Following the success of deep learning in a wide range of applications, neural network-based machine-learning techniques have received significant interest for accelerating magnetic resonance imaging (MRI) acquisition and reconstruction strategies. A number of ideas inspired by deep learning techniques for computer vision and image processing have been successfully applied to nonlinear image reconstruction in the spirit of compressed sensing for accelerated MRI. Given the rapidly growing nature of the field, it is imperative to consolidate and summarize the large number of deep learning methods that have been reported in the literature, to obtain a better understanding of the field in general. This article provides an overview of the recent developments in neural-network based approaches that have been proposed specifically for improving parallel imaging. A general background and introduction to parallel MRI is also given from a classical view of k-space based reconstruction methods. Image domain based techniques that introduce improved regularizers are covered along with k-space based methods which focus on better interpolation strategies using neural networks. While the field is rapidly evolving with plenty of papers published each year, in this review, we attempt to cover broad categories of methods that have shown good performance on publicly available data sets. Limitations and open problems are also discussed and recent efforts for producing open data sets and benchmarks for the community are examined.
The main objective of this work is to develop and evaluate an artificial intelligence system based on deep learning capable of automatically identifying, quantifying, and characterizing COVID-19 pneumonia patterns in order to assess disease severity and predict clinical outcomes, and to compare the prediction performance with respect to human reader severity assessment and whole lung radiomics. We propose a deep learning based scheme to automatically segment the different lesion subtypes in nonenhanced CT scans. The automatic lesion quantification was used to predict clinical outcomes. The proposed technique has been independently tested in a multicentric cohort of 103 patients, retrospectively collected between March and July of 2020. Segmentation of lesion subtypes was evaluated using both overlapping (Dice) and distance-based (Hausdorff and average surface) metrics, while the proposed system to predict clinically relevant outcomes was assessed using the area under the curve (AUC). Additionally, other metrics including sensitivity, specificity, positive predictive value and negative predictive value were estimated. 95% confidence intervals were properly calculated. The agreement between the automatic estimate of parenchymal damage (%) and the radiologists' severity scoring was strong, with a Spearman correlation coefficient (R) of 0.83. The automatic quantification of lesion subtypes was able to predict patient mortality, admission to the Intensive Care Units (ICU) and need for mechanical ventilation with an AUC of 0.87, 0.73 and 0.68 respectively. The proposed artificial intelligence system enabled a better prediction of those clinically relevant outcomes when compared to the radiologists' interpretation and to whole lung radiomics. In conclusion, deep learning lesion subtyping in COVID-19 pneumonia from noncontrast chest CT enables quantitative assessment of disease severity and better prediction of clinical outcomes with respect to whole lung radiomics or radiologists' severity score.
PURPOSE: Although both relaxation and diffusion imaging are sensitive to tissue microstructure, studies have reported limited sensitivity and robustness of using relaxation or conventional diffusion alone to characterize tissue microstructure. Recently, it has been shown that tensor-valued diffusion encoding and joint relaxation-diffusion quantification enable more reliable quantification of compartment-specific microstructural properties. However, scan times to acquire such data can be prohibitive. Here, we aim to simultaneously quantify relaxation and diffusion using MR fingerprinting (MRF) and b-tensor encoding in a clinically feasible time.
METHODS: We developed multidimensional MRF scans (mdMRF) with linear and spherical b-tensor encoding (LTE and STE) to simultaneously quantify T1, T2, and ADC maps from a single scan. The image quality, accuracy, and scan efficiency were compared between the mdMRF using LTE and STE. Moreover, we investigated the robustness of different sequence designs to signal errors and their impact on the maps.
RESULTS: T1 and T2 maps derived from the mdMRF scans have consistently high image quality, while ADC maps are sensitive to different sequence designs. Notably, the fast imaging steady state precession (FISP)-based mdMRF scan with peripheral pulse gating provides the best ADC maps that are free of image distortion and shading artifacts.
CONCLUSION: We demonstrated the feasibility of quantifying T1, T2, and ADC maps simultaneously from a single mdMRF scan in around 24 s/slice. The map quality and quantitative values are consistent with the reference scans.
INTRODUCTION: A plasma-derived factor VIII product (pdFVIII, Factane 100 or 200 IU/mL) and a plasma-derived von Willebrand factor product (pdVWF, Wilfactin 100 IU/mL) are approved for replacement therapy by intravenous bolus injections in haemophilia A (HA) and von Willebrand disease (VWD), respectively. However, in situations requiring intensive treatment, continuous infusion (CI) may be desirable to better control target plasma factor levels.
AIM: To evaluate the perioperative haemostatic efficacy and safety of these concentrates administered by CI.
METHODS: Three phase III trials were conducted. Adults with HA (FVIII:C <1%) (Studies 1 and 2) or VWD (VWF:RCo <20%) (Study 3) received a preoperative bolus followed by CI of undiluted concentrate for at least 6 days. Bolus doses and CI rates were based on individual recovery and clearance, respectively. Initial infusion rate had to be higher for 48 hours for HA and 24 hours for VWD patients to anticipate potential fluctuations of factor concentrations during major surgery. Target levels of FVIII:C in HA and VWF:RCo in VWD were 80 and 70 IU/dL, respectively. Efficacy was assessed using a global haemostatic efficacy score.
RESULTS: Studies 1, 2, and 3 included 12, 4 and 6 patients, respectively. Efficacy outcomes were excellent/good in all 22 major surgeries including 18 orthopaedic procedures. Most daily measured FVIII and VWF levels (92%) were on target. No safety concerns, thrombotic events or inhibitors were identified.
CONCLUSION: pdFVIII and pdVWF administered by CI represent an effective and safe alternative to bolus injections in patients with severe HA or VWD undergoing surgery.
Gambling Disorder (GD) is a condition constituting a public health concern, with a burden of harm which is much greater than that of drug addiction. Patients with GD are generally reluctant to pharmacologic treatment and seem to prefer nonpharmacological interventions. Therefore, this proof-of-concept study aimed to investigate the feasibility of continuous Theta Burst Stimulation (cTBS) on the pre-SMA in six patients (5 males, 1 female), aged 30-64 years, with a DSM-5 diagnosis of Gambling Disorder and no comorbid mood disorders. Participants received over 10 sessions of Continuous TBS (cTBS) over pre-SMA bilaterally and have been evaluated using rating scales, including the PG-YBOCS and the CGI, before treatment (T0), at day 10 of treatment (T1) and at day 30 after treatment (T2); cTBS intervention was safe and without side effects. Since the design of our study does not allow us to draw conclusions on the effectiveness of the intervention with respect to the improvement of the functioning of the subject with GD, a more in-depth study, including a sham condition, neurocognitive measures of disinhibition and decision making, and collecting follow-up data on the sustained effect of TBS over a longer period is ongoing.
Background: Tumor-related hyperintensities in high b-value diffusion-weighted imaging (DWI) are radiologically important in the workup of gliomas. However, the white matter may also appear as hyperintense, which may conflate interpretation.
Purpose: To investigate whether DWI with spherical b-tensor encoding (STE) can be used to suppress white matter and enhance the conspicuity of glioma hyperintensities unrelated to white matter.
Materials and Methods: Twenty-five patients with a glioma tumor and at least one pathology-related hyperintensity on DWI underwent conventional MRI at 3 T. The DWI was performed both with linear and spherical tensor encoding (LTE-DWI and STE-DWI). The LTE-DWI here refers to the DWI obtained with conventional diffusion encoding and averaged across diffusion-encoding directions. Retrospectively, the differences in contrast between LTE-DWI and STE-DWI, obtained at a b-value of 2,000 s/mm2, were evaluated by comparing hyperintensities and contralateral normal-appearing white matter (NAWM) both visually and quantitatively in terms of the signal intensity ratio (SIR) and contrast-to-noise ratio efficiency (CNReff).
Results: The spherical tensor encoding DWI was more effective than LTE-DWI at suppressing signals from white matter and improved conspicuity of pathology-related hyperintensities. The median SIR improved in all cases and on average by 28%. The median (interquartile range) SIR was 1.9 (1.6 - 2.1) for STE and 1.4 (1.3 - 1.7) for LTE, with a significant difference of 0.4 (0.3 -0.5) (p < 10-4, paired U-test). In 40% of the patients, the SIR was above 2 for STE-DWI, but with LTE-DWI, the SIR was below 2 for all patients. The CNReff of STE-DWI was significantly higher than of LTE-DWI: 2.5 (2 - 3.5) vs. 2.3 (1.7 - 3.1), with a significant difference of 0.4 (-0.1 -0.6) (p < 10-3, paired U-test). The STE improved CNReff in 70% of the cases. We illustrate the benefits of STE-DWI in three patients, where STE-DWI may facilitate an improved radiological description of tumor-related hyperintensity, including one case that could have been missed out if only LTE-DWI was inspected.
Conclusion: The contrast mechanism of high b-value STE-DWI results in a stronger suppression of white matter than conventional LTE-DWI, and may, therefore, be more sensitive and specific for assessment of glioma tumors and DWI-hyperintensities.
BACKGROUND: Plasma-derived von Willebrand factor (VWF) (Wilfactin®, LFB, France) was developed for prophylaxis and treatment of haemorrhages in both adults and adolescents with von Willebrand disease (VWD). Replacement therapy in paediatric patients is a key element of the clinical trial programme. MATERIAL AND METHODS: Patients aged <6 years with severe VWD were enrolled in a multinational, open-label study to evaluate the in vivo recovery for Wilfactin®, and its efficacy in preventing and treating bleeding episodes and during surgery. Overall haemostatic efficacy based on a 4-point scale was assessed by investigators. The treatment period ≥18 months investigated the long-term safety. RESULTS: Nine patients, including 7 with type 3 VWD were exposed to treatment with Wilfactin® for up to 4.2 years. Recovery of VWF in 7 patients (n=5 type 3, n=1 type 2, n=1 type 1) was 1.8±0.4 IU/dL per IU/kg. Of the 62 bleeds, 89% were controlled with one (73%) or two (16%) infusions of Wilfactin®. The median dose per infusion was 54 IU/kg. A factor VIII dose was co-administered in 1.6% of bleeds. "Excellent"/"Good" haemostatic efficacy was achieved in 90.3% of episodes. Six patients underwent 11 minor surgical interventions. Treatment duration was 1 day (range: 1-6 days) with a dose administered 30-60 minutes before procedure of 56 IU/kg (range: 41-106 IU/kg). Haemostasis was rated as "Excellent" in all surgeries. During 4-year prophylactic treatment in one patient, breakthrough bleeds were reported in 2.2% of infusions. No VWF inhibitors, thromboembolic events or allergic/anaphylactic-type reactions were observed following a total exposure of 770 days. DISCUSSION: The results show that Wilfactin® provides a safe and effective treatment in patients <6 years of age with severe VWD.
Extracellular free water (FW) increases are suggested to better provide pathophysiological information in brain aging than conventional biomarkers such as fractional anisotropy. The aim of the present study was to determine the relationship between conventional biomarkers, FW in white matter hyperintensities (WMH), FW in normal appearing white matter (NAWM) and in white matter tracts and executive functions (EF) with a speed component in elderly persons. We examined 226 healthy elderly participants (median age 69.83 years, IQR: 56.99-74.42) who underwent brain MRI and neuropsychological examination. FW in WMH and in NAWM as well as FW corrected diffusion metrics and measures derived from conventional MRI (white matter hyperintensities, brain volume, lacunes) were used in partial correlation (adjusted for age) to assess their correlation with EF with a speed component. Random forest analysis was used to assess the relative importance of these variables as determinants. Lastly, linear regression analyses of FW in white matter tracts corrected for risk factors of cognitive and white matter deterioration, were used to examine the role of specific tracts on EF with a speed component, which were then ranked with random forest regression. Partial correlation analyses revealed that almost all imaging metrics showed a significant association with EF with a speed component (r=-0.213 - 0.266). Random forest regression highlighted FW in WMH and in NAWM as most important among all diffusion and structural MRI metrics. The fornix (R2=0.421, p=0.018) and the corpus callosum (genu (R2=0.418, p=0.021), prefrontal (R2=0.416, p=0.026), premotor (R2=0.418, p=0.021)) were associated with EF with a speed component in tract based regression analyses and had highest variables importance. In a normal aging population FW in WMH and NAWM is more closely related to EF with a speed component than standard DTI and brain structural measures. Higher amounts of FW in the fornix and the frontal part of the corpus callosum leads to deteriorating EF with a speed component.
Research suggests electroconvulsive therapy (ECT) induces an acute neuroinflammatory response and changes in white matter (WM) structural connectivity. However, whether these processes are related, either to each other or to eventual treatment outcomes, has yet to be determined. We examined the relationship between levels of peripheral pro-inflammatory cytokines and diffusion imaging-indexed changes in WM microstructure in individuals with treatment-resistant depression (TRD) who underwent ECT. Forty-two patients were assessed at baseline, after their second ECT (T2), and after completion of ECT (T3). A Montgomery Åsberg Depression Rating Scale improvement of >50% post-ECT defined ECT-responders (n = 19) from non-responders (n = 23). Thirty-four controls were also examined. Tissue-specific fractional anisotropy (FAt) was estimated using diffusion imaging data and the Free-Water method in 17 WM tracts. Inflammatory panels were evaluated from peripheral blood. Cytokines were examined to characterize the association between potential ECT-induced changes in an inflammatory state and WM microstructure. Longitudinal trajectories of both measures were also examined separately for ECT-responders and non-responders. Patients exhibited elevated Interleukin-8 (IL-8) levels at baseline compared to controls. In patients, correlations between IL-8 and FAt changes from baseline to T2 were significant in the positive direction in the right superior longitudinal fasciculus (R-SLF) and right cingulum (R-CB) (psig = 0.003). In these tracts, linear mixed-effects models revealed that trajectories of IL-8 and FAt were significantly positively correlated across all time points in responders, but not non-responders (R-CB-p = .001; R-SLF-p = 0.008). Our results suggest that response to ECT in TRD may be mediated by IL-8 and WM microstructure.
PURPOSE: NousNav is a complete low-cost neuronavigation system that aims to democratize access to higher-quality healthcare in lower-resource settings. NousNav's goal is to provide a model for local actors to be able to reproduce, build and operate a fully functional neuronavigation system at an affordable cost. METHODS: NousNav is entirely open source and relies on low-cost off-the-shelf components, which makes it easy to reproduce and deploy in any region. NousNav's software is also specifically devised with the low-resource setting in mind. RESULTS: It offers means for intuitive intraoperative control. The designed interface is also clean and simple. This allows for easy intraoperative use by either the practicing clinician or a nurse. It thus alleviates the need for a dedicated technician for operation. CONCLUSION: A prototype implementation of the design was built. Hardware and algorithms were designed for robustness, ruggedness, modularity, to be standalone and data-agnostic. The built prototype demonstrates feasibility of the objectives.
There are many ways to acquire and process diffusion MRI (dMRI) data for group studies, but it is unknown which maximizes the sensitivity to white matter (WM) pathology. Inspired by this question, we analyzed data acquired for diffusion tensor imaging (DTI) and diffusion kurtosis imaging (DKI) at 3T (3T-DTI and 3T-DKI) and DTI at 7T in patients with systemic lupus erythematosus (SLE) and healthy controls (HC). Parameter estimates in 72 WM tracts were obtained using TractSeg. The impact on the sensitivity to WM pathology was evaluated for the diffusion protocol, the magnetic field strength, and the processing pipeline. Sensitivity was quantified in terms of Cohen's d for group comparison. Results showed that the choice of diffusion protocol had the largest impact on the effect size. The effect size in fractional anisotropy (FA) across all WM tracts was 0.26 higher when derived by DTI than by DKI and 0.20 higher in 3T compared with 7T. The difference due to the diffusion protocol was larger than the difference due to magnetic field strength for the majority of diffusion parameters. In contrast, the difference between including or excluding different processing steps was near negligible, except for the correction of distortions from eddy currents and motion which had a clearly positive impact. For example, effect sizes increased on average by 0.07 by including motion and eddy correction for FA derived from 3T-DTI. Effect sizes were slightly reduced by the incorporation of denoising and Gibbs-ringing removal (on average by 0.011 and 0.005, respectively). Smoothing prior to diffusion model fitting generally reduced effect sizes. In summary, 3T-DTI in combination with eddy current and motion correction yielded the highest sensitivity to WM pathology in patients with SLE. However, our results also indicated that the 3T-DKI and 7T-DTI protocols used here may be adjusted to increase effect sizes.
Background: The subthalamic nucleus (STN) is an effective neurosurgical target to improve motor symptoms in Parkinson's Disease (PD) patients. MR-guided Focused Ultrasound (MRgFUS) subthalamotomy is being explored as a therapeutic alternative to Deep Brain Stimulation (DBS) of the STN. The hyperdirect pathway provides a direct connection between the cortex and the STN and is likely to play a key role in the therapeutic effects of MRgFUS intervention in PD patients.
Objective: This study aims to investigate the topography and somatotopy of hyperdirect pathway projections from the primary motor cortex (M1).
Methods: We used advanced multi-fiber tractography and high-resolution diffusion MRI data acquired on five subjects of the Human Connectome Project (HCP) to reconstruct hyperdirect pathway projections from M1. Two neuroanatomy experts reviewed the anatomical accuracy of the tracts. We extracted the fascicles arising from the trunk, arm, hand, face and tongue area from the reconstructed pathways. We assessed the variability among subjects based on the fractional anisotropy (FA) and mean diffusivity (MD) of the fibers. We evaluated the spatial arrangement of the different fascicles using the Dice Similarity Coefficient (DSC) of spatial overlap and the centroids of the bundles.
Results: We successfully reconstructed hyperdirect pathway projections from M1 in all five subjects. The tracts were in agreement with the expected anatomy. We identified hyperdirect pathway fascicles projecting from the trunk, arm, hand, face and tongue area in all subjects. Tract-derived measurements showed low variability among subjects, and similar distributions of FA and MD values among the fascicles projecting from different M1 areas. We found an anterolateral somatotopic arrangement of the fascicles in the corona radiata, and an average overlap of 0.63 in the internal capsule and 0.65 in the zona incerta.
Conclusion: Multi-fiber tractography combined with high-resolution diffusion MRI data enables the identification of the somatotopic organization of the hyperdirect pathway. Our preliminary results suggest that the subdivisions of the hyperdirect pathway projecting from the trunk, arm, hand, face, and tongue motor area are intermixed at the level of the zona incerta and posterior limb of the internal capsule, with a predominantly overlapping topographical organization in both regions. Subject-specific knowledge of the hyperdirect pathway somatotopy could help optimize target definition in MRgFUS intervention.
Deep brain stimulation (DBS) is a well-established neurosurgical procedure for movement disorders that is also being explored for treatment-resistant psychiatric conditions. This review highlights important consideration for DBS simulation and data analysis. The literature on DBS has expanded considerably in recent years, and this article aims to identify important trends in the field. During DBS planning, surgery, and follow up sessions, several large data sets are created for each patient, and it becomes clear that any group analysis of such data is a big data analysis problem and has to be handled with care. The aim of this review is to provide an update and overview from a neuroengineering perspective of the current DBS techniques, technical aids, and emerging tools with the focus on patient-specific electric field (EF) simulations, group analysis, and visualization in the DBS domain. Examples are given from the state-of-the-art literature including our own research. This work reviews different analysis methods for EF simulations, tractography, deep brain anatomical templates, and group analysis. Our analysis highlights that group analysis in DBS is a complex multi-level problem and selected parameters will highly influence the result. DBS analysis can only provide clinically relevant information if the EF simulations, tractography results, and derived brain atlases are based on as much patient-specific data as possible. A trend in DBS research is creation of more advanced and intuitive visualization of the complex analysis results suitable for the clinical environment.
In cerebral small vessel disease (CSVD), both white matter hyperintensities (WMH) of presumed vascular origin and the normal-appearing white matter (NAWM) contain microstructural brain alterations on diffusion-weighted MRI (DWI). Contamination of DWI-derived metrics by extracellular free-water can be corrected with free-water (FW) imaging. We investigated the alterations in FW and FW-corrected fractional anisotropy (FA-t) in WMH and surrounding tissue and their association with cerebrovascular risk factors. We analysed 1,000 MRI datasets from the Hamburg City Health Study. DWI was used to generate FW and FA-t maps. WMH masks were segmented on FLAIR and T1-weighted MRI and dilated repeatedly to create 8 NAWM masks representing increasing distance from WMH. Linear models were applied to compare FW and FA-t across WMH and NAWM masks and in association with cerebrovascular risk. Median age was 64 ± 14 years. FW and FA-t were altered 8 mm and 12 mm beyond WMH, respectively. Smoking was significantly associated with FW in NAWM (p = 0.008) and FA-t in WMH (p = 0.008) and in NAWM (p = 0.003) while diabetes and hypertension were not. Further research is necessary to examine whether FW and FA-t alterations in NAWM are predictors for developing WMH.
BACKGROUND: Despite improvement in treatment strategies of atrial fibrillation (AF), a considerable number of patients still experience recurrence of atrial tachyarrhythmia (ATA) following catheter ablation (CA). This study aimed to investigate the prognostic value of left atrial (LA) deformation analysis in a large group of patients undergoing CA for AF. METHODS: This study included 678 patients with AF. Echocardiography including two-dimensional speckle tracking echocardiography (2DSTE) was performed in all patients prior to CA. Logistic regression analysis was used to assess the association between ATA recurrence and LA strain during reservoir phase (LASr), LA strain during contraction phase (LASct), and LA strain during conduit phase (LAScd). RESULTS: During one-year follow-up, 274 (40%) experienced ATA recurrence. Median age of the included study population was 63.2 years (IQR: 55.5, 69.5) and 485 (72%) were male. Patients with recurrence had lower LASr (22.6% vs. 25.1%, p = 0.001) and LASct (10.7% vs. 12.4%, p < 0.001). No difference in LAScd was observed. After adjusting for potential clinical and echocardiographic confounders LASr (OR = 1.04, CI95% [1.01; 1.07], p = 0.015, per 1% decrease) and LASct (OR = 1.06, CI95% [1.02; 1.11], p = 0.007, per 1% decrease) remained independent predictors of recurrence. However, in patients with a normal-sized LA (LA volume index<34 mL/m2), only LASct remained an independent predictor of recurrence (OR = 1.07, CI95% [1.01; 1.12], p = 0.012, per 1% decrease). CONCLUSION: In patients undergoing CA for AF, LA deformation analysis by 2DSTE could be of use in risk stratification in clinical practice regarding ATA recurrence, even in patients with a normal-sized LA.
BACKGROUND: Chronic obstructive pulmonary disease (COPD) and idiopathic pulmonary fibrosis (IPF) are characterized by shared exposures and clinical features, but distinct genetic and pathologic features exist. These features have not been well-studied using large-scale gene expression datasets. We hypothesized that there are divergent gene, pathway, and cellular signatures between COPD and IPF. METHODS: We performed RNA-sequencing on lung tissues from individuals with IPF (n = 231) and COPD (n = 377) compared to control (n = 267), defined as individuals with normal spirometry. We grouped the overlapping differential expression gene sets based on direction of expression and examined the resultant sets for genes of interest, pathway enrichment, and cell composition. Using gene set variation analysis, we validated the overlap group gene sets in independent COPD and IPF data sets. RESULTS: We found 5010 genes differentially expressed between COPD and control, and 11,454 genes differentially expressed between IPF and control (1% false discovery rate). 3846 genes overlapped between IPF and COPD. Several pathways were enriched for genes upregulated in COPD and downregulated in IPF; however, no pathways were enriched for genes downregulated in COPD and upregulated in IPF. There were many myeloid cell genes with increased expression in COPD but decreased in IPF. We found that the genes upregulated in COPD but downregulated in IPF were associated with lower lung function in the independent validation cohorts. CONCLUSIONS: We identified a divergent gene expression signature between COPD and IPF, with increased expression in COPD and decreased in IPF. This signature is associated with worse lung function in both COPD and IPF.
Patients with fibrous dysplasia (FD) often present with craniofacial lesions that affect the trigeminal nerve system. Debilitating pain, headache, and migraine are frequently experienced by FD patients with poor prognosis, while some individuals with similar bone lesions are asymptomatic. The clinical and biological factors that contribute to the etiopathogenesis of pain in craniofacial FD are largely unknown. We present two adult females with comparable craniofacial FD lesion size and location, as measured by 18F-sodium fluoride positron emission tomography/computed tomography (PET/CT), yet their respective pain phenotypes differed significantly. Over 4 weeks, the average pain reported by Patient A was 0.4/0-10 scale. Patient B reported average pain of 7.8/0-10 scale distributed across the entire skull and left facial region. Patient B did not experience pain relief from analgesics or more aggressive treatments (denosumab). In both patients, evaluation of trigeminal nerve divisions (V1, V2, and V3) with CT and magnetic resonance imaging (MRI) revealed nerve compression and displacement with more involvement of the left trigeminal branches relative to the right. First-time employment of diffusion MRI and tractography suggested reduced apparent fiber density within the cisternal segment of the trigeminal nerve, particularly for Patient B and in the left hemisphere. These cases highlight heterogeneous clinical presentation and neurobiological properties in craniofacial FD and also, the disconnect between peripheral pathology and pain severity. We hypothesize that a detailed phenotypic characterization of patients that incorporates an advanced imaging approach probing the trigeminal system may provide enhanced insights into the variable experiences with pain in craniofacial FD.
Almost 25% of all older adults experience difficulty walking. Mobility difficulties for older adults are more pronounced when they perform a simultaneous cognitive task while walking (i.e., dual task walking). Although it is known that aging results in widespread brain atrophy, few studies have integrated across more than one neuroimaging modality to comprehensively examine the structural neural correlates that may underlie dual task walking in older age. We collected spatiotemporal gait data during single and dual task walking for 37 young (18-34 years) and 23 older adults (66-86 years). We also collected T 1-weighted and diffusion-weighted MRI scans to determine how brain structure differs in older age and relates to dual task walking. We addressed two aims: (1) to characterize age differences in brain structure across a range of metrics including volumetric, surface, and white matter microstructure; and (2) to test for age group differences in the relationship between brain structure and the dual task cost (DTcost) of gait speed and variability. Key findings included widespread brain atrophy for the older adults, with the most pronounced age differences in brain regions related to sensorimotor processing. We also found multiple associations between regional brain atrophy and greater DTcost of gait speed and variability for the older adults. The older adults showed a relationship of both thinner temporal cortex and shallower sulcal depth in the frontal, sensorimotor, and parietal cortices with greater DTcost of gait. Additionally, the older adults showed a relationship of ventricular volume and superior longitudinal fasciculus free-water corrected axial and radial diffusivity with greater DTcost of gait. These relationships were not present for the young adults. Stepwise multiple regression found sulcal depth in the left precentral gyrus, axial diffusivity in the superior longitudinal fasciculus, and sex to best predict DTcost of gait speed, and cortical thickness in the superior temporal gyrus to best predict DTcost of gait variability for older adults. These results contribute to scientific understanding of how individual variations in brain structure are associated with mobility function in aging. This has implications for uncovering mechanisms of brain aging and for identifying target regions for mobility interventions for aging populations.
We present a novel weakly-supervised framework for classifying whole slide images (WSIs). WSIs, due to their gigapixel resolution, are commonly processed by patch-wise classification with patch-level labels. However, patch-level labels require precise annotations, which is expensive and usually unavailable on clinical data. With image-level labels only, patch-wise classification would be sub-optimal due to inconsistency between the patch appearance and image-level label. To address this issue, we posit that WSI analysis can be effectively conducted by integrating information at both high magnification (local) and low magnification (regional) levels. We auto-encode the visual signals in each patch into a latent embedding vector representing local information, and down-sample the raw WSI to hardware-acceptable thumbnails representing regional information. The WSI label is then predicted with a Dual-Stream Network (DSNet), which takes the transformed local patch embeddings and multi-scale thumbnail images as inputs and can be trained by the image-level label only. Experiments conducted on three large-scale public datasets demonstrate that our method outperforms all recent state-of-the-art weakly-supervised WSI classification methods.