Neuroinformatics最新文献

NeuroHackademy ( https://neurohackademy.org ) is a two-week event designed to train early-career neuroscience researchers in data science methods and their application to neuroimaging. The event seeks to bridge the big data skills gap by introducing participants to data science methods and skills that are often ignored in traditional curricula. Such skills are needed for the analysis and interpretation of the kinds of large and complex datasets that have become increasingly important to neuroimaging research due to concerted data collection efforts. In 2020, the event rapidly pivoted from an in-person event to an online event that included hundreds of participants from all over the world. This experience and those of the participants substantially changed our valuation of large online-accessible events. In subsequent events held in 2022 and 2023, we have developed a "hybrid" format that includes both online and in-person participants. We discuss the technical and sociotechnical elements of hybrid events and discuss some of the lessons we have learned while organizing them. We emphasize in particular the role that these events can play in creating a global and inclusive community of practice in the intersection of neuroimaging and data science.

Diffusion properties from diffusion tensor imaging (DTI) are exquisitely sensitive to white matter abnormalities incurred during traumatic brain injury (TBI), especially for those patients with chronic post-TBI symptoms such as headaches, dizziness, fatigue, etc. The evaluation of structural and functional connectivity using DTI has become a promising method for identifying subtle alterations in brain connectivity associated with TBI that are otherwise not visible with conventional imaging. This study assessed whether TBI patients with (n = 17) or without (n = 16) chronic symptoms (TBIcs/TBIncs) exhibit any changes in structural connectivity (SC) and mean fractional anisotropy (mFA) of intra- and inter-hemispheric connections when compared to a control group (CG) (n = 13). Reductions in SC and mFA were observed for TBIcs compared to CG, but not for TBIncs. More connections were found to have mFA reductions than SC reductions. On the whole, SC is dominated by ipsilateral connections for all the groups after the comparison of contralateral and ipsilateral connections. More contra-ipsi reductions of mFA were found for TBIcs than TBIncs compared to CG. These findings suggest that TBI patients with chronic symptoms not only demonstrate decreased global and regional mFA but also reduced structural network connectivity.

Subdural electrocorticography (ECoG) is a valuable technique for neuroscientific research and for emerging neurotechnological clinical applications. As ECoG grids accommodate increasing numbers of electrodes and higher densities with new manufacturing methods, the question arises at what point the benefit of higher density ECoG is outweighed by spatial oversampling. To clarify the optimal spacing between ECoG electrodes, in the current study we evaluate how ECoG grid density relates to the amount of non-shared neurophysiological information between electrode pairs, focusing on the sensorimotor cortex. We simultaneously recorded high-density (HD, 3 mm pitch) and ultra-high-density (UHD, 0.9 mm pitch) ECoG, obtained intraoperatively from six participants. We developed a new metric, the normalized differential root mean square (ndRMS), to quantify the information that is not shared between electrode pairs. The ndRMS increases with inter-electrode center-to-center distance up to 15 mm, after which it plateaus. We observed differences in ndRMS between frequency bands, which we interpret in terms of oscillations in frequencies below 32 Hz with phase differences between pairs, versus (un)correlated signal fluctuations in the frequency range above 64 Hz. The finding that UHD recordings yield significantly higher ndRMS than HD recordings is attributed to the amount of tissue sampled by each electrode. These results suggest that ECoG densities with submillimeter electrode distances are likely justified.

Attention Deficit Hyperactivity Disorder (ADHD) is a widespread neurobehavioral disorder affecting children and adolescents, requiring early detection for effective treatment. EEG connectivity measures can reveal the interdependencies between EEG recordings, highlighting brain network patterns and functional behavior that improve diagnostic accuracy. This study introduces a novel ADHD diagnostic method by combining linear and nonlinear brain connectivity maps with an attention-based convolutional neural network (Att-CNN). Pearson Correlation Coefficient (PCC) and Phase-Locking Value (PLV) are used to create fused connectivity maps (FCMs) from various EEG frequency subbands, which are then inputted into the Att-CNN. The attention module is strategically placed after the latest convolutional layer in the CNN. The performance of different optimizers (Adam and SGD) and learning rates are assessed. The suggested model obtained 98.88%, 98.41%, 98.19%, and 98.30% for accuracy, precision, recall, and F1 Score, respectively, using the SGD optimizer in the FCM of the theta band with a learning rate of 1e-1. With the use of FCM, Att-CNN, and advanced optimizers, the proposed technique has the potential to produce trustworthy instruments for the early diagnosis of ADHD, greatly enhancing both patient outcomes and diagnostic accuracy.

Tensor-based representations are being increasingly used to represent complex data types such as imaging data, due to their appealing properties such as dimension reduction and the preservation of spatial information. Recently, there is a growing literature on using Bayesian scalar-on-tensor regression techniques that use tensor-based representations for high-dimensional and spatially distributed covariates to predict continuous outcomes. However surprisingly, there is limited development on corresponding Bayesian classification methods relying on tensor-valued covariates. Standard approaches that vectorize the image are not desirable due to the loss of spatial structure, and alternate methods that use extracted features from the image in the predictive model may suffer from information loss. We propose a novel data augmentation-based Bayesian classification approach relying on tensor-valued covariates, with a focus on imaging predictors. We propose two data augmentation schemes, one resulting in a support vector machine (SVM) type of classifier, and another yielding a logistic regression classifier. While both types of classifiers have been proposed independently in literature, our contribution is to extend such existing methodology to accommodate high-dimensional tensor valued predictors that involve low rank decompositions of the coefficient matrix while preserving the spatial information in the image. An efficient Markov chain Monte Carlo (MCMC) algorithm is developed for implementing these methods. Simulation studies show significant improvements in classification accuracy and parameter estimation compared to routinely used classification methods. We further illustrate our method in a neuroimaging application using cortical thickness MRI data from Alzheimer's Disease Neuroimaging Initiative, with results displaying better classification accuracy throughout several classification tasks, including classification on pairs of the three diagnostic groups: normal control, AD patients, and MCI patients; gender classification (males vs females); and cognitive performance based on high and low levels of MMSE scores.

US Food and Drug Administration (FDA) cleared a Transcranial Magnetic Stimulation (TMS) system with functional Magnetic Resonance Imaging-guided (fMRI) individualized treatment protocol for major depressive disorder, which employs resting state-fMRI (RS-fMRI) functional connectivity (FC) to pinpoint the target individually to increase the accuracy and effeteness of the stimulation. Furthermore, task activation-guided TMS, as well as the use of RS-fMRI local metrics for targeted the specific abnormal brain regions, are considered a precise scheme for TMS targeting. Since 1.5 T MRI is more available in hospitals, systematic evaluation of the test-retest reliability and sensitivity of fMRI metrics on 1.5 T and 3 T MRI may provide reference for the application of fMRI-guided individualized-precise TMS stimulation. Twenty participants underwent three RS-fMRI scans and one scan of finger-tapping task fMRI with self-initiated (SI) and visual-guided (VG) conditions at both 3 T and 1.5 T. Then the location reliability derived by FC (with three seed regions) and peak activation were assessed by intra-individual distance. The test-retest reliability and sensitivity of five RS-fMRI local metrics were evaluated using intra-class correlation and effect size, separately. The intra-individual distance of peak activation location between 1.5 T and 3 T was 15.8 mm and 19 mm for two conditions, respectively. The intra-individual distance for the FC derived targets at 1.5 T was 9.6-31.2 mm, compared to that of 3 T (7.6-31.1 mm). The test-retest reliability and sensitivity of RS-fMRI local metrics showed similar trends on 1.5 T and 3 T. These findings hasten the application of fMRI-guided individualized TMS treatment in clinical practice.

Over the past decade, the intricacies of sports-related concussions among female athletes have become readily apparent. Traditional clinical methods for diagnosing concussions suffer limitations when applied to female athletes, often failing to capture subtle changes in brain structure and function. Advanced neuroinformatics techniques and machine learning models have become invaluable assets in this endeavor. While these technologies have been extensively employed in understanding concussion in male athletes, there remains a significant gap in our comprehension of their effectiveness for female athletes. With its remarkable data analysis capacity, machine learning offers a promising avenue to bridge this deficit. By harnessing the power of machine learning, researchers can link observed phenotypic neuroimaging data to sex-specific biological mechanisms, unraveling the mysteries of concussions in female athletes. Furthermore, embedding methods within machine learning enable examining brain architecture and its alterations beyond the conventional anatomical reference frame. In turn, allows researchers to gain deeper insights into the dynamics of concussions, treatment responses, and recovery processes. This paper endeavors to address the crucial issue of sex differences in multimodal neuroimaging experimental design and machine learning approaches within female athlete populations, ultimately ensuring that they receive the tailored care they require when facing the challenges of concussions. Through better data integration, feature identification, knowledge representation, validation, etc., neuroinformaticists, are ideally suited to bring clarity, context, and explainabilty to the study of sports-related head injuries in males and in females, and helping to define recovery.

Leveraging vast neuroimaging and electrophysiological datasets, AI algorithms are uncovering patterns that offer unprecedented insights into brain structure and function. Neuroinformatics, the fusion of neuroscience and AI, is advancing technologies like brain-computer interfaces, AI-driven cognitive enhancement, and personalized neuromodulation for treating neurological disorders. These developments hold potential to improve cognitive functions, restore motor abilities, and create human-machine collaborative systems. Looking ahead, the convergence of neuroscience and AI is set to transform cognitive modeling, decision-making, and mental health interventions. This fusion mirrors the quest for nuclear fusion energy, both driven by the need to unlock profound sources of understanding. As STEM disciplines continue to drive core developments of foundational models of the brain, neuroinformatics promises to lead innovations in augmented intelligence, personalized healthcare, and effective decision-making systems.

Annotation of multiple regions of interest across the whole mouse brain is an indispensable process for quantitative evaluation of a multitude of study endpoints in neuroscience digital pathology. Prior experience and domain expert knowledge are the key aspects for image annotation quality and consistency. At present, image annotation is often achieved manually by certified pathologists or trained technicians, limiting the total throughput of studies performed at neuroscience digital pathology labs. It may also mean that simpler and quicker methods of examining tissue samples are used by non-pathologists, especially in the early stages of research and preclinical studies. To address these limitations and to meet the growing demand for image analysis in a pharmaceutical setting, we developed AnNoBrainer, an open-source software tool that leverages deep learning, image registration, and standard cortical brain templates to automatically annotate individual brain regions on 2D pathology slides. Application of AnNoBrainer to a published set of pathology slides from transgenic mice models of synucleinopathy revealed comparable accuracy, increased reproducibility, and a significant reduction (~ 50%) in time spent on brain annotation, quality control and labelling compared to trained scientists in pathology. Taken together, AnNoBrainer offers a rapid, accurate, and reproducible automated annotation of mouse brain images that largely meets the experts' histopathological assessment standards (> 85% of cases) and enables high-throughput image analysis workflows in digital pathology labs.

Brain reconstruction, specially of the cerebral cortex, is a challenging task and even more so when it comes to highly gyrified brained animals. Here, we present Stitcher, a novel tool capable of generating such surfaces utilizing MRI data and manual segmentation. Stitcher makes a triangulation between consecutive brain slice segmentations by recursively adding edges that minimize the total length and simultaneously avoid self-intersection. We applied this new method to build the cortical surfaces of two dolphins: Guiana dolphin (Sotalia guianensis), Franciscana dolphin (Pontoporia blainvillei); and one pinniped: Steller sea lion (Eumetopias jubatus). Specifically in the case of P. blainvillei, two reconstructions at two different resolutions were made. Additionally, we also performed reconstructions for sub and non-cortical structures of Guiana dolphin. All our cortical mesh results show remarkable resemblance with the real anatomy of the brains, except P. blainvillei with low-resolution data. Sub and non-cortical meshes were also properly reconstructed and the spatial positioning of structures was preserved with respect to S. guianensis cerebral cortex. In a comparative perspective between methods, Stitcher presents compatible results for volumetric measurements when contrasted with other anatomical standard tools. In this way, Stitcher seems to be a viable pipeline for new neuroanatomical analysis, enhancing visualization and descriptions of non-primates species, and broadening the scope of compared neuroanatomy.