Publications

PURPOSE: To illustrate the potential bias caused by imaging gradients in correlation MRI sequences using longitudinal magnetization storage (LS) and examine the case of filter exchange imaging (FEXI) yielding maps of the apparent exchange rate (AXR). METHODS: The effects of imaging gradients in FEXI were observed on yeast cells. To analyze the AXR bias, signal evolution was calculated by applying matrix exponential operators. RESULTS: A sharp threshold for the slice thickness was identified, below which the AXR is increasingly underestimated. The bias can be understood in terms of an extended low-pass diffusion filtering during the LS interval, which is more pronounced at lower exchange rates. For a total exchange rate constant larger than 1 s, the AXR bias is expected to be negligible when slices thicker than 2.5 mm are used. CONCLUSION: In correlation experiments like FEXI, relying on LS with variable duration, imaging gradients may cause disrupting effects that cannot be easily mitigated and should be carefully considered for unbiased results. In typical clinical applications of FEXI, the imaging gradients are expected to cause a negligible AXR bias. However, the AXR bias may be significant in preclinical settings or whenever thin imaging slices are used. Magn Reson Med 79:2228-2235, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

OBJECTIVE: The semantic variant of primary progressive aphasia (svPPA) is typically associated with frontotemporal lobar degeneration (FTLD) with longTAR DNA-binding protein (TDP)-43-positive neuropil threads and dystrophic neurites (type C), and is only rarely due to a primary tauopathy or Alzheimer’s disease. We undertook this study to investigate the localisation and magnitude of the presumed tau Positron Emission Tomography (PET) tracer [F]Flortaucipir (FTP; also known as T807 or AV1451) in patients with svPPA, hypothesising that most patients would not show tracer uptake different from controls. METHODS: FTP and [C]Pittsburgh compound B PET imaging as well as MRI were performed in seven patients with svPPA and in 20 controls. FTP signal was analysed by visual inspection and by quantitative comparison to controls, with and without partial volume correction. RESULTS: All seven patients showed elevated FTP uptake in the anterior temporal lobe with a leftward asymmetry that was not observed in healthy controls. This elevated FTP signal, largely co-localised with atrophy, was evident on both visual inspection and quantitative cortical surface-based analysis. Five patients were amyloid negative, one was amyloid positive and one has an unknown amyloid status. CONCLUSIONS: In this series of patients with clinical profiles, structural MRI and amyloid PET imaging typical for svPPA, FTP signal was unexpectedly elevated with a spatial pattern localised to areas of atrophy. This raises questions about the possible off-target binding of this tracer to non-tau molecules associated with neurodegeneration. Further investigation with autopsy analysis will help illuminate the binding target(s) of FTP in cases of suspected FTLD-TDP neuropathology.

BACKGROUND: Links between dissociation and functional neurological disorder (FND)/conversion disorder are well-established, yet the pathophysiology of dissociation remains poorly understood. This MRI study investigated structural alterations associated with somatoform and psychological dissociation in FND. We hypothesized that multimodal, paralimbic cingulo-insular regions would relate to the severity of somatoform dissociation in patients with FND. METHODS: FreeSurfer cortical thickness and subcortical volumetric analyses were performed in 26 patients with motor FND and 27 matched healthy controls. Patients with high dissociation as measured by the Somatoform Dissociation Questionnaire-20 (SDQ) or Dissociative Experiences Scale (DES) were compared to controls in stratified analyses. Within-group analyses were also performed with SDQ and DES scores in patients with FND. All cortical thickness analyses were whole-brain corrected at the cluster-wise level. RESULTS: Patients with FND and high somatoform dissociation (SDQ > 35) showed reduced left caudal anterior cingulate cortex (ACC) cortical thickness compared to controls. In within-group analyses, SDQ scores inversely correlated with left caudal ACC cortical thickness in patients with FND. Depersonalization/derealization scores positively correlated with right lateral occipital cortical thickness. Both within-group findings remained statistically significant controlling for trait anxiety/depression, borderline personality disorder and post-traumatic stress disorder, adverse life events, and motor FND subtypes in post-hoc analyses. CONCLUSION: Using complementary between-group and within-group analyses, an inverse association between somatoform dissociation and left caudal ACC cortical thickness was observed in patients with FND. A positive relationship was also appreciated between depersonalization/derealization severity and cortical thickness in visual association areas. These findings advance our neuropathobiological understanding of dissociation in FND. Hum Brain Mapp 39:428-439, 2018. © 2017 Wiley Periodicals, Inc.

BACKGROUND: In smokers, the lung parenchyma is characterized by inflammation and emphysema, processes that can result in local gain and loss of lung tissue. CT measures of lung density might reflect lung tissue changes; however, longitudinal data regarding the effects of CT lung tissue on FEVin smokers with and without COPD are scarce. METHODS: The 15th percentile of CT lung density was obtained from the scans of 3,390 smokers who completed baseline and 5-year follow-up of the Genetic Epidemiology of COPD (COPDGene) study visits. The longitudinal relationship between total lung capacity-adjusted lung density (TLC-PD15) and FEVwas assessed by using multivariable mixed models. Separate models were performed in smokers at risk, smokers with preserved ratio and impaired spirometry (PRISm), and smokers with COPD according to the Global Initiative for Chronic Obstructive Lung Disease (GOLD) staging system.

RATIONALE: Deep learning is a powerful tool that may allow for improved outcome prediction. OBJECTIVES: To determine if deep learning, specifically convolutional neural network (CNN) analysis, could detect and stage chronic obstructive pulmonary disease (COPD) and predict acute respiratory disease (ARD) events and mortality in smokers. METHODS: A CNN was trained using computed tomography scans from 7,983 COPDGene participants and evaluated using 1,000 nonoverlapping COPDGene participants and 1,672 ECLIPSE participants. Logistic regression (C statistic and the Hosmer-Lemeshow test) was used to assess COPD diagnosis and ARD prediction. Cox regression (C index and the Greenwood-Nam-D’Agnostino test) was used to assess mortality. MEASUREMENTS AND MAIN RESULTS: In COPDGene, the C statistic for the detection of COPD was 0.856. A total of 51.1% of participants in COPDGene were accurately staged and 74.95% were within one stage. In ECLIPSE, 29.4% were accurately staged and 74.6% were within one stage. In COPDGene and ECLIPSE, the C statistics for ARD events were 0.64 and 0.55, respectively, and the Hosmer-Lemeshow P values were 0.502 and 0.380, respectively, suggesting no evidence of poor calibration. In COPDGene and ECLIPSE, CNN predicted mortality with fair discrimination (C indices, 0.72 and 0.60, respectively), and without evidence of poor calibration (Greenwood-Nam-D’Agnostino P values, 0.307 and 0.331, respectively). CONCLUSIONS: A deep-learning approach that uses only computed tomography imaging data can identify those smokers who have COPD and predict who are most likely to have ARD events and those with the highest mortality. At a population level CNN analysis may be a powerful tool for risk assessment.

Coherence and causality measures are often used to analyze the influence of one region on another during analysis of functional brain networks. The analysis methods usually involve a regression problem, where the signal of interest is decomposed into a mixture of regressor and a residual signal. In this paper, we revisit this basic problem and present solutions that provide the minimal-entropy residuals for different types of regression filters, such as causal, instantaneously causal, and noncausal filters. Using optimal prediction theory, we derive several novel frequency-domain expressions for partial coherence, causality, and conditional causality analysis. In particular, our solution provides a more accurate estimation of the frequency-domain causality compared with the classical Geweke causality measure. Using synthetic examples and in vivo resting-state functional magnetic resonance imaging data from the human connectome project, we show that the proposed solution is more accurate at revealing frequency-domain linear dependence among high-dimensional signals.

INTRODUCTION: Cortical mean diffusivity (MD) and free water fraction (FW) changes are proposed biomarkers for Alzheimer’s disease (AD). METHODS: We included healthy control subjects (N = 254), mild cognitive impairment (N = 41), and AD dementia (N = 31) patients. Participants underwent a lumbar puncture and a 3 T magnetic resonance imaging. Healthy control subjects were classified following National Institute on Aging-Alzheimer’s Association stages (stage 0, N = 220; stage 1, N = 25; and stage 2/3, N = 9). We assessed the cortical MD, cortical FW, and cortical thickness (CTh) changes along the AD continuum. RESULTS: Microstructural and macrostructural changes show a biphasic trajectory. Stage 1 subjects showed increased CTh and decreased MD and FW with respect the stage 0 subjects. Stage 2/3 subjects showed decreased CTh and increased cortical MD and FW, changes that were more widespread in symptomatic stages. DISCUSSION: These results support a biphasic model of changes in AD, which could affect the selection of patients for clinical trials and the use of magnetic resonance imaging as a surrogate marker of disease modification.

Neuroimaging studies demonstrate gray matter (GM) macrostructural abnormalities in patients with schizophrenia (SCZ). While ex-vivo and genetic studies suggest cellular pathology associated with abnormal neurodevelopmental processes in SCZ, few in-vivo measures have been proposed to target microstructural GM organization. Here, we use diffusion heterogeneity- to study GM microstructure in SCZ. Structural and diffusion magnetic resonance imaging (MRI) were acquired on a 3 Tesla scanner in 46 patients with SCZ and 37 matched healthy controls (HC). After correction for free water, diffusion heterogeneity as well as commonly used diffusion measures FA and MD and volume were calculated for the four cortical lobes on each hemisphere, and compared between groups. Patients with early course SCZ exhibited higher diffusion heterogeneity in the GM of the frontal lobes compared to controls. Diffusion heterogeneity of the frontal lobe showed excellent discrimination between patients and HC, while none of the commonly used diffusion measures such as FA or MD did. Higher diffusion heterogeneity in the frontal lobes in early SCZ may be due to abnormal brain maturation (migration, pruning) before and during adolescence and early adulthood. Further studies are needed to investigate the role of heterogeneity as potential biomarker for SCZ risk.

Diffusion tensor imaging studies report childhood adversity (CA) is associated with reduced fractional anisotropy (FA) in multiple white matter tracts in adults. Reduced FA may result from changes in tissue, suggesting myelin/axonal damage, and/or from increased levels of extracellular free-water, suggesting atrophy or neuroinflammation. Free-water imaging can separately identify FA in tissue (FA) and the fractional volume of free-water (FW). We tested whether CA was associated with altered FA, FA, and FW in seven white matter regions of interest (ROI), in which FA changes had been previously linked to CA (corona radiata, corpus callosum, fornix, cingulum bundle: hippocampal projection, inferior fronto-occipital fasciculus, superior longitudinal fasciculus, uncinate fasciculus). Tract-based spatial statistics were performed in 147 psychiatrically healthy adults who had completed a self-report questionnaire on CA primarily stemming from parental maltreatment. ROI were extracted according to the protocol provided by the ENIGMA-DTI working group. Analyses were performed both treating CA as a continuous and a categorical variable. CA was associated with reduced FA in all ROI (although categorical analyses failed to find an association in the fornix). In contrast, CA was only associated with reduced FAin the corona radiata, corpus callosum, and uncinate fasciculus (with the continuous measure of CA finding evidence of a negative relation between CA and FAin the fornix). There was no association between CA on FW in any ROI. These results provide preliminary evidence that childhood adversity is associated with changes to the microstructure of white matter itself in adulthood. However, these results should be treated with caution until they can be replicated by future studies which address the limitations of the present study.

We examined whether abnormal volumes of several brain regions as well as their mutual associations that have been observed in patients with schizophrenia, are also present in individuals at clinical high-risk (CHR) for developing psychosis. 3T magnetic resonance imaging was acquired in 19 CHR and 20 age- and handedness-matched controls. Volumes were measured for the body and temporal horns of the lateral ventricles, hippocampus and amygdala as well as total brain, cortical gray matter, white matter, and subcortical gray matter volumes. Relationships between volumes as well as correlations between volumes and cognitive and clinical measures were explored. Ratios of lateral ventricular volume to total brain volume and temporal horn volume to total brain volume were calculated. Volumetric abnormalities were lateralized to the left hemisphere. Volumes of the left temporal horn, and marginally, of the body of the left lateral ventricle were larger, while left amygdala but not hippocampal volume was significantly smaller in CHR participants compared to controls. Total brain volume was also significantly smaller and the ratio of the temporal horn/total brain volume was significantly higher in CHR than in controls. White matter volume correlated positively with higher verbal fluency score while temporal horn volume correlated positively with a greater number of perseverative errors. Together with the finding of larger temporal horns and smaller amygdala volumes in the left hemisphere, these results indicate that the ratio of temporal horns volume to brain volume is abnormal in CHR compared to controls. These abnormalities present in CHR individuals may constitute the biological basis for at least some of the CHR syndrome.