Differential associations of adolescent versus young adult cannabis initiation with longitudinal brain change and behavior.

Leveraging ~10 years of prospective longitudinal data on 704 participants, we examined the effects of adolescent versus young adult cannabis initiation on MRI-assessed cortical thickness development and behavior. Data were obtained from the IMAGEN study conducted across eight European sites. We identified IMAGEN participants who reported being cannabis-naïve at baseline and had data available at baseline, 5-year, and 9-year follow-up visits. Cannabis use was assessed with the European School Survey Project on Alcohol and Drugs. T1-weighted MR images were processed through the CIVET pipeline. Cannabis initiation occurring during adolescence (14-19 years) and young adulthood (19-22 years) was associated with differing patterns of longitudinal cortical thickness change. Associations between adolescent cannabis initiation and cortical thickness change were observed primarily in dorso- and ventrolateral portions of the prefrontal cortex. In contrast, cannabis initiation occurring between 19 and 22 years of age was associated with thickness change in temporal and cortical midline areas. Follow-up analysis revealed that longitudinal brain change related to adolescent initiation persisted into young adulthood and partially mediated the association between adolescent cannabis use and past-month cocaine, ecstasy, and cannabis use at age 22. Extent of cannabis initiation during young adulthood (from 19 to 22 years) had an indirect effect on psychotic symptoms at age 22 through thickness change in temporal areas. Results suggest that developmental timing of cannabis exposure may have a marked effect on neuroanatomical correlates of cannabis use as well as associated behavioral sequelae. Critically, this work provides a foundation for neurodevelopmentally informed models of cannabis exposure in humans.

Comparing perturbation models for evaluating stability of neuroimaging pipelines.

With an increase in awareness regarding a troubling lack of reproducibility in analytical software tools, the degree of validity in scientific derivatives and their downstream results has become unclear. The nature of reproducibility issues may vary across domains, tools, data sets, and computational infrastructures, but numerical instabilities are thought to be a core contributor. In neuroimaging, unexpected deviations have been observed when varying operating systems, software implementations, or adding negligible quantities of noise. In the field of numerical analysis, these issues have recently been explored through Monte Carlo Arithmetic, a method involving the instrumentation of floating-point operations with probabilistic noise injections at a target precision. Exploring multiple simulations in this context allows the characterization of the result space for a given tool or operation. In this article, we compare various perturbation models to introduce instabilities within a typical neuroimaging pipeline, including (i) targeted noise, (ii) Monte Carlo Arithmetic, and (iii) operating system variation, to identify the significance and quality of their impact on the resulting derivatives. We demonstrate that even low-order models in neuroimaging such as the structural connectome estimation pipeline evaluated here are sensitive to numerical instabilities, suggesting that stability is a relevant axis upon which tools are compared, alongside more traditional criteria such as biological feasibility, computational efficiency, or, when possible, accuracy. Heterogeneity was observed across participants which clearly illustrates a strong interaction between the tool and data set being processed, requiring that the stability of a given tool be evaluated with respect to a given cohort. We identify use cases for each perturbation method tested, including quality assurance, pipeline error detection, and local sensitivity analysis, and make recommendations for the evaluation of stability in a practical and analytically focused setting. Identifying how these relationships and recommendations scale to higher order computational tools, distinct data sets, and their implication on biological feasibility remain exciting avenues for future work.

The MNI data-sharing and processing ecosystem.

Neuroimaging has been facing a data deluge characterized by the exponential growth of both raw and processed data. As a result, mining the massive quantities of digital data collected in these studies offers unprecedented opportunities and has become paramount for today's research. As the neuroimaging community enters the world of "Big Data", there has been a concerted push for enhanced sharing initiatives, whether within a multisite study, across studies, or federated and shared publicly. This article will focus on the database and processing ecosystem developed at the Montreal Neurological Institute (MNI) to support multicenter data acquisition both nationally and internationally, create database repositories, facilitate data-sharing initiatives, and leverage existing software toolkits for large-scale data processing.

Mitochondrial genome annotation with MFannot: a critical analysis of gene identification and gene model prediction.

Compared to nuclear genomes, mitochondrial genomes (mitogenomes) are small and usually code for only a few dozen genes. Still, identifying genes and their structure can be challenging and time-consuming. Even automated tools for mitochondrial genome annotation often require manual analysis and curation by skilled experts. The most difficult steps are (i) the structural modelling of intron-containing genes; (ii) the identification and delineation of Group I and II introns; and (iii) the identification of moderately conserved, non-coding RNA (ncRNA) genes specifying 5S rRNAs, tmRNAs and RNase P RNAs. Additional challenges arise through genetic code evolution which can redefine the translational identity of both start and stop codons, thus obscuring protein-coding genes. Further, RNA editing can render gene identification difficult, if not impossible, without additional RNA sequence data. Current automated mito- and plastid-genome annotators are limited as they are typically tailored to specific eukaryotic groups. The MFannot annotator we developed is unique in its applicability to a broad taxonomic scope, its accuracy in gene model inference, and its capabilities in intron identification and classification. The pipeline leverages curated profile Hidden Markov Models (HMMs), covariance (CMs) and ERPIN models to better capture evolutionarily conserved signatures in the primary sequence (HMMs and CMs) as well as secondary structure (CMs and ERPIN). Here we formally describe MFannot, which has been available as a web-accessible service (https://megasun.bch.umontreal.ca/apps/mfannot/) to the research community for nearly 16 years. Further, we report its performance on particularly intron-rich mitogenomes and describe ongoing and future developments.

Web-based processing of physiological noise in fMRI: addition of the PhysIO toolbox to CBRAIN.

Neuroimaging research requires sophisticated tools for analyzing complex data, but efficiently leveraging these tools can be a major challenge, especially on large datasets. CBRAIN is a web-based platform designed to simplify the use and accessibility of neuroimaging research tools for large-scale, collaborative studies. In this paper, we describe how CBRAIN's unique features and infrastructure were leveraged to integrate TAPAS PhysIO, an open-source MATLAB toolbox for physiological noise modeling in fMRI data. This case study highlights three key elements of CBRAIN's infrastructure that enable streamlined, multimodal tool integration: a user-friendly GUI, a Brain Imaging Data Structure (BIDS) data-entry schema, and convenient in-browser visualization of results. By incorporating PhysIO into CBRAIN, we achieved significant improvements in the speed, ease of use, and scalability of physiological preprocessing. Researchers now have access to a uniform and intuitive interface for analyzing data, which facilitates remote and collaborative evaluation of results. With these improvements, CBRAIN aims to become an essential open-science tool for integrative neuroimaging research, supporting FAIR principles and enabling efficient workflows for complex analysis pipelines.

Conduct problems are associated with accelerated thinning of emotion-related cortical regions in a community-based sample of adolescents.

Few studies have examined the association between conduct problems and cerebral cortical development. Herein, we characterize the association between age-related brain change and conduct problems in a large longitudinal, community-based sample of adolescents. 1,039 participants from the IMAGEN study possessed psychopathology and surface-based morphometric data at study baseline (M = 14.42 years, SD = 0.40; 559 females) and 5-year follow-up. Self-reports of conduct problems were obtained using the Strengths and Difficulties Questionnaire (SDQ). Vertex-level linear mixed effects models were implemented using the Matlab toolbox, SurfStat. To investigate the extent to which cortical thickness maturation was qualified by dimensional measures of conduct problems, we tested for an interaction between age and SDQ Conduct Problems (CP) score. There was no main effect of CP score on cortical thickness; however, a significant "Age by CP" interaction was revealed in bilateral insulae, left inferior frontal gyrus, left rostral anterior cingulate, left posterior cingulate, and bilateral inferior parietal cortices. Across regions, follow-up analysis revealed higher levels of CP were associated with accelerated age-related thinning. Findings were not meaningfully altered when controlling for alcohol use, co-occurring psychopathology, and socioeconomic status. Results may help to further elucidate neurodevelopmental patterns linking adolescent conduct problems with adverse adult outcomes.

Data and Tools Integration in the Canadian Open Neuroscience Platform.

We present the Canadian Open Neuroscience Platform (CONP) portal to answer the research community's need for flexible data sharing resources and provide advanced tools for search and processing infrastructure capacity. This portal differs from previous data sharing projects as it integrates datasets originating from a number of already existing platforms or databases through DataLad, a file level data integrity and access layer. The portal is also an entry point for searching and accessing a large number of standardized and containerized software and links to a computing infrastructure. It leverages community standards to help document and facilitate reuse of both datasets and tools, and already shows a growing community adoption giving access to more than 60 neuroscience datasets and over 70 tools. The CONP portal demonstrates the feasibility and offers a model of a distributed data and tool management system across 17 institutions throughout Canada.

In situ quantification of mitochondrial respiration in permeabilized fibers of a marine invertebrate with low aerobic capacity.

The aim of this study was to design a protocol to allow the assessment of normal and alternative pathways for electron transport in mitochondria using an in situ approach (on permeabilized fibers) in high-resolution respirometry. We measured the oxygen consumption of permeabilized fibers from Nereis (Neanthes) virens with different substrates and the presence of ADP. To estimate the alternative oxidase (AOX) activity, antimycin A was introduced in order to inhibit complex III. Moreover, the apparent complex IV (COX) excess capacity was evaluated using different substrates to assess the implication of this complex in the partitioning of electrons during its progressive inhibition. Our in situ method enabled to quantify the activity of the normal COX pathway as well as the AOX pathway when different substrates were oxidized by either complex I, complex II or both. Using this approach, we confirmed that according to the substrates used, each pathway has a different role and consequently is otherwise involved in the partitioning of electrons through the electron transport system, and suggested that the AOX activity is triggered not only by the redox state of the cell but also by the type of substrates provided to mitochondria.

Refining Mitochondrial Intron Classification With ERPIN: Identification Based on Conservation of Sequence Plus Secondary Structure Motifs.

Mitochondrial genomes-in particular those of fungi-often encode genes with a large number of Group I and Group II introns that are conserved at both the sequence and the RNA structure level. They provide a rich resource for the investigation of intron and gene structure, self- and protein-guided splicing mechanisms, and intron evolution. Yet, the degree of sequence conservation of introns is limited, and the primary sequence differs considerably among the distinct intron sub-groups. It makes intron identification, classification, structural modeling, and the inference of gene models a most challenging and error-prone task-frequently passed on to an "expert" for manual intervention. To reduce the need for manual curation of intron structures and mitochondrial gene models, computational methods using ERPIN sequence profiles were initially developed in 2007. Here we present a refinement of search models and alignments using the now abundant publicly available fungal mtDNA sequences. In addition, we have tested in how far members of the originally proposed sub-groups are clearly distinguished and validated by our computational approach. We confirm clearly distinct mitochondrial Group I sub-groups IA1, IA3, IB3, IC1, IC2, and ID. Yet, IB1, IB2, and IB4 ERPIN models are overlapping substantially in predictions, and are therefore combined and reported as IB. We have further explored the conversion of our ERPIN profiles into covariance models (CM). Current limitations and prospects of the CM approach will be discussed.

A Simulation Toolkit for Testing the Sensitivity and Accuracy of Corticometry Pipelines.

In recent years, the replicability of neuroimaging findings has become an important concern to the research community. Neuroimaging pipelines consist of myriad numerical procedures, which can have a cumulative effect on the accuracy of findings. To address this problem, we propose a method for simulating artificial lesions in the brain in order to estimate the sensitivity and specificity of lesion detection, using different automated corticometry pipelines. We have applied this method to different versions of two widely used neuroimaging pipelines (CIVET and FreeSurfer), in terms of coefficients of variation; sensitivity and specificity of detecting lesions in 4 different regions of interest in the cortex, while introducing variations to the lesion size, the blurring kernel used prior to statistical analyses, and different thickness metrics (in CIVET). These variations are tested in a between-subject design (in two random groups, with and without lesions, using T1-weigted MRIs of 152 individuals from the International Consortium of Brain Mapping (ICBM) dataset) and in a within-subject pre-/post-lesion design [using 21 T1-Weighted MRIs of a single adult individual, scanned in the Infant Brain Imaging Study (IBIS)]. The simulation method is sensitive to partial volume effect and lesion size. Comparisons between pipelines illustrate the ability of this method to uncover differences in sensitivity and specificity of lesion detection. We propose that this method be adopted in the workflow of software development and release.

Association of Cannabis Use During Adolescence With Neurodevelopment.

IMPORTANCE: Animal studies have shown that the adolescent brain is sensitive to disruptions in endocannabinoid signaling, resulting in altered neurodevelopment and lasting behavioral effects. However, few studies have investigated ties between cannabis use and adolescent brain development in humans. OBJECTIVE: To examine the degree to which magnetic resonance (MR) imaging-assessed cerebral cortical thickness development is associated with cannabis use in a longitudinal sample of adolescents. DESIGN, SETTING, AND PARTICIPANTS: Data were obtained from the community-based IMAGEN cohort study, conducted across 8 European sites. Baseline data used in the present study were acquired from March 1, 2008, to December 31, 2011, and follow-up data were acquired from January 1, 2013, to December 31, 2016. A total of 799 IMAGEN participants were identified who reported being cannabis naive at study baseline and had behavioral and neuroimaging data available at baseline and 5-year follow-up. Statistical analysis was performed from October 1, 2019, to August 31, 2020. MAIN OUTCOMES AND MEASURES: Cannabis use was assessed at baseline and 5-year follow-up with the European School Survey Project on Alcohol and Other Drugs. Anatomical MR images were acquired with a 3-dimensional T1-weighted magnetization prepared gradient echo sequence. Quality-controlled native MR images were processed through the CIVET pipeline, version 2.1.0. RESULTS: The study evaluated 1598 MR images from 799 participants (450 female participants [56.3%]; mean [SD] age, 14.4 [0.4] years at baseline and 19.0 [0.7] years at follow-up). At 5-year follow-up, cannabis use (from 0 to >40 uses) was negatively associated with thickness in left prefrontal (peak: t785 = -4.87, cluster size = 1558 vertices; P = 1.10 × 10-6, random field theory cluster corrected) and right prefrontal (peak: t785 = -4.27, cluster size = 1551 vertices; P = 2.81 × 10-5, random field theory cluster corrected) cortices. There were no significant associations between lifetime cannabis use at 5-year follow-up and baseline cortical thickness, suggesting that the observed neuroanatomical differences did not precede initiation of cannabis use. Longitudinal analysis revealed that age-related cortical thinning was qualified by cannabis use in a dose-dependent fashion such that greater use, from baseline to follow-up, was associated with increased thinning in left prefrontal (peak: t815.27 = -4.24, cluster size = 3643 vertices; P = 2.28 × 10-8, random field theory cluster corrected) and right prefrontal (peak: t813.30 = -4.71, cluster size = 2675 vertices; P = 3.72 × 10-8, random field theory cluster corrected) cortices. The spatial pattern of cannabis-related thinning was associated with age-related thinning in this sample (r = 0.540; P < .001), and a positron emission tomography-assessed cannabinoid 1 receptor-binding map derived from a separate sample of participants (r = -0.189; P < .001). Analysis revealed that thinning in right prefrontal cortices, from baseline to follow-up, was associated with attentional impulsiveness at follow-up. CONCLUSIONS AND RELEVANCE: Results suggest that cannabis use during adolescence is associated with altered neurodevelopment, particularly in cortices rich in cannabinoid 1 receptors and undergoing the greatest age-related thickness change in middle to late adolescence.

Not only students can express alcohol dehydrogenase: goldfish can too!

This article describes a novel approach to study the metabolic regulation of the respiratory system in vertebrates that suits physiology lessons for undergraduate students. It consists of an experimental demonstration of the goldfish's (Carassius auratus) adaptations to anoxia. The goldfish is one of the few vertebrates showing strong enzymatic plasticity for the expression of alcohol dehydrogenase (ADH), which allows it to survive long periods of severe anoxia. Therefore, we propose two simple laboratory exercises in which students are first asked to characterize the distribution of ADH isozymes in the goldfish by performing cellulose acetate electrophoresis. The second part of this laboratory lesson is the determination of liver glycogen. To further student comprehension, an interspecies comparative component is integrated, in which the same subjects are studied in an anoxia-sensitive species, the brook charr (Salvelinus fontinalis). ADH in goldfish is restricted to skeletal muscles, where it catalyzes alcoholic fermentation, permitting ethanol excretion through the gills and therefore preventing lactate acidosis caused by sustained glycolysis during anoxia. Electrophoresis also reveals the occurrence of a liver isozyme in the brook charr, which ADH catalyzes in the opposite pathway, allowing the usual ethanol degradation. As for the liver glycogen assay, it shows largely superior content in the goldfish liver compared with the brook charr, providing goldfish with a sustained energy supply during anoxia. The results of this laboratory exercise clearly demonstrate several physiological strategies developed by goldfish to cope with such a crucial environmental challenge as oxygen depletion.

White matter microstructure is associated with hyperactive/inattentive symptomatology and polygenic risk for attention-deficit/hyperactivity disorder in a population-based sample of adolescents.

Few studies have investigated the link between putative biomarkers of attention-deficit/hyperactivity disorder (ADHD) symptomatology and genetic risk for ADHD. To address this, we investigate the degree to which ADHD symptomatology is associated with white matter microstructure and cerebral cortical thickness in a large population-based sample of adolescents. Critically, we then test the extent to which multimodal correlates of ADHD symptomatology are related to ADHD polygenic risk score (PRS). Neuroimaging, genetic, and behavioral data were obtained from the IMAGEN study. A dimensional ADHD composite score was derived from multi-informant ratings of ADHD symptomatology. Using tract-based spatial statistics, whole brain voxel-wise regressions between fractional anisotropy (FA) and ADHD composite score were calculated. Local cortical thickness was regressed on ADHD composite score. ADHD PRS was based on a very recent genome-wide association study, and calculated using PRSice. ADHD composite score was negatively associated with FA in several white matter pathways, including bilateral superior and inferior longitudinal fasciculi (p < 0.05, corrected). ADHD composite score was negatively associated with orbitofrontal cortical thickness (p < 0.05, corrected). The ADHD composite score was correlated with ADHD PRS (p < 0.001). FA correlates of ADHD symptomatology were significantly associated with ADHD PRS, whereas cortical thickness correlates of ADHD symptomatology were unrelated to ADHD PRS. Variation in hyperactive/inattentive symptomatology was associated with white matter microstructure, which, in turn, was related to ADHD PRS. Results suggest that genetic risk for ADHD symptomatology may be tied to biological processes affecting white matter microstructure.

Amygdalar reactivity is associated with prefrontal cortical thickness in a large population-based sample of adolescents.

In structural neuroimaging studies, reduced cerebral cortical thickness in orbital and ventromedial prefrontal regions is frequently interpreted as reflecting an impaired ability to downregulate neuronal activity in the amygdalae. Unfortunately, little research has been conducted in order to test this conjecture. We examine the extent to which amygdalar reactivity is associated with cortical thickness in a population-based sample of adolescents. Data were obtained from the IMAGEN study, which includes 2,223 adolescents. While undergoing functional neuroimaging, participants passively viewed video clips of a face that started from a neutral expression and progressively turned angry, or, instead, turned to a second neutral expression. Left and right amygdala ROIs were used to extract mean BOLD signal change for the angry minus neutral face contrast for all subjects. T1-weighted images were processed through the CIVET pipeline (version 2.1.0). In variable-centered analyses, local cortical thickness was regressed against amygdalar reactivity using first and second-order linear models. In a follow-up person-centered analysis, we defined a "high reactive" group of participants based on mean amygdalar BOLD signal change for the angry minus neutral face contrast. Between-group differences in cortical thickness were examined ("high reactive" versus all other participants). A significant association was revealed between the continuous measure of amygdalar reactivity and bilateral ventromedial prefrontal cortical thickness in a second-order linear model (p < 0.05, corrected). The "high reactive" group, in comparison to all other participants, possessed reduced cortical thickness in bilateral orbital and ventromedial prefrontal cortices, bilateral anterior temporal cortices, left caudal middle temporal gyrus, and the left inferior and middle frontal gyri (p < 0.05, corrected). Results are consistent with non-human primate studies, and provide empirical support for an association between reduced prefrontal cortical thickness and amygdalar reactivity. Future research will likely benefit from investigating the degree to which psychopathology qualifies relations between prefrontal cortical structure and amygdalar reactivity.

Integration of "omics" Data and Phenotypic Data Within a Unified Extensible Multimodal Framework.

Analysis of "omics" data is often a long and segmented process, encompassing multiple stages from initial data collection to processing, quality control and visualization. The cross-modal nature of recent genomic analyses renders this process challenging to both automate and standardize; consequently, users often resort to manual interventions that compromise data reliability and reproducibility. This in turn can produce multiple versions of datasets across storage systems. As a result, scientists can lose significant time and resources trying to execute and monitor their analytical workflows and encounter difficulties sharing versioned data. In 2015, the Ludmer Centre for Neuroinformatics and Mental Health at McGill University brought together expertise from the Douglas Mental Health University Institute, the Lady Davis Institute and the Montreal Neurological Institute (MNI) to form a genetics/epigenetics working group. The objectives of this working group are to: (i) design an automated and seamless process for (epi)genetic data that consolidates heterogeneous datasets into the LORIS open-source data platform; (ii) streamline data analysis; (iii) integrate results with provenance information; and (iv) facilitate structured and versioned sharing of pipelines for optimized reproducibility using high-performance computing (HPC) environments via the CBRAIN processing portal. This article outlines the resulting generalizable "omics" framework and its benefits, specifically, the ability to: (i) integrate multiple types of biological and multi-modal datasets (imaging, clinical, demographics and behavioral); (ii) automate the process of launching analysis pipelines on HPC platforms; (iii) remove the bioinformatic barriers that are inherent to this process; (iv) ensure standardization and transparent sharing of processing pipelines to improve computational consistency; (v) store results in a queryable web interface; (vi) offer visualization tools to better view the data; and (vii) provide the mechanisms to ensure usability and reproducibility. This framework for workflows facilitates brain research discovery by reducing human error through automation of analysis pipelines and seamless linking of multimodal data, allowing investigators to focus on research instead of data handling.

Boutiques: a flexible framework to integrate command-line applications in computing platforms.

We present Boutiques, a system to automatically publish, integrate, and execute command-line applications across computational platforms. Boutiques applications are installed through software containers described in a rich and flexible JSON language. A set of core tools facilitates the construction, validation, import, execution, and publishing of applications. Boutiques is currently supported by several distinct virtual research platforms, and it has been used to describe dozens of applications in the neuroinformatics domain. We expect Boutiques to improve the quality of application integration in computational platforms, to reduce redundancy of effort, to contribute to computational reproducibility, and to foster Open Science.

Energetic metabolism and biochemical adaptation: A bird flight muscle model.

The main objective of this class experiment is to measure the activity of two metabolic enzymes in crude extract from bird pectoral muscle and to relate the differences to their mode of locomotion and ecology. The laboratory is adapted to stimulate the interest of wildlife management students to biochemistry. The enzymatic activities of cytochrome c oxidase and lactate dehydrogenase are measured in pectoral muscle of black duck and ring-necked pheasant. The black ducks have a high cytochrome c oxidase/lactate dehydrogenase (LDH) ratio, which reflects high aerobic capacity required for sustained and long distance flight. The low cytochrome c oxidase/LDH ratio in ring-necked pheasants and high level of LDH activity suggest that this bird can only support short bursts of flight, which may be related to his strategy of predator avoidance.

Cyberinfrastructure for Open Science at the Montreal Neurological Institute.

Data sharing is becoming more of a requirement as technologies mature and as global research and communications diversify. As a result, researchers are looking for practical solutions, not only to enhance scientific collaborations, but also to acquire larger amounts of data, and to access specialized datasets. In many cases, the realities of data acquisition present a significant burden, therefore gaining access to public datasets allows for more robust analyses and broadly enriched data exploration. To answer this demand, the Montreal Neurological Institute has announced its commitment to Open Science, harnessing the power of making both clinical and research data available to the world (Owens, 2016a,b). As such, the LORIS and CBRAIN (Das et al., 2016) platforms have been tasked with the technical challenges specific to the institutional-level implementation of open data sharing, including: Comprehensive linking of multimodal data (phenotypic, clinical, neuroimaging, biobanking, and genomics, etc.)Secure database encryption, specifically designed for institutional and multi-project data sharing, ensuring subject confidentiality (using multi-tiered identifiers).Querying capabilities with multiple levels of single study and institutional permissions, allowing public data sharing for all consented and de-identified subject data.Configurable pipelines and flags to facilitate acquisition and analysis, as well as access to High Performance Computing clusters for rapid data processing and sharing of software tools.Robust Workflows and Quality Control mechanisms ensuring transparency and consistency in best practices.Long term storage (and web access) of data, reducing loss of institutional data assets.Enhanced web-based visualization of imaging, genomic, and phenotypic data, allowing for real-time viewing and manipulation of data from anywhere in the world.Numerous modules for data filtering, summary statistics, and personalized and configurable dashboards. Implementing the vision of Open Science at the Montreal Neurological Institute will be a concerted undertaking that seeks to facilitate data sharing for the global research community. Our goal is to utilize the years of experience in multi-site collaborative research infrastructure to implement the technical requirements to achieve this level of public data sharing in a practical yet robust manner, in support of accelerating scientific discovery.

Reproducibility of neuroimaging analyses across operating systems.

Neuroimaging pipelines are known to generate different results depending on the computing platform where they are compiled and executed. We quantify these differences for brain tissue classification, fMRI analysis, and cortical thickness (CT) extraction, using three of the main neuroimaging packages (FSL, Freesurfer and CIVET) and different versions of GNU/Linux. We also identify some causes of these differences using library and system call interception. We find that these packages use mathematical functions based on single-precision floating-point arithmetic whose implementations in operating systems continue to evolve. While these differences have little or no impact on simple analysis pipelines such as brain extraction and cortical tissue classification, their accumulation creates important differences in longer pipelines such as subcortical tissue classification, fMRI analysis, and cortical thickness extraction. With FSL, most Dice coefficients between subcortical classifications obtained on different operating systems remain above 0.9, but values as low as 0.59 are observed. Independent component analyses (ICA) of fMRI data differ between operating systems in one third of the tested subjects, due to differences in motion correction. With Freesurfer and CIVET, in some brain regions we find an effect of build or operating system on cortical thickness. A first step to correct these reproducibility issues would be to use more precise representations of floating-point numbers in the critical sections of the pipelines. The numerical stability of pipelines should also be reviewed.

CBRAIN: a web-based, distributed computing platform for collaborative neuroimaging research.

The Canadian Brain Imaging Research Platform (CBRAIN) is a web-based collaborative research platform developed in response to the challenges raised by data-heavy, compute-intensive neuroimaging research. CBRAIN offers transparent access to remote data sources, distributed computing sites, and an array of processing and visualization tools within a controlled, secure environment. Its web interface is accessible through any modern browser and uses graphical interface idioms to reduce the technical expertise required to perform large-scale computational analyses. CBRAIN's flexible meta-scheduling has allowed the incorporation of a wide range of heterogeneous computing sites, currently including nine national research High Performance Computing (HPC) centers in Canada, one in Korea, one in Germany, and several local research servers. CBRAIN leverages remote computing cycles and facilitates resource-interoperability in a transparent manner for the end-user. Compared with typical grid solutions available, our architecture was designed to be easily extendable and deployed on existing remote computing sites with no tool modification, administrative intervention, or special software/hardware configuration. As October 2013, CBRAIN serves over 200 users spread across 53 cities in 17 countries. The platform is built as a generic framework that can accept data and analysis tools from any discipline. However, its current focus is primarily on neuroimaging research and studies of neurological diseases such as Autism, Parkinson's and Alzheimer's diseases, Multiple Sclerosis as well as on normal brain structure and development. This technical report presents the CBRAIN Platform, its current deployment and usage and future direction.