From visual perception to language, sensory stimuli change their meaning depending on previous experience. Recurrent neural dynamics can interpret stimuli based on externally cued context, but it is unknown whether they can compute and employ internal hypotheses to resolve ambiguities. Here we show that mouse retrosplenial cortex (RSC) can form several hypotheses over time and perform spatial reasoning through recurrent dynamics. In our task, mice navigated using ambiguous landmarks that are identified through their mutual spatial relationship, requiring sequential refinement of hypotheses. Neurons in RSC and in artificial neural networks encoded mixtures of hypotheses, location and sensory information, and were constrained by robust low-dimensional dynamics. RSC encoded hypotheses as locations in activity space with divergent trajectories for identical sensory inputs, enabling their correct interpretation. Our results indicate that interactions between internal hypotheses and external sensory data in recurrent circuits can provide a substrate for complex sequential cognitive reasoning.

Preprint: https://doi.org/10.1101/2022.04.12.488024v1

Cellular heterogeneity within complex tissues and organs is essential to coordinate biological processes across biological scales. The effect of local cues and tissue microenvironments on cell heterogeneity has been mainly studied at the transcriptional level. However, it is within the subcellular scale - the organelles - that lays the machinery to conduct most metabolic reactions and maintain cells alive, ensuring proper tissue function. How changes in subcellular organization under different microenvironments define the functional diversity of cells within organs remains largely unexplored. Here we determine how organelles adapt to different microenvironments using the mouse liver as model system, in combination with computational approaches and machine-learning. To understand organelle adaptation in response to changing nutritional conditions, we analyzed 3D fluorescent microscopy volumes of liver samples labeled to simultaneously visualize mitochondria, peroxisomes, and lipid droplets from mice subjected to different diets: a control diet, a high-fat diet, and a control diet plus fasting. A Cellpose based pipeline was implemented for cell and organelle segmentation, which allowed us to measure 100 different organelle metrics and helped us define subcellular architectures in liver samples at the single cell level. Our results showed that hepatocytes display distinct subcellular architectures within different regions of the liver-close to the central vein, in the middle region, and near the portal vein- and across the various diet groups, thus reflecting their adaptation to specific nutritional inputs. Principal component analysis and clustering of hepatocytes based on organelle signatures revealed 12 different hepatocyte categories within the different experimental groups, highlighting a reduction in hepatocyte heterogeneity under nutritional perturbations. Finally, using single cell organelle signatures exclusively, we generated machine learning models that were able to predict with high accuracy different hepatocyte categories, diet groups, and the stages of MASLD. Our results demonstrate how organelle signatures can be used as hallmarks to define hepatocyte heterogeneity and their adaptation to different nutritional conditions. In the future, our strategy, which combines subcellular resolution imaging of liver volumes and machine learning, could help establish protocols to better define and predict liver disease progression.

The human cerebral cortex, pivotal for advanced cognitive functions, is composed of six distinct layers and dozens of functionally specialized areas. The layers and areas are distinguished both molecularly, by diverse neuronal and glial cell subtypes, and structurally, through intricate spatial organization3,4. While single-cell transcriptomics studies have advanced molecular characterization of human cortical development, a critical gap exists due to the loss of spatial context during cell dissociation. Here, we utilized multiplexed error-robust fluorescence in situ hybridization (MERFISH)9, augmented with deep-learning-based cell segmentation, to examine the molecular, cellular, and cytoarchitectural development of human fetal cortex with spatially resolved single-cell resolution. Our extensive spatial atlas, encompassing 16 million single cells, spans eight cortical areas across four time points in the second and third trimesters. We uncovered an early establishment of the six-layer structure, identifiable in the laminar distribution of excitatory neuronal subtypes by mid-gestation, long before the emergence of cytoarchitectural layers. Notably, while anterior-posterior gradients of neuronal subtypes were generally observed in most cortical areas, a striking exception was the sharp molecular border between primary (V1) and secondary visual cortices (V2) at gestational week 20. Here we discovered an abrupt binary shift in neuronal subtype specification at the earliest stages, challenging the notion that continuous morphogen gradients dictate mid-gestation cortical arealization. Moreover, integrating single-nuclei RNA-sequencing and in situ whole transcriptomics revealed an early upregulation of synaptogenesis in V1-specific Layer 4 neurons, suggesting a role of synaptogenesis in this discrete border formation. Collectively, our findings underscore the crucial role of spatial relationships in determining the molecular specification of cortical layers and areas. This work not only provides a valuable resource for the field, but also establishes a spatially resolved single-cell analysis paradigm that paves the way for a comprehensive developmental atlas of the human brain.

Preprint: https://www.biorxiv.org/content/early/2024/06/10/2024.06.05.597673

The claustrum is a functionally and structurally complex brain region, whose very spatial extent remains debated. Histochemical-based approaches typically treat the claustrum as a relatively narrow anatomical region that primarily projects to the neocortex, whereas circuit-based approaches can suggest a broader claustrum region containing projections to the neocortex and other regions. Here, in the mouse, we took a bottom-up and cell-type-specific approach to complement and possibly unite these seemingly disparate conclusions. Using single-cell RNA-sequencing, we found that the claustrum comprises two excitatory neuron subtypes that are differentiable from the surrounding cortex. Multicolor retrograde tracing in conjunction with 12-channel multiplexed in situ hybridization revealed a core-shell spatial arrangement of these subtypes, as well as differential downstream targets. Thus, the claustrum comprises excitatory neuron subtypes with distinct molecular and projection properties, whose spatial patterns reflect the narrower and broader claustral extents debated in previous research. This subtype-specific heterogeneity likely shapes the functional complexity of the claustrum.

Optogenetics allows manipulations of genetically and spatially defined neuronal populations with excellent temporal control. However, neurons are coupled with other neurons over multiple length scales, and the effects of localized manipulations thus spread beyond the targeted neurons. We benchmarked several optogenetic methods to inactivate small regions of neocortex. Optogenetic excitation of GABAergic neurons produced more effective inactivation than light-gated ion pumps. Transgenic mice expressing the light-dependent chloride channel GtACR1 produced the most potent inactivation. Generally, inactivation spread substantially beyond the photostimulation light, caused by strong coupling between cortical neurons. Over some range of light intensity, optogenetic excitation of inhibitory neurons reduced activity in these neurons, together with pyramidal neurons, a signature of inhibition-stabilized neural networks ('paradoxical effect'). The offset of optogenetic inactivation was followed by rebound excitation in a light dose-dependent manner, limiting temporal resolution. Our data offer guidance for the design of optogenetics experiments.

Transcription initiation by RNA polymerase II (RNA Pol II) requires preinitiation complex (PIC) assembly at gene promoters. In the dynamic nucleus, where thousands of promoters are broadly distributed in chromatin, it is unclear how multiple individual components converge on any target to establish the PIC. Here we use live-cell, single-molecule tracking in S. cerevisiae to visualize constrained exploration of the nucleoplasm by PIC components and Mediator's key role in guiding this process. On chromatin, TFIID/TATA-binding protein (TBP), Mediator, and RNA Pol II instruct assembly of a short-lived PIC, which occurs infrequently but efficiently within a few seconds on average. Moreover, PIC exclusion by nucleosome encroachment underscores regulated promoter accessibility by chromatin remodeling. Thus, coordinated nuclear exploration and recruitment to accessible targets underlies dynamic PIC establishment in yeast. Our study provides a global spatiotemporal model for transcription initiation in live cells.

Environmental influences on immune phenotypes are well-documented, but our understanding of which elements of the environment affect immune systems, and how, remains vague. Behaviors, including socializing with others, are central to an individual's interaction with its environment. We therefore tracked behavior of rewilded laboratory mice of three inbred strains in outdoor enclosures and examined contributions of behavior, including associations measured from spatiotemporal co-occurrences, to immune phenotypes. We found extensive variation in individual and social behavior among and within mouse strains upon rewilding. In addition, we found that the more associated two individuals were, the more similar their immune phenotypes were. Spatiotemporal association was particularly predictive of similar memory T and B cell profiles and was more influential than sibling relationships or shared infection status. These results highlight the importance of shared spatiotemporal activity patterns and/or social networks for immune phenotype and suggest potential immunological correlates of social life.

Focal adhesions (FAs) connect inner workings of the cell to the extracellular matrix to control cell adhesion, migration, and mechanosensing1,2. Previous studies demonstrated that FAs contain three vertical layers, which connect extracellular matrix to the cytoskeleton3,4,5. However, cellular processes rely on precisely-regulated FA turnover, but the molecular machineries that control FA assembly and disassembly have remained elusive. By using super-resolution iPALM microscopy, we identified two unprecedented nanoscale layers within FAs, specified by actin filaments bound to tropomyosin isoforms Tpm1.6 and Tpm3.2. The Tpm1.6-actin filaments beneath the previously identified ‘actin-regulatory layer’ are critical for adhesion maturation and controlled cell motility, whereas the Tpm3.2-actin filament layer towards the bottom of FA facilitates adhesion disassembly. Mechanistically, Tpm3.2 stabilizes KANK-family proteins at adhesions, and hence targets microtubule plus-ends to FAs to catalyse their disassembly. Loss of Tpm3.2 leads to disorganized microtubule network, abnormally stable FAs, and defects in tail retraction during cell migration. Thus, FAs are composed of at least three distinct actin filament layers, each having specific roles in coupling of adhesion to the cytoskeleton, or in controlling adhesion dynamics. In a broader context, these findings demonstrate how distinct actin filament populations can co-exist and perform specific functions within a defined cellular compartment.

The most fundamental choice an animal has to make when it detects a threat is whether to freeze, reducing its chances of being noticed, or to flee to safety. Here we show that Drosophila melanogaster exposed to looming stimuli in a confined arena either freeze or flee. The probability of freezing versus fleeing is modulated by the fly's walking speed at the time of threat, demonstrating that freeze/flee decisions depend on behavioral state. We describe a pair of descending neurons crucially implicated in freezing. Genetic silencing of DNp09 descending neurons disrupts freezing yet does not prevent fleeing. Optogenetic activation of both DNp09 neurons induces running and freezing in a state-dependent manner. Our findings establish walking speed as a key factor in defensive response choices and reveal a pair of descending neurons as a critical component in the circuitry mediating selection and execution of freezing or fleeing behaviors.

Spike sorting is the computational process of extracting the firing times of single neurons from recordings of local electrical fields. This is an important but hard problem in neuroscience, made complicated by the nonstationarity of the recordings and the dense overlap in electrical fields between nearby neurons. To address the spike-sorting problem, we have been openly developing the Kilosort framework. Here we describe the various algorithmic steps introduced in different versions of Kilosort. We also report the development of Kilosort4, a version with substantially improved performance due to clustering algorithms inspired by graph-based approaches. To test the performance of Kilosort, we developed a realistic simulation framework that uses densely sampled electrical fields from real experiments to generate nonstationary spike waveforms and realistic noise. We found that nearly all versions of Kilosort outperformed other algorithms on a variety of simulated conditions and that Kilosort4 performed best in all cases, correctly identifying even neurons with low amplitudes and small spatial extents in high drift conditions.