High-precision coding in visual cortex.

Individual neurons in visual cortex provide the brain with unreliable estimates of visual features. It is not known whether the single-neuron variability is correlated across large neural populations, thus impairing the global encoding of stimuli. We recorded simultaneously from up to 50,000 neurons in mouse primary visual cortex (V1) and in higher order visual areas and measured stimulus discrimination thresholds of 0.35° and 0.37°, respectively, in an orientation decoding task. These neural thresholds were almost 100 times smaller than the behavioral discrimination thresholds reported in mice. This discrepancy could not be explained by stimulus properties or arousal states. Furthermore, behavioral variability during a sensory discrimination task could not be explained by neural variability in V1. Instead, behavior-related neural activity arose dynamically across a network of non-sensory brain areas. These results imply that perceptual discrimination in mice is limited by downstream decoders, not by neural noise in sensory representations.

Toroidal topology of population activity in grid cells.

The medial entorhinal cortex is part of a neural system for mapping the position of an individual within a physical environment1. Grid cells, a key component of this system, fire in a characteristic hexagonal pattern of locations2, and are organized in modules3 that collectively form a population code for the animal's allocentric position1. The invariance of the correlation structure of this population code across environments4,5 and behavioural states6,7, independent of specific sensory inputs, has pointed to intrinsic, recurrently connected continuous attractor networks (CANs) as a possible substrate of the grid pattern1,8-11. However, whether grid cell networks show continuous attractor dynamics, and how they interface with inputs from the environment, has remained unclear owing to the small samples of cells obtained so far. Here, using simultaneous recordings from many hundreds of grid cells and subsequent topological data analysis, we show that the joint activity of grid cells from an individual module resides on a toroidal manifold, as expected in a two-dimensional CAN. Positions on the torus correspond to positions of the moving animal in the environment. Individual cells are preferentially active at singular positions on the torus. Their positions are maintained between environments and from wakefulness to sleep, as predicted by CAN models for grid cells but not by alternative feedforward models12. This demonstration of network dynamics on a toroidal manifold provides a population-level visualization of CAN dynamics in grid cells.

Spike sorting with Kilosort4.

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.

Cellpose 2.0: how to train your own model.

Pretrained neural network models for biological segmentation can provide good out-of-the-box results for many image types. However, such models do not allow users to adapt the segmentation style to their specific needs and can perform suboptimally for test images that are very different from the training images. Here we introduce Cellpose 2.0, a new package that includes an ensemble of diverse pretrained models as well as a human-in-the-loop pipeline for rapid prototyping of new custom models. We show that models pretrained on the Cellpose dataset can be fine-tuned with only 500-1,000 user-annotated regions of interest (ROI) to perform nearly as well as models trained on entire datasets with up to 200,000 ROI. A human-in-the-loop approach further reduced the required user annotation to 100-200 ROI, while maintaining high-quality segmentations. We provide software tools such as an annotation graphical user interface, a model zoo and a human-in-the-loop pipeline to facilitate the adoption of Cellpose 2.0.

High-dimensional geometry of population responses in visual cortex.

A neuronal population encodes information most efficiently when its stimulus responses are high-dimensional and uncorrelated, and most robustly when they are lower-dimensional and correlated. Here we analysed the dimensionality of the encoding of natural images by large populations of neurons in the visual cortex of awake mice. The evoked population activity was high-dimensional, and correlations obeyed an unexpected power law: the nth principal component variance scaled as 1/n. This scaling was not inherited from the power law spectrum of natural images, because it persisted after stimulus whitening. We proved mathematically that if the variance spectrum was to decay more slowly then the population code could not be smooth, allowing small changes in input to dominate population activity. The theory also predicts larger power-law exponents for lower-dimensional stimulus ensembles, which we validated experimentally. These results suggest that coding smoothness may represent a fundamental constraint that determines correlations in neural population codes.

Cellpose: a generalist algorithm for cellular segmentation.

Many biological applications require the segmentation of cell bodies, membranes and nuclei from microscopy images. Deep learning has enabled great progress on this problem, but current methods are specialized for images that have large training datasets. Here we introduce a generalist, deep learning-based segmentation method called Cellpose, which can precisely segment cells from a wide range of image types and does not require model retraining or parameter adjustments. Cellpose was trained on a new dataset of highly varied images of cells, containing over 70,000 segmented objects. We also demonstrate a three-dimensional (3D) extension of Cellpose that reuses the two-dimensional (2D) model and does not require 3D-labeled data. To support community contributions to the training data, we developed software for manual labeling and for curation of the automated results. Periodically retraining the model on the community-contributed data will ensure that Cellpose improves constantly.

Structured random receptive fields enable informative sensory encodings.

Brains must represent the outside world so that animals survive and thrive. In early sensory systems, neural populations have diverse receptive fields structured to detect important features in inputs, yet significant variability has been ignored in classical models of sensory neurons. We model neuronal receptive fields as random, variable samples from parameterized distributions and demonstrate this model in two sensory modalities using data from insect mechanosensors and mammalian primary visual cortex. Our approach leads to a significant theoretical connection between the foundational concepts of receptive fields and random features, a leading theory for understanding artificial neural networks. The modeled neurons perform a randomized wavelet transform on inputs, which removes high frequency noise and boosts the signal. Further, these random feature neurons enable learning from fewer training samples and with smaller networks in artificial tasks. This structured random model of receptive fields provides a unifying, mathematically tractable framework to understand sensory encodings across both spatial and temporal domains.

Fully integrated silicon probes for high-density recording of neural activity.

Sensory, motor and cognitive operations involve the coordinated action of large neuronal populations across multiple brain regions in both superficial and deep structures. Existing extracellular probes record neural activity with excellent spatial and temporal (sub-millisecond) resolution, but from only a few dozen neurons per shank. Optical Ca2+ imaging offers more coverage but lacks the temporal resolution needed to distinguish individual spikes reliably and does not measure local field potentials. Until now, no technology compatible with use in unrestrained animals has combined high spatiotemporal resolution with large volume coverage. Here we design, fabricate and test a new silicon probe known as Neuropixels to meet this need. Each probe has 384 recording channels that can programmably address 960 complementary metal-oxide-semiconductor (CMOS) processing-compatible low-impedance TiN sites that tile a single 10-mm long, 70 × 20-µm cross-section shank. The 6 × 9-mm probe base is fabricated with the shank on a single chip. Voltage signals are filtered, amplified, multiplexed and digitized on the base, allowing the direct transmission of noise-free digital data from the probe. The combination of dense recording sites and high channel count yielded well-isolated spiking activity from hundreds of neurons per probe implanted in mice and rats. Using two probes, more than 700 well-isolated single neurons were recorded simultaneously from five brain structures in an awake mouse. The fully integrated functionality and small size of Neuropixels probes allowed large populations of neurons from several brain structures to be recorded in freely moving animals. This combination of high-performance electrode technology and scalable chip fabrication methods opens a path towards recording of brain-wide neural activity during behaviour.

Robustness of Spike Deconvolution for Neuronal Calcium Imaging.

Calcium imaging is a powerful method to record the activity of neural populations in many species, but inferring spike times from calcium signals is a challenging problem. We compared multiple approaches using multiple datasets with ground truth electrophysiology and found that simple non-negative deconvolution (NND) outperformed all other algorithms on out-of-sample test data. We introduce a novel benchmark applicable to recordings without electrophysiological ground truth, based on the correlation of responses to two stimulus repeats, and used this to show that unconstrained NND also outperformed the other algorithms when run on "zoomed out" datasets of ∼10,000 cell recordings from the visual cortex of mice of either sex. Finally, we show that NND-based methods match the performance of a supervised method based on convolutional neural networks while avoiding some of the biases of such methods, and at much faster running times. We therefore recommend that spikes be inferred from calcium traces using simple NND because of its simplicity, efficiency, and accuracy.SIGNIFICANCE STATEMENT The experimental method that currently allows for recordings of the largest numbers of cells simultaneously is two-photon calcium imaging. However, use of this powerful method requires that neuronal firing times be inferred correctly from the large resulting datasets. Previous studies have claimed that complex supervised learning algorithms outperform simple deconvolution methods at this task. Unfortunately, these studies suffered from several problems and biases. When we repeated the analysis, using the same data and correcting these problems, we found that simpler spike inference methods perform better. Even more importantly, we found that supervised learning methods can introduce artifactual structure into spike trains, which can in turn lead to erroneous scientific conclusions. Of the algorithms we evaluated, we found that an extremely simple method performed best in all circumstances tested, was much faster to run, and was insensitive to parameter choices, making incorrect scientific conclusions much less likely.

Community-based benchmarking improves spike rate inference from two-photon calcium imaging data.

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.

State-dependent population coding in primary auditory cortex.

Sensory function is mediated by interactions between external stimuli and intrinsic cortical dynamics that are evident in the modulation of evoked responses by cortical state. A number of recent studies across different modalities have demonstrated that the patterns of activity in neuronal populations can vary strongly between synchronized and desynchronized cortical states, i.e., in the presence or absence of intrinsically generated up and down states. Here we investigated the impact of cortical state on the population coding of tones and speech in the primary auditory cortex (A1) of gerbils, and found that responses were qualitatively different in synchronized and desynchronized cortical states. Activity in synchronized A1 was only weakly modulated by sensory input, and the spike patterns evoked by tones and speech were unreliable and constrained to a small range of patterns. In contrast, responses to tones and speech in desynchronized A1 were temporally precise and reliable across trials, and different speech tokens evoked diverse spike patterns with extremely weak noise correlations, allowing responses to be decoded with nearly perfect accuracy. Restricting the analysis of synchronized A1 to activity within up states yielded similar results, suggesting that up states are not equivalent to brief periods of desynchronization. These findings demonstrate that the representational capacity of A1 depends strongly on cortical state, and suggest that cortical state should be considered as an explicit variable in all studies of sensory processing.

Facemap: a framework for modeling neural activity based on orofacial tracking.

Recent studies in mice have shown that orofacial behaviors drive a large fraction of neural activity across the brain. To understand the nature and function of these signals, we need better computational models to characterize the behaviors and relate them to neural activity. Here we developed Facemap, a framework consisting of a keypoint tracker and a deep neural network encoder for predicting neural activity. Our algorithm for tracking mouse orofacial behaviors was more accurate than existing pose estimation tools, while the processing speed was several times faster, making it a powerful tool for real-time experimental interventions. The Facemap tracker was easy to adapt to data from new labs, requiring as few as 10 annotated frames for near-optimal performance. We used the keypoints as inputs to a deep neural network which predicts the activity of ~50,000 simultaneously-recorded neurons and, in visual cortex, we doubled the amount of explained variance compared to previous methods. Using this model, we found that the neuronal activity clusters that were well predicted from behavior were more spatially spread out across cortex. We also found that the deep behavioral features from the model had stereotypical, sequential dynamics that were not reversible in time. In summary, Facemap provides a stepping stone toward understanding the function of the brain-wide neural signals and their relation to behavior.

Anatomically distributed neural representations of instincts in the hypothalamus.

Artificial activation of anatomically localized, genetically defined hypothalamic neuron populations is known to trigger distinct innate behaviors, suggesting a hypothalamic nucleus-centered organization of behavior control. To assess whether the encoding of behavior is similarly anatomically confined, we performed simultaneous neuron recordings across twenty hypothalamic regions in freely moving animals. Here we show that distinct but anatomically distributed neuron ensembles encode the social and fear behavior classes, primarily through mixed selectivity. While behavior class-encoding ensembles were spatially distributed, individual ensembles exhibited strong localization bias. Encoding models identified that behavior actions, but not motion-related variables, explained a large fraction of hypothalamic neuron activity variance. These results identify unexpected complexity in the hypothalamic encoding of instincts and provide a foundation for understanding the role of distributed neural representations in the expression of behaviors driven by hardwired circuits.

Electrode pooling can boost the yield of extracellular recordings with switchable silicon probes.

State-of-the-art silicon probes for electrical recording from neurons have thousands of recording sites. However, due to volume limitations there are typically many fewer wires carrying signals off the probe, which restricts the number of channels that can be recorded simultaneously. To overcome this fundamental constraint, we propose a method called electrode pooling that uses a single wire to serve many recording sites through a set of controllable switches. Here we present the framework behind this method and an experimental strategy to support it. We then demonstrate its feasibility by implementing electrode pooling on the Neuropixels 1.0 electrode array and characterizing its effect on signal and noise. Finally we use simulations to explore the conditions under which electrode pooling saves wires without compromising the content of the recordings. We make recommendations on the design of future devices to take advantage of this strategy.

Hunger or thirst state uncertainty is resolved by outcome evaluation in medial prefrontal cortex to guide decision-making.

Physiological need states direct decision-making toward re-establishing homeostasis. Using a two-alternative forced choice task for mice that models elements of human decisions, we found that varying hunger and thirst states caused need-inappropriate choices, such as food seeking when thirsty. These results show limits on interoceptive knowledge of hunger and thirst states to guide decision-making. Instead, need states were identified after food and water consumption by outcome evaluation, which depended on the medial prefrontal cortex.

Neuropixels 2.0: A miniaturized high-density probe for stable, long-term brain recordings.

Measuring the dynamics of neural processing across time scales requires following the spiking of thousands of individual neurons over milliseconds and months. To address this need, we introduce the Neuropixels 2.0 probe together with newly designed analysis algorithms. The probe has more than 5000 sites and is miniaturized to facilitate chronic implants in small mammals and recording during unrestrained behavior. High-quality recordings over long time scales were reliably obtained in mice and rats in six laboratories. Improved site density and arrangement combined with newly created data processing methods enable automatic post hoc correction for brain movements, allowing recording from the same neurons for more than 2 months. These probes and algorithms enable stable recordings from thousands of sites during free behavior, even in small animals such as mice.

Arousal Modulates Retinal Output.

At various stages of the visual system, visual responses are modulated by arousal. Here, we find that in mice this modulation operates as early as in the first synapse from the retina and even in retinal axons. To measure retinal activity in the awake, intact brain, we imaged the synaptic boutons of retinal axons in the superior colliculus. Their activity depended not only on vision but also on running speed and pupil size, regardless of retinal illumination. Arousal typically reduced their visual responses and selectivity for direction and orientation. Recordings from retinal axons in the optic tract revealed that arousal modulates the firing of some retinal ganglion cells. Arousal had similar effects postsynaptically in colliculus neurons, independent of activity in the other main source of visual inputs to the colliculus, the primary visual cortex. These results indicate that arousal modulates activity at every stage of the mouse visual system.

Computational processing of neural recordings from calcium imaging data.

Electrophysiology has long been the workhorse of neuroscience, allowing scientists to record with millisecond precision the action potentials generated by neurons in vivo. Recently, calcium imaging of fluorescent indicators has emerged as a powerful alternative. This technique has its own strengths and weaknesses and unique data processing problems and interpretation confounds. Here we review the computational methods that convert raw calcium movies to estimates of single neuron spike times with minimal human supervision. By computationally addressing the weaknesses of calcium imaging, these methods hold the promise of significantly improving data quality. We also introduce a new metric to evaluate the output of these processing pipelines, which is based on the cluster isolation distance routinely used in electrophysiology.

Spontaneous behaviors drive multidimensional, brainwide activity.

Neuronal populations in sensory cortex produce variable responses to sensory stimuli and exhibit intricate spontaneous activity even without external sensory input. Cortical variability and spontaneous activity have been variously proposed to represent random noise, recall of prior experience, or encoding of ongoing behavioral and cognitive variables. Recording more than 10,000 neurons in mouse visual cortex, we observed that spontaneous activity reliably encoded a high-dimensional latent state, which was partially related to the mouse's ongoing behavior and was represented not just in visual cortex but also across the forebrain. Sensory inputs did not interrupt this ongoing signal but added onto it a representation of external stimuli in orthogonal dimensions. Thus, visual cortical population activity, despite its apparently noisy structure, reliably encodes an orthogonal fusion of sensory and multidimensional behavioral information.

Thirst regulates motivated behavior through modulation of brainwide neural population dynamics.

Physiological needs produce motivational drives, such as thirst and hunger, that regulate behaviors essential to survival. Hypothalamic neurons sense these needs and must coordinate relevant brainwide neuronal activity to produce the appropriate behavior. We studied dynamics from ~24,000 neurons in 34 brain regions during thirst-motivated choice behavior in 21 mice as they consumed water and became sated. Water-predicting sensory cues elicited activity that rapidly spread throughout the brain of thirsty animals. These dynamics were gated by a brainwide mode of population activity that encoded motivational state. After satiation, focal optogenetic activation of hypothalamic thirst-sensing neurons returned global activity to the pre-satiation state. Thus, motivational states specify initial conditions that determine how a brainwide dynamical system transforms sensory input into behavioral output.