BigNeuron: a resource to benchmark and predict performance of algorithms for automated tracing of neurons in light microscopy datasets.

BigNeuron is an open community bench-testing platform with the goal of setting open standards for accurate and fast automatic neuron tracing. We gathered a diverse set of image volumes across several species that is representative of the data obtained in many neuroscience laboratories interested in neuron tracing. Here, we report generated gold standard manual annotations for a subset of the available imaging datasets and quantified tracing quality for 35 automatic tracing algorithms. The goal of generating such a hand-curated diverse dataset is to advance the development of tracing algorithms and enable generalizable benchmarking. Together with image quality features, we pooled the data in an interactive web application that enables users and developers to perform principal component analysis, t-distributed stochastic neighbor embedding, correlation and clustering, visualization of imaging and tracing data, and benchmarking of automatic tracing algorithms in user-defined data subsets. The image quality metrics explain most of the variance in the data, followed by neuromorphological features related to neuron size. We observed that diverse algorithms can provide complementary information to obtain accurate results and developed a method to iteratively combine methods and generate consensus reconstructions. The consensus trees obtained provide estimates of the neuron structure ground truth that typically outperform single algorithms in noisy datasets. However, specific algorithms may outperform the consensus tree strategy in specific imaging conditions. Finally, to aid users in predicting the most accurate automatic tracing results without manual annotations for comparison, we used support vector machine regression to predict reconstruction quality given an image volume and a set of automatic tracings.

Modeling the impact of neuromorphological alterations in Down syndrome on fast neural oscillations.

Cognitive disorders, including Down syndrome (DS), present significant morphological alterations in neuron architectural complexity. However, the relationship between neuromorphological alterations and impaired brain function is not fully understood. To address this gap, we propose a novel computational model that accounts for the observed cell deformations in DS. The model consists of a cross-sectional layer of the mouse motor cortex, composed of 3000 neurons. The network connectivity is obtained by accounting explicitly for two single-neuron morphological parameters: the mean dendritic tree radius and the spine density in excitatory pyramidal cells. We obtained these values by fitting reconstructed neuron data corresponding to three mouse models: wild-type (WT), transgenic (TgDyrk1A), and trisomic (Ts65Dn). Our findings reveal a dynamic interplay between pyramidal and fast-spiking interneurons leading to the emergence of gamma activity (∼40 Hz). In the DS models this gamma activity is diminished, corroborating experimental observations and validating our computational methodology. We further explore the impact of disrupted excitation-inhibition balance by mimicking the reduction recurrent inhibition present in DS. In this case, gamma power exhibits variable responses as a function of the external input to the network. Finally, we perform a numerical exploration of the morphological parameter space, unveiling the direct influence of each structural parameter on gamma frequency and power. Our research demonstrates a clear link between changes in morphology and the disruption of gamma oscillations in DS. This work underscores the potential of computational modeling to elucidate the relationship between neuron architecture and brain function, and ultimately improve our understanding of cognitive disorders.

Deficits in neuronal architecture but not over-inhibition are main determinants of reduced neuronal network activity in a mouse model of overexpression of Dyrk1A.

In this study, we investigated the impact of Dual specificity tyrosine-phosphorylation-regulated kinase 1A (Dyrk1A) overexpression, a gene associated with Down syndrome, on hippocampal neuronal deficits in mice. Our findings revealed that mice overexpressing Dyrk1A (TgDyrk1A; TG) exhibited impaired hippocampal recognition memory, disrupted excitation-inhibition balance, and deficits in long-term potentiation (LTP). Specifically, we observed layer-specific deficits in dendritic arborization of TG CA1 pyramidal neurons in the stratum radiatum. Through computational modeling, we determined that these alterations resulted in reduced storage capacity and compromised integration of inputs, with decreased high γ oscillations. Contrary to prevailing assumptions, our model suggests that deficits in neuronal architecture, rather than over-inhibition, primarily contribute to the reduced network. We explored the potential of environmental enrichment (EE) as a therapeutic intervention and found that it normalized the excitation-inhibition balance, restored LTP, and improved short-term recognition memory. Interestingly, we observed transient significant dendritic remodeling, leading to recovered high γ. However, these effects were not sustained after EE discontinuation. Based on our findings, we conclude that Dyrk1A overexpression-induced layer-specific neuromorphological disturbances impair the encoding of place and temporal context. These findings contribute to our understanding of the underlying mechanisms of Dyrk1A-related hippocampal deficits and highlight the challenges associated with long-term therapeutic interventions for cognitive impairments.

Non-homogenous axonal bouton distribution in whole-brain single-cell neuronal networks.

We examined the distribution of pre-synaptic contacts in axons of mouse neurons and constructed whole-brain single-cell neuronal networks using an extensive dataset of 1,891 fully reconstructed neurons. We found that bouton locations were not homogeneous throughout the axon and among brain regions. As our algorithm was able to generate whole-brain single-cell connectivity matrices from full morphology reconstruction datasets, we further found that non-homogeneous bouton locations have a significant impact on network wiring, including degree distribution, triad census, and community structure. By perturbing neuronal morphology, we further explored the link between anatomical details and network topology. In our in silico exploration, we found that dendritic and axonal tree span would have the greatest impact on network wiring, followed by synaptic contact deletion. Our results suggest that neuroanatomical details must be carefully addressed in studies of whole-brain networks at the single-cell level.

DSM: Deep sequential model for complete neuronal morphology representation and feature extraction.

The full morphology of single neurons is indispensable for understanding cell types, the basic building blocks in brains. Projecting trajectories are critical to extracting biologically relevant information from neuron morphologies, as they provide valuable information for both connectivity and cell identity. We developed an artificial intelligence method, deep sequential model (DSM), to extract concise, cell-type-defining features from projections across brain regions. DSM achieves more than 90% accuracy in classifying 12 major neuron projection types without compromising performance when spatial noise is present. Such remarkable robustness enabled us to efficiently manage and analyze several major full-morphology data sources, showcasing how characteristic long projections can define cell identities. We also succeeded in applying our model to both discovering previously unknown neuron subtypes and analyzing exceptional co-expressed genes involved in neuron projection circuits.

Neuronal Connectivity as a Determinant of Cell Types and Subtypes.

Classifications of single neurons at brain-wide scale is a powerful way to characterize the structural and functional organization of a brain. We acquired and standardized a large morphology database of 20,158 mouse neurons, and generated a whole-brain scale potential connectivity map of single neurons based on their dendritic and axonal arbors. With such an anatomy-morphology-connectivity mapping, we defined neuron connectivity types and subtypes (both called "c-types" for simplicity) for neurons in 31 brain regions. We found that neuronal subtypes defined by connectivity in the same regions may share statistically higher correlation in their dendritic and axonal features than neurons having contrary connectivity patterns. Subtypes defined by connectivity show distinct separation with each other, which cannot be recapitulated by morphology features, population projections, transcriptomic, and electrophysiological data produced to date. Within this paradigm, we were able to characterize the diversity in secondary motor cortical neurons, and subtype connectivity patterns in thalamocortical pathways. Our finding underscores the importance of connectivity in characterizing the modularity of brain anatomy, as well as the cell types and their subtypes. These results highlight that c-types supplement conventionally recognized transcriptional cell types (t-types), electrophysiological cell types (e-types), and morphological cell types (m-types) as an important determinant of cell classes and their identities.

Deficits in neuronal architecture but not over-inhibition are main determinants of reduced neuronal network activity in a mouse model of overexpression of Dyrk1A.

Abnormal dendritic arbors, dendritic spine "dysgenesis" and excitation inhibition imbalance are main traits assumed to underlie impaired cognition and behavioral adaptation in intellectual disability. However, how these modifications actually contribute to functional properties of neuronal networks, such as signal integration or storage capacity is unknown. Here, we used a mouse model overexpressing Dyrk1A (Dual-specificity tyrosine [Y]-regulated kinase), one of the most relevant Down syndrome (DS) candidate genes, to gather quantitative data regarding hippocampal neuronal deficits produced by the overexpression of Dyrk1A in mice (TgDyrk1A; TG). TG mice showed impaired hippocampal recognition memory, altered excitation-inhibition balance and deficits in hippocampal CA1 LTP. We also detected for the first time that deficits in dendritic arborization in TG CA1 pyramidal neurons are layer-specific, with a reduction in the width of the stratum radiatum, the postsynaptic target site of CA3 excitatory neurons, but not in the stratum lacunosum-moleculare, which receives temporo-ammonic projections. To interrogate about the functional impact of layer-specific TG dendritic deficits we developed tailored computational multicompartmental models. Computational modelling revealed that neuronal microarchitecture alterations in TG mice lead to deficits in storage capacity, altered the integration of inputs from entorhinal cortex and hippocampal CA3 region onto CA1 pyramidal cells, important for coding place and temporal context and on connectivity and activity dynamics, with impaired the ability to reach high γ oscillations. Contrary to what is assumed in the field, the reduced network activity in TG is mainly contributed by the deficits in neuronal architecture and to a lesser extent by over-inhibition. Finally, given that therapies aimed at improving cognition have also been tested for their capability to recover dendritic spine deficits and excitation-inhibition imbalance, we also tested the short- and long-term changes produced by exposure to environmental enrichment (EE). Exposure to EE normalized the excitation inhibition imbalance and LTP, and had beneficial effects on short-term recognition memory. Importantly, it produced massive but transient dendritic remodeling of hippocampal CA1, that led to recovery of high γ oscillations, the main readout of synchronization of CA1 neurons, in our simulations. However, those effects where not stable and were lost after EE discontinuation. We conclude that layer-specific neuromorphological disturbances produced by Dyrk1A overexpression impair coding place and temporal context. Our results also suggest that treatments targeting structural plasticity, such as EE, even though hold promise towards improved treatment of intellectual disabilities, only produce temporary recovery, due to transient dendritic remodeling.

Full-Spectrum Neuronal Diversity and Stereotypy through Whole Brain Morphometry.

We conducted a large-scale study of whole-brain morphometry, analyzing 3.7 peta-voxels of mouse brain images at the single-cell resolution, producing one of the largest multi-morphometry databases of mammalian brains to date. We spatially registered 205 mouse brains and associated data from six Brain Initiative Cell Census Network (BICCN) data sources covering three major imaging modalities from five collaborative projects to the Allen Common Coordinate Framework (CCF) atlas, annotated 3D locations of cell bodies of 227,581 neurons, modeled 15,441 dendritic microenvironments, characterized the full morphology of 1,891 neurons along with their axonal motifs, and detected 2.58 million putative synaptic boutons. Our analysis covers six levels of information related to neuronal populations, dendritic microenvironments, single-cell full morphology, sub-neuronal dendritic and axonal arborization, axonal boutons, and structural motifs, along with a quantitative characterization of the diversity and stereotypy of patterns at each level. We identified 16 modules consisting of highly intercorrelated brain regions in 13 functional brain areas corresponding to 314 anatomical regions in CCF. Our analysis revealed the dendritic microenvironment as a powerful method for delineating brain regions of cell types and potential subtypes. We also found that full neuronal morphologies can be categorized into four distinct classes based on spatially tuned morphological features, with substantial cross-areal diversity in apical dendrites, basal dendrites, and axonal arbors, along with quantified stereotypy within cortical, thalamic and striatal regions. The lamination of somas was found to be more effective in differentiating neuron arbors within the cortex. Further analysis of diverging and converging projections of individual neurons in 25 regions throughout the brain reveals branching preferences in the brain-wide and local distributions of axonal boutons. Overall, our study provides a comprehensive description of key anatomical structures of neurons and their types, covering a wide range of scales and features, and contributes to our understanding of neuronal diversity and its function in the mammalian brain.

Fluorescent transgenic mouse models for whole-brain imaging in health and disease.

A paradigm shift is occurring in neuroscience and in general in life sciences converting biomedical research from a descriptive discipline into a quantitative, predictive, actionable science. Living systems are becoming amenable to quantitative description, with profound consequences for our ability to predict biological phenomena. New experimental tools such as tissue clearing, whole-brain imaging, and genetic engineering technologies have opened the opportunity to embrace this new paradigm, allowing to extract anatomical features such as cell number, their full morphology, and even their structural connectivity. These tools will also allow the exploration of new features such as their geometrical arrangement, within and across brain regions. This would be especially important to better characterize brain function and pathological alterations in neurological, neurodevelopmental, and neurodegenerative disorders. New animal models for mapping fluorescent protein-expressing neurons and axon pathways in adult mice are key to this aim. As a result of both developments, relevant cell populations with endogenous fluorescence signals can be comprehensively and quantitatively mapped to whole-brain images acquired at submicron resolution. However, they present intrinsic limitations: weak fluorescent signals, unequal signal strength across the same cell type, lack of specificity of fluorescent labels, overlapping signals in cell types with dense labeling, or undetectable signal at distal parts of the neurons, among others. In this review, we discuss the recent advances in the development of fluorescent transgenic mouse models that overcome to some extent the technical and conceptual limitations and tradeoffs between different strategies. We also discuss the potential use of these strains for understanding disease.