJECTIVE: To identify dysfunctional brain regions critical for impaired reading/spelling of words/pseudowords by evaluating acute stroke patients on lexical tests and magnetic resonance imaging, before recovery or reorganization of structure-function relationships. METHODS: A series of 106 consenting patients were administered oral reading and spelling tests within 24 hours of left supratentorial stroke onset. Patients underwent diffusion- and perfusion-weighted magnetic resonance examination the same day to identify regions of hypoperfusion/infarct of 16 Brodmann areas. RESULTS: Simultaneous logistic regression analysis demonstrated that dysfunction of left Brodmann areas 40 (supramarginal gyrus) and 37 (posterior-inferior temporal/fusiform gyrus) best predicted impairment in reading words (odds ratio [OR], 6.20 [95% confidence interval (CI), 1.54-24.96] and 2.71 [95% CI, 0.87-8.45], respectively), reading pseudowords (OR, 39.65 [95% CI 3.9-400.78] and 4.41 [95% CI, 1.1-17.51], respectively), spelling words (OR, 14.11 [95% CI 1.37-144.93] and 7.41 [95% CI, 1.48-37.24], respectively), and spelling pseudowords (OR, 4.84 [95% CI, 0.73-32.13] and 7.74 [95% CI, 1.56-38.51], respectively). Whole-brain voxel-wise analyses demonstrated voxel clusters within these regions that were most strongly associated with task deficits. INTERPRETATION: Results indicate that a shared network of regions including parts of left Brodmann areas 37 and 40 is necessary for reading and spelling of words and pseudowords. Further studies may define the precise roles of these brain regions in language. Identification of any neural regions specific to one of these tasks or one type of stimuli will require study of more patients with selective deficits.

Science to Practice: IT Solutions to Drive Standardized Report Recommendations for Abdominal Aortic Aneurysm Surveillance.

Inadequate imaging surveillance has been identified as the most significant contributor to abdominal aortic aneurysm (AAA) rupture. Radiologists can contribute value to patient care and reduce morbidity and mortality related to AAA by incorporating evidence-based management recommendations from the ACR and Society of Vascular Surgery into their report impression. The challenges lie in achieving 100% radiologist compliance to incorporate the recommendations and ensuring that the patient is notified by their provider, the follow-up examination is scheduled, and the patient returns for an imaging test that may be scheduled 3 to 5 years in the future. To address these barriers, radiology quality and informatics leads have harnessed IT solutions to facilitate integration of content, communication of results, and patient follow-up.

Multicenter Research Studies in Radiology.

RATIONALE AND OBJECTIVES: Here we review the current state of multicenter radiology research (MRR), and utilize a survey of experienced researchers to identify common advantages, barriers, and resources to guide future investigators. MATERIALS AND METHODS: The Association of University Radiologists established a Radiology Research Alliance task force, Multi-center Research Studies in Radiology, composed of 12 society members to review MRR. A REDCap survey was designed to gain more insight from experienced researchers. Recipients were authors identified from a PubMed database search, utilizing search terms "multicenter" or "multisite" and "radiology." The survey included investigator background information, reasons why, barriers to, and resources that investigators found helpful in conducting or participating in MRR. RESULTS: The survey was completed by 23 of 80 recipients (29%), the majority (76%) of whom served as a primary investigator on at least one MRR project. Respondents reported meeting collaborators at national or international (74%) and society (39%) meetings. The most common perceived advantages of MRR were increased sample size (100%) and improved generalizability (91%). External funding was considered the most significant barrier to MRR, reported by 26% of respondents. Institutional funding, setting up a central picture archiving and communication system, and setting up a central database were considered a significant barrier by 30%, 22%, and 22% of respondents, respectively. Resources for overcoming barriers included motivated staff (74%), strong leadership (70%), regular conference calls (57%), and at least one face-to-face meeting (57%). CONCLUSIONS: Barriers to MRR include funding and establishing a central database and a picture archiving and communication system. Upon embarking on an MRR project, forming a motivated team who meets and speaks regularly is essential.

Clinical diagnosis based on bayesian classification of functional magnetic-resonance data.

We describe a method for classifying subjects based on functional magnetic-resonance (fMR) data, using a method combining a Bayesian-network classifier with inverse-tree structure (BNCIT), and ensemble learning. The central challenge is to generate a classifier from a small sample of high-dimensional data. The principal strengths of our method include the nonparametric multivariate Bayesian-network representation, and joint performance of feature selection and classification. Preliminary results indicate that this method can detect regions characterizing group differences, and can, on the basis of activation levels in these regions, accurately classify new subjects.

Brain morphometric analysis predicts decline of intelligence quotient in children with sickle cell disease: A preliminary study.

PURPOSE: For children with sickle cell disease (SCD) and at low risk category of stroke, we aim to build a predictive model to differentiate those with decline of intelligence-quotient (IQ) from counterparts without decline, based on structural magnetic-resonance (MR) imaging volumetric analysis. MATERIALS AND METHODS: This preliminary prospective cohort study included 25 children with SCD, homozygous for hemoglobin S, with no history of stroke and transcranial Doppler mean velocities below 170cm/s at baseline. We administered the Kaufman Brief Intelligence Test (K-BIT) to each child at yearly intervals for 2-4 years. Each child underwent MR examination within 30 days of the baseline K-BIT evaluation date. We calculated K-BIT change rates, and used rate of change in K-BIT to classify children into two groups: a decline group and a non-decline group. We then generated predictive models to predict K-BIT decline/non-decline based on regional gray-matter (GM) volumes computed from structural MR images. RESULTS: We identified six structures (the left median cingulate gyrus, the right middle occipital gyrus, the left inferior occipital gyrus, the right fusiform gyrus, the right middle temporal gyrus, the right inferior temporal gyrus) that, when assessed for volume at baseline, are jointly predictive of whether a child would suffer subsequent K-BIT decline. Based on these six regional GM volumes and the baseline K-BIT, we built a prognostic model using the K* algorithm. The accuracy, sensitivity and specificity were 0.84, 0.78 and 0.86, respectively. CONCLUSIONS: GM volumetric analysis predicts subsequent IQ decline for children with SCD.

Anatomy of spatial attention: insights from perfusion imaging and hemispatial neglect in acute stroke.

The site of lesion responsible for left hemispatial neglect after stroke has been intensely debated recently. Some studies provide evidence that right angular lesions are most likely to cause left neglect, whereas others indicate that right superior temporal lesions are most likely to cause neglect. We examine two potential accounts of the conflicting results: (1) neglect could result from cortical dysfunction beyond the structural lesion in some studies; and (2) different forms of neglect with separate neural correlates have been included in different proportions in separate studies. To evaluate these proposals, we studied 50 patients with acute right subcortical infarcts using tests of hemispatial neglect and magnetic resonance diffusion-weighted and perfusion-weighted imaging performed within 48 h of onset of symptoms. Left "allocentric" neglect (errors on the left sides of individual stimuli, regardless of location with respect to the viewer) was most strongly associated with hypoperfusion of right superior temporal gyrus (Fisher's exact test; p < 0.0001), whereas left "egocentric" neglect (errors on the left of the viewer) was most strongly associated with hypoperfusion of the right angular gyrus (p < 0.0001). Patients without cortical hypoperfusion showed no hemispatial neglect. Because the patients did not have cortical infarcts, our data show that neglect can be caused by hypoperfused dysfunctional tissue not detectable by structural magnetic resonance imaging. Moreover, different forms of neglect were associated with different sites of cortical hypoperfusion. Results help explain conflicting results in the literature and contribute to the understanding of spatial attention and representation in the human brain.

Statistical atlas of acute stroke from magnetic resonance diffusion-weighted-images of the brain.

This study presents a method for computing a probabilistic atlas that describes the spatial distributions of acute infarcts of the brain. The data consisted of diffusion-weighted-images (DWI) and high-resolution T1-weighted MR images representing 22 studies from 22 subjects. All DWI data sets contained high-intensity lesions on B-1000 maps, known from clinical history to be related to acute stroke. To compute the atlas, manually segmented infarcts on original DWI were spatially transformed and registered to a common coordinate system. This coordinate system allowed combining all lesions into a statistical atlas in the model space. As a result, the computed probabilistic map showed mild left-sided predominance of brain infarcts, which likely represents asymmetry in eloquence of brain regions. In our opinion, the statistical atlas of acute brain infarcts can facilitate computer-based detection of stroke in large image data sets.

Routine Cross-Sectional Head Imaging Before Electroconvulsive Therapy: A Tertiary Center Experience.

PURPOSE: Electroconvulsive therapy (ECT) is generally contraindicated in patients with intracranial mass lesions or in the presence of increased intracranial pressure. The purpose of this study was to determine the prevalence of incidental abnormalities on routine cross-sectional head imaging, including CT and MRI, that would preclude subsequent ECT. METHODS: This retrospective study involved a review of the electronic medical records of 105 patients (totaling 108 imaging studies) between April 27, 2007, and March 20, 2015, referred for cranial CT or MRI with the primary indication of pre-ECT evaluation. The probability of occurrence of imaging findings that would preclude ECT was computed. A cost analysis was also performed on the practice of routine pre-ECT imaging. RESULTS: Of the 105 patients who presented with the primary indication of ECT clearance (totaling 108 scans), 1 scan (0.93%) revealed findings that precluded ECT. None of the studies demonstrated findings that indicated increased intracranial pressure. A cost analysis revealed that at least $18,662.70 and 521.97 relative value units must be expended to identify one patient with intracranial pathology precluding ECT. CONCLUSIONS: The findings of this study demonstrate an extremely low prevalence of findings that preclude ECT on routine cross-sectional head imaging. The costs incurred in identifying a potential contraindication are high. The authors suggest that the performance of pre-ECT neuroimaging be driven by the clinical examination.

Connectivity-Based Brain Parcellation: A Connectivity-Based Atlas for Schizophrenia Research.

Defining brain structures of interest is an important preliminary step in brain-connectivity analysis. Researchers interested in connectivity patterns among brain structures typically employ manually delineated volumes of interest, or regions in a readily available atlas, to limit the scope of connectivity analysis to relevant regions. However, most structural brain atlases, and manually delineated volumes of interest, do not take voxel-wise connectivity patterns into consideration, and therefore may not be ideal for anatomic connectivity analysis. We herein propose a method to parcellate the brain into regions of interest based on connectivity. We formulate connectivity-based parcellation as a graph-cut problem, which we solve approximately using a novel multi-class Hopfield network algorithm. We demonstrate the application of this approach using diffusion tensor imaging data from an ongoing study of schizophrenia. Compared to a standard anatomic atlas, the connectivity-based atlas supports better classification performance when distinguishing schizophrenic from normal subjects. Comparing connectivity patterns averaged across the normal and schizophrenic subjects, we note significant systematic differences between the two atlases.