Frontiers in Neuroinformatics最新文献

In recent years, advances in microscopy and the development of novel fluorescent probes have significantly improved neuronal imaging. Many neuropsychiatric disorders are characterized by alterations in neuronal arborization, neuronal loss-as seen in Parkinson's disease-or synaptic loss, as in Alzheimer's disease. Neurodevelopmental disorders can also impact dendritic spine morphogenesis, as observed in autism spectrum disorders and schizophrenia. In this review, we provide an overview of the various labeling and microscopy techniques available to visualize neuronal structure, including dendritic spines and synapses. Particular attention is given to available fluorescent probes, recent technological advances in super-resolution microscopy (SIM, STED, STORM, MINFLUX), and segmentation methods. Aimed at biologists, this review presents both classical segmentation approaches and recent tools based on deep learning methods, with the goal of remaining accessible to readers without programming expertise.

Advancements in methodologies for efficient large-scale acquisition of high-resolution serial microscopy image data have opened new possibilities for experimental studies of cellular and subcellular features across whole brains in animal models. There is a high demand for open-source software and workflows for automated or semi-automated analysis of such data, facilitating anatomical, functional, and molecular mapping in healthy and diseased brains. These studies share a common need to consistently identify, visualize, and quantify the location of observations within anatomically defined regions, ensuring reproducible interpretation of anatomical locations, and thereby allowing meaningful comparisons of results across multiple independent studies. Addressing this need, we have developed a suite of desktop and web-applications for registration of serial brain section images to three-dimensional brain reference atlases (QuickNII, VisuAlign, WebAlign, WebWarp, and DeepSlice) and for performing data analysis in a spatial context provided by an atlas (Nutil, QCAlign, SeriesZoom, LocaliZoom, and MeshView). The software can be utilized in various combinations, creating customized analytical pipelines suited to specific research needs. The web-applications are integrated in the EBRAINS research infrastructure and coupled to the EBRAINS data platform, establishing the foundation for an online analytical workbench. We here present our software ecosystem, exemplify its use by the research community, and discuss possible directions for future developments.

In large public multi-site fMRI datasets, the sample characteristics, data acquisition methods, and MRI scanner models vary across sites and datasets. This non-neural variability obscures neural differences between groups and leads to poor machine learning based diagnostic classification of neurodevelopmental conditions. This could be potentially addressed by domain adaptation, which aims to improve classification performance in a given target domain by utilizing the knowledge learned from a different source domain by making data distributions of the two domains as similar as possible. In order to demonstrate the utility of domain adaptation for multi-site fMRI data, this research developed a variational autoencoder-maximum mean discrepancy (VAE-MMD) deep learning model for three-way diagnostic classification: (i) Autism, (ii) Asperger's syndrome, and (iii) typically developing controls. This study chooses ABIDE-II (Autism Brain Imaging Data Exchange) dataset as the target domain and ABIDE-I as the source domain. The results show that domain adaptation from ABIDE-I to ABIDE-II provides superior test accuracy of ABIDE-II compared to just using ABIDE-II for classification. Further, augmenting the source domain with additional healthy control subjects from Healthy Brain Network (HBN) and Amsterdam Open MRI Collection (AOMIC) datasets enables transfer learning and improves ABIDE-II classification performance. Finally, a comparison with statistical data harmonization techniques, such as ComBat, reveals that domain adaptation using VAE-MMD achieves comparable performance, and incorporating transfer learning (TL) with additional healthy control data substantially improves classification accuracy beyond that achieved by statistical methods (such as ComBat) alone. The dataset and the model used in this study are publicly available. The neuroimaging community can explore the possibility of further improving the model by utilizing the ever-increasing amount of healthy control fMRI data in the public domain.

Introduction: Neuroinflammation, a pathophysiological process involved in numerous disorders, is typically imaged using [¹¹C]PBR28 (or TSPO) PET. However, this technique is limited by high costs and ionizing radiation, restricting its widespread clinical use. MRI, a more accessible alternative, is commonly used for structural or functional imaging, but when used using traditional approaches has limited sensitivity to specific molecular processes. This study aims to develop a deep learning model to generate TSPO PET images from structural MRI data collected in human subjects.

Methods: A total of 204 scans, from participants with knee osteoarthritis (n = 15 scanned once, 15 scanned twice, 14 scanned three times), back pain (n = 40 scanned twice, 3 scanned three times), and healthy controls (n = 28, scanned once), underwent simultaneous 3 T MRI and [¹¹C]PBR28 TSPO PET scans. A 3D U-Net model was trained on 80% of these PET-MRI pairs and validated using 5-fold cross-validation. The model's accuracy in reconstructed PET from MRI only was assessed using various intensity and noise metrics.

Results: The model achieved a low voxel-wise mean squared error (0.0033 ± 0.0010) across all folds and a median contrast-to-noise ratio of 0.0640 ± 0.2500 when comparing true to reconstructed PET images. The synthesized PET images accurately replicated the spatial patterns observed in the original PET data. Additionally, the reconstruction accuracy was maintained even after spatial normalization.

Discussion: This study demonstrates that deep learning can accurately synthesize TSPO PET images from conventional, T1-weighted MRI. This approach could enable low-cost, noninvasive neuroinflammation imaging, expanding the clinical applicability of this imaging method.

Introduction: Accurate localization of the epileptogenic zone is essential for surgical treatment of drug-resistant epilepsy. Standard presurgical evaluations rely on multimodal neuroimaging techniques, but these may be limited by availability and interpretive challenges. This study aimed to assess the concordance between zones identified by ictal semiology and a novel distributed electrical source localization technique, CLARA, and to evaluate their impact on postsurgical outcomes.

Methods: This retrospective study included 16 patients with at least three recorded seizures. Ictal semiology was analyzed subjectively using video electroencephalography (VEEG) by a multidisciplinary team of neurologists, neurophysiologists, and radiologists, who determined the presumed epileptogenic zone at the lobar level. CLARA was subsequently applied to identify the computed zone based on ictal and/or interictal biomarker activities. The concordance between the presumed and computed zones was assessed qualitatively. Postsurgical outcomes were examined in relation to the extent of resection of the CLARA-defined zones.

Results: Among thirteen patients with sufficient data for analysis, qualitative comparison showed 77% concordance and 23% partial concordance between the presumed and computed zones. Postsurgical follow-up revealed seizure freedom in one patient with cavernoma following complete resection of the CLARA-defined zone. In contrast, patients with incomplete resection of this region continued to experience seizures.

Discussion: The findings support the potential value of CLARA as an adjunctive neuroimaging technique in the presurgical evaluation of epilepsy. By providing an additional layer of verification, CLARA may improve the accuracy of epileptogenic zone localization when used alongside established modalities such as PET, SPECT, fMRI, and MRI. Its adaptability and lower resource requirements suggest particular utility in centers with limited access to advanced medical equipment and specialized personnel. Broader implementation of CLARA could enhance presurgical decision-making and contribute to improved surgical outcomes for epilepsy patients.

Introduction: The advent of super-resolution microscopy revealed the membrane-associated periodic skeleton (MPS), a specialized neuronal cytoskeletal structure composed of actin rings spaced 190 nm apart by two spectrin dimers. While numerous ion channels, cell adhesion molecules, and signaling proteins have been shown to associate with the MPS, tools for accurate and unbiased quantification of their periodic localization remain scarce.

Methods: We developed Napari-WaveBreaker (https://github.com/SamKVs/napari-k2-WaveBreaker), an open-source plugin for the Napari image viewer. The tool quantifies MPS periodicity using autocorrelation and assesses periodic co-distribution between targets using cross-correlation. Performance was evaluated using both simulated datasets and STED microscopy images of periodic and non-periodic axonal proteins.

Results: Napari-WaveBreaker output parameters accurately reflected the visually observed periodicity and detected spatial shifts between two periodic targets. The approach was robust across varying image qualities and reliably distinguished periodic from non-periodic protein distributions.

Discussion: Napari-WaveBreaker provides an unbiased, quantitative framework for analyzing MPS-associated periodicity and co-distribution enabling new insights into the molecular organization and modulation of the MPS.

We show that recent (mid-to-late 2024) commercial large language models (LLMs) are capable of good quality metadata extraction and annotation with very little work on the part of investigators for several exemplar real-world annotation tasks in the neuroimaging literature. We investigated the GPT-4o LLM from OpenAI which performed comparably with several groups of specially trained and supervised human annotators. The LLM achieves similar performance to humans, between 0.91 and 0.97 on zero-shot prompts without feedback to the LLM. Reviewing the disagreements between LLM and gold standard human annotations we note that actual LLM errors are comparable to human errors in most cases, and in many cases these disagreements are not errors. Based on the specific types of annotations we tested, with exceptionally reviewed gold-standard correct values, the LLM performance is usable for metadata annotation at scale. We encourage other research groups to develop and make available more specialized "micro-benchmarks," like the ones we provide here, for testing both LLMs, and more complex agent systems annotation performance in real-world metadata annotation tasks.

Electroencephalogram (EEG) signal analysis plays a vital role in diagnosing and monitoring alcoholism, where accurate classification of individuals into alcoholic and control groups is essential. However, the inherent noise and complexity of EEG signals pose significant challenges. This study investigates the impact of three signal denoising techniques' Discrete Wavelet Transform(DWT), Discrete Fourier Transform(DFT), and Discrete Cosine Transform (DCT) Non EEG signal classification performance. The motivation behind this study is to identify the most effective preprocessing method for enhancing deep learning model performance in this domain. A novel DWT-CNN-BiGRU model is proposed, which leverages CNN layers for spatial feature extraction and BiGRU layers for capturing temporal dependencies. Experimental results show that the DWT-based approach, combined with standard scaling, achieves the highest accuracy of 94%, with a precision of 0.94, a recall of 0.95, and an F1-score of 0.94. Compared to the baseline DWT-CNN-BiLSTM model, the proposed method provides a modest yet meaningful improvement of approximately 17% in classification accuracy. These findings highlight the superiority of DWT as a preprocessing method and validate the proposed model's effectiveness for EEG-based classification, contributing to the development of more reliable medical diagnostic tools.

Introduction: Understanding the cognitive phenotypes of elite athletes offers a unique perspective on the intricate interplay between neurological traits and high-performance behaviors. This study aligns with advancing neuroinformatics by proposing a novel framework designed to capture and analyze the multi-dimensional dependencies of cognitive phenotypes using systems neuroscience methodologies. Traditional approaches often face limitations in disentangling the latent factors influencing cognitive variability or in preserving interpretable data structures.

Methods: To address these challenges, we developed the Latent Cognitive Embedding Network (LCEN), an innovative model that combines biologically inspired constraints with state-of-the-art neural architectures. The model features a specialized embedding mechanism for disentangling latent factors and a tailored optimization strategy incorporating domain-specific priors and regularization techniques.

Results: Experimental evaluations demonstrate LCEN's superiority in predicting and interpreting cognitive phenotypes across diverse datasets, providing deeper insights into the neural underpinnings of elite performance.

Discussion: This work bridges computational modeling, neuroscience, and psychology, contributing to the broader understanding of cognitive variability in specialized populations.

Introduction: Mental health care is undermined by fragmented data collection, as incomplete datasets can compromise treatment efficacy and research. The BrainHealth Databank (BHDB) at the Centre for Addiction and Mental Health (CAMH) establishes the governance and infrastructure for a Learning Mental Health System that integrates digital tools, measurement-based care, artificial intelligence (AI), and open science to deliver personalized, data-driven care.

Methods: Central to the BHDB's approach is its comprehensive governance framework, which actively engages clinicians, researchers, data scientists, privacy and ethics experts, and patient and family partners. This codesigned approach ensures that digital health technologies are deployed ethically, securely, and effectively within clinical settings.

Results: By aligning data collection with clinical and research goals and harmonizing over 12 million data points from 33,000 patient trajectories, the BHDB enhances data quality, enables real-time decision support, and fosters continuous improvement.

Discussion: The BHDB provides a model for integrating AI and digital tools into mental health care, as well as research data collection, analyses, storage, and sharing through the BHDB Portal (https://bhdb.camh.ca).