Dataset size considerations for robust acoustic and phonetic speech encoding models in EEG

In many experiments that investigate auditory and speech processing in the brain using electroencephalography (EEG), the experimental paradigm is often lengthy and tedious. Typically, the experimenter errs on the side of including more data, more trials, and therefore conducting a longer task to ensure that the data are robust and effects are measurable. Recent studies used naturalistic stimuli to investigate the brain's response to individual or a combination of multiple speech features using system identification techniques, such as multivariate temporal receptive field (mTRF) analyses. The neural data collected from such experiments must be divided into a training set and a test set to fit and validate the mTRF weights. While a good strategy is clearly to collect as much data as is feasible, it is unclear how much data are needed to achieve stable results. Furthermore, it is unclear whether the specific stimulus used for mTRF fitting and the choice of feature representation affects how much data would be required for robust and generalizable results. Here, we used previously collected EEG data from our lab using sentence stimuli and movie stimuli as well as EEG data from an open-source dataset using audiobook stimuli to better understand how much data needs to be collected for naturalistic speech experiments measuring acoustic and phonetic tuning. We found that the EEG receptive field structure tested here stabilizes after collecting a training dataset of approximately 200 s of TIMIT sentences, around 600 s of movie trailers training set data, and approximately 460 s of audiobook training set data. Thus, we provide suggestions on the minimum amount of data that would be necessary for fitting mTRFs from naturalistic listening data. Our findings are motivated by highly practical concerns when working with children, patient populations, or others who may not tolerate long study sessions. These findings will aid future researchers who wish to study naturalistic speech processing in healthy and clinical populations while minimizing participant fatigue and retaining signal quality.

A highly selective response to food in human visual cortex revealed by hypothesis-free voxel decomposition

Prior work has identified cortical regions selectively responsive to specific categories of visual stimuli. However, this hypothesis-driven work cannot reveal how prominent these category selectivities are in the overall functional organization of the visual cortex, or what others might exist that scientists have not thought to look for. Furthermore, standard voxel-wise tests cannot detect distinct neural selectivities that coexist within voxels. To overcome these limitations, we used data-driven voxel decomposition methods to identify the main components underlying fMRI responses to thousands of complex photographic images. Our hypothesis-neutral analysis rediscovered components selective for faces, places, bodies, and words, validating our method and showing that these selectivities are dominant features of the ventral visual pathway. The analysis also revealed an unexpected component with a distinct anatomical distribution that responded highly selectively to images of food. Alternative accounts based on low- to mid-level visual features, such as color, shape, or texture, failed to account for the food selectivity of this component. High-throughput testing and control experiments with matched stimuli on a highly accurate computational model of this component confirm its selectivity for food. We registered our methods and hypotheses before replicating them on held-out participants and in a novel dataset. These findings demonstrate the power of data-driven methods and show that the dominant neural responses of the ventral visual pathway include not only selectivities for faces, scenes, bodies, and words but also the visually heterogeneous category of food, thus constraining accounts of when and why functional specialization arises in the cortex.

Generalizable EEG Encoding Models with Naturalistic Audiovisual Stimuli

In natural conversations, listeners must attend to what others are saying while ignoring extraneous background sounds. Recent studies have used encoding models to predict electroencephalography (EEG) responses to speech in noise-free listening situations, sometimes referred to as "speech tracking." Researchers have analyzed how speech tracking changes with different types of background noise. It is unclear, however, whether neural responses from acoustically rich, naturalistic environments with and without background noise can be generalized to more controlled stimuli. If encoding models for acoustically rich, naturalistic stimuli are generalizable to other tasks, this could aid in data collection from populations of individuals who may not tolerate listening to more controlled and less engaging stimuli for long periods of time. We recorded noninvasive scalp EEG while 17 human participants (8 male/9 female) listened to speech without noise and audiovisual speech stimuli containing overlapping speakers and background sounds. We fit multivariate temporal receptive field encoding models to predict EEG responses to pitch, the acoustic envelope, phonological features, and visual cues in both stimulus conditions. Our results suggested that neural responses to naturalistic stimuli were generalizable to more controlled datasets. EEG responses to speech in isolation were predicted accurately using phonological features alone, while responses to speech in a rich acoustic background were more accurate when including both phonological and acoustic features. Our findings suggest that naturalistic audiovisual stimuli can be used to measure receptive fields that are comparable and generalizable to more controlled audio-only stimuli.SIGNIFICANCE STATEMENT Understanding spoken language in natural environments requires listeners to parse acoustic and linguistic information in the presence of other distracting stimuli. However, most studies of auditory processing rely on highly controlled stimuli with no background noise, or with background noise inserted at specific times. Here, we compare models where EEG data are predicted based on a combination of acoustic, phonetic, and visual features in highly disparate stimuli-sentences from a speech corpus and speech embedded within movie trailers. We show that modeling neural responses to highly noisy, audiovisual movies can uncover tuning for acoustic and phonetic information that generalizes to simpler stimuli typically used in sensory neuroscience experiments.

Population Coding of Natural Electrosensory Stimuli by Midbrain Neurons

Natural stimuli display spatiotemporal characteristics that typically vary over orders of magnitude, and their encoding by sensory neurons remains poorly understood. We investigated population coding of highly heterogeneous natural electrocommunication stimuli in Apteronotus leptorhynchus of either sex. Neuronal activities were positively correlated with one another in the absence of stimulation, and correlation magnitude decayed with increasing distance between recording sites. Under stimulation, we found that correlations between trial-averaged neuronal responses (i.e., signal correlations) were positive and higher in magnitude for neurons located close to another, but that correlations between the trial-to-trial variability (i.e., noise correlations) were independent of physical distance. Overall, signal and noise correlations were independent of stimulus waveform as well as of one another. To investigate how neuronal populations encoded natural electrocommunication stimuli, we considered a nonlinear decoder for which the activities were combined. Decoding performance was best for a timescale of 6 ms, indicating that midbrain neurons transmit information via precise spike timing. A simple summation of neuronal activities (equally weighted sum) revealed that noise correlations limited decoding performance by introducing redundancy. Using an evolution algorithm to optimize performance when considering instead unequally weighted sums of neuronal activities revealed much greater performance values, indicating that midbrain neuron populations transmit information that reliably enable discrimination between different stimulus waveforms. Interestingly, we found that different weight combinations gave rise to similar discriminability, suggesting robustness. Our results have important implications for understanding how natural stimuli are integrated by downstream brain areas to give rise to behavioral responses.SIGNIFICANCE STATEMENT We show that midbrain electrosensory neurons display correlations between their activities and that these can significantly impact performance of decoders. While noise correlations limited discrimination performance by introducing redundancy, considering unequally weighted sums of neuronal activities gave rise to much improved performance and mitigated the deleterious effects of noise correlations. Further analysis revealed that increased discriminability was achieved by making trial-averaged responses more separable, as well as by reducing trial-to-trial variability by eliminating noise correlations. We further found that multiple combinations of weights could give rise to similar discrimination performances, which suggests that such combinatorial codes could be achieved in the brain. We conclude that the activities of midbrain neuronal populations can be used to reliably discriminate between highly heterogeneous stimulus waveforms.

Nonlinear spatial integration in retinal bipolar cells shapes the encoding of artificial and natural stimuli

The retina dissects the visual scene into parallel information channels, which extract specific visual features through nonlinear processing. The first nonlinear stage is typically considered to occur at the output of bipolar cells, resulting from nonlinear transmitter release from synaptic terminals. In contrast, we show here that bipolar cells themselves can act as nonlinear processing elements at the level of their somatic membrane potential. Intracellular recordings from bipolar cells in the salamander retina revealed frequent nonlinear integration of visual signals within bipolar cell receptive field centers, affecting the encoding of artificial and natural stimuli. These nonlinearities provide sensitivity to spatial structure below the scale of bipolar cell receptive fields in both bipolar and downstream ganglion cells and appear to arise at the excitatory input into bipolar cells. Thus, our data suggest that nonlinear signal pooling starts earlier than previously thought: that is, at the input stage of bipolar cells.

•

Some retinal bipolar cells represent visual contrast in a nonlinear fashion

•

These bipolar cells also nonlinearly integrate visual signals over space

•

The spatial nonlinearity affects the encoding of natural stimuli by bipolar cells

•

The nonlinearity results from feedforward input, not from feedback inhibition

Supervised Learning Occurs in Visual Perceptual Learning of Complex Natural Images

There have been long-standing debates regarding whether supervised or unsupervised learning mechanisms are involved in visual perceptual learning (VPL) [1-14]. However, these debates have been based on the effects of simple feedback only about response accuracy in detection or discrimination tasks of low-level visual features such as orientation [15-22]. Here, we examined whether the content of response feedback plays a critical role for the acquisition and long-term retention of VPL of complex natural images. We trained three groups of human subjects (n = 72 in total) to better detect "grouped microcalcifications" or "architectural distortion" lesions (referred to as calcification and distortion in the following) in mammograms either with no trial-by-trial feedback, partial trial-by-trial feedback (response correctness only), or detailed trial-by-trial feedback (response correctness and target location). Distortion lesions consist of more complex visual structures than calcification lesions [23-26]. We found that partial feedback is necessary for VPL of calcifications, whereas detailed feedback is required for VPL of distortions. Furthermore, detailed feedback during training is necessary for VPL of distortion and calcification lesions to be retained for 6 months. These results show that although supervised learning is heavily involved in VPL of complex natural images, the extent of supervision for VPL varies across different types of complex natural images. Such differential requirements for VPL to improve the detectability of lesions in mammograms are potentially informative for the professional training of radiologists.

In Praise of Artifice Reloaded: Caution With Natural Image Databases in Modeling Vision

Subjective image quality databases are a major source of raw data on how the visual system works in naturalistic environments. These databases describe the sensitivity of many observers to a wide range of distortions of different nature and intensity seen on top of a variety of natural images. Data of this kind seems to open a number of possibilities for the vision scientist to check the models in realistic scenarios. However, while these natural databases are great benchmarks for models developed in some other way (e.g., by using the well-controlled artificial stimuli of traditional psychophysics), they should be carefully used when trying to fit vision models. Given the high dimensionality of the image space, it is very likely that some basic phenomena are under-represented in the database. Therefore, a model fitted on these large-scale natural databases will not reproduce these under-represented basic phenomena that could otherwise be easily illustrated with well selected artificial stimuli. In this work we study a specific example of the above statement. A standard cortical model using wavelets and divisive normalization tuned to reproduce subjective opinion on a large image quality dataset fails to reproduce basic cross-masking. Here we outline a solution for this problem by using artificial stimuli and by proposing a modification that makes the model easier to tune. Then, we show that the modified model is still competitive in the large-scale database. Our simulations with these artificial stimuli show that when using steerable wavelets, the conventional unit norm Gaussian kernels in divisive normalization should be multiplied by high-pass filters to reproduce basic trends in masking. Basic visual phenomena may be misrepresented in large natural image datasets but this can be solved with model-interpretable stimuli. This is an additional argument in praise of artifice in line with Rust and Movshon (2005).

Evidence Integration in Natural Acoustic Textures during Active and Passive Listening

Many natural sounds can be well described on a statistical level, for example, wind, rain, or applause. Even though the spectro-temporal profile of these acoustic textures is highly dynamic, changes in their statistics are indicative of relevant changes in the environment. Here, we investigated the neural representation of change detection in natural textures in humans, and specifically addressed whether active task engagement is required for the neural representation of this change in statistics. Subjects listened to natural textures whose spectro-temporal statistics were modified at variable times by a variable amount. Subjects were instructed to either report the detection of changes (active) or to passively listen to the stimuli. A subset of passive subjects had performed the active task before (passive-aware vs passive-naive). Psychophysically, longer exposure to pre-change statistics was correlated with faster reaction times and better discrimination performance. EEG recordings revealed that the build-up rate and size of parieto-occipital (PO) potentials reflected change size and change time. Reduced effects were observed in the passive conditions. While P2 responses were comparable across conditions, slope and height of PO potentials scaled with task involvement. Neural source localization identified a parietal source as the main contributor of change-specific potentials, in addition to more limited contributions from auditory and frontal sources. In summary, the detection of statistical changes in natural acoustic textures is predominantly reflected in parietal locations both on the skull and source level. The scaling in magnitude across different levels of task involvement suggests a context-dependent degree of evidence integration.

Encoding and Decoding Models in Cognitive Electrophysiology

Cognitive neuroscience has seen rapid growth in the size and complexity of data recorded from the human brain as well as in the computational tools available to analyze this data. This data explosion has resulted in an increased use of multivariate, model-based methods for asking neuroscience questions, allowing scientists to investigate multiple hypotheses with a single dataset, to use complex, time-varying stimuli, and to study the human brain under more naturalistic conditions. These tools come in the form of "Encoding" models, in which stimulus features are used to model brain activity, and "Decoding" models, in which neural features are used to generated a stimulus output. Here we review the current state of encoding and decoding models in cognitive electrophysiology and provide a practical guide toward conducting experiments and analyses in this emerging field. Our examples focus on using linear models in the study of human language and audition. We show how to calculate auditory receptive fields from natural sounds as well as how to decode neural recordings to predict speech. The paper aims to be a useful tutorial to these approaches, and a practical introduction to using machine learning and applied statistics to build models of neural activity. The data analytic approaches we discuss may also be applied to other sensory modalities, motor systems, and cognitive systems, and we cover some examples in these areas. In addition, a collection of Jupyter notebooks is publicly available as a complement to the material covered in this paper, providing code examples and tutorials for predictive modeling in python. The aim is to provide a practical understanding of predictive modeling of human brain data and to propose best-practices in conducting these analyses.

Olfactory receptor neurons use gain control and complementary kinetics to encode intermittent odorant stimuli

Insects find food and mates by navigating odorant plumes that can be highly intermittent, with intensities and durations that vary rapidly over orders of magnitude. Much is known about olfactory responses to pulses and steps, but it remains unclear how olfactory receptor neurons (ORNs) detect the intensity and timing of natural stimuli, where the absence of scale in the signal makes detection a formidable olfactory task. By stimulating Drosophila ORNs in vivo with naturalistic and Gaussian stimuli, we show that ORNs adapt to stimulus mean and variance, and that adaptation and saturation contribute to naturalistic sensing. Mean-dependent gain control followed the Weber-Fechner relation and occurred primarily at odor transduction, while variance-dependent gain control occurred at both transduction and spiking. Transduction and spike generation possessed complementary kinetic properties, that together preserved the timing of odorant encounters in ORN spiking, regardless of intensity. Such scale-invariance could be critical during odor plume navigation.

The Hierarchical Cortical Organization of Human Speech Processing

Speech comprehension requires that the brain extract semantic meaning from the spectral features represented at the cochlea. To investigate this process, we performed an fMRI experiment in which five men and two women passively listened to several hours of natural narrative speech. We then used voxelwise modeling to predict BOLD responses based on three different feature spaces that represent the spectral, articulatory, and semantic properties of speech. The amount of variance explained by each feature space was then assessed using a separate validation dataset. Because some responses might be explained equally well by more than one feature space, we used a variance partitioning analysis to determine the fraction of the variance that was uniquely explained by each feature space. Consistent with previous studies, we found that speech comprehension involves hierarchical representations starting in primary auditory areas and moving laterally on the temporal lobe: spectral features are found in the core of A1, mixtures of spectral and articulatory in STG, mixtures of articulatory and semantic in STS, and semantic in STS and beyond. Our data also show that both hemispheres are equally and actively involved in speech perception and interpretation. Further, responses as early in the auditory hierarchy as in STS are more correlated with semantic than spectral representations. These results illustrate the importance of using natural speech in neurolinguistic research. Our methodology also provides an efficient way to simultaneously test multiple specific hypotheses about the representations of speech without using block designs and segmented or synthetic speech.SIGNIFICANCE STATEMENT To investigate the processing steps performed by the human brain to transform natural speech sound into meaningful language, we used models based on a hierarchical set of speech features to predict BOLD responses of individual voxels recorded in an fMRI experiment while subjects listened to natural speech. Both cerebral hemispheres were actively involved in speech processing in large and equal amounts. Also, the transformation from spectral features to semantic elements occurs early in the cortical speech-processing stream. Our experimental and analytical approaches are important alternatives and complements to standard approaches that use segmented speech and block designs, which report more laterality in speech processing and associated semantic processing to higher levels of cortex than reported here.

The statistics of the vestibular input experienced during natural self-motion differ between rodents and primates

KEY POINTS

In order to understand how the brain's coding strategies are adapted to the statistics of the sensory stimuli experienced during everyday life, the use of animal models is essential. Mice and non-human primates have become common models for furthering our knowledge of the neuronal coding of natural stimuli, but differences in their natural environments and behavioural repertoire may impact optimal coding strategies. Here we investigated the structure and statistics of the vestibular input experienced by mice versus non-human primates during natural behaviours, and found important differences. Our data establish that the structure and statistics of natural signals in non-human primates more closely resemble those observed previously in humans, suggesting similar coding strategies for incoming vestibular input. These results help us understand how the effects of active sensing and biomechanics will differentially shape the statistics of vestibular stimuli across species, and have important implications for sensory coding in other systems.

ABSTRACT

It is widely believed that sensory systems are adapted to the statistical structure of natural stimuli, thereby optimizing coding. Recent evidence suggests that this is also the case for the vestibular system, which senses self-motion and in turn contributes to essential brain functions ranging from the most automatic reflexes to spatial perception and motor coordination. However, little is known about the statistics of self-motion stimuli actually experienced by freely moving animals in their natural environments. Accordingly, here we examined the natural self-motion signals experienced by mice and monkeys: two species commonly used to study vestibular neural coding. First, we found that probability distributions for all six dimensions of motion (three rotations, three translations) in both species deviated from normality due to long tails. Interestingly, the power spectra of natural rotational stimuli displayed similar structure for both species and were not well fitted by power laws. This result contrasts with reports that the natural spectra of other sensory modalities (i.e. vision, auditory and tactile) instead show a power-law relationship with frequency, which indicates scale invariance. Analysis of natural translational stimuli revealed important species differences as power spectra deviated from scale invariance for monkeys but not for mice. By comparing our results to previously published data for humans, we found the statistical structure of natural self-motion stimuli in monkeys and humans more closely resemble one another. Our results thus predict that, overall, neural coding strategies used by vestibular pathways to encode natural self-motion stimuli are fundamentally different in rodents and primates.

Scale-Free Neural and Physiological Dynamics in Naturalistic Stimuli Processing

Neural activity recorded at multiple spatiotemporal scales is dominated by arrhythmic fluctuations without a characteristic temporal periodicity. Such activity often exhibits a 1/f-type power spectrum, in which power falls off with increasing frequency following a power-law function: [Formula: see text], which is indicative of scale-free dynamics. Two extensively studied forms of scale-free neural dynamics in the human brain are slow cortical potentials (SCPs)-the low-frequency (<5 Hz) component of brain field potentials-and the amplitude fluctuations of α oscillations, both of which have been shown to carry important functional roles. In addition, scale-free dynamics characterize normal human physiology such as heartbeat dynamics. However, the exact relationships among these scale-free neural and physiological dynamics remain unclear. We recorded simultaneous magnetoencephalography and electrocardiography in healthy subjects in the resting state and while performing a discrimination task on scale-free dynamical auditory stimuli that followed different scale-free statistics. We observed that long-range temporal correlation (captured by the power-law exponent β) in SCPs positively correlated with that of heartbeat dynamics across time within an individual and negatively correlated with that of α-amplitude fluctuations across individuals. In addition, across individuals, long-range temporal correlation of both SCP and α-oscillation amplitude predicted subjects' discrimination performance in the auditory task, albeit through antagonistic relationships. These findings reveal interrelations among different scale-free neural and physiological dynamics and initial evidence for the involvement of scale-free neural dynamics in the processing of natural stimuli, which often exhibit scale-free dynamics.

Phase Locking of Multiple Single Neurons to the Local Field Potential in Cat V1

The local field potential (LFP) is thought to reflect a temporal reference for neuronal spiking, which may facilitate information coding and orchestrate the communication between neural populations. To explore this proposed role, we recorded the LFP and simultaneously the spike activity of one to three nearby neurons in V1 of anesthetized cats during the presentation of drifting sinusoidal gratings, binary dense noise stimuli, and natural movies. In all stimulus conditions and during spontaneous activity, the average LFP power at frequencies >20 Hz was higher when neurons were spiking versus not spiking. The spikes were weakly but significantly phase locked to all frequencies of the LFP. The average spike phase of the LFP was stable across high and low levels of LFP power, but the strength of phase locking at low frequencies (≤10 Hz) increased with increasing LFP power. In a next step, we studied how strong stimulus responses of single neurons are reflected in the LFP and the LFP-spike relationship. We found that LFP power was slightly increased and phase locking was slightly stronger during strong compared with weak stimulus-locked responses. In summary, the coupling strength between high frequencies of the LFP and spikes was not strongly modulated by LFP power, which is thought to reflect spiking synchrony, nor was it strongly influenced by how strongly the neuron was driven by the stimulus. Furthermore, a comparison between neighboring neurons showed no clustering of preferred LFP phase. We argue that hypotheses on the relevance of phase locking in their current form are inconsistent with our findings.

Measuring Preference for Supernormal Over Natural Rewards : A Two-Dimensional Anticipatory Pleasure Scale

Supernormal (SN) stimuli are artificial products that activate reward pathways and approach behavior more so than naturally occurring stimuli for which these systems were intended. Many modern consumer products (e.g., snack foods, alcohol, and pornography) appear to incorporate SN features, leading to excessive consumption, in preference to naturally occurring alternatives. No measure currently exists for the self-report assessment of individual differences or changes in susceptibility to such stimuli. Therefore, an anticipatory pleasure scale was modified to include items that represented both SN and natural (N) classes of rewarding stimuli. Exploratory factor analysis yielded a two-factor solution, and as predicted, N and SN items reliably loaded on separate dimensions. Internal reliability for the two scales was high, ρ =.93 and ρ =.90, respectively. The two-dimensional measure was evaluated via regression using the N and SN scale means as predictors and self-reports of daily consumption of 21 products with SN features as outcomes. As expected, SN pleasure ratings were related to higher SN product consumption, while N pleasure ratings had either negative or neutral associations to consumption of these products. We conclude that the resulting two-dimensional measure is a potentially reliable and valid self-report measure of differential preference for SN stimuli. While further evaluation is needed (e.g., using experimental measures), the proposed scale may play a useful role in the study of both trait- and state-based variation in human susceptibility to SN stimuli.

Subthreshold membrane responses underlying sparse spiking to natural vocal signals in auditory cortex

Natural acoustic communication signals, such as speech, are typically high-dimensional with a wide range of co-varying spectro-temporal features at multiple timescales. The synaptic and network mechanisms for encoding these complex signals are largely unknown. We are investigating these mechanisms in high-level sensory regions of the songbird auditory forebrain, where single neurons show sparse, object-selective spiking responses to conspecific songs. Using whole-cell in vivo patch clamp techniques in the caudal mesopallium and the caudal nidopallium of starlings, we examine song-driven subthreshold and spiking activity. We find that both the subthreshold and the spiking activity are reliable (i.e. the same song drives a similar response each time it is presented) and specific (i.e. responses to different songs are distinct). Surprisingly, however, the reliability and specificity of the subthreshold response was uniformly high regardless of when the cell spiked, even for song stimuli that drove no spikes. We conclude that despite a selective and sparse spiking response, high-level auditory cortical neurons are under continuous, non-selective, stimulus-specific synaptic control. To investigate the role of local network inhibition in this synaptic control, we then recorded extracellularly while pharmacologically blocking local GABAergic transmission. This manipulation modulated the strength and the reliability of stimulus-driven spiking, consistent with a role for local inhibition in regulating the reliability of network activity and the stimulus specificity of the subthreshold response in single cells. We discuss these results in the context of underlying computations that could generate sparse, stimulus-selective spiking responses, and models for hierarchical pooling.

A thesaurus for a neural population code

Information is carried in the brain by the joint spiking patterns of large groups of noisy, unreliable neurons. This noise limits the capacity of the neural code and determines how information can be transmitted and read-out. To accurately decode, the brain must overcome this noise and identify which patterns are semantically similar. We use models of network encoding noise to learn a thesaurus for populations of neurons in the vertebrate retina responding to artificial and natural videos, measuring the similarity between population responses to visual stimuli based on the information they carry. This thesaurus reveals that the code is organized in clusters of synonymous activity patterns that are similar in meaning but may differ considerably in their structure. This organization is highly reminiscent of the design of engineered codes. We suggest that the brain may use this structure and show how it allows accurate decoding of novel stimuli from novel spiking patterns.

DOI:http://dx.doi.org/10.7554/eLife.06134.001

Our ability to perceive the world is dependent on information from our senses being passed between different parts of the brain. The information is encoded as patterns of electrical pulses or ‘spikes’, which other brain regions must be able to decipher. Cracking this code would thus enable us to predict the patterns of nerve impulses that would occur in response to specific stimuli, and ‘decode’ which stimuli had produced particular patterns of impulses.

This task is challenging in part because of its scale—vast numbers of stimuli are encoded by huge numbers of neurons that can send their spikes in many different combinations. Furthermore, neurons are inherently noisy and their response to identical stimuli may vary considerably in the number of spikes and their timing. This means that the brain cannot simply link a single unchanging pattern of firing with each stimulus, because these firing patterns are often distorted by biophysical noise.

Ganmor et al. have now modeled the effects of noise in a network of neurons in the retina (found at the back of the eye), and, in doing so, have provided insights into how the brain solves this problem. This has brought us a step closer to cracking the neural code. First, 10 second video clips of natural scenes and artificial stimuli were played on a loop to a sample of retina taken from a salamander, and the responses of nearly 100 neurons in the sample were recorded for two hours. Dividing the 10 second clip into short segments provided a series of 500 stimuli, which the network had been exposed to more than 600 times.

Ganmor et al. analyzed the responses of groups of 20 cells to each stimulus and found that physically similar firing patterns were not particularly likely to encode the same stimulus. This can be likened to the way that words such as ‘light’ and ‘night’ have similar structures but different meanings. Instead, the model reveals that each stimulus was represented by a cluster of firing patterns that bore little physical resemblance to one another, but which nevertheless conveyed the same meaning. To continue on with the previous example, this is similar to way that ‘light’ and ‘illumination’ have the same meaning but different structures.

Ganmor et al. use these new data to map the organization of the ‘vocabulary’ of populations of cells the retina, and put together a kind of ‘thesaurus’ that enables new activity patterns of the retina to be decoded and could be used to crack the neural code. Furthermore, the organization of ‘synonyms’ is strikingly similar to codes that are favored in many forms of telecommunication. In these man-made codes, codewords that represent different items are chosen to be so distinct from each other that even if they were corrupted by noise, they could be correctly deciphered. Correspondingly, in the retina, patterns that carry the same meaning occupy a distinct area, and new patterns can be interpreted based on their proximity to these clusters.

DOI:http://dx.doi.org/10.7554/eLife.06134.002

Five key factors determining pairwise correlations in visual cortex

The responses of cortical neurons to repeated presentation of a stimulus are highly variable, yet correlated. These "noise correlations" reflect a low-dimensional structure of population dynamics. Here, we examine noise correlations in 22,705 pairs of neurons in primary visual cortex (V1) of anesthetized cats, during ongoing activity and in response to artificial and natural visual stimuli. We measured how noise correlations depend on 11 factors. Because these factors are themselves not independent, we distinguished their influences using a nonlinear additive model. The model revealed that five key factors play a predominant role in determining pairwise correlations. Two of these are distance in cortex and difference in sensory tuning: these are known to decrease correlation. A third factor is firing rate: confirming most earlier observations, it markedly increased pairwise correlations. A fourth factor is spike width: cells with a broad spike were more strongly correlated amongst each other. A fifth factor is spike isolation: neurons with worse isolation were more correlated, even if they were recorded on different electrodes. For pairs of neurons with poor isolation, this last factor was the main determinant of correlations. These results were generally independent of stimulus type and timescale of analysis, but there were exceptions. For instance, pairwise correlations depended on difference in orientation tuning more during responses to gratings than to natural stimuli. These results consolidate disjoint observations in a vast literature on pairwise correlations and point towards regularities of population coding in sensory cortex.

The increased sensitivity of irregular peripheral canal and otolith vestibular afferents optimizes their encoding of natural stimuli

Efficient processing of incoming sensory input is essential for an organism's survival. A growing body of evidence suggests that sensory systems have developed coding strategies that are constrained by the statistics of the natural environment. Consequently, it is necessary to first characterize neural responses to natural stimuli to uncover the coding strategies used by a given sensory system. Here we report for the first time the statistics of vestibular rotational and translational stimuli experienced by rhesus monkeys during natural (e.g., walking, grooming) behaviors. We find that these stimuli can reach intensities as high as 1500 deg/s and 8 G. Recordings from afferents during naturalistic rotational and linear motion further revealed strongly nonlinear responses in the form of rectification and saturation, which could not be accurately predicted by traditional linear models of vestibular processing. Accordingly, we used linear-nonlinear cascade models and found that these could accurately predict responses to naturalistic stimuli. Finally, we tested whether the statistics of natural vestibular signals constrain the neural coding strategies used by peripheral afferents. We found that both irregular otolith and semicircular canal afferents, because of their higher sensitivities, were more optimized for processing natural vestibular stimuli as compared with their regular counterparts. Our results therefore provide the first evidence supporting the hypothesis that the neural coding strategies used by the vestibular system are matched to the statistics of natural stimuli.

Accurately decoding visual information from fMRI data obtained in a realistic virtual environment

Three-dimensional interactive virtual environments (VEs) are a powerful tool for brain-imaging based cognitive neuroscience that are presently under-utilized. This paper presents machine-learning based methods for identifying brain states induced by realistic VEs with improved accuracy as well as the capability for mapping their spatial topography on the neocortex. VEs provide the ability to study the brain under conditions closer to the environment in which humans evolved, and thus to probe deeper into the complexities of human cognition. As a test case, we designed a stimulus to reflect a military combat situation in the Middle East, motivated by the potential of using real-time functional magnetic resonance imaging (fMRI) in the treatment of post-traumatic stress disorder. Each subject experienced moving through the virtual town where they encountered 1-6 animated combatants at different locations, while fMRI data was collected. To analyze the data from what is, compared to most studies, more complex and less controlled stimuli, we employed statistical machine learning in the form of Multi-Voxel Pattern Analysis (MVPA) with special attention given to artificial Neural Networks (NN). Extensions to NN that exploit the block structure of the stimulus were developed to improve the accuracy of the classification, achieving performances from 58 to 93% (chance was 16.7%) with six subjects. This demonstrates that MVPA can decode a complex cognitive state, viewing a number of characters, in a dynamic virtual environment. To better understand the source of this information in the brain, a novel form of sensitivity analysis was developed to use NN to quantify the degree to which each voxel contributed to classification. Compared with maps produced by general linear models and the searchlight approach, these sensitivity maps revealed a more diverse pattern of information relevant to the classification of cognitive state.

Saliency and saccade encoding in the frontal eye field during natural scene search

The frontal eye field (FEF) plays a central role in saccade selection and execution. Using artificial stimuli, many studies have shown that the activity of neurons in the FEF is affected by both visually salient stimuli in a neuron's receptive field and upcoming saccades in a certain direction. However, the extent to which visual and motor information is represented in the FEF in the context of the cluttered natural scenes we encounter during everyday life has not been explored. Here, we model the activities of neurons in the FEF, recorded while monkeys were searching natural scenes, using both visual and saccade information. We compare the contribution of bottom-up visual saliency (based on low-level features such as brightness, orientation, and color) and saccade direction. We find that, while saliency is correlated with the activities of some neurons, this relationship is ultimately driven by activities related to movement. Although bottom-up visual saliency contributes to the choice of saccade targets, it does not appear that FEF neurons actively encode the kind of saliency posited by popular saliency map theories. Instead, our results emphasize the FEF's role in the stages of saccade planning directly related to movement generation.

Statistics of the vestibular input experienced during natural self-motion: implications for neural processing

It is widely believed that sensory systems are optimized for processing stimuli occurring in the natural environment. However, it remains unknown whether this principle applies to the vestibular system, which contributes to essential brain functions ranging from the most automatic reflexes to spatial perception and motor coordination. Here we quantified, for the first time, the statistics of natural vestibular inputs experienced by freely moving human subjects during typical everyday activities. Although previous studies have found that the power spectra of natural signals across sensory modalities decay as a power law (i.e., as 1/f(α)), we found that this did not apply to natural vestibular stimuli. Instead, power decreased slowly at lower and more rapidly at higher frequencies for all motion dimensions. We further establish that this unique stimulus structure is the result of active motion as well as passive biomechanical filtering occurring before any neural processing. Notably, the transition frequency (i.e., frequency at which power starts to decrease rapidly) was lower when subjects passively experienced sensory stimulation than when they actively controlled stimulation through their own movement. In contrast to signals measured at the head, the spectral content of externally generated (i.e., passive) environmental motion did follow a power law. Specifically, transformations caused by both motor control and biomechanics shape the statistics of natural vestibular stimuli before neural processing. We suggest that the unique structure of natural vestibular stimuli will have important consequences on the neural coding strategies used by this essential sensory system to represent self-motion in everyday life.

A practical, intuitive brain-computer interface for communicating 'yes' or 'no' by listening

OBJECTIVE

Previous work has shown that it is possible to build an EEG-based binary brain-computer interface system (BCI) driven purely by shifts of attention to auditory stimuli. However, previous studies used abrupt, abstract stimuli that are often perceived as harsh and unpleasant, and whose lack of inherent meaning may make the interface unintuitive and difficult for beginners. We aimed to establish whether we could transition to a system based on more natural, intuitive stimuli (spoken words 'yes' and 'no') without loss of performance, and whether the system could be used by people in the locked-in state.

APPROACH

We performed a counterbalanced, interleaved within-subject comparison between an auditory streaming BCI that used beep stimuli, and one that used word stimuli. Fourteen healthy volunteers performed two sessions each, on separate days. We also collected preliminary data from two subjects with advanced amyotrophic lateral sclerosis (ALS), who used the word-based system to answer a set of simple yes-no questions.

MAIN RESULTS

The N1, N2 and P3 event-related potentials elicited by words varied more between subjects than those elicited by beeps. However, the difference between responses to attended and unattended stimuli was more consistent with words than beeps. Healthy subjects' performance with word stimuli (mean 77% ± 3.3 s.e.) was slightly but not significantly better than their performance with beep stimuli (mean 73% ± 2.8 s.e.). The two subjects with ALS used the word-based BCI to answer questions with a level of accuracy similar to that of the healthy subjects.

SIGNIFICANCE

Since performance using word stimuli was at least as good as performance using beeps, we recommend that auditory streaming BCI systems be built with word stimuli to make the system more pleasant and intuitive. Our preliminary data show that word-based streaming BCI is a promising tool for communication by people who are locked in.

Octopaminergic modulation of a fly visual motion-sensitive neuron during stimulation with naturalistic optic flow

In a variety of species locomotor activity, like walking or flying, has been demonstrated to alter visual information processing. The neuromodulator octopamine was shown to change the response characteristics of optic flow processing neurons in the fly's visual system in a similar way as locomotor activity. This modulation resulted in enhanced neuronal responses, in particular during sustained stimulation with high temporal frequencies, and in shorter latencies of responses to abrupt onsets of pattern motion. These state-dependent changes were interpreted to adjust neuronal tuning to the range of high velocities encountered during locomotion. Here we assess the significance of these changes for the processing of optic flow as experienced during flight. Naturalistic image sequences were reconstructed based on measurements of the head position and gaze direction of Calliphora vicina flying in an arena. We recorded the responses of the V1 neuron during presentation of these image sequences on a panoramic stimulus device ("FliMax"). Consistent with previous accounts, we found that spontaneous as well as stimulus-induced spike rates were increased by an octopamine agonist and decreased by an antagonist. Moreover, a small but consistent decrease in response latency upon octopaminergic activation was present, which might support fast responses to optic flow cues and limit instabilities during closed-loop optomotor regulation. However, apart from these effects the similarities between the dynamic response properties in the different pharmacologically induced states were surprisingly high, indicating that the processing of naturalistic optic flow is not fundamentally altered by octopaminergic modulation.

Trade-off between curvature tuning and position invariance in visual area V4

Humans can rapidly recognize a multitude of objects despite differences in their appearance. The neural mechanisms that endow high-level sensory neurons with both selectivity to complex stimulus features and "tolerance" or invariance to identity-preserving transformations, such as spatial translation, remain poorly understood. Previous studies have demonstrated that both tolerance and selectivity to conjunctions of features are increased at successive stages of the ventral visual stream that mediates visual recognition. Within a given area, such as visual area V4 or the inferotemporal cortex, tolerance has been found to be inversely related to the sparseness of neural responses, which in turn was positively correlated with conjunction selectivity. However, the direct relationship between tolerance and conjunction selectivity has been difficult to establish, with different studies reporting either an inverse or no significant relationship. To resolve this, we measured V4 responses to natural scenes, and using recently developed statistical techniques, we estimated both the relevant stimulus features and the range of translation invariance for each neuron. Focusing the analysis on tuning to curvature, a tractable example of conjunction selectivity, we found that neurons that were tuned to more curved contours had smaller ranges of position invariance and produced sparser responses to natural stimuli. These trade-offs provide empirical support for recent theories of how the visual system estimates 3D shapes from shading and texture flows, as well as the tiling hypothesis of the visual space for different curvature values.

Retinal oscillations carry visual information to cortex

Thalamic relay cells fire action potentials that transmit information from retina to cortex. The amount of information that spike trains encode is usually estimated from the precision of spike timing with respect to the stimulus. Sensory input, however, is only one factor that influences neural activity. For example, intrinsic dynamics, such as oscillations of networks of neurons, also modulate firing pattern. Here, we asked if retinal oscillations might help to convey information to neurons downstream. Specifically, we made whole-cell recordings from relay cells to reveal retinal inputs (EPSPs) and thalamic outputs (spikes) and then analyzed these events with information theory. Our results show that thalamic spike trains operate as two multiplexed channels. One channel, which occupies a low frequency band (<30 Hz), is encoded by average firing rate with respect to the stimulus and carries information about local changes in the visual field over time. The other operates in the gamma frequency band (40–80 Hz) and is encoded by spike timing relative to retinal oscillations. At times, the second channel conveyed even more information than the first. Because retinal oscillations involve extensive networks of ganglion cells, it is likely that the second channel transmits information about global features of the visual scene.

On the computations analyzing natural optic flow: quantitative model analysis of the blowfly motion vision pathway

For many animals, including humans, the optic flow generated on the eyes during locomotion is an important source of information about self-motion and the structure of the environment. The blowfly has been used frequently as a model system for experimental analysis of optic flow processing at the microcircuit level. Here, we describe a model of the computational mechanisms implemented by these circuits in the blowfly motion vision pathway. Although this model was originally proposed based on simple experimenter-designed stimuli, we show that it is also capable to quantitatively predict the responses to the complex dynamic stimuli a blowfly encounters in free flight. In particular, the model visual system exploits the active saccadic gaze and flight strategy of blowflies in a similar way, as does its neuronal counterpart. The model circuit extracts information about translation velocity in the intersaccadic intervals and thus, indirectly, about the three-dimensional layout of the environment. By stepwise dissection of the model circuit, we determine which of its components are essential for these remarkable features. When accounting for the responses to complex natural stimuli, the model is much more robust against parameter changes than when explaining the neuronal responses to simple experimenter-defined stimuli. In contrast to conclusions drawn from experiments with simple stimuli, optimization of the parameter set for different segments of natural optic flow stimuli do not indicate pronounced adaptational changes of these parameters during long-lasting stimulation.

Function and coding in the blowfly H1 neuron during naturalistic optic flow

Naturalistic stimuli, reconstructed from measured eye movements of flying blowflies, were replayed on a panoramic stimulus device. The directional movement-sensitive H1 neuron was recorded from blowflies watching these stimuli. The response of the H1 neuron is dominated by the response to fast saccadic turns into one direction. The response between saccades is mostly inhibited by the front-to-back optic flow caused by the forward translation during flight. To unravel the functional significance of the H1 neuron, we replayed, in addition to the original behaviorally generated stimulus, two targeted stimulus modifications: (1) a stimulus in which flow resulting from translation was removed (this stimulus produced strong intersaccadic responses); and (2) a stimulus in which the saccades were removed by assuming that the head follows the smooth flight trajectory (this stimulus produced alternating zero or nearly saturating spike rates). The responses to the two modified stimuli are strongly different from the response to the original stimulus, showing the importance of translation and saccades for the H1 response to natural optic flow. The response to the original stimulus thus suggests a double function for the H1 neuron, assisting two major classes of movement-sensitive output neurons targeted by H1. First, its strong response to saccades may function as a saccadic suppressor (via one of its target neurons) for cells involved in figure-ground discrimination. Second, its intersaccadic response may increase the signal-to-noise ratio (SNR) of wide-field neurons involved in detecting translational optic flow between saccades, in particular when flying speeds are low or when object distances are large.

Processing of natural temporal stimuli by macaque retinal ganglion cells

This study quantifies the performance of primate retinal ganglion cells in response to natural stimuli. Stimuli were confined to the temporal and chromatic domains and were derived from two contrasting environments, one typically northern European and the other a flower show. The performance of the cells was evaluated by investigating variability of cell responses to repeated stimulus presentations and by comparing measured to model responses. Both analyses yielded a quantity called the coherence rate (in bits per second), which is related to the information rate. Magnocellular (MC) cells yielded coherence rates of up to 100 bits/sec, rates of parvocellular (PC) cells were much lower, and short wavelength (S)-cone-driven ganglion cells yielded intermediate rates. The modeling approach showed that for MC cells, coherence rates were generated almost exclusively by the luminance content of the stimulus. Coherence rates of PC cells were also dominated by achromatic content. This is a consequence of the stimulus structure; luminance varied much more in the natural environment than chromaticity. Only approximately one-sixth of the coherence rate of the PC cells derived from chromatic content, and it was dominated by frequencies below 10 Hz. S-cone-driven ganglion cells also yielded coherence rates dominated by low frequencies. Below 2-3 Hz, PC cell signals contained more power than those of MC cells. Response variation between individual ganglion cells of a particular class was analyzed by constructing generic cells, the properties of which may be relevant for performance higher in the visual system. The approach used here helps define retinal modules useful for studies of higher visual processing of natural stimuli.

Redundancy reduction and sustained firing with stochastic depressing synapses

Many synapses in the CNS transmit only a fraction of the action potentials that reach them. Although unreliable, such synapses do not transmit completely randomly, because the probability of transmission depends on the recent history of synaptic activity. We examine how a variety of spike trains, including examples recorded from area V1 of monkeys freely viewing natural scenes, are transmitted through a realistic model synapse with activity-dependent depression arising from vesicle depletion or postrelease refractoriness. The resulting sequences of transmitted spikes are significantly less correlated, and hence less redundant, than the presynaptic spike trains that generate them. The spike trains we analyze, which are typical of those recorded in a variety of brain regions, have positive autocorrelations because of the occurrence of variable length periods of sustained firing at approximately constant rates. Sustained firing may, at first, seem inconsistent with input from depressing synapses. We show, however, that such a pattern of activity can arise if the postsynaptic neuron is driven by a fixed population of direct, "feedforward" inputs accompanied by a variable number of delayed, "reverberatory" inputs. This leads to a prediction for the number and latency distribution of the inputs that typically drive a cortical neuron.

Representation of acoustic communication signals by insect auditory receptor neurons

Despite their simple auditory systems, some insect species recognize certain temporal aspects of acoustic stimuli with an acuity equal to that of vertebrates; however, the underlying neural mechanisms and coding schemes are only partially understood. In this study, we analyze the response characteristics of the peripheral auditory system of grasshoppers with special emphasis on the representation of species-specific communication signals. We use both natural calling songs and artificial random stimuli designed to focus on two low-order statistical properties of the songs: their typical time scales and the distribution of their modulation amplitudes. Based on stimulus reconstruction techniques and quantified within an information-theoretic framework, our data show that artificial stimuli with typical time scales of >40 msec can be read from single spike trains with high accuracy. Faster stimulus variations can be reconstructed only for behaviorally relevant amplitude distributions. The highest rates of information transmission (180 bits/sec) and the highest coding efficiencies (40%) are obtained for stimuli that capture both the time scales and amplitude distributions of natural songs. Use of multiple spike trains significantly improves the reconstruction of stimuli that vary on time scales <40 msec or feature amplitude distributions as occur when several grasshopper songs overlap. Signal-to-noise ratios obtained from the reconstructions of natural songs do not exceed those obtained from artificial stimuli with the same low-order statistical properties. We conclude that auditory receptor neurons are optimized to extract both the time scales and the amplitude distribution of natural songs. They are not optimized, however, to extract higher-order statistical properties of the song-specific rhythmic patterns.