T1-weighted Brain MRI Research Articles

Predicting brain age with complex networks: From adolescence to adulthood

In recent years, several studies have demonstrated that machine learning and deep learning systems can be very useful to accurately predict brain age. In this work, we propose a novel approach based on complex networks using 1016 T1-weighted MRI brain scans (in the age range 7−64years). We introduce a structural connectivity model of the human brain: MRI scans are divided in rectangular boxes and Pearson’s correlation is measured among them in order to obtain a complex network model. Brain connectivity is then characterized through few and easy-to-interpret centrality measures; finally, brain age is predicted by feeding a compact deep neural network. The proposed approach is accurate, robust and computationally efficient, despite the large and heterogeneous dataset used. Age prediction accuracy, in terms of correlation between predicted and actual age r=0.89and Mean Absolute Error MAE =2.19years, compares favorably with results from state-of-the-art approaches. On an independent test set including 262 subjects, whose scans were acquired with different scanners and protocols we found MAE =2.52. The only imaging analysis steps required in the proposed framework are brain extraction and linear registration, hence robust results are obtained with a low computational cost. In addition, the network model provides a novel insight on aging patterns within the brain and specific information about anatomical districts displaying relevant changes with aging.

Magnetic Resonance-Guided Laser Interstitial Thermal Therapy for the Treatment of Nonlesional Insular Epilepsy in Pediatric Patients: Technical Considerations

Introduction: The insula presents anatomic challenges to surgical exploration and intervention. Open neurosurgical intervention is associated with high rates of complications despite improved seizure control. Minimally invasive techniques using novel energy delivery methods have gained popularity due to their relative safety and ability to overcome access-related barriers. The goal of this paper is to present an operative technical report and methodological considerations on the application of magnetic resonance-guided laser interstitial thermal therapy (MRgLITT) for the treatment of nonlesional, medically refractory, insular epilepsy in pediatric patients. Methods: Visualase laser probe(s) were implanted using ROSA robotic stereotactic guidance into the insula using a parasagittal trajectory. After confirmation of placement using intraoperative MRI, thermal energy was delivered under real-time MR guidance. Laser wire pullback was performed when the initial dose of thermal energy was insufficient to ablate the target in its entirety. Thermal ablation within the intended target was confirmed using gadolinium-enhanced brain MRI. Following removal of laser wires, a final T1-weighted axial brain MRI was performed to confirm no evidence of hemorrhage. Results: Three patients underwent MRgLITT of nonlesional insular epilepsy over an 11-month period. The epileptogenic focus was localized to the insula using stereoelectroencephalography. The anterior and middle portions of the insula were accessed using a parasagittal trajectory. Laser ablation was performed for up to 3 min using an output of 10.5 W. No complications were encountered, and all patients were discharged within 24 h after the surgery. At the most recent follow-up, all patients had an Engel I outcome without any new neurologic deficits. Conclusion: This small cohort shows that insular ablation can be achieved safely with promising seizure outcomes in the short term.

Binary Classification of Alzheimer's Disease Using sMRI Imaging Modality and Deep Learning.

Alzheimer's disease (AD) is an irreversible devastative neurodegenerative disorder associated with progressive impairment of memory and cognitive functions. Its early diagnosis is crucial for the development of possible future treatment option(s). Structural magnetic resonance images (sMRI) play an important role to help in understanding the anatomical changes related to AD especially in its early stages. Conventional methods require the expertise of domain experts and extract hand-picked features such as gray matter substructures and train a classifier to distinguish AD subjects from healthy subjects. Different from these methods, this paper proposes to construct multiple deep 2D convolutional neural networks (2D-CNNs) to learn the various features from local brain images which are combined to make the final classification for AD diagnosis. The whole brain image was passed through two transfer learning architectures; Inception version 3 and Xception, as well as a custom Convolutional Neural Network (CNN) built with the help of separable convolutional layers which can automatically learn the generic features from imaging data for classification. Our study is conducted using cross-sectional T1-weighted structural MRI brain images from Open Access Series of Imaging Studies (OASIS) database to maintain the size and contrast over different MRI scans. Experimental results show that the transfer learning approaches exceed the performance of non-transfer learning-based approaches demonstrating the effectiveness of these approaches for the binary AD classification task.

Inpainting as a Technique for Estimation of Missing Voxels in Brain Imaging.

Issues with model fitting (i.e. suboptimal standard deviation, linewidth/full-width-at-half-maximum, and/or signal-to-noise ratio) in multi-voxel MRI spectroscopy, or chemical shift imaging (CSI) can result in the significant loss of usable voxels. A potential solution to minimize this problem is to estimate the value of unusable voxels by utilizing information from reliable voxels in the same image. We assessed an image restoration method called inpainting as a tool to restore unusable voxels, and compared it with traditional interpolation methods (nearest neighbor, trilinear interpolation and tricubic interpolation). In order to evaluate the performance across varying image contrasts and spatial resolutions, we applied the same techniques to a T1-weighted MRI brain dataset, and N-acetylaspartate (NAA) spectroscopy maps from a CSI dataset. For all image types, inpainting exhibited superior performance (lower normalized root-mean-square errors, NRMSE) compared to all other methods considered (p's < 0.001). Inpainting maintained an average NRMSE of less than 5% even with 50% missing voxels, whereas the other techniques demonstrated up to three times that value, depending on the nature of the image. For CSI maps, inpainting maintained its superiority whether the previously unusable voxels were randomly distributed, or located in regions most commonly affected by voxel loss in real-world data. Inpainting is a promising approach for recovering unusable or missing voxels in voxel-wise analyses, particularly in imaging modalities characterized by low SNR such as CSI. We hypothesize that this technique may also be applicable for datasets from other imaging modalities, such as positron emission tomography, or dynamic susceptibility contrast MRI.

SUN-684 The Dance of Diabetes

Nonketotic hyperglycemic chorea-ballism (NKHCB) is a rare movement disorder typically associated with uncontrolled type 2 diabetes. It is often the result of a focal lesion of the contralateral basal ganglia. Here we present a case of a patient who presented with hyperglycemic chorea as the initial manifestation of their diabetes. A 73 year old caucasian female with no known past medical history was brought to the emergency room with involuntary movements of the left face, arm and leg. These movements were sudden in onset and not suppressible. Her vitals were stable on arrival and exam was benign apart from choreiform movements of the left arm, left leg and left perioral muscles. Her labs were significant for a random blood glucose level of 418 mg/dl and HbA1c of 12%. As part of her workup she underwent a CT head which showed an asymmetric hyperdensity involving the right putamen. She underwent additional workup of chorea to exclude infection, vitamin deficiencies, heavy metal poisoning, electrolyte imbalances, liver dysfunction, all of which returned negative. She eventually underwent an MRI of the brain which showed symmetric, mildly bright T1 signal within both putamen suggestive of mild bilateral calcium deposition bilaterally. The patient was further evaluated by neurology who felt that her overall clinical presentation and imaging findings were consistent with a diagnosis of NKHCB. She was placed on an insulin regimen with significant improvement in blood glucose levels. Her symptoms resolved within two weeks as noted on outpatient follow up. NKHCB is a rare condition that manifests in the setting of uncontrolled nonketotic diabetes mellitus. The exact underlying pathophysiology of changes seen on imaging of patients with NKHCB is not fully understood. The diagnosis is based on typical clinical and radiological features including the presence of ballistic or choreiform movements in the setting of marked hyperglycemia and the absence of ketoacidosis. This case is unusual as her chorea was the presenting symptom of her diabetes which was previously undiagnosed. The characteristic radiological feature is a high signal intensity basal ganglia lesion on the T1-weighted brain MRI. The mainstay of treatment includes normalization of the blood glucose, although in some cases antipsychotic use has been described. In most cases, complete resolution of symptoms has been reported within 10 months. Given the excellent prognosis with management of blood glucose, non ketotic hyperglycemia is an important differential to consider in a patient presenting with chorea-hemiballismus.

Semi-Supervised Learning for Fetal Brain MRI Quality Assessment with ROI consistency.

Fetal brain MRI is useful for diagnosing brain abnormalities but is challenged by fetal motion. The current protocol for T2-weighted fetal brain MRI is not robust to motion so image volumes are degraded by inter- and intra-slice motion artifacts. Besides, manual annotation for fetal MR image quality assessment are usually time-consuming. Therefore, in this work, a semi-supervised deep learning method that detects slices with artifacts during the brain volume scan is proposed. Our method is based on the mean teacher model, where we not only enforce consistency between student and teacher models on the whole image, but also adopt an ROI consistency loss to guide the network to focus on the brain region. The proposed method is evaluated on a fetal brain MR dataset with 11,223 labeled images and more than 200,000 unlabeled images. Results show that compared with supervised learning, the proposed method can improve model accuracy by about 6% and outperform other state-of-the-art semi-supervised learning methods. The proposed method is also implemented and evaluated on an MR scanner, which demonstrates the feasibility of online image quality assessment and image reacquisition during fetal MR scans.

368. Incidental Findings on Brain MRI in People Living with HIV

BackgroundHIV infection is associated with an array of neurocognitive changes, collectively referred to as HIV-Associated Neurocognitive Disorder (HAND). These changes have been the subject of a great deal of study, often including structural MRI of the brain. Incidental findings (IF) are a well-known complication of imaging studies done for both diagnostic and research indications, and can pose important ethical and clinical dilemmas. Little is known about the rates and types of IF found on brain MRI in patients with HIV infection. We identified and characterized such findings in participants who participated in a study of neurophysiological markers of HAND.MethodsThe parent study included 108 HIV-infected adults and 125 demographically matched uninfected controls without cognitive impairment who had undergone T1-weighted structural brain MRI for research purposes. Demographic and diagnostic data were abstracted from the research records. Each MRI study was read by the same neuroradiologist, blind to the participant’s HIV status. IF were classified as vascular, neoplastic, congenital, other neurologic, or non-neurologic. Categorical measures were compared using Pearson chi-square tests while continuous measures were compared using t-tests.ResultsAmong HIV-infected participants, 36/108 (33.3%) had IF compared with 33/125 (26.4%) of controls (P = 0.248). Rates of IF were significantly correlated with increased age in both HIV-infected and control participants. We found no correlation among presence or absence of IF and sex, race/ethnicity, or CD4 count and HAND status for the HIV-infected cohort. The most common categories were neurologic (27), followed by non-neurologic (8), vascular (6), and neoplastic (2) (Table 1).ConclusionIf were common in both HIV-infected participants and controls, at higher rates than previously reported, possibly because of increased sensitivity of MRI machines over time. Surprisingly, we found no significant difference between the groups and no correlation with HAND status or CD4 count. Age was the only factor correlated with rates of IF in either HIV-infected participants or controls. To our knowledge, this study is the first of its kind to characterize incidental findings in HIV-infected patients. Disclosures All authors: No reported disclosures.

LMNB1-Related Adult-Onset Autosomal Dominant Leukodystrophy Presenting as Movement Disorder: A Case Report and Review of the Literature.

Adult-onset autosomal dominant leukodystrophy (ADLD) is a lately described rare form of leukodystrophy with only one family report from China. As the only disease associated with increased lamina B1 encoded by LMNB1, ADLDs have different clinical presentations, ranging from autonomic to pyramidal tract and cerebellar ataxia. Here, we report a case of ADLD that presented with positional tremor as the initial symptom. T2-weighted brain MRI showed brain atrophy and diffuse high signal intensity of the cerebral white matter and the brain stem. The precise diagnosis was made by identification of the mutated gene. To the best of our knowledge, this is perhaps the first case report of ADLD presenting as tremor in China.

Using 3D-Printed Mesh-Like Brain Cortex with Deep Structures for Planning Intracranial EEG Electrode Placement.

Surgical evaluation of medically refractory epilepsy frequently necessitates implantation of multiple intracranial electrodes for the identification of the seizure focus. Knowledge of the individual brain's surface anatomy and deep structures is crucial for planning the electrode implantation. We present a novel method of 3D printing a brain that allows for the simulation of placement of all types of intracranial electrodes. We used a DICOM dataset of a T1-weighted 3D-FSPGR brain MRI from one subject. The segmentation tools of Materialise Mimics 21.0 were used to remove the osseous anatomy from brain parenchyma. Materialise 3-matic 13.0 was then utilized in order to transform the cortex of the segmented brain parenchyma into a mesh-like surface. Using 3-matic tools, the model was modified to incorporate deep brain structures and create an opening in the medial aspect. The final model was then 3D printed as a cerebral hemisphere with nylon material using selective laser sintering technology. The final model was light and durable and reflected accurate details of the surface anatomy and some deep structures. Additionally, standard surgical depth electrodes could be passed through the model to reach deep structures without damaging the model. This novel 3D-printed brain model provides a unique combination of visualizing both the surface anatomy and deep structures through the mesh-like surface while allowing repeated needle insertions. This relatively low-cost technique can be implemented for interdisciplinary preprocedural planning in patients requiring intracranial EEG monitoring and for any intervention that requires needle insertion into a solid organ with unique anatomy and internal targets.

An adaptive differential evolution algorithm to optimal multi-level thresholding for MRI brain image segmentation

Segmentation is an important method for MRI medical image analysis as it can provide the radiologists with noninvasive information about a patient that is crucial to the diagnostic process. The efficiency of such a computer-aided diagnosis system relies on the accuracy of an adopted image segmentation method. Multi-level thresholding is a segmentation method that has been widely adopted in medical image analysis in recent studies, where selecting the optimal thresholds has a pivotal role in determining the efficiency and the accuracy of the segmentation algorithm. While some well-known methods, such as Kapur’s and Otsu’s, are proven effective for bi-level thresholding, multi-level thresholding remains a challenge as it is computationally expensive. Evolutionary algorithms, such as Differential Evolution (DE), have the potential to address this problem, as they can find sufficiently good solutions with manageable computational effort. While a number of DE solutions have been proposed for multi-level thresholding, they are not stable, in that, when the number of thresholds increases, the algorithm efficiency decreases due to the imbalance between exploration and exploitation. In this paper, we propose a DE solution that achieves a good balance between exploration and exploitation through a new adaptive approach and new mutation strategies. The new adaptive approach can generate optimal solutions in assigning populations by measuring the quality of candidate solutions to evaluate the efficiency of different parts of the proposed DE algorithm. The new mutation methods harness Mantegna Lévy and Cauchy distributions, as well as Cotes’ Spiral to improve global search, and to further balance between exploitation and exploration. We further experimentally compare the proposed DE algorithm, referred to as Adaptive Differential Evolution with Lévy Distribution (ALDE), against three DE benchmark algorithms on T2 weighted MRI brain images. Our results show that ALDE can, not only obtain optimal thresholds at a reasonable computational cost, but more importantly, clearly outperforms the benchmark algorithms.

Comparison of the WEKA and SVM-light based on support vector machine in classifying Alzheimer’s disease using structural features from brain MR imaging

The aim was to compare the WEKA and SVM-light software based on support vector machine (SVM) algorithm using features from brain T1-weighted MRI for differentiating AD patients and normal elderly subjects. The FreeSurfer software was used to extract cerebral volumes and thicknesses from T1-weighted brain MRI (100 AD patients and 100 normal elderly subjects). Seven structures were selected based on literature reviews consisting of hippocampus and amygdala volume, entorhinal cortex thickness of both hemispheres, and total gray matter volume. Relative volume of hippocampus, amygdala, and total gray matter were normalized by total intracranial volume (TIV). Fifteen combinations of seven structures were applied as input features to WEKA and SVM-light. The receiver operating characteristic (ROC) analysis and area under the curve (AUC) were used to evaluate the classification performance. The combination of hippocampus relative volume and entorhinal cortex thickness provided the highest classification performance and the AUC values were 0.913 and 0.918 for WEKA and SVM-light, respectively. There was no statistically difference of the AUC values (p-value > 0.05) between two software using the same input features. In conclusion, there was no statistically difference between the use of WEKA and SVM-light software for differentiating AD patients and normal elderly subjects.

Age-specific optimization of T1-weighted brain MRI throughout infancy

The infant brain undergoes drastic morphological and functional development during the first year of life. Three-dimensional T1-weighted Magnetic Resonance Imaging (3D T1w-MRI) is a major tool to characterize the brain anatomy, which however, manifests inherently low and rapidly changing contrast between white matter (WM) and gray matter (GM) in the infant brains (0-12 month-old). Despite the prior efforts made to maximize tissue contrast in the neonatal brains (≤1 months), optimization of imaging methods in the rest of the infancy (1–12 months) is not fully addressed, while brains in the latter period exhibit even more challenging contrast. Here, we performed a systematic investigation to improve the contrast between cortical GM and subcortical WM throughout the infancy. We first performed simultaneous T1 and proton density mapping in a normally developing infant cohort at 3T (n = 57). Based on the evolution of T1 relaxation times, we defined three age groups and simulated the relative tissue contrast between WM and GM in each group. Age-specific imaging strategies were proposed according to the Bloch simulation: inversion time (TI) around 800 ms for the 0-3 month-old group, dual TI at 500 ms and 700 ms for the 3-7 month-old group, and TI around 700 ms for 7-12 month-old group, using a centrically encoded 3D-MPRAGE sequence at 3T. Experimental results with varying TIs in each group confirmed improved contrast at the proposed optimal TIs, even in 3–7 month-old infants who had nearly isointense contrast. We further demonstrated the advantage of improved relative contrast in segmenting the neonatal brains using a multi-atlas segmentation method. The proposed age-specific optimization strategies can be easily adapted to routine clinical examinations, and the improved image contrast would facilitate quantitative analysis of the infant brain development.

Deep Learning and Multiplex Networks for Accurate Modeling of Brain Age

Recent works have extensively investigated the possibility to predict brain aging from T1-weighted MRI brain scans. The main purposes of these studies are the investigation of subject-specific aging mechanisms and the development of accurate models for age prediction. Deviations between predicted and chronological age are known to occur in several neurodegenerative diseases; as a consequence, reaching higher levels of age prediction accuracy is of paramount importance to develop diagnostic tools. In this work, we propose a novel complex network model for brain based on segmenting T1-weighted MRI scans in rectangular boxes, called patches, and measuring pairwise similarities using Pearson's correlation to define a subject-specific network. We fed a deep neural network with nodal metrics, evaluating both the intensity and the uniformity of connections, to predict subjects' ages. Our model reaches high accuracies which compare favorably with state-of-the-art approaches. We observe that the complex relationships involved in this brain description cannot be accurately modeled with standard machine learning approaches, such as Ridge and Lasso regression, Random Forest, and Support Vector Machines, instead a deep neural network has to be used.

Association between white matter lesions and cerebral glucose metabolism in patients with cognitive impairment

AimWhite matter lesions (WMLs), detected as hyperintensities in T2-weighted MRI, represent small vessel disease in the brain and are considered a potential risk factor for memory and cognitive impairment. It has not been sufficiently evidenced that cognitive impairment in patients with Alzheimer's disease is caused by WMLs as well as β-amyloid (Aβ) pathology. The aim of this study was to evaluate the relationship between WMLs and cerebral glucose metabolism in patients with cognitive impairment after adjustment of cerebral Aβ burden. Materials and methodsEighty-three subjects with cognitive performance ranging from normal to dementia, who underwent brain MRI and 18F-florbetaben positron emission tomography (PET) and 18F-fluorodeoxyglucose PET, were included in this cross-sectional study. The Fazekas scale was used to quantify WMLs on brain T2-weighted MRI. The cerebral Aβ burden and cerebral glucose metabolism were quantitatively estimated using volume-of-interest analysis. Differences in the regional cerebral glucose metabolism were evaluated between low-WML (Fazekas scale<2) and high-WML (Fazekas scale≥2) groups. Multiple linear regression analysis adjusted for age, sex and cerebral Aβ burden was performed to evaluate the relationship between the Fazekas scale score and cerebral glucose metabolism. ResultsThe regional cerebral glucose metabolism for the bilateral frontal, temporal, and parietal cortices, and limbic lobes in the high-WML group were significantly lower than those in the low-WML group. There were significant negative correlations between the Fazekas scale score and regional cerebral glucose metabolism in the bilateral frontal, bilateral temporal and left parietal cortices, and bilateral limbic lobes. Multiple linear regression analysis revealed that the Fazekas scale score was an independent determinant of the glucose metabolism in the bilateral frontal and temporal cortices and limbic lobes. ConclusionsWMLs are associated with decreased cerebral glucose metabolism. Our findings suggest that small vessel disease, as well as Aβ pathology, may contribute to cognitive impairment in patients with Alzheimer's disease.

3D T1-weighted contrast-enhanced brain MRI in children using a fat-suppressed golden angle radial acquisition: an alternative to Cartesian inversion-recovery imaging

BackgroundT1-weighted post-contrast MRI is essential in brain protocols. We demonstrate the feasibility and utility of a 3D non-Cartesian radial acquisition in children. PurposeTo compare bulk motion artifacts, image quality, and lesion conspicuity in 3D T1-weighted post-contrast brain MRI between a new fat-suppressed radial gradient-echo and a traditional non-fat-suppressed inversion-recovery Cartesian gradient-echo sequence. Material and methodsImages from 53 patients acquired at 3 Tesla were compared. Three radiologists rated the images in three categories, including the presence of bulk motion and whether it impacted diagnosis, whether one sequence was preferred over the other in overall image quality and conspicuity of vascular structures and lesions, and whether diagnosis was possible if only the new fat-suppressed radial acquisition was obtained. ResultsThe Fleiss' kappa for inter-rater agreement was 0.67 for bulk motion and 0.54 for sequence preference. Of the 53 cases, 56% were identified to have significant motion on conventional imaging, while only 13% had motion artifacts on the radial acquisition (p < 0.05). There were no cases where motion was seen on the radial acquisition but not on conventional imaging. Both sequences were equally preferred in 87% of the cases. All radiologists agreed that the radial approach had lower gray-white matter contrast than the conventional inversion-recovery method, but preferred the former for making diagnosis in uncooperative patients. ConclusionWe demonstrate the potential utility of a fat-suppressed 3D T1-weighted post-contrast brain gradient-echo sequence in children. The technique is useful in non-sedate pediatric imaging due to its reduced sensitivity to bulk motion.

Anatomical characterization of the inferior petrosal sinus and adjacent cerebellopontine angle cistern for development of an endovascular transdural cerebrospinal fluid shunt

Background and purposeWe evaluated the inferior petrosal sinus (IPS) and adjacent cerebellopontine angle (CPA) cistern as a potential implantation site for a novel venous endovascular transdural CSF shunt concept to...

Cross-sectional and longitudinal assessment of the upper cervical spinal cord in motor neuron disease.

BackgroundAmyotrophic lateral sclerosis (ALS) is a progressive neuromuscular disease characterized by both upper and lower motor neuron degeneration. While neuroimaging studies of the brain can detect upper motor neuron degeneration, these brain MRI scans also include the upper part of the cervical spinal cord, which offers the possibility to expand the focus also towards lower motor neuron degeneration. Here, we set out to investigate cross-sectional and longitudinal disease effects in the upper cervical spinal cord in patients with ALS, progressive muscular atrophy (PMA: primarily lower motor neuron involvement) and primary lateral sclerosis (PLS: primarily upper motor neuron involvement), and their relation to disease severity and grey and white matter brain measurements.MethodsWe enrolled 108 ALS patients without C9orf72 repeat expansion (ALS C9–), 26 ALS patients with C9orf72 repeat expansion (ALS C9+), 28 PLS patients, 56 PMA patients and 114 controls. During up to five visits, longitudinal T1-weighted brain MRI data were acquired and used to segment the upper cervical spinal cord (UCSC, up to C3) and individual cervical segments (C1 to C4) to calculate cross-sectional areas (CSA). Using linear (mixed-effects) models, the CSA differences were assessed between groups and correlated with disease severity. Furthermore, a relationship between CSA and brain measurements was examined in terms of cortical thickness of the precentral gyrus and white matter integrity of the corticospinal tract.ResultsCompared to controls, CSAs at baseline showed significantly thinner UCSC in all groups in the MND spectrum. Over time, ALS C9– patients demonstrated significant thinning of the UCSC and, more specifically, of segment C3 compared to controls. Progressive thinning over time was also observed in C1 of PMA patients, while ALS C9+ and PLS patients did not show any longitudinal changes. Longitudinal spinal cord measurements showed a significant relationship with disease severity and we found a significant correlation between spinal cord and motor cortex thickness or corticospinal tract integrity for PLS and PMA, but not for ALS patients.DiscussionOur findings demonstrate atrophy of the upper cervical spinal cord in the motor neuron disease spectrum, which was progressive over time for all but PLS patients. Cervical spinal cord imaging in ALS seems to capture different disease effects than brain neuroimaging. Atrophy of the cervical spinal cord is therefore a promising additional biomarker for both diagnosis and disease progression and could help in the monitoring of treatment effects in future clinical trials.

Subcortical brain volumes, cortical thickness and cortical surface area in families genetically enriched for social anxiety disorder - A multiplex multigenerational neuroimaging study.

BackgroundSocial anxiety disorder (SAD) is a disabling psychiatric condition with a genetic background. Brain alterations in gray matter (GM) related to SAD have been previously reported, but it remains to be elucidated whether GM measures are candidate endophenotypes of SAD. Endophenotypes are measurable characteristics on the causal pathway from genotype to phenotype, providing insight in genetically-based disease mechanisms. Based on a review of existing evidence, we examined whether GM characteristics meet two endophenotype criteria, using data from a unique sample of SAD-patients and their family-members of two generations. First, we investigated whether GM characteristics co-segregate with social anxiety within families genetically enriched for SAD. Secondly, heritability of the GM characteristics was estimated.MethodsFamilies with a genetic predisposition for SAD participated in the Leiden Family Lab study on SAD; T1-weighted MRI brain scans were acquired (n = 110, 8 families). Subcortical volumes, cortical thickness and cortical surface area were determined for a-priori determined regions of interest (ROIs). Next, associations with social anxiety and heritabilities were estimated.FindingsSeveral subcortical and cortical GM characteristics, derived from frontal, parietal and temporal ROIs, co-segregated with social anxiety within families (uncorrected p-level) and showed moderate to high heritability.InterpretationThese findings provide preliminary evidence that GM characteristics of multiple ROIs, which are distributed over the brain, are candidate endophenotypes of SAD. Thereby, they shed light on the genetic vulnerability for SAD. Future research is needed to confirm these results and to link them to functional brain alterations and to genetic variations underlying these GM changes.FundLeiden University Research Profile ‘Health, Prevention and the Human Life Cycle’.

Extinction and Renewal of Conditioned Eyeblink Responses in Focal Cerebellar Disease.

Extinction of conditioned aversive responses (CR) has been shown to be context-dependent. The hippocampus and prefrontal cortex are of particular importance. The cerebellum may contribute to context-related processes because of its known connections with the hippocampus and prefrontal cortex. Context dependency of extinction can be demonstrated by the renewal effect. When CR acquisition takes place in context A and is extinguished in context B, renewal refers to the recovery of the CR in context A (A-B-A paradigm). In the present study acquisition, extinction and renewal of classically conditioned eyeblink responses were tested in 18 patients with subacute focal cerebellar lesions and 18 age- and sex-matched healthy controls. Standard delay eyeblink conditioning was performed using an A-B-A paradigm. All cerebellar patients underwent a high-resolution T1-weighted brain MRI scan to perform lesion-symptom mapping. CR acquisition was not significantly different between cerebellar and control participants allowing to draw conclusions on extinction. CR extinction was significantly less in cerebellar patients. Reduction of CR extinction tended to be more likely in patients with lesions in the lateral parts of lobule VI and Crus I. A significant renewal effect was present in controls only. The present data provide further evidence that the cerebellum contributes to extinction of conditioned eyeblink responses. Because acquisition was preserved and extinction took place in another context than acquisition, more lateral parts of the cerebellar hemisphere may contribute to context-related processes. Furthermore, lack of renewal in cerebellar patients suggest a contribution of the cerebellum to context-related processes.

Real-Time Deep Pose Estimation With Geodesic Loss for Image-to-Template Rigid Registration.

With an aim to increase the capture range and accelerate the performance of state-of-the-art inter-subject and subject-to-template 3-D rigid registration, we propose deep learning-based methods that are trained to find the 3-D position of arbitrarily-oriented subjects or anatomy in a canonical space based on slices or volumes of medical images. For this, we propose regression convolutional neural networks (CNNs) that learn to predict the angle-axis representation of 3-D rotations and translations using image features. We use and compare mean square error and geodesic loss to train regression CNNs for 3-D pose estimation used in two different scenarios: slice-to-volume registration and volume-to-volume registration. As an exemplary application, we applied the proposed methods to register arbitrarily oriented reconstructed images of fetuses scanned in-utero at a wide gestational age range to a standard atlas space. Our results show that in such registration applications that are amendable to learning, the proposed deep learning methods with geodesic loss minimization achieved 3-D pose estimation with a wide capture range in real-time (<100ms). We also tested the generalization capability of the trained CNNs on an expanded age range and on images of newborn subjects with similar and different MR image contrasts. We trained our models on T2-weighted fetal brain MRI scans and used them to predict the 3-D pose of newborn brains based on T1-weighted MRI scans. We showed that the trained models generalized well for the new domain when we performed image contrast transfer through a conditional generative adversarial network. This indicates that the domain of application of the trained deep regression CNNs can be further expanded to image modalities and contrasts other than those used in training. A combination of our proposed methods with accelerated optimization-based registration algorithms can dramatically enhance the performance of automatic imaging devices and image processing methods of the future.