Complexity of frequency receptive fields predicts tonotopic variability across species

Universal statistics of hippocampal place fields across species and dimensionalities

Hippocampal place cells have single, bell-shaped place fields in small environments. Recent experiments, however, reveal that, in large environments, place cells have multiple fields with heterogeneous shapes and sizes. We show that this diversity is explained by a surprisingly simple mathematical model, in which place fields are generated by thresholding a realization of a random Gaussian process. The model captures the statistics of field arrangements and generates new quantitative predictions about the statistics of field shapes and topologies. These predictions are quantitatively verified in bats and rodents, in one, two, and three dimensions, in both small and large environments. These results imply that common mechanisms underlie the diverse statistics observed in different experiments and further suggest that synaptic projections to CA1 are predominantly random.

Random compressed coding with neurons

Classical models of efficient coding in neurons assume simple mean responses-"tuning curves"- such as bell-shaped or monotonic functions of a stimulus feature. Real neurons, however, can be more complex: grid cells, for example, exhibit periodic responses that impart the neural population code with high accuracy. But do highly accurate codes require fine-tuning of the response properties? We address this question with the use of a simple model: a population of neurons with random, spatially extended, and irregular tuning curves. Irregularity enhances the local resolution of the code but gives rise to catastrophic, global errors. For optimal smoothness of the tuning curves, when local and global errors balance out, the neural population compresses information about a continuous stimulus into a low-dimensional representation, and the resulting distributed code achieves exponential accuracy. An analysis of recordings from monkey motor cortex points to such "compressed efficient coding." Efficient codes do not require a finely tuned design-they emerge robustly from irregularity or randomness.

Dendrites endow artificial neural networks with accurate, robust and parameter-efficient learning

Artificial neural networks (ANNs) are at the core of most Deep Learning (DL) algorithms that successfully tackle complex problems like image recognition, autonomous driving, and natural language processing. However, unlike biological brains who tackle similar problems in a very efficient manner, DL algorithms require a large number of trainable parameters, making them energy-intensive and prone to overfitting. Here, we show that a new ANN architecture that incorporates the structured connectivity and restricted sampling properties of biological dendrites counteracts these limitations. We find that dendritic ANNs are more robust to overfitting and match or outperform traditional ANNs on several image classification tasks while using significantly fewer trainable parameters. These advantages are likely the result of a different learning strategy, whereby most of the nodes in dendritic ANNs respond to multiple classes, unlike classical ANNs that strive for class-specificity. Our findings suggest that the incorporation of dendritic properties can make learning in ANNs more precise, resilient, and parameter-efficient and shed new light on how biological features can impact the learning strategies of ANNs.

Ferret contributions to the business of sensory neurobiology

In this brief review, we will highlight the ferret Mustela putorius furo as an increasingly utilized animal model for sensory systems and cognitive neuroscience research. In particular, the human like hearing range of the ferret, coupled with their amenability to training, make them an especially useful model for auditory and multisensory neuroscience. These factors, combined with the increasing availability of virally mediated circuit dissection methods, mean they occupy a unique niche as a versatile and valuable research model.

Analysis methods for large-scale neuronal recordings

Simultaneous recordings from hundreds or thousands of neurons are becoming routine because of innovations in instrumentation, molecular tools, and data processing software. Such recordings can be analyzed with data science methods, but it is not immediately clear what methods to use or how to adapt them for neuroscience applications. We review, categorize, and illustrate diverse analysis methods for neural population recordings and describe how these methods have been used to make progress on longstanding questions in neuroscience. We review a variety of approaches, ranging from the mathematically simple to the complex, from exploratory to hypothesis-driven, and from recently developed to more established methods. We also illustrate some of the common statistical pitfalls in analyzing large-scale neural data.

Dendrites endow artificial neural networks with accurate, robust and parameter-efficient learning

Artificial neural networks (ANNs) are at the core of most Deep learning (DL) algorithms that successfully tackle complex problems like image recognition, autonomous driving, and natural language processing. However, unlike biological brains who tackle similar problems in a very efficient manner, DL algorithms require a large number of trainable parameters, making them energy-intensive and prone to overfitting. Here, we show that a new ANN architecture that incorporates the structured connectivity and restricted sampling properties of biological dendrites counteracts these limitations. We find that dendritic ANNs are more robust to overfitting and outperform traditional ANNs on several image classification tasks while using significantly fewer trainable parameters. These advantages are likely the result of a different learning strategy, whereby most of the nodes in dendritic ANNs respond to multiple classes, unlike classical ANNs that strive for class-specificity. Our findings suggest that the incorporation of dendritic properties can make learning in ANNs more precise, resilient, and parameter-efficient and shed new light on how biological features can impact the learning strategies of ANNs.

Topography and Ensemble Activity in the Auditory Cortex of a Mouse Model of Fragile X Syndrome

Autism spectrum disorder (ASD) is often associated with social communication impairments and specific sound processing deficits, for example, problems in following speech in noisy environments. To investigate underlying neuronal processing defects located in the auditory cortex (AC), we performed two-photon Ca2+ imaging in FMR1 (fragile X messenger ribonucleoprotein 1) knock-out (KO) mice, a model for fragile X syndrome (FXS), the most common cause of hereditary ASD in humans. For primary AC (A1) and the anterior auditory field (AAF), topographic frequency representation was less ordered compared with control animals. We additionally analyzed ensemble AC activity in response to various sounds and found subfield-specific differences. In A1, ensemble correlations were lower in general, while in secondary AC (A2), correlations were higher in response to complex sounds, but not to pure tones. Furthermore, sound specificity of ensemble activity was decreased in AAF. Repeating these experiments 1 week later revealed no major differences regarding representational drift. Nevertheless, we found subfield- and genotype-specific changes in ensemble correlation values between the two times points, hinting at alterations in network stability in FMR1 KO mice. These detailed insights into AC network activity and topography in FMR1 KO mice add to the understanding of auditory processing defects in FXS.

Cross-hemispheric communication: Insights on lateralized brain functions

On the surface, the two hemispheres of vertebrate brains look almost perfectly symmetrical, but several motor, sensory, and cognitive systems show a deeply lateralized organization. Importantly, the two hemispheres are connected by various commissures, white matter tracts that cross the brain's midline and enable cross-hemispheric communication. Cross-hemispheric communication has been suggested to play an important role in the emergence of lateralized brain functions. Here, we review current advances in understanding cross-hemispheric communication that have been made using modern neuroscientific tools in rodents and other model species, such as genetic labeling, large-scale recordings of neuronal activity, spatiotemporally precise perturbation, and quantitative behavior analyses. These findings suggest that the emergence of lateralized brain functions cannot be fully explained by largely static factors such as genetic variation and differences in structural brain asymmetries. In addition, learning-dependent asymmetric interactions between the left and right hemispheres shape lateralized brain functions.

Fractured columnar small-world functional network organization in volumes of L2/3 of mouse auditory cortex

The sensory cortices of the brain exhibit large-scale functional topographic organization, such as the tonotopic organization of the primary auditory cortex (A1) according to sound frequency. However, at the level of individual neurons, layer 2/3 (L2/3) A1 appears functionally heterogeneous. To identify if there exists a higher-order functional organization of meso-scale neuronal networks within L2/3 that bridges order and disorder, we used in vivo two-photon calcium imaging of pyramidal neurons to identify networks in three-dimensional volumes of L2/3 A1 in awake mice. Using tonal stimuli, we found diverse receptive fields with measurable colocalization of similarly tuned neurons across depth but less so across L2/3 sublayers. These results indicate a fractured microcolumnar organization with a column radius of ∼50 µm, with a more random organization of the receptive field over larger radii. We further characterized the functional networks formed within L2/3 by analyzing the spatial distribution of signal correlations (SCs). Networks show evidence of Rentian scaling in physical space, suggesting effective spatial embedding of subnetworks. Indeed, functional networks have characteristics of small-world topology, implying that there are clusters of functionally similar neurons with sparse connections between differently tuned neurons. These results indicate that underlying the regularity of the tonotopic map on large scales in L2/3 is significant tuning diversity arranged in a hybrid organization with microcolumnar structures and efficient network topologies.

Large-scale single-neuron speech sound encoding across the depth of human cortex

Understanding the neural basis of speech perception requires that we study the human brain both at the scale of the fundamental computational unit of neurons and in their organization across the depth of cortex. Here we used high-density Neuropixels arrays1-3 to record from 685 neurons across cortical layers at nine sites in a high-level auditory region that is critical for speech, the superior temporal gyrus4,5, while participants listened to spoken sentences. Single neurons encoded a wide range of speech sound cues, including features of consonants and vowels, relative vocal pitch, onsets, amplitude envelope and sequence statistics. Neurons at each cross-laminar recording exhibited dominant tuning to a primary speech feature while also containing a substantial proportion of neurons that encoded other features contributing to heterogeneous selectivity. Spatially, neurons at similar cortical depths tended to encode similar speech features. Activity across all cortical layers was predictive of high-frequency field potentials (electrocorticography), providing a neuronal origin for macroelectrode recordings from the cortical surface. Together, these results establish single-neuron tuning across the cortical laminae as an important dimension of speech encoding in human superior temporal gyrus.

Distinct topographic organization and network activity patterns of corticocollicular neurons within layer 5 auditory cortex

The auditory cortex (AC) modulates the activity of upstream pathways in the auditory brainstem via descending (corticofugal) projections. This feedback system plays an important role in the plasticity of the auditory system by shaping response properties of neurons in many subcortical nuclei. The majority of layer (L) 5 corticofugal neurons project to the inferior colliculus (IC). This corticocollicular (CC) pathway is involved in processing of complex sounds, auditory-related learning, and defense behavior. Partly due to their location in deep cortical layers, CC neuron population activity patterns within neuronal AC ensembles remain poorly understood. We employed two-photon imaging to record the activity of hundreds of L5 neurons in anesthetized as well as awake animals. CC neurons are broader tuned than other L5 pyramidal neurons and display weaker topographic order in core AC subfields. Network activity analyses revealed stronger clusters of CC neurons compared to non-CC neurons, which respond more reliable and integrate information over larger distances. However, results obtained from secondary auditory cortex (A2) differed considerably. Here CC neurons displayed similar or higher topography, depending on the subset of neurons analyzed. Furthermore, specifically in A2, CC activity clusters formed in response to complex sounds were spatially more restricted compared to other L5 neurons. Our findings indicate distinct network mechanism of CC neurons in analyzing sound properties with pronounced subfield differences, demonstrating that the topography of sound-evoked responses within AC is neuron-type dependent.

Altered population activity and local tuning heterogeneity in auditory cortex of Cacna2d3-deficient mice

The α2δ3 auxiliary subunit of voltage-activated calcium channels is required for normal synaptic transmission and precise temporal processing of sounds in the auditory brainstem. In mice its loss additionally leads to an inability to distinguish amplitude-modulated tones. Furthermore, loss of function of α2δ3 has been associated with autism spectrum disorder in humans. To investigate possible alterations of network activity in the higher-order auditory system in α2δ3 knockout mice, we analyzed neuronal activity patterns and topography of frequency tuning within networks of the auditory cortex (AC) using two-photon Ca2+ imaging. Compared to wild-type mice we found distinct subfield-specific alterations in the primary auditory cortex, expressed in overall lower correlations between the network activity patterns in response to different sounds as well as lower reliability of these patterns upon repetitions of the same sound. Higher AC subfields did not display these alterations but showed a higher amount of well-tuned neurons along with lower local heterogeneity of the neurons' frequency tuning. Our results provide new insight into AC network activity alterations in an autism spectrum disorder-associated mouse model.

High-density single-unit human cortical recordings using the Neuropixels probe

The action potential is a fundamental unit of neural computation. Even though significant advances have been made in recording large numbers of individual neurons in animal models, translation of these methodologies to humans has been limited because of clinical constraints and electrode reliability. Here, we present a reliable method for intraoperative recording of dozens of neurons in humans using the Neuropixels probe, yielding up to ∼100 simultaneously recorded single units. Most single units were active within 1 min of reaching target depth. The motion of the electrode array had a strong inverse correlation with yield, identifying a major challenge and opportunity to further increase the probe utility. Cell pairs active close in time were spatially closer in most recordings, demonstrating the power to resolve complex cortical dynamics. Altogether, this approach provides access to population single-unit activity across the depth of human neocortex at scales previously only accessible in animal models.

A silent two-photon imaging system for studying in vivo auditory neuronal functions

Two-photon laser-scanning microscopy has become an essential tool for imaging neuronal functions in vivo and has been applied to different parts of the neural system, including the auditory system. However, many components of a two-photon microscope, such as galvanometer-based laser scanners, generate mechanical vibrations and thus acoustic artifacts, making it difficult to interpret auditory responses from recorded neurons. Here, we report the development of a silent two-photon imaging system and its applications in the common marmoset (Callithrix Jacchus), a non-human primate species sharing a similar hearing range with humans. By utilizing an orthogonal pair of acousto-optical deflectors (AODs), full-frame raster scanning at video rate was achieved without introducing mechanical vibrations. Imaging depth can be optically controlled by adjusting the chirping speed on the AODs without any mechanical motion along the Z-axis. Furthermore, all other sound-generating components of the system were acoustically isolated, leaving the noise floor of the working system below the marmoset's hearing threshold. Imaging with the system in awake marmosets revealed many auditory cortex neurons that exhibited maximal responses at low sound levels, which were not possible to study using traditional two-photon imaging systems. This is the first demonstration of a silent two-photon imaging system that is capable of imaging auditory neuronal functions in vivo without acoustic artifacts. This capacity opens new opportunities for a better understanding of auditory functions in the brain and helps isolate animal behavior from microscope-generated acoustic interference.

Auditory thalamus dysfunction and pathophysiology in tinnitus: a predictive network hypothesis

Tinnitus is the perception of a 'ringing' sound without an acoustic source. It is generally accepted that tinnitus develops after peripheral hearing loss and is associated with altered auditory processing. The thalamus is a crucial relay in the underlying pathways that actively shapes processing of auditory signals before the respective information reaches the cerebral cortex. Here, we review animal and human evidence to define thalamic function in tinnitus. Overall increased spontaneous firing patterns and altered coherence between the thalamic medial geniculate body (MGB) and auditory cortices is observed in animal models of tinnitus. It is likely that the functional connectivity between the MGB and primary and secondary auditory cortices is reduced in humans. Conversely, there are indications for increased connectivity between the MGB and several areas in the cingulate cortex and posterior cerebellar regions, as well as variability in connectivity between the MGB and frontal areas regarding laterality and orientation in the inferior, medial and superior frontal gyrus. We suggest that these changes affect adaptive sensory gating of temporal and spectral sound features along the auditory pathway, reflecting dysfunction in an extensive thalamo-cortical network implicated in predictive temporal adaptation to the auditory environment. Modulation of temporal characteristics of input signals might hence factor into a thalamo-cortical dysrhythmia profile of tinnitus, but could ultimately also establish new directions for treatment options for persons with tinnitus.

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.

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.

Enhanced representation of natural sound sequences in the ventral auditory midbrain

The auditory midbrain (inferior colliculus, IC) plays an important role in sound processing, acting as hub for acoustic information extraction and for the implementation of fast audio-motor behaviors. IC neurons are topographically organized according to their sound frequency preference: dorsal IC regions encode low frequencies while ventral areas respond best to high frequencies, a type of sensory map defined as tonotopy. Tonotopic maps have been studied extensively using artificial stimuli (pure tones) but our knowledge of how these maps represent information about sequences of natural, spectro-temporally rich sounds is sparse. We studied this question by conducting simultaneous extracellular recordings across IC depths in awake bats (Carollia perspicillata) that listened to sequences of natural communication and echolocation sounds. The hypothesis was that information about these two types of sound streams is represented at different IC depths since they exhibit large differences in spectral composition, i.e., echolocation covers the high-frequency portion of the bat soundscape (> 45 kHz), while communication sounds are broadband and carry most power at low frequencies (20–25 kHz). Our results showed that mutual information between neuronal responses and acoustic stimuli, as well as response redundancy in pairs of neurons recorded simultaneously, increase exponentially with IC depth. The latter occurs regardless of the sound type presented to the bats (echolocation or communication). Taken together, our results indicate the existence of mutual information and redundancy maps at the midbrain level whose response cannot be predicted based on the frequency composition of natural sounds and classic neuronal tuning curves.

Complexity of frequency receptive fields predicts tonotopic variability across species

Primary cortical areas contain maps of sensory features, including sound frequency in primary auditory cortex (A1). Two-photon calcium imaging in mice has confirmed the presence of these global tonotopic maps, while uncovering an unexpected local variability in the stimulus preferences of individual neurons in A1 and other primary regions. Here we show that local heterogeneity of frequency preferences is not unique to rodents. Using two-photon calcium imaging in layers 2/3, we found that local variance in frequency preferences is equivalent in ferrets and mice. Neurons with multipeaked frequency tuning are less spatially organized than those tuned to a single frequency in both species. Furthermore, we show that microelectrode recordings may describe a smoother tonotopic arrangement due to a sampling bias towards neurons with simple frequency tuning. These results help explain previous inconsistencies in cortical topography across species and recording techniques.