We describe an engineered family of highly antigenic molecules based on GFP-like fluorescent proteins. These molecules contain numerous copies of peptide epitopes and simultaneously bind IgG antibodies at each location. These 'spaghetti monster' fluorescent proteins (smFPs) distributed well in neurons, notably into small dendrites, spines and axons. smFP immunolabeling localized weakly expressed proteins not well resolved with traditional epitope tags. By varying epitope and scaffold, we generated a diverse family of mutually orthogonal antigens. In cultured neurons and mouse and fly brains, smFP probes allowed robust, orthogonal multicolor visualization of proteins, cell populations and neuropil. smFP variants complement existing tracers and greatly increase the number of simultaneous imaging channels, and they performed well in advanced preparations such as array tomography, super-resolution fluorescence imaging and electron microscopy. In living cells, the probes improved single-molecule image tracking and increased yield for RNA-seq. These probes facilitate new experiments in connectomics, transcriptomics and protein localization.

Neural circuits within the frontal cortex support the flexible selection of goal-directed behaviors by integrating input from brain regions associated with sensory, emotional, episodic, and semantic memory functions. From a connectomics perspective, determining how these disparate afferent inputs target their synapses to specific cell types in the frontal cortex may prove crucial in understanding circuit-level information processing. Here, we used monosynaptic retrograde rabies mapping to examine the distribution of afferent neurons targeting four distinct classes of local inhibitory interneurons and four distinct classes of excitatory projection neurons in mouse infralimbic cortex. Interneurons expressing parvalbumin, somatostatin, or vasoactive intestinal peptide received a large proportion of inputs from hippocampal regions, while interneurons expressing neuron-derived neurotrophic factor received a large proportion of inputs from thalamic regions. A more moderate hippocampal-thalamic dichotomy was found among the inputs targeting excitatory neurons that project to the basolateral amygdala, lateral entorhinal cortex, nucleus reuniens of the thalamus, and the periaqueductal gray. Together, these results show a prominent bias among hippocampal and thalamic afferent systems in their targeting to genetically or anatomically defined sets of frontal cortical neurons. Moreover, they suggest the presence of two distinct local microcircuits that control how different inputs govern frontal cortical information processing.

Relating the function of neuronal cell types to information processing and behavior is a central goal of neuroscience. In the hippocampus, pyramidal cells in CA1 and the subiculum process sensory and motor cues to form a cognitive map encoding spatial, contextual, and emotional information, which they transmit throughout the brain. Do these cells constitute a single class or are there multiple cell types with specialized functions? Using unbiased cluster analysis, we show that there are two morphologically and electrophysiologically distinct principal cell types that carry hippocampal output. We show further that these two cell types are inversely modulated by the synergistic action of glutamate and acetylcholine acting on metabotropic receptors that are central to hippocampal function. Combined with prior connectivity studies, our results support a model of hippocampal processing in which the two pyramidal cell types are predominantly segregated into two parallel pathways that process distinct modalities of information.

Clarifying gene expression in narrowly defined neuronal populations can provide insight into cellular identity, computation, and functionality. Here, we used next-generation RNA sequencing (RNA-seq) to produce a quantitative, whole genome characterization of gene expression for the major excitatory neuronal classes of the hippocampus; namely, granule cells and mossy cells of the dentate gyrus, and pyramidal cells of areas CA3, CA2, and CA1. Moreover, for the canonical cell classes of the trisynaptic loop, we profiled transcriptomes at both dorsal and ventral poles, producing a cell-class- and region-specific transcriptional description for these populations. This dataset clarifies the transcriptional properties and identities of lesser-known cell classes, and moreover reveals unexpected variation in the trisynaptic loop across the dorsal-ventral axis. We have created a public resource, Hipposeq (http://hipposeq.janelia.org), which provides analysis and visualization of these data and will act as a roadmap relating molecules to cells, circuits, and computation in the hippocampus.

Context plays a foundational role in determining how to interpret potentially fear-producing stimuli, yet the precise neurobiological substrates of context are poorly understood. In this issue of Cell, Xu et al. elegantly show that parallel neuronal circuits are necessary for two distinct roles of context in fear conditioning.

Spatial and temporal features of synaptic inputs engage integration mechanisms on multiple scales, including presynaptic release sites, postsynaptic dendrites, and networks of inhibitory interneurons. Here we investigate how these mechanisms cooperate to filter synaptic input in hippocampal area CA1. Dendritic recordings from CA1 pyramidal neurons reveal that proximal inputs from CA3 as well as distal inputs from entorhinal cortex layer III (ECIII) sum sublinearly or linearly at low firing rates due to feedforward inhibition, but sum supralinearly at high firing rates due to synaptic facilitation, producing a high-pass filter. However, during ECIII and CA3 input comparison, supralinear dendritic integration is dynamically balanced by feedforward and feedback inhibition, resulting in suppression of dendritic complex spiking. We find that a particular subpopulation of CA1 interneurons expressing neuropeptide Y (NPY) contributes prominently to this dynamic filter by integrating both ECIII and CA3 input pathways and potently inhibiting CA1 pyramidal neuron dendrites.

Elucidating the diversity and spatial organization of cell types in the brain is an essential goal of neuroscience, with many emerging technologies helping to advance this endeavor. Using a new in situ hybridization method that can measure the expression of hundreds of genes in a given mouse brain section (amplified seqFISH), Shah et al. (2016) describe a spatial organization of hippocampal cell types that differs from previous reports. In seeking to understand this discrepancy, we find that many of the barcoded genes used by seqFISH to characterize this spatial organization, when cross-validated by other sensitive methodologies, exhibit negligible expression in the hippocampus. Additionally, the results of Shah et al. (2016) do not recapitulate canonical cellular hierarchies and improperly classify major neuronal cell types. We suggest that, when describing the spatial organization of brain regions, cross-validation using multiple techniques should be used to yield robust and informative cellular classification. This Matters Arising paper is in response to Shah et al. (2016), published in Neuron. See also the response by Shah et al. (2017), published in this issue.

The anatomical and electrophysiological properties of neurons in the stratum lucidum of the CA3 subfield of the hippocampus were examined by using patch-pipette recordings combined with biocytin staining. This method facilitated the analysis of the morphological features and passive and active properties of a recently described class of spiny neurons in the stratum lucidum, as well as aspiny neurons in this region. Some, but not all, synaptic inputs of both types of neurons were found to arise from the mossy fiber system. The axons of spiny neurons in the stratum lucidum were heavily collateralized, terminating primarily in the stratum lucidum and stratum radiatum of CA3, and to a lesser extent in the stratum pyramidale and stratum oriens. Only a few axonal projections were found that extended beyond the CA3 region into CA1 and the hilus. Aspiny neurons fell into two classes: those projecting axons to the stratum lucidum and stratum radiatum of CA3 and those with axon terminations mainly in the stratum pyramidale and stratum oriens. The electrophysiological properties of spiny and aspiny neurons in the stratum lucidum were similar, but on average, the aspiny neurons had significantly higher maximal firing rates and narrower action potential half-widths. The results demonstrate that a diverse population of neurons exists in the region of mossy fiber termination in area CA3. These neurons may be involved in local-circuit feedback, or feed-forward systems controlling the flow of information through the hippocampus.

The hippocampus has been used extensively as a model to study plastic changes in the brain's neural circuitry. Immediately after high-frequency stimulation to hippocampal Schaffer collateral axons, a dramatic change occurs in the relationship between the presynaptic CA3 and the postsynaptic CA1 pyramidal neurons. For a fixed excitatory postsynaptic potential (EPSP), there arises an increased likelihood of action potential generation in the CA1 pyramidal neuron. This phenomenon is called EPSP-spike (E-S) potentiation. We explored E-S potentiation, using patch-clamp techniques in the hippocampal slice preparation. A specific protocol was developed to measure the action potential probability for a given synaptic strength, which allowed us to quantify the amount of E-S potentiation for a single neuron. E-S potentiation was greatest when gamma-aminobutyric acid (GABA)ergic inhibition was intact, suggesting that modulation of inhibition is a major aspect of E-S potentiation. Expression of E-S potentiation also correlated with a reduced action-potential threshold, which was greatest when GABAergic inhibition was intact. Conditioning stimuli produced a smaller threshold reduction when inhibition was blocked, but some reduction also occurred in the absence of a conditioning stimulus. Together, these results suggest that E-S potentiation is caused primarily through a reduction of GABAergic inhibition, leading to larger EPSPs and reduced action potential threshold. Our findings do not rule out, however, the possibility that modulation of voltage-gated conductances also contributes to E-S potentiation.

Cognitive maps confer animals with flexible intelligence by representing spatial, temporal, and abstract relationships that can be used to shape thought, planning, and behavior. Cognitive maps have been observed in the hippocampus, but their algorithmic form and the processes by which they are learned remain obscure. Here, we employed large-scale, longitudinal two-photon calcium imaging to record activity from thousands of neurons in the CA1 region of the hippocampus while mice learned to efficiently collect rewards from two subtly different versions of linear tracks in virtual reality. The results provide a detailed view of the formation of a cognitive map in the hippocampus. Throughout learning, both the animal behavior and hippocampal neural activity progressed through multiple intermediate stages, gradually revealing improved task understanding and behavioral efficiency. The learning process led to progressive decorrelations in initially similar hippocampal neural activity within and across tracks, ultimately resulting in orthogonalized representations resembling a state machine capturing the inherent structure of the task. We show that a Hidden Markov Model (HMM) and a biologically plausible recurrent neural network trained using Hebbian learning can both capture core aspects of the learning dynamics and the orthogonalized representational structure in neural activity. In contrast, we show that gradient-based learning of sequence models such as Long Short-Term Memory networks (LSTMs) and Transformers do not naturally produce such representations. We further demonstrate that mice exhibited adaptive behavior in novel task settings, with neural activity reflecting flexible deployment of the state machine. These findings shed light on the mathematical form of cognitive maps, the learning rules that sculpt them, and the algorithms that promote adaptive behavior in animals. The work thus charts a course toward a deeper understanding of biological intelligence and offers insights toward developing more robust learning algorithms in artificial intelligence.