Hasil untuk "Neurosciences. Biological psychiatry. Neuropsychiatry"

Diya Prasanth, Matthew Tivnan

We present BioNIC, a multi-layer feedforward neural network for emotion classification, inspired by detailed synaptic connectivity graphs from the MICrONs dataset. At a structural level, we incorporate architectural constraints derived from a single cortical column of the mouse Primary Visual Cortex(V1): connectivity imposed via adjacency masks, laminar organization, and graded inhibition representing inhibitory neurons. At the functional level, we implement biologically inspired learning: Hebbian synaptic plasticity with homeostatic regulation, Layer Normalization, data augmentation to model exposure to natural variability in sensory input, and synaptic noise to model neural stochasticity. We also include convolutional layers for spatial processing, mimicking retinotopic mapping. The model performance is evaluated on the Facial Emotion Recognition task FER-2013 and compared with a conventional baseline. Additionally, we investigate the impacts of each biological feature through a series of ablation experiments. While connectivity was limited to a single cortical column and biologically relevant connections, BioNIC achieved performance comparable to that of conventional models, with an accuracy of 59.77 $\pm$ 0.27% on FER-2013. Our findings demonstrate that integrating constraints derived from connectomics is a computationally plausible approach to developing biologically inspired artificial intelligence systems. This work also highlights the potential of new generation peta-scale connectomics data in advancing both neuroscience modeling and artificial intelligence.

Rikuto Kotoge, Ziwei Yang, Zheng Chen et al.

Retrieving targeted pathways in biological knowledge bases, particularly when incorporating wet-lab experimental data, remains a challenging task and often requires downstream analyses and specialized expertise. In this paper, we frame this challenge as a solvable graph learning and explaining task and propose a novel subgraph inference framework, ExPAth, that explicitly integrates experimental data to classify various graphs (bio-networks) in biological databases. The links (representing pathways) that contribute more to classification can be considered as targeted pathways. Our framework can seamlessly integrate biological foundation models to encode the experimental molecular data. We propose ML-oriented biological evaluations and a new metric. The experiments involving 301 bio-networks evaluations demonstrate that pathways inferred by ExPath are biologically meaningful, achieving up to 4.5x higher Fidelity+ (necessity) and 14x lower Fidelity- (sufficiency) than explainer baselines, while preserving signaling chains up to 4x longer.

Aswin Paul, Moein Khajehnejad, Forough Habibollahi et al.

With recent and rapid advancements in artificial intelligence (AI), understanding the foundation of purposeful behaviour in autonomous agents is crucial for developing safe and efficient systems. While artificial neural networks have dominated the path to AI, recent studies are exploring the potential of biologically based systems, such as networks of living biological neuronal networks. Along with promises of high power and data efficiency, these systems may also inform more explainable and biologically plausible models. In this work, we propose a framework rooted in active inference, a general theory of behaviour, to model decision-making in embodied agents. Using experiment-informed generative models, we simulate decision-making processes in a simulated game-play environment, mirroring experimental setups that use biological neurons. Our results demonstrate learning in these agents, providing insights into the role of memory-based learning and predictive planning in intelligent decision-making. This work contributes to the growing field of explainable AI by offering a biologically grounded and scalable approach to understanding purposeful behaviour in agents.

Fang Wang, Ying Liang, Qin-Wen Wang

Abstract Background Alzheimer’s disease (AD) is a complex neurodegenerative disorder with a largely unexplored epigenetic landscape. Objective This study employs an innovative approach that integrates multi-omics analysis and explainable machine learning to explore the epigenetic regulatory mechanisms underlying the epigenetic signature of PRRT1 implicated in AD. Methods Through comprehensive DNA methylation and transcriptomic profiling, we identified distinct epigenetic signatures associated with gene PRRT1 expression in AD patient samples compared to healthy controls. Utilizing interpretable machine learning models and ELMAR analysis, we dissected the complex relationships between these epigenetic signatures and gene expression patterns, revealing novel regulatory elements and pathways. Finally, the epigenetic mechanisms of these genes were investigated experimentally. Results This study identified ten epigenetic signatures, constructed an interpretable AD diagnostic model, and utilized various bioinformatics methods to create an epigenomic map. Subsequently, the ELMAR R package was used to integrate multi-omics data and identify the upstream transcription factor MAZ for PRRT1. Finally, experiments confirmed the interaction between MAZ and PRRT1, which mediated apoptosis and autophagy in AD. Conclusion This study adopts a strategy that integrates bioinformatics analysis with molecular experiments, providing new insights into the epigenetic regulatory mechanisms of PRRT1 in AD and demonstrating the importance of explainable machine learning in elucidating complex disease mechanisms.

Paolo Bonato, David Reinkensmeyer, Mario Manto

Abstract The Journal of NeuroEngineering and Rehabilitation (JNER) has become a major actor for the dissemination of knowledge in the scientific community, bridging the gaps between innovative neuroengineering and rehabilitation. Major fields of innovations have emerged these last 25 years, such as machine learning and the ongoing AI revolution, wearable technologies, human machine interfaces, robotics, advanced prosthetics, functional electrical stimulation and various neuromodulation techniques. With the major burden of disorders impacting on the central/peripheral nervous system and the musculoskeletal system both in adults and in children, successful tailored neurorehabilitation has become a major objective for the research and clinical community at a world scale. JNER contributes to this challenging goal, publishing groundbreaking cutting-edge research using the open access model. The multidisciplinary approaches at the crossroads of biomedical engineering, neuroscience, physical medicine and rehabilitation make of the journal a unique growing platform welcoming breakthrough discoveries to reshape the field and restore function.

Tina Howischer, Lukas Gattermeyer-Kell, Stephanie Hirschbichler et al.

Background/Objectives: Anti-IgLON5 disease is a neurological disorder characterized by the presence of autoantibodies directed against the neuronal cell adhesion protein IgLON5. Pathophysiology involves both autoimmune inflammation and neurodegenerative processes. The most common causes of death are sudden death, central hypoventilation, dysphagia, and aspiration. However, the high rate of largely unclear sudden deaths calls for further research in this area. Methods: We performed a systematic search of the literature on causes of death in anti-IgLON5 disease following the PRISMA guidelines. In addition, we present a new case that was followed up in our clinic until death. Results: Of 258 publications with anti-IgLON5 disease, 21 publications comprising 61 cases that reported the causes of death were included in the analysis. The most common cause of death was death due to complications at 36.1%, followed by sudden death, accounting for 32.8% of the cases. Other causes include respiratory, cardiac, and unknown causes. The patient presented here as a case report was also diagnosed with cardiac amyloidosis and died from a cardiac cause of sudden death. Conclusions: Sudden death in anti-IgLON5 disease is one of the most common causes of death in the literature. A progressive neurodegenerative process in the brain stem causing central hypoventilation is generally assumed as a major causative factor. The case reported here had concomitant cardiac amyloidosis, which may raise the question as to whether unrecognized cardiac causes, which are not routinely screened for in this population, might represent another cause of sudden death, which would have important therapeutic implications.

Simone Azeglio, Olivier Marre, Peter Neri et al.

We propose a novel approach to image classification inspired by complex nonlinear biological visual processing, whereby classical convolutional neural networks (CNNs) are equipped with learnable higher-order convolutions. Our model incorporates a Volterra-like expansion of the convolution operator, capturing multiplicative interactions akin to those observed in early and advanced stages of biological visual processing. We evaluated this approach on synthetic datasets by measuring sensitivity to testing higher-order correlations and performance in standard benchmarks (MNIST, FashionMNIST, CIFAR10, CIFAR100 and Imagenette). Our architecture outperforms traditional CNN baselines, and achieves optimal performance with expansions up to 3rd/4th order, aligning remarkably well with the distribution of pixel intensities in natural images. Through systematic perturbation analysis, we validate this alignment by isolating the contributions of specific image statistics to model performance, demonstrating how different orders of convolution process distinct aspects of visual information. Furthermore, Representational Similarity Analysis reveals distinct geometries across network layers, indicating qualitatively different modes of visual information processing. Our work bridges neuroscience and deep learning, offering a path towards more effective, biologically inspired computer vision models. It provides insights into visual information processing and lays the groundwork for neural networks that better capture complex visual patterns, particularly in resource-constrained scenarios.

Kian Kenyon-Dean, Zitong Jerry Wang, John Urbanik et al.

Large-scale cell microscopy screens are used in drug discovery and molecular biology research to study the effects of millions of chemical and genetic perturbations on cells. To use these images in downstream analysis, we need models that can map each image into a feature space that represents diverse biological phenotypes consistently, in the sense that perturbations with similar biological effects have similar representations. In this work, we present the largest foundation model for cell microscopy data to date, a new 1.9 billion-parameter ViT-G/8 MAE trained on over 8 billion microscopy image crops. Compared to a previous published ViT-L/8 MAE, our new model achieves a 60% improvement in linear separability of genetic perturbations and obtains the best overall performance on whole-genome biological relationship recall and replicate consistency benchmarks. Beyond scaling, we developed two key methods that improve performance: (1) training on a curated and diverse dataset; and, (2) using biologically motivated linear probing tasks to search across each transformer block for the best candidate representation of whole-genome screens. We find that many self-supervised vision transformers, pretrained on either natural or microscopy images, yield significantly more biologically meaningful representations of microscopy images in their intermediate blocks than in their typically used final blocks. More broadly, our approach and results provide insights toward a general strategy for successfully building foundation models for large-scale biological data.

Zahra Babaiee, Peyman M. Kiasari, Daniela Rus et al.

In this study, we present evidence suggesting that depthwise convolutional kernels are effectively replicating the structural intricacies of the biological receptive fields observed in the mammalian retina. We provide analytics of trained kernels from various state-of-the-art models substantiating this evidence. Inspired by this intriguing discovery, we propose an initialization scheme that draws inspiration from the biological receptive fields. Experimental analysis of the ImageNet dataset with multiple CNN architectures featuring depthwise convolutions reveals a marked enhancement in the accuracy of the learned model when initialized with biologically derived weights. This underlies the potential for biologically inspired computational models to further our understanding of vision processing systems and to improve the efficacy of convolutional networks.

Jenny Drnevich, Frederick J. Tan, Fabricio Almeida-Silva et al.

Modern biological research is increasingly data-intensive, leading to a growing demand for effective training in biological data science. In this article, we provide an overview of key resources and best practices available within the Bioconductor project - an open-source software community focused on omics data analysis. This guide serves as a valuable reference for both learners and educators in the field.

Jörn von Döhren

IntroductionNeurodevelopment in larval stages of non-model organisms, with a focus on the serotonin- and FMRFamide-immunoreactive components, has been in the focus of research in the recent past. However, some taxonomic groups remain understudied. Nemertea (ribbon worms) represent such an understudied clade with only few reports on nervous system development mostly from phylogenetically or developmentally derived species. It would be insightful to explore neurodevelopment in additional species to be able to document the diversity and deduce common patterns to trace the evolution of nervous system development.MethodsFluorescent immunohistochemical labeling with polyclonal primary antibodies against serotonin and FMRF-amide and a monoclonal antibody against synapsin performed on series of fixed larval stages of two nemertean species Cephalothrix rufifrons (Archinemertea, Palaeonemertea) and Emplectonema gracile (Monostilifera, Hoplonemertea) were analyzed with confocal laser scanning microscopy.ResultsThis contribution gives detailed accounts on the development of the serotonin- and FMRFamide-immunoreactive subsets of the nervous system in two nemertean species from the first appearance of the respective signals. Additionally, data on synapsin-like immunoreactivity illustrates the general structure of neuropil components. Events common to both investigated species are the appearance of serotonin-like immunoreactive signals before the appearance of FMRF-like immunoreactive signals and the strict progression of the development of the lateral nerve cords from the anteriorly located, ring-shaped brain toward the posterior pole of the larva. Notable differences are (1) the proboscis nervous system that is developing much earlier in investigated larval stages of E. gracile and (2) distinct early, but apparently transient, serotonergic neurons on the frontal and caudal pole of the larva in E. gracile that seem to be absent in C. rufifrons.DiscussionAccording to the results from this investigation and in line with previously published accounts on nervous system development, the hypothetical last common ancestor of Nemertea had a ring-shaped brain arranged around the proboscis opening, from which a pair of ventro-lateral nerve cords develops in anterior to posterior progression. Early frontal and caudal serotonergic neurons that later degenerate or cease to express serotonin are an ancestral character of Nemertea that they share with several other spiralian clades.

Wu Jintao, Zhao Yue

The ventral bed nucleus of the stria terminalis (vBNST) plays a key role in cocaine addiction, especially relapse. However, the direct effects of cocaine on corticotropin-releasing hormone (CRH) neurons in the vBNST remain unclear. Here, we identify that cocaine exposure can remarkably attenuate the intrinsic excitability of CRH neurons in the vBNST in vitro. Accumulating studies reveal the crucial role of Sigma-1 receptors (Sig-1Rs) in modulating cocaine addiction. However, to the authors’ best knowledge no investigations have explored the role of Sig-1Rs in the vBNST, let alone CRH neurons. Given that cocaine acts as a type of Sig-1Rs agonist, and the dramatic role of Sig-1Rs played in intrinsic excitability of neurons as well as cocaine addiction, we employ BD1063 a canonical Sig-1Rs antagonist to block the effects of cocaine, and significantly recover the excitability of CRH neurons. Together, we suggest that cocaine exposure leads to the firing rate depression of CRH neurons in the vBNST via binding to Sig-1Rs.

Shervin Safavi

Finding general principles underlying brain function has been appealing to scientists. Indeed, in some branches of science like physics and chemistry (and to some degree biology) a general theory often can capture the essence of a wide range of phenomena. Whether we can find such principles in neuroscience, and [assuming they do exist] what those principles are, are important questions. Abstracting the brain as a complex system is one of the perspectives that may help us answer this question. While it is commonly accepted that the brain is a (or even the) prominent example of a complex system, the far reaching implications of this are still arguably overlooked in our approaches to neuroscientific questions. One of the reasons for the lack of attention could be the apparent difference in foci of investigations in these two fields -- neuroscience and complex systems. This thesis is an effort toward providing a bridge between systems neuroscience and complex systems by harnessing systems neuroscience tools & notions for building empirical approaches toward the brain as a complex system. Perhaps, in the spirit of searching for principles, we should abstract and approach the brain as a complex adaptive system as the more complete perspective (rather than just a complex system). In the end, the brain, even the most "complex system", need to survive in the environment. Indeed, in the field of complex adaptive systems, the intention is understanding very similar questions in nature. As an outlook, we also touch on some research directions pertaining to the adaptivity of the brain as well.

Meng-Yi Chen, Meng-Yi Chen, Qinge Zhang et al.

BackgroundA growing number of studies has implicated oxidative stress in the pathophysiology of psychiatric disorders including schizophrenia. The aim of this study was to explore the field of schizophrenia and oxidative stress-related research from a bibliometric perspective.MethodsAll relevant publications on schizophrenia and oxidative stress were obtained from Web of Science Core Collection (WOSCC) database from its inception date to November 8, 2022. VOSviewer software was used to examine co-authorships and co-occurring keywords. R software was used to present the main characteristics of publications and cooperation frequency among countries. CiteSpace was used to investigate keywords with the strongest citation bursts.ResultsA total of 3,510 publications on schizophrenia and oxidative stress were included. The United States had the largest number of publications (26.1%), and international collaborations. University of Melbourne was the most productive institution, while Schizophrenia Research was the most productive journal in this field. Apart from “schizophrenia” and “oxidative stress”, the terms “prefrontal cortex”, “brain” and “nitric oxide” were among the most frequently used keywords.ConclusionsIn conclusion, research on the association between oxidative stress and schizophrenia has received growing attention in the academic literature that is expected to continue its upward trajectory during the next two decades. Existing research suggests there has been a transition from research focused on pathways to animal models, and subsequently to clinical applications. Intervention studies on oxidative stress and schizophrenia are likely to be an important focus of related work in the near future.

Sheilagh Hodgins, Sheilagh Hodgins

The percentage of forensic psychiatric patients who are female varies from 5 to 13% in Europe, rises to 18% in England and Wales, and sits at 15% in Canada. Similarly, many fewer women than men are incarcerated in correctional facilities. While these statistics supposedly reflect less antisocial and aggressive behavior (AAB) among females than males, not all findings support this supposition. Data from prospective longitudinal studies show that aggressive and antisocial behavior onsets in childhood, and in a small group of females it remains stable across the life-span. Unlike similar males, few of these females are convicted of crimes. This article begins with a review of descriptive studies of females sentenced by criminal courts to treatment in forensic psychiatric hospitals and moves on to present evidence showing that most female AAB does not lead to criminal prosecution. Next, studies of female AAB are reviewed, noting that it onsets in early childhood and, that in a small group remains stable across the life-span. Subsequent sections of the article focus on the two most common mental disorders presented by female forensic patients, schizophrenia and borderline personality disorder, highlighting what is known about the sub-groups of women with these disorders who present AAB. The article concludes with recommendations for earlier identification by psychiatric services of women presenting mental disorders and AAB, treatments to reduce both the symptoms of their mental disorders and their life-long AAB, and the research that is needed in order to improve the effectiveness of these treatments. The real possibilities of prevention of the development of AAB, and even perhaps aspects of the mental disorders that plague female forensic patients, are described.

Alfonso Blázquez-Castro, Angel García-Cabañes, Mercedes Carrascosa

The study and applications of ferroelectric materials in the biomedical and biotechnological fields is a novel and very promising scientific area that spans roughly one decade. However, some groups have already provided experimental proof of very interesting biological modulation when living systems are exposed to different ferroelectrics and excitation mechanisms. These materials should offer several advantages in the field of bioelectricity, such as no need of an external electric power source or circuits, scalable size of the electroactive regions, flexible and reconfigurable virtual electrodes, or fully proved biocompatibility. In this focused review we provide the underlying physics of ferroelectric activity and a recount of the research reports already published, along with some tentative biophysical mechanisms that can explain the observed results. More specifically, we focused on the biological actions of domain ferroelectrics, and ferroelectrics excited by the bulk photovoltaic effect or the pyroelectric effect. It is our goal to provide a comprehensive account of the published material so far, and to set the stage for a vigorous expansion of the field, with envisioned applications that span from cell biology and signaling to cell and tissue regeneration, antitumoral action, or cell bioengineering to name a few.

Mehmet Akçakaya, Burhaneddin Yaman, Hyungjin Chung et al.

Recently, deep learning approaches have become the main research frontier for biological image reconstruction and enhancement problems thanks to their high performance, along with their ultra-fast inference times. However, due to the difficulty of obtaining matched reference data for supervised learning, there has been increasing interest in unsupervised learning approaches that do not need paired reference data. In particular, self-supervised learning and generative models have been successfully used for various biological imaging applications. In this paper, we overview these approaches from a coherent perspective in the context of classical inverse problems, and discuss their applications to biological imaging, including electron, fluorescence and deconvolution microscopy, optical diffraction tomography and functional neuroimaging.

Emma A. Honkanen, Mikael Eklund, Simo Nuuttila et al.

Objective: To evaluate possible differences between brain dopamine transporter (DAT) binding in a group of symptomatic parkinsonism patients without dopaminergic degeneration and healthy individuals. Background: Dopaminergic neuroimaging studies of Parkinson’s disease (PD) have often used control groups formed from symptomatic patients with apparently normal striatal dopamine function. We sought to investigate whether symptomatic patients can be used to represent dopaminergically normal healthy controls. Methods: Forty healthy elderly individuals were scanned with DAT [123I]FP-CIT SPECT and compared to 69 age- and sex-matched symptomatic patients with nondegenerative conditions (including essential tremor, drug-induced parkinsonism and vascular parkinsonism). An automated region-of-interest based analysis of the caudate nucleus and the anterior/posterior putamen was performed. Specific binding ratios (SBR = [ROI-occ]/occ) were compared between the groups. Results: DAT binding in symptomatic patients was 8.6% higher in the posterior putamen than in healthy controls (p = 0.03). Binding correlated negatively with age in both groups but not with motor symptom severity, cognitive function or depression ratings. Conclusions: Putaminal DAT binding, as measured with [123I]FP-CIT SPECT, was higher in symptomatic controls than in healthy individuals. The reason for the difference is unclear but can include selection bias when DAT binding is used to aid clinical diagnosis and possible self-selection bias in healthy volunteerism. This effect should be taken into consideration when designing and interpreting neuroimaging trials investigating the dopamine system with [123I]FP-CIT SPECT.

Oriol Grau-Rivera, Irene Navalpotro-Gomez, Gonzalo Sánchez-Benavides et al.

Abstract Background Recognizing clinical manifestations heralding the development of Alzheimer’s disease (AD)-related cognitive impairment could improve the identification of individuals at higher risk of AD who may benefit from potential prevention strategies targeting preclinical population. We aim to characterize the association of body weight change with cognitive changes and AD biomarkers in cognitively unimpaired middle-aged adults. Methods This prospective cohort study included data from cognitively unimpaired adults from the ALFA study (n = 2743), a research platform focused on preclinical AD. Cognitive and anthropometric data were collected at baseline between April 2013 and November 2014. Between October 2016 and February 2020, 450 participants were visited in the context of the nested ALFA+ study and underwent cerebrospinal fluid (CSF) extraction and acquisition of positron emission tomography images with [18F]flutemetamol (FTM-PET). From these, 408 (90.1%) were included in the present study. We used data from two visits (average interval 4.1 years) to compute rates of change in weight and cognitive performance. We tested associations between these variables and between weight change and categorical and continuous measures of CSF and neuroimaging AD biomarkers obtained at follow-up. We classified participants with CSF data according to the AT (amyloid, tau) system and assessed between-group differences in weight change. Results Weight loss predicted a higher likelihood of positive FTM-PET visual read (OR 1.27, 95% CI 1.00–1.61, p = 0.049), abnormal CSF p-tau levels (OR 1.50, 95% CI 1.19–1.89, p = 0.001), and an A+T+ profile (OR 1.64, 95% CI 1.25–2.20, p = 0.001) and was greater among participants with an A+T+ profile (p < 0.01) at follow-up. Weight change was positively associated with CSF Aβ42/40 ratio (β = 0.099, p = 0.032) and negatively associated with CSF p-tau (β = − 0.141, p = 0.005), t-tau (β = − 0.147 p = 0.004) and neurogranin levels (β = − 0.158, p = 0.002). In stratified analyses, weight loss was significantly associated with higher t-tau, p-tau, neurofilament light, and neurogranin, as well as faster cognitive decline in A+ participants only. Conclusions Weight loss predicts AD CSF and PET biomarker results and may occur downstream to amyloid-β accumulation in preclinical AD, paralleling cognitive decline. Accordingly, it should be considered as an indicator of increased risk of AD-related cognitive impairment. Trial registration NCT01835717 , NCT02485730 , NCT02685969 .

Monica Daibert-Nido, Monica Daibert-Nido, Yulia Pyatova et al.

Background/Objectives: Visual field loss is frequent in patients with brain tumors, worsening their daily life and exacerbating the burden of disease, and no supportive care strategies exist. In this case series, we sought to characterize the feasibility and potential effectiveness of a home-based visual rehabilitation program in hemianopia patients using immersive virtual-reality stimulation.Subjects/Methods: Two patients, one with homonymous hemianopia and the other with bitemporal hemianopia, consecutive to pediatric brain tumors, with no prior visual rehabilitation performed 15 min of home-based audiovisual stimulation every 2 days for 6 weeks (case 2) and 7 weeks (case 1) between February and August 2020. Patients used a virtual-reality, stand-alone, and remotely controlled device loaded with a non-commercial audiovisual stimulation program managed in real time from the laboratory. Standard visual outcomes assessed in usual care in visual rehabilitation were measured at the clinic. Following a mixed method approach in this pragmatic study of two cases, we collected quantitative and qualitative data on feasibility and potential effectiveness and compared the results pre- and post-treatment.Results: Implementation and wireless delivery of the audiovisual stimulation, remote data collection, and analysis for cases 1 and 2 who completed 19/20 and 20/20 audiovisual stimulation sessions at home, respectively, altogether indicated feasibility. Contrast sensitivity increased in both eyes for cases 1 and 2. Visual fields, measured by binocular Esterman and monocular Humphrey full-field analyses, improved in case 1. A minor increase was observed in case 2. Cases 1 and 2 enhanced reading speed. Case 2 strongly improved quality of life scores.Conclusion: This is the first report of a home-based virtual-reality visual rehabilitation program for adult patients with hemianopia consecutive to a pediatric brain tumor. We show the feasibility in real-world conditions and potential effectiveness of such technology on visual perception and quality of life.