Spontaneous Necker-cube reversals may not be that spontaneous.

Introduction: During observation of the ambiguous Necker cube, our perception suddenly reverses between two about equally possible 3D interpretations. During passive observation, perceptual reversals seem to be sudden and spontaneous. A number of theoretical approaches postulate destabilization of neural representations as a pre-condition for reversals of ambiguous figures. In the current study, we focused on possible Electroencephalogram (EEG) correlates of perceptual destabilization, that may allow prediction of an upcoming perceptual reversal. Methods: We presented ambiguous Necker cube stimuli in an onset-paradigm and investigated the neural processes underlying endogenous reversals as compared to perceptual stability across two consecutive stimulus presentations. In a separate experimental condition, disambiguated cube variants were alternated randomly, to exogenously induce perceptual reversals. We compared the EEG immediately before and during endogenous Necker cube reversals with corresponding time windows during exogenously induced perceptual reversals of disambiguated cube variants. Results: For the ambiguous Necker cube stimuli, we found the earliest differences in the EEG between reversal trials and stability trials already 1 s before a reversal occurred, at bilateral parietal electrodes. The traces remained similar until approximately 1100 ms before a perceived reversal, became maximally different at around 890 ms (p = 7.59 × 10-6, Cohen's d = 1.35) and remained different until shortly before offset of the stimulus preceding the reversal. No such patterns were found in the case of disambiguated cube variants. Discussion: The identified EEG effects may reflect destabilized states of neural representations, related to destabilized perceptual states preceding a perceptual reversal. They further indicate that spontaneous Necker cube reversals are most probably not as spontaneous as generally thought. Rather, the destabilization may occur over a longer time scale, at least 1 s before a reversal event, despite the reversal event as such being perceived as spontaneous by the viewer.

Feed-forward and noise-tolerant detection of feature homogeneity in spiking networks with a latency code.

In studies of the visual system as well as in computer vision, the focus is often on contrast edges. However, the primate visual system contains a large number of cells that are insensitive to spatial contrast and, instead, respond to uniform homogeneous illumination of their visual field. The purpose of this information remains unclear. Here, we propose a mechanism that detects feature homogeneity in visual areas, based on latency coding and spike time coincidence, in a purely feed-forward and therefore rapid manner. We demonstrate how homogeneity information can interact with information on contrast edges to potentially support rapid image segmentation. Furthermore, we analyze how neuronal crosstalk (noise) affects the mechanism's performance. We show that the detrimental effects of crosstalk can be partly mitigated through delayed feed-forward inhibition that shapes bi-phasic post-synaptic events. The delay of the feed-forward inhibition allows effectively controlling the size of the temporal integration window and, thereby, the coincidence threshold. The proposed model is based on single-spike latency codes in a purely feed-forward architecture that supports low-latency processing, making it an attractive scheme of computation in spiking neuronal networks where rapid responses and low spike counts are desired.

Facilitating the propagation of spiking activity in feedforward networks by including feedback.

Transient oscillations in network activity upon sensory stimulation have been reported in different sensory areas of the brain. These evoked oscillations are the generic response of networks of excitatory and inhibitory neurons (EI-networks) to a transient external input. Recently, it has been shown that this resonance property of EI-networks can be exploited for communication in modular neuronal networks by enabling the transmission of sequences of synchronous spike volleys ('pulse packets'), despite the sparse and weak connectivity between the modules. The condition for successful transmission is that the pulse packet (PP) intervals match the period of the modules' resonance frequency. Hence, the mechanism was termed communication through resonance (CTR). This mechanism has three severe constraints, though. First, it needs periodic trains of PPs, whereas single PPs fail to propagate. Second, the inter-PP interval needs to match the network resonance. Third, transmission is very slow, because in each module, the network resonance needs to build up over multiple oscillation cycles. Here, we show that, by adding appropriate feedback connections to the network, the CTR mechanism can be improved and the aforementioned constraints relaxed. Specifically, we show that adding feedback connections between two upstream modules, called the resonance pair, in an otherwise feedforward modular network can support successful propagation of a single PP throughout the entire network. The key condition for successful transmission is that the sum of the forward and backward delays in the resonance pair matches the resonance frequency of the network modules. The transmission is much faster, by more than a factor of two, than in the original CTR mechanism. Moreover, it distinctly lowers the threshold for successful communication by synchronous spiking in modular networks of weakly coupled networks. Thus, our results suggest a new functional role of bidirectional connectivity for the communication in cortical area networks.

Differential Coding Strategies in Glutamatergic and GABAergic Neurons in the Medial Cerebellar Nucleus.

The cerebellum drives motor coordination and sequencing of actions at the millisecond timescale through adaptive control of cerebellar nuclear output. Cerebellar nuclei integrate high-frequency information from both the cerebellar cortex and the two main excitatory inputs of the cerebellum: the mossy fibers and the climbing fiber collaterals. However, how nuclear cells process rate and timing of inputs carried by these inputs is still debated. Here, we investigate the influence of the cerebellar cortical output, the Purkinje cells, on identified cerebellar nuclei neurons in vivo in male mice. Using transgenic mice expressing Channelrhodopsin2 specifically in Purkinje cells and tetrode recordings in the medial nucleus, we identified two main groups of neurons based on the waveform of their action potentials. These two groups of neurons coincide with glutamatergic and GABAergic neurons identified by optotagging after Chrimson expression in VGLUT2-cre and GAD-cre mice, respectively. The glutamatergic-like neurons fire at high rate and respond to both rate and timing of Purkinje cell population inputs, whereas GABAergic-like neurons only respond to the mean population firing rate of Purkinje cells at high frequencies. Moreover, synchronous activation of Purkinje cells can entrain the glutamatergic-like, but not the GABAergic-like, cells over a wide range of frequencies. Our results suggest that the downstream effect of synchronous and rhythmic Purkinje cell discharges depends on the type of cerebellar nuclei neurons targeted.SIGNIFICANCE STATEMENT Motor coordination and skilled movements are driven by the permanent discharge of neurons from the cerebellar nuclei that communicate cerebellar computation to other brain areas. Here, we set out to study how specific subtypes of cerebellar nuclear neurons of the medial nucleus are controlled by Purkinje cells, the sole output of the cerebellar cortex. We could isolate different subtypes of nuclear cell that differentially encode Purkinje cell inhibition. Purkinje cell stimulation entrains glutamatergic projection cells at their firing frequency, whereas GABAergic neurons are only inhibited. These differential coding strategies may favor temporal precision of cerebellar excitatory outputs associated with specific features of movement control while setting the global level of cerebellar activity through inhibition via rate coding mechanisms.

From space to time: Spatial inhomogeneities lead to the emergence of spatiotemporal sequences in spiking neuronal networks.

Spatio-temporal sequences of neuronal activity are observed in many brain regions in a variety of tasks and are thought to form the basis of meaningful behavior. However, mechanisms by which a neuronal network can generate spatio-temporal activity sequences have remained obscure. Existing models are biologically untenable because they either require manual embedding of a feedforward network within a random network or supervised learning to train the connectivity of a network to generate sequences. Here, we propose a biologically plausible, generative rule to create spatio-temporal activity sequences in a network of spiking neurons with distance-dependent connectivity. We show that the emergence of spatio-temporal activity sequences requires: (1) individual neurons preferentially project a small fraction of their axons in a specific direction, and (2) the preferential projection direction of neighboring neurons is similar. Thus, an anisotropic but correlated connectivity of neuron groups suffices to generate spatio-temporal activity sequences in an otherwise random neuronal network model.

Direct pathway neurons in mouse dorsolateral striatum in vivo receive stronger synaptic input than indirect pathway neurons.

Striatal projection neurons, the medium spiny neurons (MSNs), play a crucial role in various motor and cognitive functions. MSNs express either D1- or D2-type dopamine receptors and initiate the direct-pathway (dMSNs) or indirect pathways (iMSNs) of the basal ganglia, respectively. dMSNs have been shown to receive more inhibition than iMSNs from intrastriatal sources. Based on these findings, computational modeling of the striatal network has predicted that under healthy conditions dMSNs should receive more total input than iMSNs. To test this prediction, we analyzed in vivo whole cell recordings from dMSNs and iMSNs in healthy and dopamine-depleted (6OHDA) anaesthetized mice. By comparing their membrane potential fluctuations, we found that dMSNs exhibited considerably larger membrane potential fluctuations over a wide frequency range. Furthermore, by comparing the spike-triggered average membrane potentials, we found that dMSNs depolarized toward the spike threshold significantly faster than iMSNs did. Together, these findings (in particular the STA analysis) corroborate the theoretical prediction that direct-pathway MSNs receive stronger total input than indirect-pathway neurons. Finally, we found that dopamine-depleted mice exhibited no difference between the membrane potential fluctuations of dMSNs and iMSNs. These data provide new insights into the question of how the lack of dopamine may lead to behavioral deficits associated with Parkinson's disease.NEW & NOTEWORTHY The direct and indirect pathways of the basal ganglia originate from the D1- and D2-type dopamine receptor expressing medium spiny neurons (dMSNs and iMSNs). Theoretical results have predicted that dMSNs should receive stronger synaptic input than iMSNs. Using in vivo intracellular membrane potential data, we provide evidence that dMSNs indeed receive stronger input than iMSNs, as has been predicted by the computational model.

Portraits of communication in neuronal networks.

The brain is organized as a network of highly specialized networks of spiking neurons. To exploit such a modular architecture for computation, the brain has to be able to regulate the flow of spiking activity between these specialized networks. In this Opinion article, we review various prominent mechanisms that may underlie communication between neuronal networks. We show that communication between neuronal networks can be understood as trajectories in a two-dimensional state space, spanned by the properties of the input. Thus, we propose a common framework to understand neuronal communication mediated by seemingly different mechanisms. We also suggest that the nesting of slow (for example, alpha-band and theta-band) oscillations and fast (gamma-band) oscillations can serve as an important control mechanism that allows or prevents spiking signals to be routed between specific networks. We argue that slow oscillations can modulate the time required to establish network resonance or entrainment and, thereby, regulate communication between neuronal networks.

Real-life speech production and perception have a shared premotor-cortical substrate.

Motor-cognitive accounts assume that the articulatory cortex is involved in language comprehension, but previous studies may have observed such an involvement as an artefact of experimental procedures. Here, we employed electrocorticography (ECoG) during natural, non-experimental behavior combined with electrocortical stimulation mapping to study the neural basis of real-life human verbal communication. We took advantage of ECoG's ability to capture high-gamma activity (70-350 Hz) as a spatially and temporally precise index of cortical activation during unconstrained, naturalistic speech production and perception conditions. Our findings show that an electrostimulation-defined mouth motor region located in the superior ventral premotor cortex is consistently activated during both conditions. This region became active early relative to the onset of speech production and was recruited during speech perception regardless of acoustic background noise. Our study thus pinpoints a shared ventral premotor substrate for real-life speech production and perception with its basic properties.

Hyperactivity of Anterior Cingulate Cortex Areas 24a/24b Drives Chronic Pain-Induced Anxiodepressive-like Consequences.

Pain associates both sensory and emotional aversive components, and often leads to anxiety and depression when it becomes chronic. Here, we characterized, in a mouse model, the long-term development of these sensory and aversive components as well as anxiodepressive-like consequences of neuropathic pain and determined their electrophysiological impact on the anterior cingulate cortex (ACC, cortical areas 24a/24b). We show that these symptoms of neuropathic pain evolve and recover in different time courses following nerve injury in male mice. In vivo electrophysiological recordings evidence an increased firing rate and bursting activity within the ACC when anxiodepressive-like consequences developed, and this hyperactivity persists beyond the period of mechanical hypersensitivity. Whole-cell patch-clamp recordings also support ACC hyperactivity, as shown by increased excitatory postsynaptic transmission and contribution of NMDA receptors. Optogenetic inhibition of the ACC hyperactivity was sufficient to alleviate the aversive and anxiodepressive-like consequences of neuropathic pain, indicating that these consequences are underpinned by ACC hyperactivity.SIGNIFICANCE STATEMENT Chronic pain is frequently comorbid with mood disorders, such as anxiety and depression. It has been shown that it is possible to model this comorbidity in animal models by taking into consideration the time factor. In this study, we aimed at determining the dynamic of different components and consequences of chronic pain, and correlated them with electrophysiological alterations. By combining electrophysiological, optogenetic, and behavioral analyses in a mouse model of neuropathic pain, we show that the mechanical hypersensitivity, ongoing pain, anxiodepressive consequences, and their recoveries do not necessarily exhibit temporal synchrony during chronic pain processing, and that the hyperactivity of the anterior cingulate cortex is essential for driving the emotional impact of neuropathic pain.

Sensorimotor Processing in the Basal Ganglia Leads to Transient Beta Oscillations during Behavior.

Brief epochs of beta oscillations have been implicated in sensorimotor control in the basal ganglia of task-performing healthy animals. However, which neural processes underlie their generation and how they are affected by sensorimotor processing remains unclear. To determine the mechanisms underlying transient beta oscillations in the LFP, we combined computational modeling of the subthalamo-pallidal network for the generation of beta oscillations with realistic stimulation patterns derived from single-unit data recorded from different basal ganglia subregions in rats performing a cued choice task. In the recordings, we found distinct firing patterns in the striatum, globus pallidus, and subthalamic nucleus related to sensory and motor events during the behavioral task. Using these firing patterns to generate realistic inputs to our network model led to transient beta oscillations with the same time course as the rat LFP data. In addition, our model can account for further nonintuitive aspects of beta modulation, including beta phase resets after sensory cues and correlations with reaction time. Overall, our model can explain how the combination of temporally regulated sensory responses of the subthalamic nucleus, ramping activity of the subthalamic nucleus, and movement-related activity of the globus pallidus leads to transient beta oscillations during behavior.SIGNIFICANCE STATEMENT Transient beta oscillations emerge in the normal functioning cortico-basal ganglia loop during behavior. Here, we used a unique approach connecting a computational model closely with experimental data. In this way, we achieved a simulation environment for our model that mimics natural input patterns in awake, behaving animals. We demonstrate that a computational model for beta oscillations in Parkinson's disease (PD) can also account for complex patterns of transient beta oscillations in healthy animals. Therefore, we propose that transient beta oscillations in healthy animals share the same mechanism with pathological beta oscillations in PD. This important result connects functional and pathological roles of beta oscillations in the basal ganglia.

Activity Dynamics and Signal Representation in a Striatal Network Model with Distance-Dependent Connectivity.

The striatum is the main input nucleus of the basal ganglia. Characterizing striatal activity dynamics is crucial to understanding mechanisms underlying action selection, initiation, and execution. Here, we studied the effects of spatial network connectivity on the spatiotemporal structure of striatal activity. We show that a striatal network with nonmonotonically changing distance-dependent connectivity (according to a gamma distribution) can exhibit a wide repertoire of spatiotemporal dynamics, ranging from spatially homogeneous, asynchronous-irregular (AI) activity to a state with stable, spatially localized activity bumps, as in "winner-take-all" (WTA) dynamics. Among these regimes, the unstable activity bumps [transition activity (TA)] regime closely resembles the experimentally observed spatiotemporal activity dynamics and neuronal assemblies in the striatum. In contrast, striatal networks with monotonically decreasing distance-dependent connectivity (in a Gaussian fashion) can exhibit only an AI state. Thus, given the observation of spatially compact neuronal clusters in the striatum, our model suggests that recurrent connectivity among striatal projection neurons should vary nonmonotonically. In brain disorders such as Parkinson's disease, increased cortical inputs and high striatal firing rates are associated with a reduction in stimulus sensitivity. Consistent with this, our model suggests that strong cortical inputs drive the striatum to a WTA state, leading to low stimulus sensitivity and high variability. In contrast, the AI and TA states show high stimulus sensitivity and reliability. Thus, based on these results, we propose that in a healthy state the striatum operates in a AI/TA state and that lack of dopamine pushes it into a WTA state.

Electrical Stimulation of the Human Cerebral Cortex by Extracranial Muscle Activity: Effect Quantification With Intracranial EEG and FEM Simulations.

OBJECTIVE: Electric fields (EF) of approx. 0.2 V/m have been shown to be sufficiently strong to both modulate neuronal activity in the cerebral cortex and have measurable effects on cognitive performance. We hypothesized that the EF caused by the electrical activity of extracranial muscles during natural chewing may reach similar strength in the cerebral cortex and hence might act as an endogenous modality of brain stimulation. Here, we present first steps toward validating this hypothesis. METHODS: Using a realistic volume conductor head model of an epilepsy patient having undergone intracranial electrode placement and utilizing simultaneous intracranial and extracranial electrical recordings during chewing, we derive predictions about the chewing-related cortical EF strength to be expected in healthy individuals. RESULTS: We find that in the region of the temporal poles, the expected EF strength may reach amplitudes in the order of 0.1-1 V/m. CONCLUSION: The cortical EF caused by natural chewing could be large enough to modulate ongoing neural activity in the cerebral cortex and influence cognitive performance. SIGNIFICANCE: The present study lends first support for the assumption that extracranial muscle activity might represent an endogenous source of electrical brain stimulation. This offers a new potential explanation for the puzzling effects of gum chewing on cognition, which have been repeatedly reported in the literature.

Push-Pull Receptive Field Organization and Synaptic Depression: Mechanisms for Reliably Encoding Naturalistic Stimuli in V1.

Neurons in the primary visual cortex are known for responding vigorously but with high variability to classical stimuli such as drifting bars or gratings. By contrast, natural scenes are encoded more efficiently by sparse and temporal precise spiking responses. We used a conductance-based model of the visual system in higher mammals to investigate how two specific features of the thalamo-cortical pathway, namely push-pull receptive field organization and fast synaptic depression, can contribute to this contextual reshaping of V1 responses. By comparing cortical dynamics evoked respectively by natural vs. artificial stimuli in a comprehensive parametric space analysis, we demonstrate that the reliability and sparseness of the spiking responses during natural vision is not a mere consequence of the increased bandwidth in the sensory input spectrum. Rather, it results from the combined impacts of fast synaptic depression and push-pull inhibition, the later acting for natural scenes as a form of "effective" feed-forward inhibition as demonstrated in other sensory systems. Thus, the combination of feedforward-like inhibition with fast thalamo-cortical synaptic depression by simple cells receiving a direct structured input from thalamus composes a generic computational mechanism for generating a sparse and reliable encoding of natural sensory events.

Dynamical state of the network determines the efficacy of single neuron properties in shaping the network activity.

Spike patterns are among the most common electrophysiological descriptors of neuron types. Surprisingly, it is not clear how the diversity in firing patterns of the neurons in a network affects its activity dynamics. Here, we introduce the state-dependent stochastic bursting neuron model allowing for a change in its firing patterns independent of changes in its input-output firing rate relationship. Using this model, we show that the effect of single neuron spiking on the network dynamics is contingent on the network activity state. While spike bursting can both generate and disrupt oscillations, these patterns are ineffective in large regions of the network state space in changing the network activity qualitatively. Finally, we show that when single-neuron properties are made dependent on the population activity, a hysteresis like dynamics emerges. This novel phenomenon has important implications for determining the network response to time-varying inputs and for the network sensitivity at different operating points.

Predominance of Movement Speed Over Direction in Neuronal Population Signals of Motor Cortex: Intracranial EEG Data and A Simple Explanatory Model.

How neuronal activity of motor cortex is related to movement is a central topic in motor neuroscience. Motor-cortical single neurons are more closely related to hand movement velocity than speed, that is, the magnitude of the (directional) velocity vector. Recently, there is also increasing interest in the representation of movement parameters in neuronal population activity, such as reflected in the intracranial EEG (iEEG). We show that in iEEG, contrasting to what has been previously found on the single neuron level, speed predominates over velocity. The predominant speed representation was present in nearly all iEEG signal features, up to the 600-1000 Hz range. Using a model of motor-cortical signals arising from neuronal populations with realistic single neuron tuning properties, we show how this reversal can be understood as a consequence of increasing population size. Our findings demonstrate that the information profile in large population signals may systematically differ from the single neuron level, a principle that may be helpful in the interpretation of neuronal population signals in general, including, for example, EEG and functional magnetic resonance imaging. Taking advantage of the robust speed population signal may help in developing brain-machine interfaces exploiting population signals.

Recovery of Dynamics and Function in Spiking Neural Networks with Closed-Loop Control.

There is a growing interest in developing novel brain stimulation methods to control disease-related aberrant neural activity and to address basic neuroscience questions. Conventional methods for manipulating brain activity rely on open-loop approaches that usually lead to excessive stimulation and, crucially, do not restore the original computations performed by the network. Thus, they are often accompanied by undesired side-effects. Here, we introduce delayed feedback control (DFC), a conceptually simple but effective method, to control pathological oscillations in spiking neural networks (SNNs). Using mathematical analysis and numerical simulations we show that DFC can restore a wide range of aberrant network dynamics either by suppressing or enhancing synchronous irregular activity. Importantly, DFC, besides steering the system back to a healthy state, also recovers the computations performed by the underlying network. Finally, using our theory we identify the role of single neuron and synapse properties in determining the stability of the closed-loop system.

Effect of edge pruning on structural controllability and observability of complex networks.

Controllability and observability of complex systems are vital concepts in many fields of science. The network structure of the system plays a crucial role in determining its controllability and observability. Because most naturally occurring complex systems show dynamic changes in their network connectivity, it is important to understand how perturbations in the connectivity affect the controllability of the system. To this end, we studied the control structure of different types of artificial, social and biological neuronal networks (BNN) as their connections were progressively pruned using four different pruning strategies. We show that the BNNs are more similar to scale-free networks than to small-world networks, when comparing the robustness of their control structure to structural perturbations. We introduce a new graph descriptor, 'the cardinality curve', to quantify the robustness of the control structure of a network to progressive edge pruning. Knowing the susceptibility of control structures to different pruning methods could help design strategies to destroy the control structures of dangerous networks such as epidemic networks. On the other hand, it could help make useful networks more resistant to edge attacks.

Invasive brain-machine interfaces: a survey of paralyzed patients' attitudes, knowledge and methods of information retrieval.

OBJECTIVE: Brain-machine interfaces (BMI) are an emerging therapeutic option that can allow paralyzed patients to gain control over assistive technology devices (ATDs). BMI approaches can be broadly classified into invasive (based on intracranially implanted electrodes) and noninvasive (based on skin electrodes or extracorporeal sensors). Invasive BMIs have a favorable signal-to-noise ratio, and thus allow for the extraction of more information than noninvasive BMIs, but they are also associated with the risks related to neurosurgical device implantation. Current noninvasive BMI approaches are typically concerned, among other issues, with long setup times and/or intensive training. Recent studies have investigated the attitudes of paralyzed patients eligible for BMIs, particularly patients affected by amyotrophic lateral sclerosis (ALS). These studies indicate that paralyzed patients are indeed interested in BMIs. Little is known, however, about the degree of knowledge among paralyzed patients concerning BMI approaches or about how patients retrieve information on ATDs. Furthermore, it is not yet clear if paralyzed patients would accept intracranial implantation of BMI electrodes with the premise of decoding improvements, and what the attitudes of a broader range of patients with diseases such as stroke or spinal cord injury are towards this new kind of treatment. APPROACH: Using a questionnaire, we surveyed 131 paralyzed patients for their opinions on invasive BMIs and their attitude toward invasive BMI treatment options. MAIN RESULTS: The majority of the patients knew about and had a positive attitude toward invasive BMI approaches. The group of ALS patients was especially open to the concept of BMIs. The acceptance of invasive BMI technology depended on the improvements expected from the technology. Furthermore, the survey revealed that for paralyzed patients, the Internet is an important source of information on ATDs. SIGNIFICANCE: Websites tailored to prospective BMI users should be further developed to provide reliable information to patients, and also to help to link prospective BMI users with researchers involved in the development of BMI technology.

Role of input correlations in shaping the variability and noise correlations of evoked activity in the neocortex.

Recent analysis of evoked activity recorded across different brain regions and tasks revealed a marked decrease in noise correlations and trial-by-trial variability. Given the importance of correlations and variability for information processing within the rate coding paradigm, several mechanisms have been proposed to explain the reduction in these quantities despite an increase in firing rates. These models suggest that anatomical clusters and/or tightly balanced excitation-inhibition can generate intrinsic network dynamics that may exhibit a reduction in noise correlations and trial-by-trial variability when perturbed by an external input. Such mechanisms based on the recurrent feedback crucially ignore the contribution of feedforward input to the statistics of the evoked activity. Therefore, we investigated how statistical properties of the feedforward input shape the statistics of the evoked activity. Specifically, we focused on the effect of input correlation structure on the noise correlations and trial-by-trial variability. We show that the ability of neurons to transfer the input firing rate, correlation, and variability to the output depends on the correlations within the presynaptic pool of a neuron, and that an input with even weak within-correlations can be sufficient to reduce noise correlations and trial-by-trial variability, without requiring any specific recurrent connectivity structure. In general, depending on the ongoing activity state, feedforward input could either increase or decrease noise correlation and trial-by-trial variability. Thus, we propose that evoked activity statistics are jointly determined by the feedforward and feedback inputs.