Limits to growth: why neuroscience needs large-scale science

Soul, butterfly, mythological nymph: psyche in philosophy and neuroscience

The term "psyche" and its derivatives - including "Psychology" and "Psychiatry" - are rooted in classical philosophy and in mythology. Over the centuries, psyche has been the subject of discourse and contemplation, and of fable; it has also come to signify, in entomology, the order of Lepidoptera. In the current surge of research on brain and mind, there is a gradual transition from the psyche (or the "soul") to the specified descriptors defined by the fields of Behavioral, Cognitive and Integrative Neuroscience.

Morphological Neuron Classification Using Machine Learning

Classification and quantitative characterization of neuronal morphologies from histological neuronal reconstruction is challenging since it is still unclear how to delineate a neuronal cell class and which are the best features to define them by. The morphological neuron characterization represents a primary source to address anatomical comparisons, morphometric analysis of cells, or brain modeling. The objectives of this paper are (i) to develop and integrate a pipeline that goes from morphological feature extraction to classification and (ii) to assess and compare the accuracy of machine learning algorithms to classify neuron morphologies. The algorithms were trained on 430 digitally reconstructed neurons subjectively classified into layers and/or m-types using young and/or adult development state population of the somatosensory cortex in rats. For supervised algorithms, linear discriminant analysis provided better classification results in comparison with others. For unsupervised algorithms, the affinity propagation and the Ward algorithms provided slightly better results.

Proposed Training to Meet Challenges of Large-Scale Data in Neuroscience

The scale of data being produced in neuroscience at present and in the future creates new and unheralded challenges, outstripping conventional ways of handling, considering, and analyzing data. As neuroinformatics enters into this big data era, a need for a highly trained and perhaps unique workforce is emerging. To determine the staffing needs created by the impending era of big data, a workshop (iNeuro Project) was convened November 13–14, 2014. Participants included data resource providers, bioinformatics/analytics trainers, computer scientists, library scientists, and neuroscience educators. These individuals provided perspectives on the challenges of big data, the preparation of a workforce to meet these challenges, and the present state of training programs. Participants discussed whether suitable training programs will need to be constructed from scratch or if existing programs can serve as models. Currently, most programs at the undergraduate and graduate levels are located in Europe—participants knew of none in the United States. The skill sets that training programs would need to provide as well as the curriculum necessary to teach them were also discussed. Consistent with Vision and Change in Undergraduate Biology Education: A Call to Action¹, proposed curricula included authentic, hands-on research experiences. Further discussions revolved around the logistics and barriers to creating such programs. The full white paper, iNeuro Project Workshop Report, is available from iNeuro Project².

Interacting with the National Database for Autism Research (NDAR) via the LONI Pipeline workflow environment

Under the umbrella of the National Database for Clinical Trials (NDCT) related to mental illnesses, the National Database for Autism Research (NDAR) seeks to gather, curate, and make openly available neuroimaging data from NIH-funded studies of autism spectrum disorder (ASD). NDAR has recently made its database accessible through the LONI Pipeline workflow design and execution environment to enable large-scale analyses of cortical architecture and function via local, cluster, or "cloud"-based computing resources. This presents a unique opportunity to overcome many of the customary limitations to fostering biomedical neuroimaging as a science of discovery. Providing open access to primary neuroimaging data, workflow methods, and high-performance computing will increase uniformity in data collection protocols, encourage greater reliability of published data, results replication, and broaden the range of researchers now able to perform larger studies than ever before. To illustrate the use of NDAR and LONI Pipeline for performing several commonly performed neuroimaging processing steps and analyses, this paper presents example workflows useful for ASD neuroimaging researchers seeking to begin using this valuable combination of online data and computational resources. We discuss the utility of such database and workflow processing interactivity as a motivation for the sharing of additional primary data in ASD research and elsewhere.

National Database for Autism Research (NDAR): Big Data Opportunities for Health Services Research and Health Technology Assessment

The National Database for Autism Research (NDAR) is a US National Institutes of Health (NIH)-funded research data repository created by integrating heterogeneous datasets through data sharing agreements between autism researchers and the NIH. To date, NDAR is considered the largest neuroscience and genomic data repository for autism research. In addition to biomedical data, NDAR contains a large collection of clinical and behavioral assessments and health outcomes from novel interventions. Importantly, NDAR has a global unique patient identifier that can be linked to aggregated individual-level data for hypothesis generation and testing, and for replicating research findings. As such, NDAR promotes collaboration and maximizes public investment in the original data collection. As screening and diagnostic technologies as well as interventions for children with autism are expensive, health services research (HSR) and health technology assessment (HTA) are needed to generate more evidence to facilitate implementation when warranted. This article describes NDAR and explains its value to health services researchers and decision scientists interested in autism and other mental health conditions. We provide a description of the scope and structure of NDAR and illustrate how data are likely to grow over time and become available for HSR and HTA.

Neurogenomics: An opportunity to integrate neuroscience, genomics and bioinformatics research in Africa

Modern genomic approaches have made enormous contributions to improving our understanding of the function, development and evolution of the nervous system, and the diversity within and between species. However, most of these research advances have been recorded in countries with advanced scientific resources and funding support systems. On the contrary, little is known about, for example, the possible interplay between different genes, non-coding elements and environmental factors in modulating neurological diseases among populations in low-income countries, including many African countries. The unique ancestry of African populations suggests that improved inclusion of these populations in neuroscience-related genomic studies would significantly help to identify novel factors that might shape the future of neuroscience research and neurological healthcare. This perspective is strongly supported by the recent identification that diseased individuals and their kindred from specific sub-Saharan African populations lack common neurological disease-associated genetic mutations. This indicates that there may be population-specific causes of neurological diseases, necessitating further investigations into the contribution of additional, presently-unknown genomic factors. Here, we discuss how the development of neurogenomics research in Africa would help to elucidate disease-related genomic variants, and also provide a good basis to develop more effective therapies. Furthermore, neurogenomics would harness African scientists' expertise in neuroscience, genomics and bioinformatics to extend our understanding of the neural basis of behaviour, development and evolution.

A blueprint for research on Shankopathies: a view from research on autism spectrum disorder

Autism spectrum disorders (ASD) are associated with mutations in a host of genes including a number that function in synaptic transmission. Phelan McDermid syndrome involves mutations in SHANK3 which encodes a protein that forms a scaffold for glutamate receptors at the synapse. SHANK3 is one of the genes that underpins the synaptic hypothesis for ASD. We discuss this hypothesis with a view to the broader context of ASD and with special emphasis on highly penetrant genetic disorders including Shankopathies. We propose a blueprint for near and longer-term goals for fundamental and translational research on Shankopathies.

Towards structured sharing of raw and derived neuroimaging data across existing resources

Data sharing efforts increasingly contribute to the acceleration of scientific discovery. Neuroimaging data is accumulating in distributed domain-specific databases and there is currently no integrated access mechanism nor an accepted format for the critically important meta-data that is necessary for making use of the combined, available neuroimaging data. In this manuscript, we present work from the Derived Data Working Group, an open-access group sponsored by the Biomedical Informatics Research Network (BIRN) and the International Neuroimaging Coordinating Facility (INCF) focused on practical tools for distributed access to neuroimaging data. The working group develops models and tools facilitating the structured interchange of neuroimaging meta-data and is making progress towards a unified set of tools for such data and meta-data exchange. We report on the key components required for integrated access to raw and derived neuroimaging data as well as associated meta-data and provenance across neuroimaging resources. The components include (1) a structured terminology that provides semantic context to data, (2) a formal data model for neuroimaging with robust tracking of data provenance, (3) a web service-based application programming interface (API) that provides a consistent mechanism to access and query the data model, and (4) a provenance library that can be used for the extraction of provenance data by image analysts and imaging software developers. We believe that the framework and set of tools outlined in this manuscript have great potential for solving many of the issues the neuroimaging community faces when sharing raw and derived neuroimaging data across the various existing database systems for the purpose of accelerating scientific discovery.

Do brain image databanks support understanding of normal ageing brain structure? A systematic review

OBJECTIVE

To document accessible magnetic resonance (MR) brain images, metadata and statistical results from normal older subjects that may be used to improve diagnoses of dementia.

METHODS

We systematically reviewed published brain image databanks (print literature and Internet) concerned with normal ageing brain structure.

RESULTS

From nine eligible databanks, there appeared to be 944 normal subjects aged ≥60 years. However, many subjects were in more than one databank and not all were fully representative of normal ageing clinical characteristics. Therefore, there were approximately 343 subjects aged ≥60 years with metadata representative of normal ageing, but only 98 subjects were openly accessible. No databank had the range of MR image sequences, e.g. T2*, fluid-attenuated inversion recovery (FLAIR), required to effectively characterise the features of brain ageing. No databank supported random subject retrieval; therefore, manual selection bias and errors may occur in studies that use these subjects as controls. Finally, no databank stored results from statistical analyses of its brain image and metadata that may be validated with analyses of further data.

CONCLUSION

Brain image databanks require open access, more subjects, metadata, MR image sequences, searchability and statistical results to improve understanding of normal ageing brain structure and diagnoses of dementia.

KEY POINTS

• We reviewed databanks with structural MR brain images of normal older people. • Among these nine databanks, 98 normal subjects ≥60 years were openly accessible. • None had all the required sequences, random subject retrieval or statistical results. • More access, subjects, sequences, metadata, searchability and results are needed. • These may improve understanding of normal brain ageing and diagnoses of dementia.

Open source tools for the information theoretic analysis of neural data

The recent and rapid development of open source software tools for the analysis of neurophysiological datasets consisting of simultaneous multiple recordings of spikes, field potentials and other neural signals holds the promise for a significant advance in the standardization, transparency, quality, reproducibility and variety of techniques used to analyze neurophysiological data and for the integration of information obtained at different spatial and temporal scales. In this review we focus on recent advances in open source toolboxes for the information theoretic analysis of neural responses. We also present examples of their use to investigate the role of spike timing precision, correlations across neurons, and field potential fluctuations in the encoding of sensory information. These information toolboxes, available both in MATLAB and Python programming environments, hold the potential to enlarge the domain of application of information theory to neuroscience and to lead to new discoveries about how neurons encode and transmit information.

Using brain-based cognitive measures to support clinical decisions in ADHD

Measures of cognition support diagnostic and treatment decisions in attention deficit hyperactivity disorder. We used an integrative neuroscience framework to assess cognition and associated brain-function correlates in large attention deficit hyperactivity disorder and healthy groups. Matched groups of 175 attention deficit hyperactivity disorder children/adolescents and 175 healthy control subjects were assessed clinically, with the touch screen-based cognitive assessment battery "IntegNeuro" (Brain Resource Ltd., Sydney, Australia) and the "LabNeuro" (Brain Resource Ltd., Sydney, Australia) platform for psychophysiologic recordings of brain function and body arousal. IntegNeuro continuous performance task measures of sustained attention classified 68% of attention deficit hyperactivity disorder patients with 76% specificity, consistent with previous reports. Our additional cognitive measures of impulsivity, intrusive errors, inhibition, and response variability improved sensitivity to 88%, and specificity to 91%. Positive predictive power was 96%, and negative predictive power, 88%. These metrics were stable across attention deficit hyperactivity disorder subtypes and age. Consistent with their brain-based validity, cognitive measures were correlated with corresponding brain-function and body-arousal measures. We propose a combination of candidate cognitive "markers" that define a signature for attention deficit hyperactivity disorder: "sustained attention," "impulsivity," "inhibition," "intrusions," and "response variability." These markers offer a frame of reference to support diagnostic and treatment decisions, and an objective benchmark for monitoring outcomes of interventions.

Data sharing for computational neuroscience

Computational neuroscience is a subfield of neuroscience that develops models to integrate complex experimental data in order to understand brain function. To constrain and test computational models, researchers need access to a wide variety of experimental data. Much of those data are not readily accessible because neuroscientists fall into separate communities that study the brain at different levels and have not been motivated to provide data to researchers outside their community. To foster sharing of neuroscience data, a workshop was held in 2007, bringing together experimental and theoretical neuroscientists, computer scientists, legal experts and governmental observers. Computational neuroscience was recommended as an ideal field for focusing data sharing, and specific methods, strategies and policies were suggested for achieving it. A new funding area in the NSF/NIH Collaborative Research in Computational Neuroscience (CRCNS) program has been established to support data sharing, guided in part by the workshop recommendations. The new funding area is dedicated to the dissemination of high quality data sets with maximum scientific value for computational neuroscience. The first round of the CRCNS data sharing program supports the preparation of data sets which will be publicly available in 2008. These include electrophysiology and behavioral (eye movement) data described towards the end of this article.

Analysis of the cerebral transcriptome in mice subjected to traumatic brain injury: importance of IL-6

Traumatic brain injury is one of the leading causes of incapacity and death among young people. Injury to the brain elicits a potent inflammatory response, comprising recruitment of inflammatory cells, reactive astrogliosis and activation of brain macrophages. Under the influence of presumably several cytokines and growth factors, a cascade of events is activated that result ultimately in increased oxidative stress and tissue damage, but also in activation of counterregulatory factors and tissue regeneration. The complexity of this response is being unraveled by high-throughput methodologies such as microarrays. The combination of these modern techniques with the comparison of normal and genetically modified mice boosts the significance of the results obtained. With this approach, we have demonstrated that a cytokine such as interleukin-6 is one of the key players in the response of the brain to injury.

Novel roles for metallothionein-I + II (MT-I + II) in defense responses, neurogenesis, and tissue restoration after traumatic brain injury: Insights from global gene expression profiling in wild-type and MT-I + II knockout mice

Traumatic injury to the brain is one of the leading causes of injury-related death or disability, especially among young people. Inflammatory processes and oxidative stress likely underlie much of the damage elicited by injury, but the full repertoire of responses involved is not well known. A genomic approach, such as the use of microarrays, provides much insight in this regard, especially if combined with the use of gene-targeted animals. We report here the results of one of these studies comparing wild-type and metallothionein-I + II knockout mice subjected to a cryolesion of the somatosensorial cortex and killed at 0, 1, 4, 8, and 16 days postlesion (dpl) using Affymetrix genechips/oligonucleotide arrays interrogating approximately 10,000 different murine genes (MG_U74Av2). Hierarchical clustering analysis of these genes readily shows an orderly pattern of gene responses at specific times consistent with the processes involved in the initial tissue injury and later regeneration of the parenchyma, as well as a prominent effect of MT-I + II deficiency. The results thoroughly confirmed the importance of the antioxidant proteins MT-I + II in the response of the brain to injury and opened new avenues that were confirmed by immunohistochemistry. Data in KO, MT-I-overexpressing, and MT-II-injected mice strongly suggest a role of these proteins in postlesional activation of neural stem cells.

SynDB: a Synapse protein DataBase based on synapse ontology

A synapse is the junction across which a nerve impulse passes from an axon terminal to a neuron, muscle cell or gland cell. The functions and building molecules of the synapse are essential to almost all neurobiological processes. To describe synaptic structures and functions, we have developed Synapse Ontology (SynO), a hierarchical representation that includes 177 terms with hundreds of synonyms and branches up to eight levels deep. associated 125 additional protein keywords and 109 InterPro domains with these SynO terms. Using a combination of automated keyword searches, domain searches and manual curation, we collected 14,000 non-redundant synapse-related proteins, including 3000 in human. We extensively annotated the proteins with information about sequence, structure, function, expression, pathways, interactions and disease associations and with hyperlinks to external databases. The data are stored and presented in the Synapse protein DataBase (SynDB, http://syndb.cbi.pku.edu.cn). SynDB can be interactively browsed by SynO, Gene Ontology (GO), domain families, species, chromosomal locations or Tribe-MCL clusters. It can also be searched by text (including Boolean operators) or by sequence similarity. SynDB is the most comprehensive database to date for synaptic proteins.

Reducing interscanner variability of activation in a multicenter fMRI study: Controlling for signal-to-fluctuation-noise-ratio (SFNR) differences

Variation in scanner performance will lead to variation in activation patterns in multicenter fMRI studies. The purpose of this investigation was to evaluate the effect of statistically covarying for scanner differences in signal-to-fluctuation-noise-ratio (SFNR) on reducing scanner differences in activation effect size as part of a multicenter fMRI project (FIRST BIRN). For SFNR, "signal" is typically the mean intensity over time and "fluctuation noise" is the temporal standard deviation. Five subjects were sent to 9 centers (10 scanners) and scanned on two consecutive days using a sensorimotor fMRI protocol. High-field (4 T and 3 T) and low-field (1.5 T) scanners from three vendors (GE, Siemens and Picker) were included. The effect size for the detection of neural activation during a sensorimotor task was evaluated as the percent of temporal variance accounted for by our model (percent of variance accounted for, or PVAF). Marked scanner effects were noted for both PVAF as well as SFNR. After covariate adjustment with one of several measures of SFNR, there were dramatic reductions in scanner-to-scanner variations in activation effect size. Variance components analyses revealed 75%-81% reductions in variance due to scanner with this method. Thus, controlling for scanner variation in SFNR may be an effective method to homogenize activation effect sizes in multicenter studies.

Modulation of anxiety circuits by serotonergic systems

Anxiety is a complex emotional state associated with sustained heightened autonomic and behavioral arousal and an increase in avoidance behavior. Anxiety-related behavior is a form of risk assessment behavior that is associated with a level of uncertainty or unpredictability regarding the outcome of emotionally salient events, often when both rewarding and aversive outcomes are possible. In this review, we highlight recent advances in our understanding of the neural circuits regulating anxiety states and anxiety-related behavior with an emphasis on the role of brainstem serotonergic systems in modulating anxiety-related circuits. In particular, we explore the possibility that the regulation of anxiety states and anxiety-related behavior by serotonergic systems is dependent on a specific, topographically organized mesolimbocortical serotonergic system that originates in the mid-rostrocaudal and caudal parts of the dorsal raphe nucleus.

Discovery and integrative neuroscience

Hypothesis driven research has been shown to be an excellent model for pursuing investigations in neuroscience. The Human Genome Project demonstrated the added value of discovery research, especially in areas where large amounts of data are produced. Neuroscience has become a data rich field, and one that would be enhanced by incorporating the discovery approach. Databases, as well as analytical, modeling and simulation tools, will have to be developed, and they will need to be interoperable and federated. This paper presents an overview of the development of the field of neuroscience databases and associate tools: Neuroinformatics. The primary focus is on the impact of NIH funding of this process. The important issues of data sharing, as viewed from the perspective of the scientist and private and public funding organizations, are discussed. Neuroinformatics will provide more than just a sophisticated array of information technologies to help scientists understand and integrate nervous system data. It will make available powerful models of neural functions and facilitate discovery, hypothesis formulation and electronic collaboration.

Integrative neuroscience: the role of a standardized database

Most brain related databases bring together specialized information, with a growing number that include neuroimaging measures. This article outlines the potential use and insights from the first entirely standardized and centralized database, which integrates information from neuroimaging measures (EEG, event related potential (ERP), structural/functional MRI), arousal (skin conductance responses (SCR)s, heart rate, respiration), neuropsychological and personality tests, genomics and demographics: The Brain Resource International Database. It comprises data from over 2000 "normative" subjects and a growing number of patients with neurological and psychiatric illnesses, acquired from over 50 laboratories (in the U.S.A, United Kingdom, Holland, South Africa, Israel and Australia), all with identical equipment and experimental procedures. Three primary goals of this database are to quantify individual differences in normative brain function, to compare an individual's performance to their database peers, and to provide a robust normative framework for clinical assessment and treatment prediction. We present three example demonstrations in relation to these goals. First, we show how consistent age differences may be quantified when large subject numbers are available, using EEG and ERP data from nearly 2000 stringently screened. normative subjects. Second, the use of a normalization technique provides a means to compare clinical subjects (50 ADHD subjects in this study) to the normative database with the effects of age and gender taken into account. Third, we show how a profile of EEG/ERP and autonomic measures potentially provides a means to predict treatment response in ADHD subjects. The example data consists of EEG under eyes open and eyes closed and ERP data for auditory oddball, working memory and Go-NoGo paradigms. Autonomic measures of skin conductance (tonic skin conductance level, SCL, and phasic skin conductance responses, SCRs) were acquired simultaneously with central EEG/ERP measures. The findings show that the power of large samples, tested using standardized protocols, allows for the quantification of individual differences that can subsequently be used to control such variation and to enhance the sensitivity and specificity of comparisons between normative and clinical groups. In terms of broader significance, the combination of size and multidimensional measures tapping the brain's core cognitive competencies, may provide a normative and evidence-based framework for individually-based assessments in "Personalized Medicine."

Brain response to traumatic brain injury in wild-type and interleukin-6 knockout mice: a microarray analysis

Traumatic injury to the brain is one of the leading causes of injury-related death or disability. Brain response to injury is orchestrated by cytokines, such as interleukin (IL)-6, but the full repertoire of responses involved is not well known. We here report the results obtained with microarrays in wild-type and IL-6 knockout mice subjected to a cryolesion of the somatosensorial cortex and killed at 0, 1, 4, 8 and 16 days post-lesion. Overall gene expression was analyzed by using Affymetrix genechips/oligonucleotide arrays with approximately 12,400 probe sets corresponding to approximately 10,000 different murine genes (MG_U74Av2). A robust, conventional statistical method (two-way anova) was employed to select the genes significantly affected. An orderly pattern of gene responses was clearly detected, with genes being up- or down-regulated at specific timings consistent with the processes involved in the initial tissue injury and later regeneration of the parenchyma. IL-6 deficiency showed a dramatic effect in the expression of many genes, especially in the 1 day post-lesion timing, which presumably underlies the poor capacity of IL-6 knockout mice to cope with brain damage. The results highlight the importance of IL-6 controlling the response of the brain to injury as well as the suitability of microarrays for identifying specific targets worthy of further study.