Neuroinformatics最新文献

Neuroscience education is challenged by rapidly evolving technology and the development of interdisciplinary approaches for brain research. The Human Brain Project (HBP) Education Programme aimed to address the need for interdisciplinary expertise in brain research by equipping a new generation of researchers with skills across neuroscience, medicine, and information technology. Over its ten year duration, the programme engaged over 1,300 experts and attracted more than 5,500 participants from various scientific disciplines in its blended learning curriculum, specialised schools and workshops, and events fostering dialogue among early-career researchers. Key principles of the programme's approach included fostering interdisciplinarity, adaptability to the evolving research landscape and infrastructure, and a collaborative environment with a focus on empowering early-career researchers. Following the programme's conclusion, we provide here an analysis and in-depth view across a diverse range of educational formats and events. Our results show that the Education Programme achieved success in its wide geographic reach, the diversity of participants, and the establishment of transversal collaborations. Building on these experiences and achievements, we describe how leveraging digital tools and platforms provides accessible and highly specialised training, which can enhance existing education programmes for the next generation of brain researchers working in decentralised European collaborative spaces. Finally, we present the lessons learnt so that similar initiatives may improve upon our experience and incorporate our suggestions into their own programme.

This article seeks to investigate the impact of aging on functional connectivity across different cognitive control scenarios, particularly emphasizing the identification of brain regions significantly associated with early aging. By conceptualizing functional connectivity within each cognitive control scenario as a graph, with brain regions as nodes, the statistical challenge revolves around devising a regression framework to predict a binary scalar outcome (aging or normal) using multiple graph predictors. Popular regression methods utilizing multiplex graph predictors often face limitations in effectively harnessing information within and across graph layers, leading to potentially less accurate inference and predictive accuracy, especially for smaller sample sizes. To address this challenge, we propose the Bayesian Multiplex Graph Classifier (BMGC). Accounting for multiplex graph topology, our method models edge coefficients at each graph layer using bilinear interactions between the latent effects associated with the two nodes connected by the edge. This approach also employs a variable selection framework on node-specific latent effects from all graph layers to identify influential nodes linked to observed outcomes. Crucially, the proposed framework is computationally efficient and quantifies the uncertainty in node identification, coefficient estimation, and binary outcome prediction. BMGC outperforms alternative methods in terms of the aforementioned metrics in simulation studies. An additional BMGC validation was completed using an fMRI study of brain networks in adults. The proposed BMGC technique identified that sensory motor brain network obeys certain lateral symmetries, whereas the default mode network exhibits significant brain asymmetries associated with early aging.

The brain is an intricate system that controls a variety of functions. It consists of a vast number of cells that exhibit diverse characteristics. To understand brain function in health and disease, it is crucial to classify neurons accurately. Recent advancements in machine learning have provided a way to classify neurons based on their electrophysiological activity. This paper presents a deep-learning framework that classifies neurons solely on this basis. The framework uses data from the Allen Cell Types database, which contains a survey of biological features derived from single-cell recordings from mice and humans. The shared information from both sources is used to classify neurons into their broad types with the help of a joint model. An accurate domain-adaptive model, integrating electrophysiological data from both mice and humans, is implemented. Furthermore, data from mouse neurons, which also includes labels of transgenic mouse lines, is further classified into subtypes using an interpretable neural network model. The framework provides state-of-the-art results in terms of accuracy and precision while also providing explanations for the predictions.

Research data management has become an indispensable skill in modern neuroscience. Researchers can benefit from following good practices as well as from having proficiency in using particular software solutions. But as these domain-agnostic skills are commonly not included in domain-specific graduate education, community efforts increasingly provide early career scientists with opportunities for organised training and materials for self-study. Investing effort in user documentation and interacting with the user base can, in turn, help developers improve quality of their software. In this work, we detail and evaluate our multi-modal teaching approach to research data management in the DataLad ecosystem, both in general and with concrete software use. Spanning an online and printed handbook, a modular course suitable for in-person and virtual teaching, and a flexible collection of research data management tips in a knowledge base, our free and open source collection of training material has made research data management and software training available to various different stakeholders over the past five years.

Neurotechnology and big data are two rapidly advancing fields that have the potential to transform our understanding of the brain and its functions. Advancements in neurotechnology have enabled researchers to investigate the function of the brain at unprecedented levels of granularity at the functional, molecular, and anatomical levels. Thus, resulting in the collection of not only more data, but also larger datasets. To fully harness the potential of big data and advancements in neurotechnology to improve our understanding of the nervous system, there is a need to train a new generation of neuroscientists capable of not only domain expertise, but also the computational and data science skills required to interrogate and integrate big data. Importantly, neuroinformatics is the subdiscipline of neuroscience devoted to the development of neuroscience data and knowledge bases together with computational models and analytical tools for sharing, integration and analysis of experimental data, and advancement of theories about the nervous system function. While there are only a few formal training programs in neuroinformatics, and since neuroinformatics is rarely incorporated into traditional neuroscience training programs, the neuroinformatics community has attempted to bridge the gap between the traditional neuroscience education programs and the needs of the next generation of neuroscience researchers through community initiatives and workshops. Thus, the purpose of this special collection is to highlight several such community efforts which span from in-person workshops to large-scale, global virtual training consortiums and from training students to training-the-trainers.

Advances in the spatiotemporal resolution and field-of-view of neuroimaging tools are driving mesoscale studies for translational neuroscience. On October 10, 2023, the Center for Mesoscale Mapping (CMM) at the Massachusetts General Hospital (MGH) Athinoula A. Martinos Center for Biomedical Imaging and the Massachusetts Institute of Technology (MIT) Health Sciences Technology based Neuroimaging Training Program (NTP) hosted a symposium exploring the state-of-the-art in this rapidly growing area of research. "Mesoscale Brain Mapping: Bridging Scales and Modalities in Neuroimaging" brought together researchers who use a broad range of imaging techniques to study brain structure and function at the convergence of the microscopic and macroscopic scales. The day-long event centered on areas in which the CMM has established expertise, including the development of emerging technologies and their application to clinical translational needs and basic neuroscience questions. The in-person symposium welcomed more than 150 attendees, including 57 faculty members, 61 postdoctoral fellows, 35 students, and four industry professionals, who represented institutions at the local, regional, and international levels. The symposium also served the training goals of both the CMM and the NTP. The event content, organization, and format were planned collaboratively by the faculty and trainees. Many CMM faculty presented or participated in a panel discussion, thus contributing to the dissemination of both the technologies they have developed under the auspices of the CMM and the findings they have obtained using those technologies. NTP trainees who benefited from the symposium included those who helped to organize the symposium and/or presented posters and gave "flash" oral presentations. In addition to gaining experience from presenting their work, they had opportunities throughout the day to engage in one-on-one discussions with visiting scientists and other faculty, potentially opening the door to future collaborations. The symposium presentations provided a deep exploration of the many technological advances enabling progress in structural and functional mesoscale brain imaging. Finally, students worked closely with the presenting faculty to develop this report summarizing the content of the symposium and putting it in the broader context of the current state of the field to share with the scientific community. We note that the references cited here include conference abstracts corresponding to the symposium poster presentations.

The black box nature of deep neural networks (DNNs) makes researchers and clinicians hesitant to rely on their findings. Saliency maps can enhance DNN explainability by suggesting the anatomic localization of relevant brain features. This study compares seven popular attribution-based saliency approaches to assign neuroanatomic interpretability to DNNs that estimate biological brain age (BA) from magnetic resonance imaging (MRI). Cognitively normal (CN) adults (N = 13,394, 5,900 males; mean age: 65.82 ± 8.89 years) are included for DNN training, testing, validation, and saliency map generation to estimate BA. To study saliency robustness to the presence of anatomic deviations from normality, saliency maps are also generated for adults with mild traumatic brain injury (mTBI, $N$ = 214, 135 males; mean age: 55.3 ± 9.9 years). We assess saliency methods' capacities to capture known anatomic features of brain aging and compare them to a surrogate ground truth whose anatomic saliency is known a priori. Anatomic aging features are identified most reliably by the integrated gradients method, which outperforms all others through its ability to localize relevant anatomic features. Gradient Shapley additive explanations, input × gradient, and masked gradient perform less consistently but still highlight ubiquitous neuroanatomic features of aging (ventricle dilation, hippocampal atrophy, sulcal widening). Saliency methods involving gradient saliency, guided backpropagation, and guided gradient-weight class attribution mapping localize saliency outside the brain, which is undesirable. Our research suggests the relative tradeoffs of saliency methods to interpret DNN findings during BA estimation in typical aging and after mTBI.

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.