Each training step for a variational autoencoder (VAE) requires us to sample from the approximate posterior, so we usually choose simple (e.g. factorised) approximate posteriors in which sampling is an efficient computation that fully exploits GPU parallelism. However, such simple approximate posteriors are often insufficient, as they eliminate statistical dependencies in the posterior. While it is possible to use normalizing flow approximate posteriors for continuous latents, some problems have discrete latents and strong statistical dependencies. The most natural approach to model these dependencies is an autoregressive distribution, but sampling from such distributions is inherently sequential and thus slow. We develop a fast, parallel sampling procedure for autoregressive distributions based on fixed-point iterations which enables efficient and accurate variational inference in discrete state-space latent variable dynamical systems. To optimize the variational bound, we considered two ways to evaluate probabilities: inserting the relaxed samples directly into the pmf for the discrete distribution, or converting to continuous logistic latent variables and interpreting the K-step fixed-point iterations as a normalizing flow. We found that converting to continuous latent variables gave considerable additional scope for mismatch between the true and approximate posteriors, which resulted in biased inferences, we thus used the former approach. Using our fast sampling procedure, we were able to realize the benefits of correlated posteriors, including accurate uncertainty estimates for one cell, and accurate connectivity estimates for multiple cells, in an order of magnitude less time.

Many forms of human and animal behavior involve head movements. A new study reveals the neural code for three-dimensional head movements in the midbrain of freely moving mice.

Highlights:

• Biological specimens introduce wavefront aberrations and deteriorate the image quality of optical microscopy.
• Adaptive optics is used in optical microscopy to recover ideal imaging performance.
• Adaptive optical imaging improves structural imaging of neurons, allowing for synaptic-level resolution at depth.
• Adaptive optical imaging leads to a more accurate characterization of the functional properties of neurons.

With the ability to correct for the aberrations introduced by biological specimens, adaptive optics—a method originally developed for astronomical telescopes—has been applied to optical microscopy to recover diffraction-limited imaging performance deep within living tissue. In particular, this technology has been used to improve image quality and provide a more accurate characterization of both structure and function of neurons in a variety of living organisms. Among its many highlights, adaptive optical microscopy has made it possible to image large volumes with diffraction-limited resolution in zebrafish larval brains, to resolve dendritic spines over 600μm deep in the mouse brain, and to more accurately characterize the orientation tuning properties of thalamic boutons in the primary visual cortex of awake mice.

Recent advances in understanding intracellular amino acid transport and mechanistic target of rapamycin complex 1 (mTORC1) signaling shed light on solute carrier 38, family A member 9 (SLC38A9), a lysosomal transporter responsible for the binding and translocation of several essential amino acids. Here we present the first crystal structure of SLC38A9 from Danio rerio in complex with arginine. As captured in the cytosol-open state, the bound arginine was locked in a transitional state stabilized by transmembrane helix 1 (TM1) of drSLC38A9, which was anchored at the groove between TM5 and TM7. These anchoring interactions were mediated by the highly conserved WNTMM motif in TM1, and mutations in this motif abolished arginine transport by drSLC38A9. The underlying mechanism of substrate binding is critical for sensitizing the mTORC1 signaling pathway to amino acids and for maintenance of lysosomal amino acid homeostasis. This study offers a first glimpse into a prototypical model for SLC38 transporters.

Repeated exposure to stressors is known to produce large-scale remodeling of neurons within the prefrontal cortex (PFC). Recent work suggests stress-related forms of structural plasticity can interact with aging to drive distinct patterns of pyramidal cell morphological changes. However, little is known about how other cellular components within PFC might be affected by these challenges. Here, we examined the effects of stress exposure and aging on medial prefrontal cortical glial subpopulations. Interestingly, we found no changes in glial morphology with stress exposure but a profound morphological change with aging. Furthermore, we found an upregulation of non-nuclear glucocorticoid receptors (GR) with aging, while nuclear levels remained largely unaffected. Both changes are selective for microglia, with no stress or aging effect found in astrocytes. Lastly, we show that the changes found within microglia inversely correlated with the density of dendritic spines on layer III pyramidal cells. These findings suggest microglia play a selective role in synaptic health within the aging brain.

In general, neurons in insects and many other invertebrate groups are individually recognizable, enabling us to assign an index number to specific neurons in a manner which is rarely possible in a vertebrate brain. This endows many studies on insect nervous systems with the opportunity to document neurons with great precision, so that in favourable cases we can return to the same neuron or neuron type repeatedly so as to recognize many separate morphological classes. The visual system of the fly's compound eye particularly provides clear examples of the accuracy of neuron wiring, allowing numerical comparisons between representatives of the same cell type, and estimates of the accuracy of their wiring.

Ventral tegmental area (VTA) glutamate neurons are important components of reward circuitry, but whether they are subject to cholinergic modulation is unknown. To study this, we used molecular, physiological, and photostimulation techniques to examine nicotinic acetylcholine receptors (nAChRs) in VTA glutamate neurons. Cells in the medial VTA, where glutamate neurons are enriched, are responsive to acetylcholine (ACh) released from cholinergic axons. VTA VGLUT2 neurons express mRNA and protein subunits known to comprise heteromeric nAChRs. Electrophysiology, coupled with two-photon microscopy and laser flash photolysis of photoactivatable nicotine, was used to demonstrate nAChR functional activity in the somatodendritic subcellular compartment of VTA VGLUT2 neurons. Finally, optogenetic isolation of intrinsic VTA glutamatergic microcircuits along with gene-editing techniques demonstrated that nicotine potently modulates excitatory transmission within the VTA via heteromeric nAChRs. These results indicate that VTA glutamate neurons are modulated by cholinergic mechanisms and participate in the cascade of physiological responses to nicotine exposure.

In recent years, two-photon calcium imaging has become a standard tool to probe the function of neural circuits and to study computations in neuronal populations. However, the acquired signal is only an indirect measurement of neural activity due to the comparatively slow dynamics of fluorescent calcium indicators. Different algorithms for estimating spike rates from noisy calcium measurements have been proposed in the past, but it is an open question how far performance can be improved. Here, we report the results of the spikefinder challenge, launched to catalyze the development of new spike rate inference algorithms through crowd-sourcing. We present ten of the submitted algorithms which show improved performance compared to previously evaluated methods. Interestingly, the top-performing algorithms are based on a wide range of principles from deep neural networks to generative models, yet provide highly correlated estimates of the neural activity. The competition shows that benchmark challenges can drive algorithmic developments in neuroscience.

A central tenet of biology is that globular proteins have a unique 3D structure under physiological conditions. Recent work has challenged this notion by demonstrating that some proteins switch folds, a process that involves remodeling of secondary structure in response to a few mutations (evolved fold switchers) or cellular stimuli (extant fold switchers). To date, extant fold switchers have been viewed as rare byproducts of evolution, but their frequency has been neither quantified nor estimated. By systematically and exhaustively searching the Protein Data Bank (PDB), we found ∼100 extant fold-switching proteins. Furthermore, we gathered multiple lines of evidence suggesting that these proteins are widespread in nature. Based on these lines of evidence, we hypothesized that the frequency of extant fold-switching proteins may be underrepresented by the structures in the PDB. Thus, we sought to identify other putative extant fold switchers with only one solved conformation. To do this, we identified two characteristic features of our ∼100 extant fold-switching proteins, incorrect secondary structure predictions and likely independent folding cooperativity, and searched the PDB for other proteins with similar features. Reassuringly, this method identified dozens of other proteins in the literature with indication of a structural change but only one solved conformation in the PDB. Thus, we used it to estimate that 0.5-4% of PDB proteins switch folds. These results demonstrate that extant fold-switching proteins are likely more common than the PDB reflects, which has implications for cell biology, genomics, and human health.