T1-weighted MRI Data Research Articles

Brain connectivity disruptions in PTSD related to early adversity: a multimodal neuroimaging study

ABSTRACT Background: Post-traumatic stress disorder (PTSD) is increasingly prevalent in individuals with adverse childhood experiences (ACE). However, the underlying neurobiology of ACE-related PTSD remains unclear. Objective: The present study investigated the brain connectivity in ACE-related PTSD using multimodal neuroimaging data. Methods: Using a total of 119 participants with ACE (70 with ACE-related PTSD and 49 ACE-exposed controls), this study acquired T1-weighted MRI, diffusion-weighted MRI, and resting-state fMRI data to examine structural and functional connectivity between groups. Joint connectivity matrix independent component analysis (Jcm-ICA) was employed to allow shared information from all modalities to be examined and assess structural and functional connectivity differences between groups. Results: Jcm-ICA revealed distinct connectivity alterations in key brain regions involved in cognitive control, self-referential processing, and social behaviour. Compared to controls, the PTSD group exhibited functional hyperconnectivity of the right medial prefrontal cortex (PFC) of the default mode network and right inferior temporal cortex, and functional hypoconnectivity in the lateral-PFC of the central executive network and structural hypoconnectivity in white matter pathways including the right orbitofrontal region (OFC) linked to social behaviour. Post-hoc analyses using the joint brain-based information revealed that the severity of ACE, the number of traumas, and PTSD symptoms later in life significantly predicted the effects of ACE-related PTSD on the brain. Notably, no direct association between brain connectivity alterations and PTSD symptoms or the number of traumas within the PTSD group was observed. Conclusion: This study offers novel insights into the neurobiology of ACE-related PTSD using multimodal data fusion. We identified alterations in key brain networks (DMN, CEN) and OFC, suggesting potential deficits in cognitive control and social behaviour alongside heightened emotional processing in individuals with PTSD. Furthermore, our findings highlight the combined influence of ACE exposure, number of traumas experienced, and PTSD severity on brain connectivity disruptions, potentially informing future interventions.

Comparison of cortical gyrification patterns in patients with panic disorder with and without comorbid generalized anxiety disorder

BackgroundLower functioning and higher symptom severity are observed when panic disorder (PD) co-occurs with generalized anxiety disorder (PD + GAD). No research on cortical gyrification patterns in the PD + GAD group has been conducted to date, which could show the alterations in brain connectivity in the extended fear network (EFN). This study aimed to investigate the characteristics of cortical gyrification in the PD + GAD group, compared to that in the PD without comorbid GAD (PD-GAD) group. MethodsThis study included 90 patients with PD, with propensity score matching between the PD + GAD (n = 30) and PD-GAD groups (n = 60), and 65 healthy controls (HC). For clinical evaluation, we assessed the anxiety symptomatology, suicidality, and harm avoidance. The local gyrification index (LGI) was obtained from T1-weighted brain MRI data using FreeSurfer. ResultsIn the PD group compared to the HC, the hypergyrification involved the EFN. In the PD + GAD group compared to the PD-GAD group, hypergyrification was shown in the pathological worry-related brain regions. Within the PD + GAD group, significant positive correlations were observed between the superior frontal gyrus LGI values and suicidality scores, as well as between the superior parietal gyrus LGI values and harm avoidance levels. LimitationsGiven the variability in cortical gyrification patterns, longitudinal studies are needed to assess the occurrence of hypergyrification in specific brain regions. ConclusionsThis study is the first to demonstrate cortical gyrification patterns in the PD + GAD group compared to those in the PD-GAD group. Notably, the EFN and pathological worry-related brain regions have been implicated in the pathology of PD + GAD.

Neuroanatomical heterogeneity drives divergent cognitive and motor trajectories in Parkinson's disease subtypes

IntroductionCognitive symptoms of Parkinson's disease (PD) may initially present subtly, often overshadowed by more noticeable motor symptoms. However, as PD progresses, predicting which individuals will experience significant cognitive decline becomes challenging due to variability, suggesting distinct PD subtypes with varying cognitive trajectories. This study aimed to identify early PD subtypes based on patterns of gray matter atrophy in brain regions associated with cognition and assess their distinct patterns of cognitive change over time. Recognizing PD primarily as a movement disorder, we also evaluated their motor symptoms. MethodsWe analyzed T1-weighted MRI data, cognitive, and motor scores from 114 de novo PD patients and 120 healthy controls. Multivariate gray matter volumetric distances (MGMV) across frontal, subcortical, parietal, temporal, and occipital regions were computed, and K-means clustering was used to identify PD subtypes. Subsequently, cognitive assessments were compared between subtypes at baseline and 48 months using linear mixed-effects models and reliable change indices. Motor-symptom changes were assessed using linear mixed-effects models. ResultsTwo PD subtypes were identified from baseline MRI. Subtype 1 showed significantly higher MGMV in frontal (p < 0.001) and subcortical (p < 0.001) regions, indicating atrophy. At 48 months, subtype 1 had poorer global cognitive performance than subtype 2 (p = 0.005) and faster progression of postural instability and gait disturbance (p = 0.04). ConclusionsPD subtypes identified early by distinct frontal and subcortical atrophy patterns exhibited divergent trajectories of cognitive decline and worsening motor symptoms over time, underscoring the neuroanatomical heterogeneity that drives clinical variability in PD.

Deep Learning-Based Segmentation and Quantitative Analysis of Lumbar Paraspinal Muscles and Fat Infiltration in Multicenter T2-Weighted MRI Data

Deep Learning-Based Segmentation and Quantitative Analysis of Lumbar Paraspinal Muscles and Fat Infiltration in Multicenter T2-Weighted MRI Data

Imaging tumor and ascites-associated macrophages in a mouse model of metastatic ovarian cancer

BackgroundTumor-Associated Macrophages (TAMs) play a critical role in the pathogenesis and progression of ovarian cancer, a lethal gynecologic malignancy. [124I]iodo-DPA-713 is a PET radiotracer that is selectively trapped within reactive macrophages. We have employed this radioligand here as well as a fluorescent analog to image TAMs associated with primary tumors, secondary pulmonary metastases and gastrointestinal tract-associated macrophages, associated with ascites accumulation in a syngeneic mouse model of metastatic ovarian cancer. Intact female C57BL/6 mice were engrafted with ID8-Defb29-VEGF tumor pieces. One month after engraftment, the mice were selected for positive bioluminescence to show primary and secondary tumor burden and were then scanned by PET/MRI with [124I]iodo-DPA-713, observing a 24 h uptake time. PET data were overlayed with T2-weighted MRI data to facilitate PET uptake tissue identity. Additionally, mice were imaged ex vivo using Near IR Fluorescence (NIRF), capturing the uptake and sequestration of DPA-713-IRDye800CW, a fluorescent analog of the radioligand used here. Additionally, cell culture uptake of DPA-713-IRDye680LT in ID8-DEFb29-VEGF, IOSE hTERT and RAW264.7 cells was conducted to measure tracer uptake in ovarian cancer cells, ovarian epithelial cells and macrophage.ResultsPET/MRI data show an intense ring of radiotracer uptake surrounding primary tumors. PET uptake is also associated with lung metastases, but not healthy lung. Mice displaying ascites also display PET uptake along the gastrointestinal tract while sham-operated mice show minimal gastrointestinal uptake. All mice show specific kidney uptake. Mice imaged by NIRF confirmed TAMs uptake mostly at the rim of primary tumors while 1 mm secondary tumors in the lungs displayed robust, homogeneous uptake of the radio- and fluorescent analog. Ex vivo biodistribution of [124I]iodo-DPA-713 showed that contralateral ovaries in middle-stage disease had the highest probe uptake with tissues sampled in mid- and late-stage disease showing increasing uptake.Conclusion[124I]iodo-DPA-713 and DPA-713-IRDye800CW sensitively identify and locate TAMs in a syngeneic mouse model of metastatic ovarian cancer.

A worldwide study of subcortical shape as a marker for clinical staging in Parkinson’s disease

Alterations in subcortical brain regions are linked to motor and non-motor symptoms in Parkinson’s disease (PD). However, associations between clinical expression and regional morphological abnormalities of the basal ganglia, thalamus, amygdala and hippocampus are not well established. We analyzed 3D T1-weighted brain MRI and clinical data from 2525 individuals with PD and 1326 controls from 22 global sources in the ENIGMA-PD consortium. We investigated disease effects using mass univariate and multivariate models on the medial thickness of 27,120 vertices of seven bilateral subcortical structures. Shape differences were observed across all Hoehn and Yahr (HY) stages, as well as correlations with motor and cognitive symptoms. Notably, we observed incrementally thinner putamen from HY1, caudate nucleus and amygdala from HY2, hippocampus, nucleus accumbens, and thalamus from HY3, and globus pallidus from HY4–5. Subregions of the thalami were thicker in HY1 and HY2. Largely congruent patterns were associated with a longer time since diagnosis and worse motor symptoms and cognitive performance. Multivariate regression revealed patterns predictive of disease stage. These cross-sectional findings provide new insights into PD subcortical degeneration by demonstrating patterns of disease stage-specific morphology, largely consistent with ongoing degeneration.

Predicting brain age with global-local attention network from multimodal neuroimaging data: Accuracy, generalizability, and behavioral associations

Brain age, an emerging biomarker for brain diseases and aging, is typically predicted using single-modality T1-weighted structural MRI data. This study investigates the benefits of integrating structural MRI with diffusion MRI to enhance brain age prediction. We propose an attention-based deep learning model that fuses global-context information from structural MRI with local details from diffusion metrics. The model was evaluated using two large datasets: the Human Connectome Project (HCP, n = 1064, age 22–37) and the Cambridge Center for Aging and Neuroscience (Cam-CAN, n = 639, age 18–88). It was tested for generalizability and robustness on three independent datasets (n = 546, age 20–86), reproducibility on a test-retest dataset (n = 44, age 22–35), and longitudinal consistency (n = 129, age 46–92). We also examined the relationship between predicted brain age and behavioral measures. Results showed that the multimodal model improved prediction accuracy, achieving mean absolute errors (MAEs) of 2.44 years in the HCP dataset (sagittal plane) and 4.36 years in the Cam-CAN dataset (axial plane). The corresponding R2 values were 0.258 and 0.914, respectively, reflecting the model's ability to explain variance in the predictions across both datasets. Compared to single-modality models, the multimodal approach showed better generalization, reducing MAEs by 10–76 % and enhancing robustness by 22–82 %. While the multimodal model exhibited superior reproducibility, the sMRI model showed slightly better longitudinal consistency. Importantly, the multimodal model revealed unique associations between predicted brain age and behavioral measures, such as walking endurance and loneliness in the HCP dataset, which were not detected with chronological age alone. In the Cam-CAN dataset, brain age and chronological age exhibited similar correlations with behavioral measures. By integrating sMRI and dMRI through an attention-based model, our proposed approach enhances predictive accuracy and provides deeper insights into the relationship between brain aging and behavior.

Automated Vertebral Bone Quality Determination from T1-Weighted Lumbar Spine MRI Data Using a Hybrid Convolutional Neural Network–Transformer Neural Network

Vertebral bone quality (VBQ) is a promising new method that can improve screening for osteoporosis. The drawback of the current method is that it requires manual determination of the regions of interest (ROIs) of vertebrae and cerebrospinal fluid (CSF) by a radiologist. In this work, an automatic method for determining the VBQ is proposed, in which the ROIs are obtained using a trained neural network model. A large, publicly available dataset of sagittal lumbar spine MRI images with ground truth segmentations was used to train a BRAU-Net++ hybrid CNN–transformer neural network. The performance of the trained model was evaluated using the dice similarity coefficient (DSC), accuracy, precision, recall and intersection-over-union (IoU) metrics. The trained model performed similarly to state-of-the-art lumbar spine segmentation models, with an average DSC value of 0.914 ± 0.007 for the vertebrae and 0.902 for the spinal canal. Four different methods of VBQ determination with automatic segmentation are presented and compared with one-way ANOVA. These methods use different algorithms for CSF extraction from the segmentation of the spinal canal using T1- and T2-weighted image data and applying erosion to the vertebral ROI to avoid a sharp change in SI at the edge of the vertebral body.

A novel MSN-II feature extracted from T1-weighted MRI for discriminating between BD patients and MDD patients.

Differentiating between patients with bipolar disorder (BD) and major depressive disorder (MDD) is clinically challenging. This study aimed to explore the potential of radiomic textural features for discriminating BD and MDD. A total 253 subjects (114 patients with BD, 139 patients with MDD) with T1-weighted MRI data were recruited. Radiomics features and gray matter volume (GMV) features were extracted from each brain region. A novel high-level MSN_II feature method based on radiomic features was proposed. And a total of 21 MSN features (5 MSN_I and 16 MSN_II) based on different combinations of the 5 types of radiomic textural feature were calculated. Classification models were constructed using various combinations of MSNs or GMV, and their performance and stability was evaluated through 2000 repeated experiments. The model built with combined features (GMV and GMV + MSN_II_GLCM_GLSZM_NGTDM) showed the best classification performance (AUC = 0.896±0.058, ACC = 0.831±0.064) in the validation cohort. After MANOVA analysis and FDR correlation, the MSN_II_GLCM_GLSZM_NGTDM values in 4 regions (right rectus gyrus, right temporal pole: middle temporal gyrus, Vermis3 and Vermis10) showed significant difference between BD and MDD. The main limitation of this study is that the data is derived from a single center without an external independent test set. Incorporating the high-level MSN_II based on radiomics features can improve the classification performance compared to models solely relying on GMV features alone. This result implied the potential application of the proposed high level MSN method and radiomics textural features on the MDD and BD clinical studies.

Multimodal brain age indicators of internalising problems in early adolescence: A longitudinal investigation

BackgroundAdolescence is a time of increased risk for the onset of internalising problems, particularly in females. However, how individual differences in brain maturation relate to the increased vulnerability for internalising problems in adolescence remains poorly understood due to a scarcity of longitudinal studies. MethodsUsing Adolescent Brain Cognitive Development (ABCD) Study data, we examined longitudinal associations between multimodal brain age and youth internalising problems. Brain age models were trained, validated, and tested independently on T1-weighted (T1; N=9523), diffusion tensor (DTI; N=8834), and resting-state functional (rs-fMRI; N=8233) MRI data at baseline (Mage= 9.9 years) and 2-year follow-up (Mage= 11.9 years). Self-reported internalising problems were measured at 3-year follow-up (Mage= 12.9 years) using the Brief Problem Monitor. ResultsLatent change score models demonstrated that although brain age gap (BAG) at baseline was not related to later internalising problems, an increase in BAG between timepoints was positively associated with internalising problems at 3-year follow-up in females but not males. This association between an increasing BAG and higher internalising problems was observed in the T1 (β = 0.067, SE = 0.050, pFDR = 0.020) and rs-fMRI β = 0.090, SE = 0.025, pFDR = 0.007) models but not DTI (β=-0.002, SE=0.053, pFDR = 0.932), and remained significant when accounting for earlier internalising problems. ConclusionsA greater increase in BAG in early adolescence may reflect the heightened vulnerability shown by female youth to internalising problems. Longitudinal research is necessary to understand if this increasing BAG signifies accelerated brain development and its relationship to the trajectory of internalising problems throughout adolescence.

Structural alterations in a rumination-related network in patients with major depressive disorder

Rumination is a common symptom in major depressive disorder (MDD). Previous work has connected individual differences in rumination to structural properties in various brain regions. Some of these, such as the dorsolateral prefrontal cortex (dlPFC), have also been highlighted as being altered in MDD, suggesting a connection between structural changes and ruminative symptoms. Although informative, such localised relations have limitations in the context of a network view of the brain. To further investigate rumination-related structural changes in depression, and to situate these within potential functional networks, we acquired T1-weighted structural MRI data from patients with MDD (n = 32) and controls (n = 69). Rumination was measured with the Rumination Response Scale. Surface-based, whole-brain analysis of cortical grey-matter identified group differences in the dlPFC that were, however, not related to rumination. Instead, rumination was correlated with grey-matter properties in the right precuneus. Using normative functional connectivity analysis on an independent sample (n = 100), we show these two regions to be interconnected. Further developing a network-based perspective, it was shown that the rumination-related precuneus region is connected with networks associated with processes such as executive function, autobiographical memory, and visual perception. Notably, these processes have been connected to rumination. These results suggest that rumination in depression may be linked to focal structural changes. The effects of these focal changes on rumination may then be connected to their influence on distributed functional networks.

Functional connectivity compensation in Parkinson's disease with freezing of gait.

Freezing of gait (FOG) is a disabling motor symptom prevalent in patients with Parkinson's disease (PD); however, its pathophysiological mechanisms are poorly understood. This study aimed to investigate whole-brain functional connectivity (FC) pattern alterations in PD patients with FOG. A total of 18 PD patients, 10 with FOG (PD-FOG) and 8 without FOG (PD-nFOG), and 10 healthy controls were enrolled. High-resolution 3D T1-weighted and resting-state functional MRI (rs-fMRI) data were obtained from all participants. The groups' internetwork connectivity differences were explored with rs-fMRI FC using seed-based analysis and graph theory. Multiple linear regression analysis estimated the relationship between FC changes and clinical measurements. Rs-fMRI analysis demonstrated alterations in FC in various brain regions between the three groups. Freezing of Gait Questionnaire severity was correlated with decreased brain functional connection between Vermis12 and the left temporal occipital fusiform cortex (r = -0.82, P < .001). Graph theory topological metrics indicated a decreased clustering coefficient in the right superior temporal gyrus in the PD-nFOG group. PD-FOG patients exhibited a compensatory increase in connectivity between the left inferior frontal gyrus language network and the postcentral gyrus compared to PD-nFOG patients. Further, the decreased connection between Vermis 12 and the left temporal occipital fusiform cortex may serve as a potential neuroimaging biomarker for tracking PD-FOG and distinguishing between PD subtypes.

Involvement of the cerebellum in structural connectivity enhancement in episodic migraine

BackgroundThe pathophysiology of migraine remains poorly understood, yet a growing number of studies have shown structural connectivity disruptions across large-scale brain networks. Although both structural and functional changes have been found in the cerebellum of migraine patients, the cerebellum has barely been assessed in previous structural connectivity studies of migraine. Our objective is to investigate the structural connectivity of the entire brain, including the cerebellum, in individuals diagnosed with episodic migraine without aura during the interictal phase, compared with healthy controls.MethodsTo that end, 14 migraine patients and 15 healthy controls were recruited (all female), and diffusion-weighted and T1-weighted MRI data were acquired. The structural connectome was estimated for each participant based on two different whole-brain parcellations, including cortical and subcortical regions as well as the cerebellum. The structural connectivity patterns, as well as global and local graph theory metrics, were compared between patients and controls, for each of the two parcellations, using network-based statistics and a generalized linear model (GLM), respectively. We also compared the number of connectome streamlines within specific white matter tracts using a GLM.ResultsWe found increased structural connectivity in migraine patients relative to healthy controls with a distinct involvement of cerebellar regions, using both parcellations. Specifically, the node degree of the posterior lobe of the cerebellum was greater in patients than in controls and patients presented a higher number of streamlines within the anterior limb of the internal capsule. Moreover, the connectomes of patients exhibited greater global efficiency and shorter characteristic path length, which correlated with the age onset of migraine.ConclusionsA distinctive pattern of heightened structural connectivity and enhanced global efficiency in migraine patients compared to controls was identified, which distinctively involves the cerebellum. These findings provide evidence for increased integration within structural brain networks in migraine and underscore the significance of the cerebellum in migraine pathophysiology.

Investigation of ComBat Harmonization on Radiomic and Deep Features from Multi-Center Abdominal MRI Data.

ComBat harmonization has been developed to remove non-biological variations for data in multi-center research applying artificial intelligence (AI). We investigated the effectiveness of ComBat harmonization on radiomic and deep features extracted from large, multi-center abdominal MRI data. A retrospective study was conducted on T2-weighted (T2W) abdominal MRI data retrieved from individual patients with suspected or known chronic liver disease at three study sites. MRI data were acquired using systems from three manufacturers and two field strengths. Radiomic features and deep features were extracted using the PyRadiomics pipeline and a Swin Transformer. ComBat was used to harmonize radiomic and deep features across different manufacturers and field strengths. Student's t-test, ANOVA test, and Cohen's F score were applied to assess the difference in individual features before and after ComBat harmonization. Between two field strengths, 76.7%, 52.9%, and 26.7% of radiomic features, and 89.0%, 56.5%, and 0.1% of deep features from three manufacturers were significantly different. Among the three manufacturers, 90.1% and 75.0% of radiomic features and 89.3% and 84.1% of deep features from two field strengths were significantly different. After ComBat harmonization, there were no significant differences in radiomic and deep features among manufacturers or field strengths based on t-tests or ANOVA tests. Reduced Cohen's F scores were consistently observed after ComBat harmonization. ComBat harmonization effectively harmonizes radiomic and deep features by removing the non-biological variations due to system manufacturers and/or field strengths in large multi-center clinical abdominal MRI datasets.

Assessing the performance of AI-assisted technicians in liver segmentation, Couinaud division, and lesion detection: a pilot study

BackgroundIn patients with primary and secondary liver cancer, the number and sizes of lesions, their locations within the Couinaud segments, and the volume and health status of the future liver remnant are key for informing treatment planning. Currently this is performed manually, generally by trained radiologists, who are seeing an inexorable growth in their workload. Integrating artificial intelligence (AI) and non-radiologist personnel into the workflow potentially addresses the increasing workload without sacrificing accuracy. This study evaluated the accuracy of non-radiologist technicians in liver cancer imaging compared with radiologists, both assisted by AI.MethodsNon-contrast T1-weighted MRI data from 18 colorectal liver metastasis patients were analyzed using an AI-enabled decision support tool that enables non-radiology trained technicians to perform key liver measurements. Three non-radiologist, experienced operators and three radiologists performed whole liver segmentation, Couinaud segment segmentation, and the detection and measurements of lesions aided by AI-generated delineations. Agreement between radiologists and non-radiologists was assessed using the intraclass correlation coefficient (ICC). Two additional radiologists adjudicated any lesion detection discrepancies.ResultsWhole liver volume showed high levels of agreement between the non-radiologist and radiologist groups (ICC = 0.99). The Couinaud segment volumetry ICC range was 0.77–0.96. Both groups identified the same 41 lesions. As well, the non-radiologist group identified seven more structures which were also confirmed as lesions by the adjudicators. Lesion diameter categorization agreement was 90%, Couinaud localization 91.9%. Within-group variability was comparable for lesion measurements.ConclusionWith AI assistance, non-radiologist experienced operators showed good agreement with radiologists for quantifying whole liver volume, Couinaud segment volume, and the detection and measurement of lesions in patients with known liver cancer. This AI-assisted non-radiologist approach has potential to reduce the stress on radiologists without compromising accuracy.Graphical abstract

Structural and functional features characterizing the brains of individuals with higher controllability of motor imagery

Motor imagery is a higher-order cognitive brain function that mentally simulates movements without performing the actual physical one. Although motor imagery has attracted the interest of many researchers, and mental practice utilizing motor imagery has been widely used in sports training and post-stroke rehabilitation, neural bases that determine individual differences in motor imagery ability are not well understood. In this study, using controllability of motor imagery (CMI) test that can objectively evaluate individual ability to manipulate one’s imaginary postures, we examined structural and functional features characterizing the brains of individuals with higher controllability of motor imagery, by analyzing T1-weighted structural MRI data obtained from 89 participants and functional MRI data obtained from 28 of 89 participants. The higher CMI test scorers had larger volume in the bilateral superior frontoparietal white matter regions. The CMI test activated the bilateral dorsal premotor cortex (PMD) and superior parietal lobule (SPL); specifically, the left PMD and/or the right SPL enhanced functional coupling with the visual body, somatosensory, and motor/kinesthetic areas in the higher scorers. Hence, controllability of motor imagery is higher for those who well-develop superior frontoparietal network, and for those whose this network accesses these sensory areas to predict the expected multisensory experiences during motor imagery. This study elucidated for the first time the structural and functional features characterizing the brains of individuals with higher controllability of motor imagery, and advanced understanding of individual differences in motor imagery ability.

Translating phenotypic prediction models from big to small anatomical MRI data using meta-matching

Abstract Individualized phenotypic prediction based on structural MRI is an important goal in neuroscience. Prediction performance increases with larger samples, but small-scale datasets with fewer than 200 participants are often unavoidable. We have previously proposed a “meta-matching” framework to translate models trained from large datasets to improve the prediction of new unseen phenotypes in small collection efforts. Meta-matching exploits correlations between phenotypes, yielding large improvement over classical machine learning when applied to prediction models using resting-state functional connectivity as input features. Here, we adapt the two best performing meta-matching variants (“meta-matching finetune” and “meta-matching stacking”) from our previous study to work with T1-weighted MRI data by changing the base neural network architecture to a 3D convolution neural network. We compare the two meta-matching variants with elastic net and classical transfer learning using the UK Biobank (N = 36,461), Human Connectome Project Young Adults (HCP-YA) dataset (N = 1,017) and HCP-Aging dataset (N = 656). We find that meta-matching outperforms elastic net and classical transfer learning by a large margin, both when translating models within the same dataset, as well as translating models across datasets with different MRI scanners, acquisition protocols and demographics. For example, when translating a UK Biobank model to 100 HCP-YA participants, meta-matching finetune yielded a 136% improvement in variance explained over transfer learning, with an average absolute gain of 2.6% (minimum = -0.9%, maximum = 17.6%) across 35 phenotypes. Overall, our results highlight the versatility of the meta-matching framework.

Enhancing Classification of Alzheimer’s Disease using Spatial Attention Mechanism

Aim This study aims to enhance the precision of Alzheimer's disease (AD) detection by integrating Spatial Attention Mechanism into a Convolutional Neural Network (CNN) architecture. Background Alzheimer's disease is a progressive neurodegenerative disorder characterized by abnormal protein deposits in the brain, leading to nerve cell loss and posing a significant global health challenge. Early and accurate detection is crucial for disease management and treatment due to the lack of a cure and the disease's severe progression. Objective The objective of this research is to improve the accuracy of Alzheimer's disease classification using MRI data by implementing a Spatial Attention Mechanism in a CNN architecture. Methods The study utilized T1-weighted MRI data from the OASIS 1 and OASIS 2 datasets. The key innovation is the Spatial Attention layer incorporated within a CNN model, which computes the average of each channel in the input feature map. This layer guides subsequent layers to focus on critical brain regions, enhancing the model's accuracy in differentiating between Alzheimer's disease stages. Results The model achieved a validation accuracy of 99.69% with a sensitivity and specificity of 1.0000, demonstrating its reliability in distinguishing between different stages of Alzheimer's disease. The adaptability of the Spatial Attention layer allows the model to assign higher weights to crucial brain regions, improving its discriminative power. Conclusion The integration of the Spatial Attention Mechanism into the CNN architecture significantly contributes to the early detection of Alzheimer's disease, enabling timely interventions. This innovative approach has the potential to revolutionize Alzheimer's diagnosis by enhancing accuracy and offering a robust solution for classification.

Muscle texture features on preoperative MRI for diagnosis and assessment of severity of congenital muscular torticollis

ObjectivesTo develop an objective method based on texture analysis on MRI for diagnosis of congenital muscular torticollis (CMT).Material and methodsThe T1- and T2-weighted imaging, Q-dixon, and T1-mapping MRI data of 38 children with CMT were retrospectively analyzed. The region of interest (ROI) was manually drawn at the level of the largest cross-sectional area of the SCM on the affected side. MaZda software was used to obtain the texture features of the T2WI sequences of the ROI in healthy and affected SCM. A radiomics diagnostic model based on muscle texture features was constructed using logistic regression analysis. Fatty infiltration grade was calculated by hematoxylin and eosin staining, and fibrosis ratio by Masson staining. Correlation between the MRI parameters and pathological indicators was analyzed.ResultsThere was positive correlation between fatty infiltration grade and mean value, standard deviation, and maximum value of the Q-dixon sequence of the affected SCM (correlation coefficients, 0.65, 0.59, and 0.58, respectively, P < 0.05).Three muscle texture features—S(2,2)SumAverg, S(3,3)SumVarnc, and T2WI extreme difference—were selected to construct the diagnostic model. The model showed significant diagnostic value for CMT (P < 0.05). The area under the curve of the multivariate conditional logistic regression model was 0.828 (95% confidence interval 0.735–0.922); the sensitivity was 0.684 and the specificity 0.868.ConclusionThe radiomics diagnostic model constructed using T2WI muscle texture features and MRI signal values appears to have good diagnostic efficiency. Q-dixon sequence can reflect the fatty infiltration grade of CMT.

Principal component analysis as an efficient method for capturing multivariate brain signatures of complex disorders-ENIGMA study in people with bipolar disorders and obesity.

Multivariate techniques better fit the anatomy of complex neuropsychiatric disorders which are characterized not by alterations in a single region, but rather by variations across distributed brain networks. Here, we used principal component analysis (PCA) to identify patterns of covariance across brain regions and relate them to clinical and demographic variables in a large generalizable dataset of individuals with bipolar disorders and controls. We then compared performance of PCA and clustering on identical sample to identify which methodology was better in capturing links between brain and clinical measures. Using data from the ENIGMA-BD working group, we investigated T1-weighted structural MRI data from 2436 participants with BD and healthy controls, and applied PCA to cortical thickness and surface area measures. We then studied the association of principal components with clinical and demographic variables using mixed regression models. We compared the PCA model with our prior clustering analyses of the same data and also tested it in a replication sample of 327 participants with BD or schizophrenia and healthy controls. The first principal component, which indexed a greater cortical thickness across all 68 cortical regions, was negatively associated with BD, BMI, antipsychotic medications, and age and was positively associated with Li treatment. PCA demonstrated superior goodness of fit to clustering when predicting diagnosis and BMI. Moreover, applying the PCA model to the replication sample yielded significant differences in cortical thickness between healthy controls and individuals with BD or schizophrenia. Cortical thickness in the same widespread regional network as determined by PCA was negatively associated with different clinical and demographic variables, including diagnosis, age, BMI, and treatment with antipsychotic medications or lithium. PCA outperformed clustering and provided an easy-to-use and interpret method to study multivariate associations between brain structure and system-level variables. PRACTITIONER POINTS: In this study of 2770 Individuals, we confirmed that cortical thickness in widespread regional networks as determined by principal component analysis (PCA) was negatively associated with relevant clinical and demographic variables, including diagnosis, age, BMI, and treatment with antipsychotic medications or lithium. Significant associations of many different system-level variables with the same brain network suggest a lack of one-to-one mapping of individual clinical and demographic factors to specific patterns of brain changes. PCA outperformed clustering analysis in the same data set when predicting group or BMI, providing a superior method for studying multivariate associations between brain structure and system-level variables.