We present a method to automatically identify and track nuclei in time-lapse microscopy recordings of entire developing embryos. The method combines deep learning and global optimization. On a mouse dataset, it reconstructs 75.8% of cell lineages spanning 1 h, as compared to 31.8% for the competing method. Our approach improves understanding of where and when cell fate decisions are made in developing embryos, tissues, and organs.

Calcium imaging has been widely adopted for its ability to record from large neuronal populations. To summarize the time course of neural activity, dimensionality reduction methods, which have been applied extensively to population spiking activity, may be particularly useful. However, it is unclear if the dimensionality reduction methods applied to spiking activity are appropriate for calcium imaging. We thus carried out a systematic study of design choices based on standard dimensionality reduction methods. We also developed a novel method to perform deconvolution and dimensionality reduction simultaneously (termed CILDS). CILDS most accurately recovered the single-trial, low-dimensional time courses from calcium imaging that would have been recovered from spiking activity. CILDS also outperformed the other methods on calcium imaging recordings from larval zebrafish and mice. More broadly, this study represents a foundation for summarizing calcium imaging recordings of large neuronal populations using dimensionality reduction in diverse experimental settings.

Individual neurons in prefrontal cortex – a key brain area involved in cognitive functions – are selective for variables such as space or time, as well as more cognitive aspects of tasks, such as learned categories. Many neurons exhibit mixed selectivity, that is, they show selectivity for multiple variables. A fundamental question is whether neurons are functionally specialized for particular variables and how selectivity for different variables intersects across the population. Here, we analyzed neural correlates of space and time in rats performing a navigational task with two behaviorally important categories – starts and goals. Using simultaneous recordings of many medial prefrontal cortex (mPFC) neurons during behavior, we found that population codes for elapsed time were invariant to different locations within categories, and subsets of neurons had functional preferences for time or space across categories. Thus, mPFC exhibits structured selectivity, which may facilitate complex behaviors by efficiently generating informative representations of multiple variables.

Animals can discriminate myriad sensory stimuli but can also generalize from learned experience. You can probably distinguish the favorite teas of your colleagues while still recognizing that all tea pales in comparison to coffee. Tradeoffs between detection, discrimination, and generalization are inherent at every layer of sensory processing. During development, specific quantitative parameters are wired into perceptual circuits and set the playing field on which plasticity mechanisms play out. A primary goal of systems neuroscience is to understand how material properties of a circuit define the logical operations— computations--that it makes, and what good these computations are for survival. A cardinal method in biology—and the mechanism of evolution--is to change a unit or variable within a system and ask how this affects organismal function. Here, we make use of our knowledge of developmental wiring mechanisms to modify hard-wired circuit parameters in the Drosophila melanogaster mushroom body and assess the functional and behavioral consequences. By altering the number of expansion layer neurons (Kenyon cells) and their dendritic complexity, we find that input number, but not cell number, tunes odor selectivity. Simple odor discrimination performance is maintained when Kenyon cell number is reduced and augmented by Kenyon cell expansion.

Dopaminergic neurons with distinct projection patterns and physiological properties compose memory subsystems in a brain. However, it is poorly understood whether or how they interact during complex learning. Here, we identify a feedforward circuit formed between dopamine subsystems and show that it is essential for second-order conditioning, an ethologically important form of higher-order associative learning. The Drosophila mushroom body comprises a series of dopaminergic compartments, each of which exhibits distinct memory dynamics. We find that a slow and stable memory compartment can serve as an effective “teacher” by instructing other faster and transient memory compartments via a single key interneuron, which we identify by connectome analysis and neurotransmitter prediction. This excitatory interneuron acquires enhanced response to reward-predicting odor after first-order conditioning and, upon activation, evokes dopamine release in the “student” compartments. These hierarchical connections between dopamine subsystems explain distinct properties of first- and second-order memory long known by behavioral psychologists.

Daily experience suggests that we perceive distances near us linearly. However, the actual geometry of spatial representation in the brain is unknown. Here we report that neurons in the CA1 region of rat hippocampus that mediate spatial perception represent space according to a non-linear hyperbolic geometry. This geometry uses an exponential scale and yields greater positional information than a linear scale. We found that the size of the representation matches the optimal predictions for the number of CA1 neurons. The representations also dynamically expanded proportional to the logarithm of time that the animal spent exploring the environment, in correspondence with the maximal mutual information that can be received. The dynamic changes tracked even small variations due to changes in the running speed of the animal. These results demonstrate how neural circuits achieve efficient representations using dynamic hyperbolic geometry.

Recent success in training artificial agents and robots derives from a combination of direct learning of behavioural policies and indirect learning through value functions. Policy learning and value learning use distinct algorithms that optimize behavioural performance and reward prediction, respectively. In animals, behavioural learning and the role of mesolimbic dopamine signalling have been extensively evaluated with respect to reward prediction; however, so far there has been little consideration of how direct policy learning might inform our understanding. Here we used a comprehensive dataset of orofacial and body movements to understand how behavioural policies evolved as naive, head-restrained mice learned a trace conditioning paradigm. Individual differences in initial dopaminergic reward responses correlated with the emergence of learned behavioural policy, but not the emergence of putative value encoding for a predictive cue. Likewise, physiologically calibrated manipulations of mesolimbic dopamine produced several effects inconsistent with value learning but predicted by a neural-network-based model that used dopamine signals to set an adaptive rate, not an error signal, for behavioural policy learning. This work provides strong evidence that phasic dopamine activity can regulate direct learning of behavioural policies, expanding the explanatory power of reinforcement learning models for animal learning.

Insects like Drosophila produce a second brain adapted to the form and behavior of a larva. Neurons for both larval and adult brains are produced by the same stem cells (neuroblasts) but the larva possesses only the earliest born neurons produced from each. To understand how a functional larval brain is made from this reduced set of neurons, we examined the origins and metamorphic fates of the neurons of the larval and adult mushroom body circuits. The adult mushroom body core is built sequentially of γ Kenyon cells, that form a medial lobe, followed by α’β’, and αβ Kenyon cells that form additional medial lobes and two vertical lobes. Extrinsic input (MBINs) and output (MBONs) neurons divide this core into computational compartments. The larval mushroom body contains only γ neurons. Its medial lobe compartments are roughly homologous to those of the adult and same MBONs are used for both. The larval vertical lobe, however, is an analogous “facsimile” that uses a larval-specific branch on the γ neurons to make up for the missing α’β’, and αβ neurons. The extrinsic cells for the facsimile are early-born neurons that trans-differentiate to serve a mushroom body function in the larva and then shift to other brain circuits in the adult. These findings are discussed in the context of the evolution of a larval brain in insects with complete metamorphosis.

The central nucleus of the amygdala (CEA) is a brain region that integrates external and internal sensory information and executes innate and adaptive behaviors through distinct output pathways. Despite its complex functions, the diversity of molecularly defined neuronal types in the CEA and their contributions to major axonal projection targets have not been examined systematically. Here, we performed single-cell RNA-sequencing (scRNA-Seq) to classify molecularly defined cell types in the CEA and identified marker-genes to map the location of these neuronal types using expansion assisted iterative fluorescence in situ hybridization (EASI-FISH). We developed new methods to integrate EASI-FISH with 5-plex retrograde axonal labeling to determine the spatial, morphological, and connectivity properties of ∼30,000 molecularly defined CEA neurons. Our study revealed spatio-molecular organization of the CEA, with medial and lateral CEA associated with distinct cell families. We also found a long-range axon projection network from the CEA, where target regions receive inputs from multiple molecularly defined cell types. Axon collateralization was found primarily among projections to hindbrain targets, which are distinct from forebrain projections. This resource reports marker-gene combinations for molecularly defined cell types and axon-projection types, which will be useful for selective interrogation of these neuronal populations to study their contributions to the diverse functions of the CEA.

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, complicated by the non-stationarity of the recordings and the dense overlap in electrical fields between nearby neurons. To solve the spike sorting problem, we have continuously developed over the past eight years a framework known as Kilosort. This paper describes the various algorithmic steps introduced in different versions of Kilosort. We also report the development of Kilosort4, a new version with substantially improved performance due to new clustering algorithms inspired by graph-based approaches. To test the performance of Kilosort, we developed a realistic simulation framework which uses densely sampled electrical fields from real experiments to generate non-stationary spike waveforms and realistic noise. We find that nearly all versions of Kilosort outperform other algorithms on a variety of simulated conditions, and Kilosort4 performs best in all cases, correctly identifying even neurons with low amplitudes and small spatial extents in high drift conditions.