Cellular data extraction from multiplexed brain imaging data using self-supervised Dual-loss Adaptive Masked Autoencoder.

Reliable large-scale cell detection and segmentation is the fundamental first step to understanding biological processes in the brain. The ability to phenotype cells at scale can accelerate preclinical drug evaluation and system-level brain histology studies. The impressive advances in deep learning offer a practical solution to cell image detection and segmentation. Unfortunately, categorizing cells and delineating their boundaries for training deep networks is an expensive process that requires skilled biologists. This paper presents a novel self-supervised Dual-Loss Adaptive Masked Autoencoder (DAMA) for learning rich features from multiplexed immunofluorescence brain images. DAMA's objective function minimizes the conditional entropy in pixel-level reconstruction and feature-level regression. Unlike existing self-supervised learning methods based on a random image masking strategy, DAMA employs a novel adaptive mask sampling strategy to maximize mutual information and effectively learn brain cell data. To the best of our knowledge, this is the first effort to develop a self-supervised learning method for multiplexed immunofluorescence brain images. Our extensive experiments demonstrate that DAMA features enable superior cell detection, segmentation, and classification performance without requiring many annotations. In addition, to examine the generalizability of DAMA, we also experimented on TissueNet, a multiplexed imaging dataset comprised of two-channel fluorescence images from six distinct tissue types, captured using six different imaging platforms. Our code is publicly available at https://github.com/hula-ai/DAMA.

Prospective cohort study of long-term neurological outcomes in retired elite athletes: the Advanced BiomaRker, Advanced Imaging and Neurocognitive (BRAIN) Health Study protocol.

INTRODUCTION: Although limited, recent research suggests that contact sport participation might have an adverse long-term effect on brain health. Further work is required to determine whether this includes an increased risk of neurodegenerative disease and/or subsequent changes in cognition and behaviour. The Advanced BiomaRker, Advanced Imaging and Neurocognitive Health Study will prospectively examine the neurological, psychiatric, psychological and general health of retired elite-level rugby union and association football/soccer players. METHODS AND ANALYSIS: 400 retired athletes will be recruited (200 rugby union and 200 association football players, male and female). Athletes will undergo a detailed clinical assessment, advanced neuroimaging, blood testing for a range of brain health outcomes and neuropsychological assessment longitudinally. Follow-up assessments will be completed at 2 and 4 years after baseline visit. 60 healthy volunteers will be recruited and undergo an aligned assessment protocol including advanced neuroimaging, blood testing and neuropsychological assessment. We will describe the previous exposure to head injuries across the cohort and investigate relationships between biomarkers of brain injury and clinical outcomes including cognitive performance, clinical diagnoses and psychiatric symptom burden. ETHICS AND DISSEMINATION: Relevant ethical approvals have been granted by the Camberwell St Giles Research Ethics Committee (Ref: 17/LO/2066). The study findings will be disseminated through manuscripts in clinical/academic journals, presentations at professional conferences and through participant and stakeholder communications.

Normative volumes and relaxation times at 3T during brain development.

While research has unveiled and quantified brain markers of abnormal neurodevelopment, clinicians still work with qualitative metrics for MRI brain investigation. The purpose of the current article is to bridge the knowledge gap between case-control cohort studies and individual patient care. Here, we provide a unique dataset of seventy-three 3-to-17 years-old healthy subjects acquired with a 6-minute MRI protocol encompassing T1 and T2 relaxation quantitative sequence that can be readily implemented in the clinical setting; MP2RAGE for T1 mapping and the prototype sequence GRAPPATINI for T2 mapping. White matter and grey matter volumes were automatically quantified. We further provide normative developmental curves based on these two imaging sequences; T1, T2 and volume normative ranges with respect to age were computed, for each ROI of a pediatric brain atlas. This open-source dataset provides normative values allowing to position individual patients acquired with the same protocol on the brain maturation curve and as such provides potentially useful quantitative biomarkers facilitating precise and personalized care.

Automatic brain-tumor diagnosis using cascaded deep convolutional neural networks with symmetric U-Net and asymmetric residual-blocks.

The use of various kinds of magnetic resonance imaging (MRI) techniques for examining brain tissue has increased significantly in recent years, and manual investigation of each of the resulting images can be a time-consuming task. This paper presents an automatic brain-tumor diagnosis system that uses a CNN for detection, classification, and segmentation of glioblastomas; the latter stage seeks to segment tumors inside glioma MRI images. The structure of the developed multi-unit system consists of two stages. The first stage is responsible for tumor detection and classification by categorizing brain MRI images into normal, high-grade glioma (glioblastoma), and low-grade glioma. The uniqueness of the proposed network lies in its use of different levels of features, including local and global paths. The second stage is responsible for tumor segmentation, and skip connections and residual units are used during this step. Using 1800 images extracted from the BraTS 2017 dataset, the detection and classification stage was found to achieve a maximum accuracy of 99%. The segmentation stage was then evaluated using the Dice score, specificity, and sensitivity. The results showed that the suggested deep-learning-based system ranks highest among a variety of different strategies reported in the literature.

MRCT and CT in the diagnosis of pediatric disease imaging: assessing imaging performance and clinical effects.

OBJECTIVE: This study focused on analyzing the clinical value and effect of magnetic resonance imaging plus computed tomography (MRCT) and CT in the clinical diagnosis of cerebral palsy in children. METHODS: From February 2021 to April 2023, 94 children diagnosed with cerebral palsy were selected from our hospital for study subjects. These patients were divided into CT and MRI groups, with CT examination given to the CT group and MRI examination given to the MRI group. The positive rate of the two examination methods in the diagnosis of cerebral palsy was compared, different imaging signs in two groups of children with cerebral palsy were compared, and the diagnostic test typing results between two groups were further analyzed. RESULTS: The diagnostic positivity rate of the children in the MRI group was 91.49%, which was significantly higher than that of the children in the CT group (70.21%) (P < 0.05). In both groups, encephalomalacia, bilateral frontal subdural effusions, and gray-white matter atrophy of the brain were the main signs, and the difference in the proportion of these three imaging signs between the two groups was not significant (P > 0.05). Differences between the two groups examined for cerebral palsy subtypes were not significant (P > 0.05). CONCLUSION: The positive rate of pediatric cerebral palsy examined by MRI is higher than that of CT diagnosis, but the clinic should organically combine the two to further improve the detection validity and accuracy.

Quantitative analysis of apparent diffusion coefficients to predict neurological prognosis in cardiac arrest survivors: an observational derivation and internal-external validation study.

BACKGROUND: This study aimed to validate apparent diffusion coefficient (ADC) values and thresholds to predict poor neurological outcomes in out-of-hospital cardiac arrest (OHCA) survivors by quantitatively analysing the ADC values via brain magnetic resonance imaging (MRI). METHODS: This observational study used prospectively collected data from two tertiary academic hospitals. The derivation cohort comprised 70% of the patients randomly selected from one hospital, whereas the internal validation cohort comprised the remaining 30%. The external validation cohort used the data from another hospital, and the MRI data were restricted to scans conducted at 3 T within 72-96 h after an OHCA experience. We analysed the percentage of brain volume below a specific ADC value at 50-step intervals ranging from 200 to 1200 × 10-6 mm2/s, identifying thresholds that differentiate between good and poor outcomes. Poor neurological outcomes were defined as cerebral performance categories 3-5, 6 months after experiencing an OHCA. RESULTS: A total of 448 brain MRI scans were evaluated, including a derivation cohort (n = 224) and internal/external validation cohorts (n = 96/128, respectively). The proportion of brain volume with ADC values below 450, 500, 550, 600, and 650 × 10-6 mm2/s demonstrated good to excellent performance in predicting poor neurological outcomes in the derivation group (area under the curve [AUC] 0.89-0.91), and there were no statistically significant differences in performances among the derivation, internal validation, and external validation groups (all P > 0.5). Among these, the proportion of brain volume with an ADC below 600 × 10-6 mm2/s predicted a poor outcome with a 0% false-positive rate (FPR) and 76% (95% confidence interval [CI] 68-83) sensitivity at a threshold of > 13.2% in the derivation cohort. In both the internal and external validation cohorts, when using the same threshold, a specificity of 100% corresponded to sensitivities of 71% (95% CI 58-81) and 78% (95% CI 66-87), respectively. CONCLUSIONS: In this validation study, by consistently restricting the MRI types and timing during quantitative analysis of ADC values in brain MRI, we observed high reproducibility and sensitivity at a 0% FPR. Prospective multicentre studies are necessary to validate these findings.

Associations of plasma neurofilament light chain with cognition and neuroimaging measures in community-dwelling early old age men.

BACKGROUND: Plasma neurofilament light chain (NfL) is a promising biomarker of neurodegeneration with potential clinical utility in monitoring the progression of neurodegenerative diseases. However, the cross-sectional associations of plasma NfL with measures of cognition and brain have been inconsistent in community-dwelling populations. METHODS: We examined these associations in a large community-dwelling sample of early old age men (N = 969, mean age = 67.57 years, range = 61-73 years), who are either cognitively unimpaired (CU) or with mild cognitive impairment (MCI). Specifically, we investigated five cognitive domains (executive function, episodic memory, verbal fluency, processing speed, visual-spatial ability), as well as neuroimaging measures of gray and white matter. RESULTS: After adjusting for age, health status, and young adult general cognitive ability, plasma NfL level was only significantly associated with processing speed and white matter hyperintensity (WMH) volume, but not with other cognitive or neuroimaging measures. The association with processing speed was driven by individuals with MCI, as it was not detected in CU individuals. CONCLUSIONS: These results suggest that in early old age men without dementia, plasma NfL does not appear to be sensitive to cross-sectional individual differences in most domains of cognition or neuroimaging measures of gray and white matter. The revealed plasma NfL associations were limited to WMH for all participants and processing speed only within the MCI cohort. Importantly, considering cognitive status in community-based samples will better inform the interpretation of the relationships of plasma NfL with cognition and brain and may help resolve mixed findings in the literature.

An executive-functions-based reading training enhances sensory-motor systems integration during reading fluency in children with dyslexia.

The Simple View of Reading model suggests that intact language processing and word decoding lead to proficient reading comprehension, with recent studies pointing at executive functions as an important component contributing to reading proficiency. Here, we aimed to determine the underlying mechanism(s) for these changes. Participants include 120 8- to 12-year-old children (n = 55 with dyslexia, n = 65 typical readers) trained on an executive functions-based reading program, including pre/postfunctional MRI and behavioral data collection. Across groups, improved word reading was related to stronger functional connections within executive functions and sensory networks. In children with dyslexia, faster and more accurate word reading was related to stronger functional connections within and between sensory networks. These results suggest greater synchronization of brain systems after the intervention, consistent with the "neural noise" hypothesis in children with dyslexia and support the consideration of including executive functions as part of the Simple View of Reading model.

Case report: Atypical case of autoimmune glial fibrillary acidic protein astrocytopathy following COVID-19 vaccination refractory to immunosuppressive treatments.

A 54-year-old Japanese man presented with headache and fever the day after SARS-CoV-2 vaccination. He became deeply unconscious within a week. Brain MRI showed periventricular linear enhancements and a few spotty lesions in the cerebral white matter. Cerebrospinal fluid (CSF) testing showed mild pleocytosis. He was treated with intravenous methylprednisolone and plasma exchange. However, the white matter lesions enlarged to involve the brainstem and cerebellum, and long cord spinal lesions appeared. Anti-glial fibrillary acidic protein (GFAP) antibody was positive in the CSF and serum, and he was therefore diagnosed as autoimmune GFAP-astrocytopathy (GFAP-A). In addition, high-dose immunoglobulin therapy was administered twice, but his symptoms did not improve; the white matter lesions enlarged further, and modified Rankin Scale score increased to 5. A brain biopsy specimen showed infiltration of macrophages and CD4 + lymphocytes together with neuron and oligodendrocytic injuries and glial scar. Although GFAP-A generally responds well to steroids, the present case developed GFAP-A following SARS-CoV-2 vaccination, with refractory to intensive immunosuppressive therapy and atypical pathologic findings of infiltration of CD4 + lymphocytes and demyelination.

Brain structural alterations associated with impulsiveness in male violent patients with schizophrenia.

BACKGROUND: Violence in schizophrenia (SCZ) is a phenomenon associated with neurobiological factors. However, the neural mechanisms of violence in patients with SCZ are not yet sufficiently understood. Thus, this study aimed to explore the structural changes associated with the high risk of violence and its association with impulsiveness in patients with SCZ to reveal the possible neurobiological basis. METHOD: The voxel-based morphometry approach and whole-brain analyses were used to measure the alteration of gray matter volume (GMV) for 45 schizophrenia patients with violence (VSC), 45 schizophrenia patients without violence (NSC), and 53 healthy controls (HC). Correlation analyses were used to examine the association of impulsiveness and brain regions associated with violence. RESULTS: The results demonstrated reduced GMV in the right insula within the VSC group compared with the NSC group, and decreased GMV in the right temporal pole and left orbital part of superior frontal gyrus only in the VSC group compared to the HC group. Spearman correlation analyses further revealed a positive correlation between impulsiveness and GMV of the left superior temporal gyrus, bilateral insula and left medial orbital part of the superior frontal gyrus in the VSC group. CONCLUSION: Our findings have provided further evidence for structural alterations in patients with SCZ who had engaged in severe violence, as well as the relationship between the specific brain alterations and impulsiveness. This work provides neural biomarkers and improves our insight into the neural underpinnings of violence in patients with SCZ.

Distinct brain network organizations between club players and novices under different difficulty levels.

SIGNIFICANT: Chunk memory is one of the essential cognitive functions for high-expertise (HE) player to make efficient decisions. However, it remains unknown how the neural mechanisms of chunk memory processes mediate or alter chess players' performance when facing different opponents. AIM: This study aimed at inspecting the significant brain networks associated with chunk memory, which would vary between club players and novices. APPROACH: Functional networks and topological features of 20 club players (HE) and 20 novice players (LE) were compared at different levels of difficulty by means of functional near-infrared spectroscopy. RESULTS: Behavioral performance indicated that the club player group was unaffected by differences in difficulty. Furthermore, the club player group demonstrated functional connectivity among the dorsolateral prefrontal cortex, the frontopolar cortex, the supramarginal gyrus, and the subcentral gyrus, as well as higher clustering coefficients and lower path lengths in the high-difficulty task. CONCLUSIONS: The club player group illustrated significant frontal-parietal functional connectivity patterns and topological characteristics, suggesting enhanced chunking processes for improved chess performance.

Childhood maltreatment influences adult brain structure through its effects on immune, metabolic, and psychosocial factors.

Childhood maltreatment (CM) leads to a lifelong susceptibility to mental ill-health which might be reflected by its effects on adult brain structure, perhaps indirectly mediated by its effects on adult metabolic, immune, and psychosocial systems. Indexing these systemic factors via body mass index (BMI), C-reactive protein (CRP), and rates of adult trauma (AT), respectively, we tested three hypotheses: (H1) CM has direct or indirect effects on adult trauma, BMI, and CRP; (H2) adult trauma, BMI, and CRP are all independently related to adult brain structure; and (H3) childhood maltreatment has indirect effects on adult brain structure mediated in parallel by BMI, CRP, and AT. Using path analysis and data from N = 116,887 participants in UK Biobank, we find that CM is related to greater BMI and AT levels, and that these two variables mediate CM's effects on CRP [H1]. Regression analyses on the UKB MRI subsample (N = 21,738) revealed that greater CRP and BMI were both independently related to a spatially convergent pattern of cortical effects (Spearman's ρ = 0.87) characterized by fronto-occipital increases and temporo-parietal reductions in thickness. Subcortically, BMI was associated with greater volume, AT with lower volume and CPR with effects in both directions [H2]. Finally, path models indicated that CM has indirect effects in a subset of brain regions mediated through its direct effects on BMI and AT and indirect effects on CRP [H3]. Results provide evidence that childhood maltreatment can influence brain structure decades after exposure by increasing individual risk toward adult trauma, obesity, and inflammation.

Joint transformer architecture in brain 3D MRI classification: its application in Alzheimer's disease classification.

Alzheimer's disease (AD), a neurodegenerative disease that mostly affects the elderly, slowly impairs memory, cognition, and daily tasks. AD has long been one of the most debilitating chronic neurological disorders, affecting mostly people over 65. In this study, we investigated the use of Vision Transformer (ViT) for Magnetic Resonance Image processing in the context of AD diagnosis. ViT was utilized to extract features from MRIs, map them to a feature sequence, perform sequence modeling to maintain interdependencies, and classify features using a time series transformer. The proposed model was evaluated using ADNI T1-weighted MRIs for binary and multiclass classification. Two data collections, Complete 1Yr 1.5T and Complete 3Yr 3T, from the ADNI database were used for training and testing. A random split approach was used, allocating 60% for training and 20% for testing and validation, resulting in sample sizes of (211, 70, 70) and (1378, 458, 458), respectively. The performance of our proposed model was compared to various deep learning models, including CNN with BiL-STM and ViT with Bi-LSTM. The suggested technique diagnoses AD with high accuracy (99.048% for binary and 99.014% for multiclass classification), precision, recall, and F-score. Our proposed method offers researchers an approach to more efficient early clinical diagnosis and interventions.

Cerebral rScO2 Measured by Near-Infrared Spectroscopy (NIRS) During Therapeutic Hypothermia in Neonates with Hypoxic-Ischemic Encephalopathy: A Systematic Review.

INTRODUCTION: Perinatal asphyxia, a leading cause of neonatal mortality and neurological sequelae, necessitates early detection of pathophysiological neurologic changes during hypoxic-ischaemic encephalopathy (HIE). This study aimed to review published data on rScO2 monitoring during hypothermia treatment in neonates with perinatal asphyxia to predict short- and long-term neurological injury. METHODS: A systematic review was performed using the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines. Study identification was performed through a search between November and December 2021 in the electronic databases PubMed, Embase, Lilacs, Scopus, Web of Science, and Cochrane Central Register of Controlled Trials (CENTRAL). The main outcome was short-term (Changes in brain magnetic resonating imaging) and long-term (In neurodevelopment) neurological injury. The study protocol was registered in PROSPERO (International Prospective Register of Systematic Reviews) with CRD42023395438. RESULTS: 380 articles were collected from databases in the initial search. Finally, 15 articles were selected for extraction and analysis of the information. An increase in rScO2 measured by NIRS (Near-infrared spectroscopy) at different moments of treatment predicts neurological injury. However, there exists a wide variability in the methods and outcomes of the studies. CONCLUSION: High rScO2 values were found to predict negative outcomes, with substantial discord among studies. NIRS is proposed as a real-time bedside tool for predicting brain injury in neonates with moderate to severe HIE.

Differences in fractional amplitude of low-frequency fluctuations (fALFF) and cognitive function between untreated major depressive disorder and schizophrenia with depressive mood patients.

BACKGROUND: Distinguishing untreated major depressive disorder without medication (MDD) from schizophrenia with depressed mood (SZDM) poses a clinical challenge. This study aims to investigate differences in fractional amplitude of low-frequency fluctuations (fALFF) and cognition in untreated MDD and SZDM patients. METHODS: The study included 42 untreated MDD cases, 30 SZDM patients, and 46 healthy controls (HC). Cognitive assessment utilized the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS). Resting-state functional magnetic resonance imaging (rs-fMRI) scans were conducted, and data were processed using fALFF in slow-4 and slow-5 bands. RESULTS: Significant fALFF changes were observed in four brain regions across MDD, SZDM, and HC groups for both slow-4 and slow-5 fALFF. Compared to SZDM, the MDD group showed increased slow-5 fALFF in the right gyrus rectus (RGR). Relative to HC, SZDM exhibited decreased slow-5 fALFF in the left gyrus rectus (LGR) and increased slow-5 fALFF in the right putamen. Changes in slow-5 fALFF in both RGR and LGR were negatively correlated with RBANS scores. No significant correlations were found between remaining fALFF (slow-4 and slow-5 bands) and RBANS scores in MDD or SZDM groups. CONCLUSIONS: Alterations in slow-5 fALFF in RGR may serve as potential biomarkers for distinguishing MDD from SZDM, providing preliminary insights into the neural mechanisms of cognitive function in schizophrenia.

An analysis of information segregation in parallel streams of a multi-stream convolutional neural network.

Visual information is processed in hierarchically organized parallel streams in the primate brain. In the present study, information segregation in parallel streams was examined by constructing a convolutional neural network with parallel architecture in all of the convolutional layers. Although filter weights for convolution were initially set to random values, color information was segregated from shape information in most model instances after training. Deletion of the color-related stream decreased recognition accuracy of animate images, whereas deletion of the shape-related stream decreased recognition accuracy of both animate and inanimate images. The results suggest that properties of filters and functions of a stream are spontaneously segregated in parallel streams of neural networks.

Brain activation and connection across resting and motor-task states in patients with generalized tonic-clonic seizures.

AIMS: Motor abnormalities have been identified as one common symptom in patients with generalized tonic-clonic seizures (GTCS) inspiring us to explore the disease in a motor execution condition, which might provide novel insight into the pathomechanism. METHODS: Resting-state and motor-task fMRI data were collected from 50 patients with GTCS, including 18 patients newly diagnosed without antiepileptic drugs (ND_GTCS) and 32 patients receiving antiepileptic drugs (AEDs_GTCS). Motor activation and its association with head motion and cerebral gradients were assessed. Whole-brain network connectivity across resting and motor states was further calculated and compared between groups. RESULTS: All patients showed over-activation in the postcentral gyrus and the ND_GTCS showed decreased activation in putamen. Specifically, activation maps of ND_GTCS showed an abnormal correlation with head motion and cerebral gradient. Moreover, we detected altered functional network connectivity in patients within states and across resting and motor states by using repeated-measures analysis of variance. Patients did not show abnormal connectivity in the resting state, while distributed abnormal connectivity in the motor-task state. Decreased across-state network connectivity was also found in all patients. CONCLUSION: Convergent findings suggested the over-response of activation and connection of the brain to motor execution in GTCS, providing new clues to uncover motor susceptibility underlying the disease.

Comprehensive analysis of synthetic learning applied to neonatal brain MRI segmentation.

Brain segmentation from neonatal MRI images is a very challenging task due to large changes in the shape of cerebral structures and variations in signal intensities reflecting the gestational process. In this context, there is a clear need for segmentation techniques that are robust to variations in image contrast and to the spatial configuration of anatomical structures. In this work, we evaluate the potential of synthetic learning, a contrast-independent model trained using synthetic images generated from the ground truth labels of very few subjects. We base our experiments on the dataset released by the developmental Human Connectome Project, for which high-quality images are available for more than 700 babies aged between 26 and 45 weeks postconception. First, we confirm the impressive performance of a standard UNet trained on a few volumes, but also confirm that such models learn intensity-related features specific to the training domain. We then confirm the robustness of the synthetic learning approach to variations in image contrast. However, we observe a clear influence of the age of the baby on the predictions. We improve the performance of this model by enriching the synthetic training set with realistic motion artifacts and over-segmentation of the white matter. Based on extensive visual assessment, we argue that the better performance of the model trained on real T2w data may be due to systematic errors in the ground truth. We propose an original experiment allowing us to show that learning from real data will reproduce any systematic bias affecting the training set, while synthetic models can avoid this limitation. Overall, our experiments confirm that synthetic learning is an effective solution for segmenting neonatal brain MRI. Our adapted synthetic learning approach combines key features that will be instrumental for large multisite studies and clinical applications.