Search

In the brain, spiking patterns live in a high-dimensional space of neurons and time. Thus, determining the intrinsic structure of this space presents a theoretical and experimental challenge. To address this challenge, we introduce a new framework for applying topological data analysis (TDA) to spike train data and use it to determine the geometry of spiking patterns in the visual cortex. Key to our approach is a parametrized family of distances based on the timing of spikes that quantifies the dissimilarity between neuronal responses. We applied TDA to visually driven single-unit and multiple single-unit spiking activity in macaque V1 and V2. TDA across timescales reveals a common geometry for spiking patterns in V1 and V2 which, among simple models, is most similar to that of a low-dimensional space endowed with Euclidean or hyperbolic geometry with modest curvature. Remarkably, the inferred geometry depends on timescale and is clearest for the timescales that are important for encoding contrast, orientation and spatial correlations., NIH grants number EY09314 and EY07977 from the National Eye Institute of the NIH. NIH grants no. EY034150 from the National Eye Institute and NS111019 from the National Institute of Neurological Disorders and Stroke. Ikerbasque. Wallenberg AI, Autonomous Systems and Software Program (WASP) funded by the Knut and Alice Wallenberg Foundation. Inria Associated Team "NeuroTransSF".

Tools assaying the neural networks that modulate consciousness may facilitate tracking of recovery after acute severe brain injury. The ABCD framework classifies resting-state EEG into categories reflecting levels of thalamocortical network function that correlate with outcome in post-cardiac arrest coma. In this longitudinal cohort study, we applied the ABCD framework to 20 patients with acute severe traumatic brain injury requiring intensive care (12 of whom were also studied at ≥6-months post-injury) and 16 healthy controls. We tested four hypotheses: 1) EEG ABCD classifications are spatially heterogeneous and temporally variable; 2) ABCD classifications improve longitudinally, commensurate with the degree of behavioral recovery; 3) ABCD classifications correlate with behavioral level of consciousness; and 4) the Coma Recovery Scale-Revised arousal facilitation protocol yields improved ABCD classifications. Channel-level EEG power spectra were classified based on spectral peaks within pre-defined frequency bands: 'A' = no peaks above delta (4 Hz) range (complete thalamocortical disruption); 'B' = theta (4-8 Hz) peak (severe thalamocortical disruption); 'C' = theta and beta (13-24 Hz) peaks (moderate thalamocortical disruption); or 'D' = alpha (8-13 Hz) and beta peaks (normal thalamocortical function). Acutely, 95% of patients demonstrated 'D' signals in at least one channel but exhibited within-session temporal variability and spatial heterogeneity in the proportion of different channel-level ABCD classifications. By contrast, healthy participants and patients at follow-up consistently demonstrated signals corresponding to intact thalamocortical network function. Patients demonstrated longitudinal improvement in ABCD classifications (p .05) and ABCD classification distinguished patients with and without command-following in the subacute-to-chronic phase of recovery (p .01). In patients studied acutely, ABCD classifications improved after the Coma Recovery Scale-Revised arousal facilitation protocol (p .05) but did not correspond with behavioral level of consciousness. These findings support the use of the ABCD framework to characterize channel-level EEG dynamics and track fluctuations in functional thalamocortical network integrity in spatial detail.

We perceive the world based on visual information acquired via oculomotor control(1), an activity intertwined with ongoing cognitive processes(2–4). Cognitive influences have been primarily studied in the context of macroscopic movements, like saccades and smooth pursuits. However, our eyes are never still, even during periods of fixation. One of the fixational eye movements, ocular drifts, shifts the stimulus over hundreds of receptors on the retina, a motion that has been argued to enhance the processing of spatial detail by translating spatial into temporal information(5). Despite their apparent randomness, ocular drifts are under neural control(6–8). However little is known about the control of drift beyond the brainstem circuitry of the vestibulo-ocular reflex(9,10). Here, we investigated the cognitive control of ocular drifts with a letter discrimination task. The experiment was designed to reveal open-loop effects, i.e., cognitive oculomotor control driven by specific prior knowledge of the task, independent of incoming sensory information. Open-loop influences were isolated by randomly presenting pure noise fields (no letters) while subjects engaged in discriminating different letter pairs. Our results show open-loop control of drift direction in human observers.

Assessing cognitive function—especially language processing—in severely brain-injured patients is critical for prognostication, care, and development of communication devices (e.g. brain–computer interfaces). In patients with diminished motor function, language processing has been probed using EEG measures of command-following in motor imagery tasks. While such tests eliminate the need for motor response, they require sustained attention. However, passive listening tasks, with an EEG response measure can reduce both motor and attentional demands. These considerations motivated the development of two assays of low-level language processing—identification of differential phoneme-class responses and tracking of the natural speech envelope. This cross-sectional study looks at a cohort of 26 severely brain-injured patient subjects and 10 healthy controls. Patients’ level of function was assessed via the coma recovery scale–revised at the bedside. Patients were also tested for command-following via EEG and/or MRI assays of motor imagery. For the present investigation, EEG was recorded while presenting a 148 s audio clip of Alice in Wonderland. Time-locked EEG responses to phoneme classes were extracted and compared to determine a differential phoneme-class response. Tracking of the natural speech envelope was assessed from the same recordings by cross-correlating the EEG response with the speech envelope. In healthy controls, the dynamics of the two measures were temporally similar but spatially different: a central parieto-occipital component of differential phoneme-class response was absent in the natural speech envelope response. The differential phoneme-class response was present in all patient subjects, including the six classified as vegetative state/unresponsive wakefulness syndrome by behavioural assessment. However, patient subjects with evidence of language processing either by behavioural assessment or motor imagery tests had an early bilateral response in the first 50 ms that was lacking in patient subjects without any evidence of language processing. The natural speech envelope tracking response was also present in all patient subjects and responses in the first 100 ms distinguished patient subjects with evidence of language processing. Specifically, patient subjects with evidence of language processing had a more global response in the first 100 ms whereas those without evidence of language processing had a frontopolar response in that period. In summary, we developed two passive EEG-based methods to probe low-level language processing in severely brain-injured patients. In our cohort, both assays showed a difference between patient subjects with evidence of command-following and those with no evidence of command-following: a more prominent early bilateral response component.

Analysis of visual texture is important for many key steps in early vision. We study visual sensitivity to image statistics in three families of textures that include multiple gray levels and correlations in two spatial dimensions. Sensitivities to positive and negative correlations are approximately independent of correlation sign, and signals from different kinds of correlations combine quadratically. We build a computational model, fully constrained by prior studies of sensitivity to uncorrelated textures and black-and-white textures with spatial correlations. The model accounts for many features of the new data, including sign-independence, quadratic combination, and the dependence on gray-level distribution.

Recent work has shown that most cells in the rostral, gustatory portion of the nucleus tractus solitarius (rNTS) in awake, freely licking rats show lick-related firing. However, the relationship between taste-related and lick-related activity in rNTS remains unclear. Here, we tested if GABA-derived inhibitory activity regulates the balance of lick- and taste-driven neuronal activity. Combinatorial viral tools were used to restrict expression of ChR2-EYFP to GAD1+ GABAergic neurons. Viral infusions were bilateral in rNTS. 2-4wks later, an optical fiber attached to 8-16 drivable microwires was implanted into the rNTS. After recovery, water-deprived rats were presented with taste stimuli in an experimental chamber. Trials were 5 consecutive taste licks [NaCl, KCl, NH4Cl, sucrose, MSG/IMP, citric acid, quinine, or artificial saliva (AS)] separated by 5 AS licks on a VR5 schedule. Each taste lick triggered a 1s train of laser light (25Hz; 473nm; 8-10mW) in a random half of the trials. In all, 113 cells were recorded in the rNTS, 50 responded to one or more taste stimuli without GABA enhancement. Selective changes in response magnitude (spike count) within cells shifted across unit patterns but preserved inter-stimulus relationships. Cells where enhanced GABAergic tone increased lick coherence conveyed more information distinguishing basic taste qualities and different salts than other cells. In addition, GABA activation significantly amplified the amount of information that discriminated palatable vs. unpalatable tastants. By dynamically regulating lick coherence and remodeling the across-unit response patterns to taste, enhancing GABAergic tone in rNTS reconfigures the neural activity reflecting sensation and movement.Significance StatementThe rostral nucleus tractus solitarius (rNTS) is the first structure in the central gustatory pathway. Electrophysiological recordings from the rNTS in awake, freely-licking animals show that cells in this area have lick- as well as taste-related activity, but the relationship between these characteristics is not well understood. Here, we showed evidence that GABA activation can dynamically regulate both of these two properties in rNTS cells to enhance the information conveyed, especially about palatable vs. unpalatable tastants. These data provide insights into the role of inhibitory activity in the rNTS.

Similarity judgments are commonly used to study mental representations and their neural correlates. This approach has been used to characterize perceptual spaces in many domains: colors, objects, images, words, and sounds. Ideally, one might want to compare estimates of perceived similarity between all pairs of stimuli, but this is often impractical. For example, if one asks a subject to compare the similarity of two items with the similarity of two other items, the number of comparisons grows with the fourth power of the stimulus set size. An alternative strategy is to ask a subject to rate similarities of isolated pairs, e.g., on a Likert scale. This is much more efficient (the number of ratings grows quadratically with set size rather than quartically), but these ratings tend to be unstable and have limited resolution, and the approach also assumes that there are no context effects.Here, a novel ranking paradigm for efficient collection of similarity judgments is presented, along with an analysis pipeline (software provided) that tests whether Euclidean distance models account for the data. Typical trials consist of eight stimuli around a central reference stimulus: the subject ranks stimuli in order of their similarity to the reference. By judicious selection of combinations of stimuli used in each trial, the approach has internal controls for consistency and context effects. The approach was validated for stimuli drawn from Euclidean spaces of up to five dimensions. The approach is illustrated with an experiment measuring similarities among 37 words. Each trial yields the results of 28 pairwise comparisons of the form, “Was A more similar to the reference than B was to the reference?” While directly comparing all pairs of pairs of stimuli would have required 221445 trials, this design enables reconstruction of the perceptual space from 5994 such comparisons obtained from 222 trials.

Few reliable biomarkers of consciousness exist for patients with acute severe brain injury. Tools assaying the neural networks that modulate consciousness may allow for tracking of recovery. The mesocircuit model, and its instantiation as the ABCD framework, classifies resting-state EEG power spectral densities into categories reflecting widely separated levels of thalamocortical network function and correlates with outcome in post-cardiac arrest coma.We applied the ABCD framework to acute severe traumatic brain injury and tested four hypotheses: 1) EEG channel-level ABCD classifications are spatially heterogeneous and temporally variable; 2) ABCD classifications improve longitudinally, commensurate with the degree of behavioural recovery; 3) ABCD classifications correlate with behavioural level of consciousness; and 4) the Coma Recovery Scale-Revised arousal facilitation protocol improves EEG dynamics along the ABCD scale. In this longitudinal cohort study, we enrolled 20 patients with acute severe traumatic brain injury requiring intensive care and 16 healthy controls. Through visual inspection, channel-level spectra from resting-state EEG were classified based on spectral peaks within frequency bands defined by the ABCD framework: ‘A’ = no peaks above delta (Acutely, 95% of patients demonstrated ‘D’ signals in at least one channel but exhibited heterogeneity in the proportion of different channel-level ABCD classifications (mean percent ‘D’ signals: 37%, range: 0-90%). By contrast, healthy participants and patients at follow-up predominantly demonstrated signals corresponding to intact thalamocortical network function (mean percent ‘D’ signals: 94%). In patients studied acutely, ABCD classifications improved after the Coma Recovery Scale-Revised arousal facilitation protocol (PPPThese findings support the use of the ABCD framework to characterize channel-level EEG dynamics and track fluctuations in functional thalamocortical network integrity in spatial detail.

Dopaminergic stimulation has been proposed as a treatment strategy for post-traumatic brain injured patients in minimally conscious state based on a clinical trial using amantadine, a weak dopamine transporter blocker. However, a specific contribution of dopaminergic neuromodulation in minimally conscious state is undemonstrated. In a phase 0 clinical trial, we evaluated 13 normal volunteers and seven post-traumatic minimally conscious state patients using 11C-raclopride PET to estimate dopamine 2-like receptors occupancy in the striatum and central thalamus before and after dopamine transporter blockade with dextroamphetamine. If a presynaptic deficit was observed, a third and a fourth 11C-raclopride PET were acquired to evaluate changes in dopamine release induced by l-DOPA and l-DOPA+dextroamphetamine. Permutation analysis showed a significant reduction of dopamine release in patients, demonstrating a presynaptic deficit in the striatum and central thalamus that could not be reversed by blocking the dopamine transporter. However, administration of the dopamine precursor l-DOPA reversed the presynaptic deficit by restoring the biosynthesis of dopamine from both ventral tegmentum and substantia nigra. The advantages of alternative pharmacodynamic approaches in post-traumatic minimally conscious state patients should be tested in clinical trials, as patients currently refractory to amantadine might benefit from them.

The gustatory system encodes information about chemical identity, nutritional value, and concentration of sensory stimuli before transmitting the signal from taste buds to central neurons that process and transform the signal. Deciphering the coding logic for taste quality requires examining responses at each level along the neural axis - from peripheral sensory organs to gustatory cortex. From the earliest single fiber recordings, it was clear that some afferent neurons respond uniquely, others to stimuli of multiple qualities. There is frequently a “best stimulus” for a given neuron, leading to the suggestion that taste exhibits “labeled line coding”. In the extreme, a strict “labeled line” requires neurons and pathways dedicated to single qualities (e.g. sweet, bitter, etc.). At the other end of the spectrum, "across-fiber”, “combinatorial”, or “ensemble” coding requires minimal specific information to be imparted by a single neuron. Instead, taste quality information is encoded by simultaneous activity in ensembles of afferent fibers. Further, “temporal coding” models have proposed that certain features of taste quality may be embedded in the cadence of impulse activity. Taste receptor proteins are often expressed in non-overlapping sets of cells in taste buds apparently supporting “labeled lines”. Yet, taste buds include both narrowly- and broadly-tuned cells. As gustatory signals proceed to the hindbrain and on to higher centers, coding become more distributed, and temporal patterns of activity become important. Here, we present the conundrum of taste coding in the light of current electrophysiological and imaging techniques at several levels of the gustatory processing pathway.&nbsp

Visual texture is an important cue to figure-ground organization. While processing of texture differences is a prerequisite for the use of this cue to extract figure-ground organization, these stages are distinct processes. One potential indicator of this distinction is the possibility that texture statistics play a different role in the figure vs. in the ground. To determine whether this is the case, we probed figure-ground processing with a family of local image statistics that specified textures that varied in the strength and spatial scale of structure, and the extent to which features are oriented. For image statistics that generated approximately isotropic textures, the threshold for identification of figure-ground structure was determined by the difference in correlation strength in figure vs. ground, independent of whether the correlations were present in figure, ground, or both. However, for image statistics with strong orientation content, thresholds were up to two times higher for correlations in the ground, vs. the figure. This held equally for texture-defined objects with convex or concave boundaries, indicating that these threshold differences are driven by border ownership, not boundary shape. Similar threshold differences were found for presentation times ranging from 125 to 500 ms. These findings identify a qualitative difference in how texture is used for figure-ground analysis, vs. texture discrimination. Additionally, it reveals a functional recursion: texture differences are needed to identify tentative boundaries and consequent scene organization into figure and ground, but then scene organization modifies sensitivity to texture differences according to the figure-ground assignment.

Summary Humans use rapid gaze shifts, known as saccades, to explore visual scenes. These movements yield abrupt luminance changes on the retina, which elicit robust neural discharges at fixation onsets. Yet little is known about the spatial content of saccade transients. Here, we show that saccades redistribute spatial information within the temporal range of retinal sensitivity following two distinct regimes: saccade modulations counterbalance (whiten) the spectral density of natural scenes at low spatial frequencies and follow the external power distribution at higher frequencies. This redistribution is a consequence of saccade dynamics, particularly the speed/amplitude/duration relation known as the main sequence. It resembles the redistribution resulting from inter-saccadic eye drifts, revealing a continuum in the modulations given by different eye movements, with oculomotor transitions primarily acting by regulating the bandwidth of whitening. Our findings suggest important computational roles for saccade transients in the establishment of spatial representations and lead to testable predictions about their consequences for visual functions and encoding mechanisms. Video Abstract Download : Download video (30MB)

Animals use stereo sampling of odor concentration to localize sources and follow odor trails. We analyze the dynamics of a bilateral model that depends on the simultaneous comparison between odor concentrations detected by left and right sensors. The general model consists of three differential equations for the positions in the plane and the heading. When the odor landscape is an infinite trail, then we reduce the dynamics to a planar system whose dynamics have just two fixed points. Using an integrable approximation (for short sensors) we estimate the basin of attraction. In the case of a radially symmetric landscape, we again can reduce the dynamics to a planar system, but the behavior is considerably richer with multi-stability, isolas, and limit cycles. As in the linear trail case, there is also an underlying integrable system when the sensors are short. In odor landscapes that consist of multiple spots and trail segments, we find periodic and chaotic dynamics and characterize the behavior on trails with gaps and that turn corners.

A perceptual space is a mental workspace of points in a sensory domain that supports similarity and difference judgments and enables further processing such as classification and naming. Perceptual spaces are present across sensory modalities; examples include colors, faces, auditory textures, and odors. Color is perhaps the best-studied perceptual space, but it is atypical in two respects. First, the dimensions of color space are directly linked to the three cone absorption spectra, but the dimensions of generic perceptual spaces are not as readily traceable to single-neuron properties. Second, generic perceptual spaces have more than three dimensions. This is important because representing each distinguishable point in a high-dimensional space by a separate neuron or population is unwieldy; combinatorial strategies may be needed to overcome this hurdle. To study the representation of a complex perceptual space, we focused on a well-characterized 10-dimensional domain of visual textures. Within this domain, we determine perceptual distances in a threshold task (segmentation) and a suprathreshold task (border salience comparison). In N = 4 human observers, we find both quantitative and qualitative differences between these sets of measurements. Quantitatively, observers’ segmentation thresholds were inconsistent with their uncertainty determined from border salience comparisons. Qualitatively, segmentation thresholds suggested that distances are determined by a coordinate representation with Euclidean geometry. Border salience comparisons, in contrast, indicated a global curvature of the space, and that distances are determined by activity patterns across broadly tuned elements. Thus, our results indicate two representations of this perceptual space, and suggest that they use differing combinatorial strategies. Significance Statement To move from sensory signals to decisions and actions, the brain carries out a sequence of transformations. An important stage in this process is the construction of a “perceptual space” – an internal workspace of sensory information that captures similarities and differences, and enables further processing, such as classification and naming. Perceptual spaces for color, faces, visual and haptic textures and shapes, sounds, and odors (among others) are known to exist. How such spaces are represented is at present unknown. Here, using visual textures as a model, we investigate this. Psychophysical measurements suggest roles for two combinatorial strategies: one based on projections onto coordinate-like axes, and one based on patterns of activity across broadly tuned elements scattered throughout the space.