Refine
Year of publication
Document Type
- Preprint (47) (remove)
Language
- English (47) (remove)
Has Fulltext
- yes (47) (remove)
Is part of the Bibliography
- no (47)
Keywords
- natural scenes (2)
- neuronal populations (2)
- primary visual cortex (2)
- stimulus encoding (2)
- visual attention (2)
- MEG (1)
- inter-individual variability (1)
- myelin (1)
- perceptual closure (1)
- precision weighting (1)
Institute
- MPI für Hirnforschung (47) (remove)
The firing pattern of ventral midbrain dopamine neurons is controlled by afferent and intrinsic activity to generate prediction error signals that are essential for reward-based learning. Given the absence of intracellular in vivo recordings in the last three decades, the subthreshold membrane potential events that cause changes in dopamine neuron firing patterns remain unknown. By establishing stable in vivo whole-cell recordings of >100 spontaneously active midbrain dopamine neurons in anaesthetized mice, we identified the repertoire of subthreshold membrane potential signatures associated with distinct in vivo firing patterns. We demonstrate that dopamine neuron in vivo activity deviates from a single spike pacemaker pattern by eliciting transient increases in firing rate generated by at least two diametrically opposing biophysical mechanisms: a transient depolarization resulting in high frequency plateau bursts associated with a reactive, depolarizing shift in action potential threshold; and a prolonged hyperpolarization preceding slower rebound bursts characterized by a predictive, hyperpolarizing shift in action potential threshold. Our findings therefore illustrate a framework for the biophysical implementation of prediction error and sensory cue coding in dopamine neurons by tuning action potential threshold dynamics.
Several studies have probed perceptual performance at different times after a self-paced motor action and found frequency-specific modulations of perceptual performance phase-locked to the action. Such action-related modulation has been reported for various frequencies and modulation strengths. In an attempt to establish a basic effect at the population level, we had a relatively large number of participants (n=50) perform a self-paced button press followed by a detection task at threshold, and we applied both fixed- and random-effects tests. The combined data of all trials and participants surprisingly did not show any significant action-related modulation. However, based on previous studies, we explored the possibility that such modulation depends on the participant’s internal state. Indeed, when we split trials based on performance in neighboring trials, then trials in periods of low performance showed an action-related modulation at ≈17 Hz. When we split trials based on the performance in the preceding trial, we found that trials following a “miss” showed an action-related modulation at ≈17 Hz. Finally, when we split participants based on their false-alarm rate, we found that participants with no false alarms showed an action-related modulation at ≈17 Hz. All these effects were significant in random-effects tests, supporting an inference on the population. Together, these findings indicate that action-related modulations are not always detectable. However, the results suggest that specific internal states such as lower attentional engagement and/or higher decision criterion are characterized by a modulation in the beta-frequency range.
Several recent studies investigated the rhythmic nature of cognitive processes that lead to perception and behavioral report. These studies used different methods, and there has not yet been an agreement on a general standard. Here, we present a way to test and quantitatively compare these methods. We simulated behavioral data from a typical experiment and analyzed these data with several methods. We applied the main methods found in the literature, namely sine-wave fitting, the Discrete Fourier Transform (DFT) and the Least Square Spectrum (LSS). DFT and LSS can be applied both on the averaged accuracy time course and on single trials. LSS is mathematically equivalent to DFT in the case of regular, but not irregular sampling - which is more common. LSS additionally offers the possibility to take into account a weighting factor which affects the strength of the rhythm, such as arousal. Statistical inferences were done either on the investigated sample (fixed-effect) or on the population (random-effect) of simulated participants. Multiple comparisons across frequencies were corrected using False-Discovery-Rate, Bonferroni, or the Max-Based approach. To perform a quantitative comparison, we calculated Sensitivity, Specificity and D-prime of the investigated analysis methods and statistical approaches. Within the investigated parameter range, single-trial methods had higher sensitivity and D-prime than the methods based on the averaged-accuracy-time-course. This effect was further increased for a simulated rhythm of higher frequency. If an additional (observable) factor influenced detection performance, adding this factor as weight in the LSS further improved Sensitivity and D-prime. For multiple comparison correction, the Max-Based approach provided the highest Specificity and D-prime, closely followed by the Bonferroni approach. Given a fixed total amount of trials, the random-effect approach had higher D-prime when trials were distributed over a larger number of participants, even though this gave less trials per participant. Finally, we present the idea of using a dampened sinusoidal oscillator instead of a simple sinusoidal function, to further improve the fit to behavioral rhythmicity observed after a reset event.
Analyzing non-invasive recordings of electroencephalography (EEG) and magnetoencephalography (MEG) directly in sensor space, using the signal from individual sensors, is a convenient and standard way of working with this type of data. However, volume conduction introduces considerable challenges for sensor space analysis. While the general idea of signal mixing due to volume conduction in EEG/MEG is recognized, the implications have not yet been clearly exemplified. Here, we illustrate how different types of activity overlap on the level of individual sensors. We show spatial mixing in the context of alpha rhythms, which are known to have generators in different areas of the brain. Using simulations with a realistic 3D head model and lead field and data analysis of a large resting-state EEG dataset, we show that electrode signals can be differentially affected by spatial mixing by computing a sensor complexity measure. While prominent occipital alpha rhythms result in less heterogeneous spatial mixing on posterior electrodes, central electrodes show a diversity of rhythms present. This makes the individual contributions, such as the sensorimotor mu-rhythm and temporal alpha rhythms, hard to disentangle from the dominant occipital alpha. Additionally, we show how strong occipital rhythms rhythms can contribute the majority of activity to frontal channels, potentially compromising analyses that are solely conducted in sensor space. We also outline specific consequences of signal mixing for frequently used assessment of power, power ratios and connectivity profiles in basic research and for neurofeedback application. With this work, we hope to illustrate the effects of volume conduction in a concrete way, such that the provided practical illustrations may be of use to EEG researchers to in order to evaluate whether sensor space is an appropriate choice for their topic of investigation.
Borders and edges are salient and behaviourally relevant features for navigating the environment. The brain forms dedicated neural representations of environmental boundaries, which are assumed to serve as a reference for spatial coding. Here we expand this border coding network to include the retrosplenial cortex (RSC) in which we identified neurons that increase their firing near all boundaries of an arena. RSC border cells specifically encode walls, but not objects, and maintain their tuning in the absence of direct sensory detection. Unlike border cells in the medial entorhinal cortex (MEC), RSC border cells are sensitive to the animal’s direction to nearby walls located contralateral to the recorded hemisphere. Pharmacogenetic inactivation of MEC led to a disruption of RSC border coding, but not vice versa, indicating network directionality. Together these data shed light on how information about distance and direction of boundaries is generated in the brain for guiding navigation behaviour.
Brookshire (2022) claims that previous analyses of periodicity in detection performance after a reset event suffer from extreme false-positive rates. Here we show that this conclusion is based on an incorrect implemention of a null-hypothesis of aperiodicity, and that a correct implementation confirms low false-positive rates. Furthermore, we clarify that the previously used method of shuffling-in-time, and thereby shuffling-in-phase, cleanly implements the null hypothesis of no temporal structure after the reset, and thereby of no phase locking to the reset. Moving from a corresponding phase-locking spectrum to an inference on the periodicity of the underlying process can be accomplished by parameterizing the spectrum. This can separate periodic from non-periodic components, and quantify the strength of periodicity.
Cognition requires the dynamic modulation of effective connectivity, i.e. the modulation of the postsynaptic neuronal response to a given input. If postsynaptic neurons are rhythmically active, this might entail rhythmic gain modulation, such that inputs synchronized to phases of high gain benefit from enhanced effective connectivity. We show that visually induced gamma-band activity in awake macaque area V4 rhythmically modulates responses to unpredictable stimulus events. This modulation exceeded a simple additive superposition of a constant response onto ongoing gamma-rhythmic firing, demonstrating the modulation of multiplicative gain. Gamma phases leading to strongest neuronal responses also led to shortest behavioral reaction times, suggesting functional relevance of the effect. Furthermore, we find that constant optogenetic stimulation of anesthetized cat area 21a produces gamma-band activity entailing a similar gain modulation. As the gamma rhythm in area 21a did not spread backwards to area 17, this suggests that postsynaptic gamma is sufficient for gain modulation.
Synchronization has been implicated in neuronal communication, but causal evidence remains indirect. We used optogenetics to generate depolarizing currents in pyramidal neurons of cat visual cortex, emulating excitatory synaptic inputs under precise temporal control, while measuring spike output. Cortex transformed constant excitation into strong gamma-band synchronization, revealing the well-known cortical resonance. Increasing excitation with ramps increased the strength and frequency of synchronization. Slow, symmetric excitation profiles revealed hysteresis of power and frequency. Crucially, white-noise input sequences enabled causal analysis of network transmission, establishing that cortical resonance selectively transmits coherent input components. Models composed of recurrently coupled excitatory and inhibitory units uncovered a crucial role of feedback inhibition and suggest that hysteresis can arise through spike-frequency adaptation. The presented approach provides a powerful means to investigate the resonance properties of local circuits and probe how these properties transform input and shape transmission.
The gamma rhythm has been implicated in neuronal communication, but causal evidence remains indirect. We measured spike output of local neuronal networks and emulated their synaptic input through optogenetics. Opsins provide currents through somato-dendritic membranes, similar to synapses, yet under experimental control with high temporal precision. We expressed Channelrhodopsin-2 in excitatory neurons of cat visual cortex and recorded neuronal responses to light with different temporal characteristics. Sine waves of different frequencies entrained neuronal responses with a reliability that peaked for input frequencies in the gamma band. Crucially, we also presented white-noise sequences, because their temporal unpredictability enables analysis of causality. Neuronal spike output was caused specifically by the input’s gamma component. This gamma-specific transfer function is likely an emergent property of in-vivo networks with feedback inhibition. The method described here could reveal the transfer function between the input to any one and the output of any other neuronal group.
Signal transfer of visual stimuli to V4 occurs in gamma-rhythmic, pulsed information packages
(2020)
Summary Selective visual attention allows the brain to focus on behaviorally relevant information while ignoring irrelevant signals. As a possible mechanism, routing by synchronization was proposed: neural populations sending attended signals align their gamma-rhythmic activities with receiving populations, such that spikes from the senders arrive at excitability peaks of the receivers, enhancing signal transfer. Conversely, the non-attended signals arrive unaligned to the receiver’s oscillation, reducing signal transfer. Therefore, visual signals should be transferred through periodically pulsed information packages, resulting in a modulation of the stimulus content within the receiver’s activity by its gamma phase and amplitude. To test this prediction, we quantified gamma phase-specific stimulus content within neural activity from area V4 of macaques performing a visual attention task. For the attended stimulus we find enhanced stimulus content reaching its maximum near excitability peaks, with effect magnitude increasing with oscillation amplitude, establishing a functional link between selective processing and gamma activity.
Afterimages result from a prolonged exposure to still visual stimuli. They are best detectable when viewed against uniform backgrounds and can persist for multiple seconds. Consequently, the dynamics of afterimages appears to be slow by their very nature. To the contrary, we report here that about 50% of an afterimage intensity can be erased rapidly—within less than a second. The prerequisite is that subjects view a rich visual content to erase the afterimage; fast erasure of afterimages does not occur if subjects view a blank screen. Moreover, we find evidence that fast removal of afterimages is a skill learned with practice as our subjects were always more effective in cleaning up afterimages in later parts of the experiment. These results can be explained by a tri-level hierarchy of adaptive mechanisms, as has been proposed by the theory of practopoiesis.
Cross-frequency coupling (CFC) has been proposed to coordinate neural dynamics across spatial and temporal scales. Despite its potential relevance for understanding healthy and pathological brain function, the standard CFC analysis and physiological interpretation come with fundamental problems. For example, apparent CFC can appear because of spectral correlations due to common non-stationarities that may arise in the total absence of interactions between neural frequency components. To provide a road map towards an improved mechanistic understanding of CFC, we organize the available and potential novel statistical/modeling approaches according to their biophysical interpretability. While we do not provide solutions for all the problems described, we provide a list of practical recommendations to avoid common errors and to enhance the interpretability of CFC analysis.
How much data do we need? Lower bounds of brain activation states to predict human cognitive ability
(2022)
Human functional brain connectivity can be temporally decomposed into states of high and low cofluctuation, defined as coactivation of brain regions over time. Despite their low frequency of occurrence, states of particularly high cofluctuation have been shown to reflect fundamentals of intrinsic functional network architecture (derived from resting-state fMRI) and to be highly subject-specific. However, it is currently unclear whether such network-defining states of high cofluctuation also contribute to individual variations in cognitive abilities – which strongly rely on the interactions among distributed brain regions. By introducing CMEP, an eigenvector-based prediction framework, we show that functional connectivity estimates from as few as 20 temporally separated time frames (< 3% of a 10 min resting-state fMRI scan) are significantly predictive of individual differences in intelligence (N = 281, p < .001). In contrast and against previous expectations, individual’s network-defining time frames of particularly high cofluctuation do not achieve significant prediction of intelligence. Multiple functional brain networks contribute to the prediction, and all results replicate in an independent sample (N = 831). Our results suggest that although fundamentals of person-specific functional connectomes can be derived from few time frames of highest brain connectivity, temporally distributed information is necessary to extract information about cognitive abilities from functional connectivity time series. This information, however, is not restricted to specific connectivity states, like network-defining high-cofluctuation states, but rather reflected across the entire length of the brain connectivity time series.
Probing the association between resting state brain network dynamics and psychological resilience
(2021)
Abstract
This study aimed at replicating a previously reported negative correlation between node flexibility and psychological resilience, i.e., the ability to retain mental health in the face of stress and adversity. To this end, we used multiband resting-state BOLD fMRI (TR = .675 sec) from 52 participants who had filled out three psychological questionnaires assessing resilience. Time-resolved functional connectivity was calculated by performing a sliding window approach on averaged time series parcellated according to different established atlases. Multilayer modularity detection was performed to track network reconfigurations over time and node flexibility was calculated as the number of times a node changes community assignment. In addition, node promiscuity (the fraction of communities a node participates in) and node degree (as proxy for time-varying connectivity) were calculated to extend previous work. We found no substantial correlations between resilience and node flexibility. We observed a small number of correlations between the two other brain measures and resilience scores, that were however very inconsistently distributed across brain measures, differences in temporal sampling, and parcellation schemes. This heterogeneity calls into question the existence of previously postulated associations between resilience and brain network flexibility and highlights how results may be influenced by specific analysis choices.
Author Summary We tested the replicability and generalizability of a previously proposed negative association between dynamic brain network reconfigurations derived from multilayer modularity detection (node flexibility) and psychological resilience. Using multiband resting-state BOLD fMRI data and exploring several parcellation schemes, sliding window approaches, and temporal resolutions of the data, we could not replicate previously reported findings regarding the association between node flexibility and resilience. By extending this work to other measures of brain dynamics (node promiscuity, degree) we observe a rather inconsistent pattern of correlations with resilience, that strongly varies across analysis choices. We conclude that further research is needed to understand the network neuroscience basis of mental health and discuss several reasons that may account for the variability in results.
Word familiarity and predictive context facilitate visual word processing, leading to faster recognition times and reduced neuronal responses. Previously, models with and without top-down connections, including lexical-semantic, pre-lexical (e.g., orthographic/ phonological), and visual processing levels were successful in accounting for these facilitation effects. Here we systematically assessed context-based facilitation with a repetition priming task and explicitly dissociated pre-lexical and lexical processing levels using a pseudoword familiarization procedure. Experiment 1 investigated the temporal dynamics of neuronal facilitation effects with magnetoencephalography (MEG; N=38 human participants) while Experiment 2 assessed behavioral facilitation effects (N=24 human participants). Across all stimulus conditions, MEG demonstrated context-based facilitation across multiple time windows starting at 100 ms, in occipital brain areas. This finding indicates context based-facilitation at an early visual processing level. In both experiments, we furthermore found an interaction of context and lexical familiarity, such that stimuli with associated meaning showed the strongest context-dependent facilitation in brain activation and behavior. Using MEG, this facilitation effect could be localized to the left anterior temporal lobe at around 400 ms, indicating within-level (i.e., exclusively lexical-semantic) facilitation but no top-down effects on earlier processing stages. Increased pre-lexical familiarity (in pseudowords familiarized utilizing training) did not enhance or reduce context effects significantly. We conclude that context based-facilitation is achieved within visual and lexical processing levels. Finally, by testing alternative hypotheses derived from mechanistic accounts of repetition suppression, we suggest that the facilitatory context effects found here are implemented using a predictive coding mechanism.
To characterize the left-ventral occipito-temporal cortex (lvOT) role during reading in a quantitatively explicit and testable manner, we propose the lexical categorization model (LCM). The LCM assumes that lvOT optimizes linguistic processing by allowing fast meaning access when words are familiar and filter out orthographic strings without meaning. The LCM successfully simulates benchmark results from functional brain imaging. Empirically, using functional magnetic resonance imaging, we demonstrate that quantitative LCM simulations predict lvOT activation across three studies better than alternative models. Besides, we found that word-likeness, which is assumed as input to LCM, is represented posterior to lvOT. In contrast, a dichotomous word/non-word contrast, which is assumed as the LCM’s output, could be localized to upstream frontal brain regions. Finally, we found that training lexical categorization results in more efficient reading. Thus, we propose a ventral-visual-stream processing framework for reading involving word-likeness extraction followed by lexical categorization, before meaning extraction.
To a crucial extent, the efficiency of reading results from the fact that visual word recognition is faster in predictive contexts. Predictive coding models suggest that this facilitation results from pre-activation of predictable stimulus features across multiple representational levels before stimulus onset. Still, it is not sufficiently understood which aspects of the rich set of linguistic representations that are activated during reading – visual, orthographic, phonological, and/or lexical-semantic – contribute to context-dependent facilitation. To investigate in detail which linguistic representations are pre-activated in a predictive context and how they affect subsequent stimulus processing, we combined a well-controlled repetition priming paradigm, including words and pseudowords (i.e., pronounceable nonwords), with behavioral and magnetoencephalography measurements. For statistical analysis, we used linear mixed modeling, which we found had a higher statistical power compared to conventional multivariate pattern decoding analysis. Behavioral data from 49 participants indicate that word predictability (i.e., context present vs. absent) facilitated orthographic and lexical-semantic, but not visual or phonological processes. Magnetoencephalography data from 38 participants show sustained activation of orthographic and lexical-semantic representations in the interval before processing the predicted stimulus, suggesting selective pre-activation at multiple levels of linguistic representation as proposed by predictive coding. However, we found more robust lexical-semantic representations when processing predictable in contrast to unpredictable letter strings, and pre-activation effects mainly resembled brain responses elicited when processing the expected letter string. This finding suggests that pre-activation did not result in ‘explaining away’ predictable stimulus features, but rather in a ‘sharpening’ of brain responses involved in word processing.
Most current models assume that the perceptual and cognitive processes of visual word recognition and reading operate upon neuronally coded domain-general low-level visual representations – typically oriented line representations. We here demonstrate, consistent with neurophysiological theories of Bayesian-like predictive neural computations, that prior visual knowledge of words may be utilized to ‘explain away’ redundant and highly expected parts of the visual percept. Subsequent processing stages, accordingly, operate upon an optimized representation of the visual input, the orthographic prediction error, highlighting only the visual information relevant for word identification. We show that this optimized representation is related to orthographic word characteristics, accounts for word recognition behavior, and is processed early in the visual processing stream, i.e., in V4 and before 200 ms after word-onset. Based on these findings, we propose that prior visual-orthographic knowledge is used to optimize the representation of visually presented words, which in turn allows for highly efficient reading processes.
How is semantic information stored in the human mind and brain? Some philosophers and cognitive scientists argue for vectorial representations of concepts, where the meaning of a word is represented as its position in a high-dimensional neural state space. At the intersection of natural language processing and artificial intelligence, a class of very successful distributional word vector models has developed that can account for classic EEG findings of language, i.e., the ease vs. difficulty of integrating a word with its sentence context. However, models of semantics have to account not only for context-based word processing, but should also describe how word meaning is represented. Here, we investigate whether distributional vector representations of word meaning can model brain activity induced by words presented without context. Using EEG activity (event-related brain potentials) collected while participants in two experiments (English, German) read isolated words, we encode and decode word vectors taken from the family of prediction-based word2vec algorithms. We find that, first, the position of a word in vector space allows the prediction of the pattern of corresponding neural activity over time, in particular during a time window of 300 to 500 ms after word onset. Second, distributional models perform better than a human-created taxonomic baseline model (WordNet), and this holds for several distinct vector-based models. Third, multiple latent semantic dimensions of word meaning can be decoded from brain activity. Combined, these results suggest that empiricist, prediction-based vectorial representations of meaning are a viable candidate for the representational architecture of human semantic knowledge.
The outstanding speed of language comprehension necessitates a highly efficient implementation of cognitive-linguistic processes. The domain-general theory of Predictive Coding suggests that our brain solves this problem by continuously forming linguistic predictions about expected upcoming input. The neurophysiological implementation of these predictive linguistic processes, however, is not yet understood. Here, we use EEG (human participants, both sexes) to investigate the existence and nature of online-generated, category-level semantic representations during sentence processing. We conducted two experiments in which some nouns – embedded in a predictive spoken sentence context – were unexpectedly delayed by 1 second. Target nouns were either abstract/concrete (Experiment 1) or animate/inanimate (Experiment 2). We hypothesized that if neural prediction error signals following (temporary) omissions carry specific information about the stimulus, the semantic category of the upcoming target word is encoded in brain activity prior to its presentation. Using time-generalized multivariate pattern analysis, we demonstrate significant decoding of word category from silent periods directly preceding the target word, in both experiments. This provides direct evidence for predictive coding during sentence processing, i.e., that information about a word can be encoded in brain activity before it is perceived. While the same semantic contrast could also be decoded from EEG activity elicited by isolated words (Experiment 1), the identified neural patterns did not generalize to pre-stimulus delay period activity in sentences. Our results not only indicate that the brain processes language predictively, but also demonstrate the nature and sentence-specificity of category-level semantic predictions preactivated during sentence comprehension.
Across languages, the speech signal is characterized by a predominant modulation of the amplitude spectrum between about 4.3-5.5Hz, reflecting the production and processing of linguistic information chunks (syllables, words) every ∼200ms. Interestingly, ∼200ms is also the typical duration of eye fixations during reading. Prompted by this observation, we demonstrate that German readers sample written text at ∼5Hz. A subsequent meta-analysis with 142 studies from 14 languages replicates this result, but also shows that sampling frequencies vary across languages between 3.9Hz and 5.2Hz, and that this variation systematically depends on the complexity of the writing systems (character-based vs. alphabetic systems, orthographic transparency). Finally, we demonstrate empirically a positive correlation between speech spectrum and eye-movement sampling in low-skilled readers. Based on this convergent evidence, we propose that during reading, our brain’s linguistic processing systems imprint a preferred processing rate, i.e., the rate of spoken language production and perception, onto the oculomotor system.
Mental imagery provides an essential simulation tool for remembering the past and planning the future, with its strength affecting both cognition and mental health. Research suggests that neural activity spanning prefrontal, parietal, temporal, and visual areas supports the generation of mental images. Exactly how this network controls the strength of visual imagery remains unknown. Here, brain imaging and transcranial magnetic phosphene data show that lower resting activity and excitability levels in early visual cortex (V1-V3) predict stronger sensory imagery. Electrically decreasing visual cortex excitability using tDCS increases imagery strength, demonstrating a causative role of visual cortex excitability in controlling visual imagery. These data suggest a neurophysiological mechanism of cortical excitability involved in controlling the strength of mental images.
Changes in the efficacies of synapses are thought to be the neurobiological basis of learning and memory. The efficacy of a synapse depends on its current number of neurotransmitter receptors. Recent experiments have shown that these receptors are highly dynamic, moving back and forth between synapses on time scales of seconds and minutes. This suggests spontaneous fluctuations in synaptic efficacies and a competition of nearby synapses for available receptors. Here we propose a mathematical model of this competition of synapses for neurotransmitter receptors from a local dendritic pool. Using minimal assumptions, the model produces a fast multiplicative scaling behavior of synapses. Furthermore, the model explains a transient form of heterosynaptic plasticity and predicts that its amount is inversely related to the size of the local receptor pool. Overall, our model reveals logistical tradeoffs during the induction of synaptic plasticity due to the rapid exchange of neurotransmitter receptors between synapses.
Natural scene responses in the primary visual cortex are modulated simultaneously by attention and by contextual signals about scene statistics stored across the connectivity of the visual processing hierarchy. We hypothesize that attentional and contextual top-down signals interact in V1, in a manner that primarily benefits the representation of natural visual stimuli, rich in high-order statistical structure. Recording from two macaques engaged in a spatial attention task, we show that attention enhances the decodability of stimulus identity from population responses evoked by natural scenes but, critically, not by synthetic stimuli in which higher-order statistical regularities were eliminated. Attentional enhancement of stimulus decodability from population responses occurs in low dimensional spaces, as revealed by principal component analysis, suggesting an alignment between the attentional and the natural stimulus variance. Moreover, natural scenes produce stimulus-specific oscillatory responses in V1, whose power undergoes a global shift from low to high frequencies with attention. We argue that attention and perception share top-down pathways, which mediate hierarchical interactions optimized for natural vision.
Reducing neuronal size results in less cell membrane and therefore lower input conductance. Smaller neurons are thus more excitable as seen in their voltage responses to current injections in the soma. However, the impact of a neuron’s size and shape on its voltage responses to synaptic activation in dendrites is much less understood. Here we use analytical cable theory to predict voltage responses to distributed synaptic inputs and show that these are entirely independent of dendritic length. For a given synaptic density, a neuron’s response depends only on the average dendritic diameter and its intrinsic conductivity. These results remain true for the entire range of possible dendritic morphologies irrespective of any particular arborisation complexity. Also, spiking models result in morphology invariant numbers of action potentials that encode the percentage of active synapses. Interestingly, in contrast to spike rate, spike times do depend on dendrite morphology. In summary, a neuron’s excitability in response to synaptic inputs is not affected by total dendrite length. It rather provides a homeostatic input-output relation that specialised synapse distributions, local non-linearities in the dendrites and synaptic plasticity can modulate. Our work reveals a new fundamental principle of dendritic constancy that has consequences for the overall computation in neural circuits.
Abstract Trial-to-trial variability and spontaneous activity of cortical recordings have been suggested to reflect intrinsic noise. This view is currently challenged by mounting evidence for structure in these phenomena: Trial-to-trial variability decreases following stimulus onset and can be predicted by previous spontaneous activity. This spontaneous activity is similar in magnitude and structure to evoked activity and can predict decisions. Allof the observed neuronal properties described above can be accounted for, at an abstract computational level, by the sampling-hypothesis, according to which response variability reflects stimulus uncertainty. However, a mechanistic explanation at the level of neural circuit dynamics is still missing.
In this study, we demonstrate that all of these phenomena can be accounted for by a noise-free self-organizing recurrent neural network model (SORN). It combines spike-timing dependent plasticity (STDP) and homeostatic mechanisms in a deterministic network of excitatory and inhibitory McCulloch-Pitts neurons. The network self-organizes to spatio-temporally varying input sequences.
We find that the key properties of neural variability mentioned above develop in this model as the network learns to perform sampling-like inference. Importantly, the model shows high trial-to-trial variability although it is fully deterministic. This suggests that the trial-to-trial variability in neural recordings may not reflect intrinsic noise. Rather, it may reflect a deterministic approximation of sampling-like learning and inference. The simplicity of the model suggests that these correlates of the sampling theory are canonical properties of recurrent networks that learn with a combination of STDP and homeostatic plasticity mechanisms.
Author Summary Neural recordings seem very noisy. If the exact same stimulus is shown to an animal multiple times, the neural response will vary. In fact, the activity of a single neuron shows many features of a stochastic process. Furthermore, in the absence of a sensory stimulus, cortical spontaneous activity has a magnitude comparable to the activity observed during stimulus presentation. These findings have led to a widespread belief that neural activity is indeed very noisy. However, recent evidence indicates that individual neurons can operate very reliably and that the spontaneous activity in the brain is highly structured, suggesting that much of the noise may in fact be signal. One hypothesis regarding this putative signal is that it reflects a form of probabilistic inference through sampling. Here we show that the key features of neural variability can be accounted for in a completely deterministic network model through self-organization. As the network learns a model of its sensory inputs, the deterministic dynamics give rise to sampling-like inference. Our findings show that the notorious variability in neural recordings does not need to be seen as evidence for a noisy brain. Instead it may reflect sampling-like inference emerging from a self-organized learning process.
An important question concerning inter-areal communication in the cortex is whether these interactions are synergistic, i.e. convey information beyond what can be performed by isolated signals. Here, we dissociated cortical interactions sharing common information from those encoding complementary information during prediction error processing. To this end, we computed co-information, an information-theoretical measure that distinguishes redundant from synergistic information among brain signals. We analyzed auditory and frontal electrocorticography (ECoG) signals in three common awake marmosets and investigated to what extent event-related-potentials (ERP) and broadband (BB) dynamics exhibit redundancy and synergy for auditory prediction error signals. We observed multiple patterns of redundancy and synergy across the entire cortical hierarchy with distinct dynamics. The information conveyed by ERPs and BB signals was highly synergistic even at lower stages of the hierarchy in the auditory cortex, as well as between lower and higher areas in the frontal cortex. These results indicate that the distributed representations of prediction error signals across the cortical hierarchy can be highly synergistic.