State-dependent alteration of respiration in a rat model of Parkinson's disease.

Parkinson's disease (PD) is the second most frequent neurodegenerative disorder. Besides major deficits in motor coordination, patients may also display sensory and cognitive impairments, which are often overlooked despite being inherently part of the PD symptomatology. Amongst those symptoms, respiration, a key mechanism involved in the regulation of multiple physiological and neuronal processes, appears to be altered. Importantly, breathing patterns are highly correlated with the animal's behavioral states. This raises the question of the potential impact of behavioral state on respiration deficits in PD. To answer this question, we first characterized the respiratory parameters in a neurotoxin-induced rat model of PD (6-OHDA) across three different vigilance states: sleep, quiet waking and exploration. We noted a significantly higher respiratory frequency in 6-OHDA rats during quiet waking compared to Sham rats. A higher respiratory amplitude was also observed in 6-OHDA rats during both quiet waking and exploration. No effect of the treatment was noted during sleep. Given the relation between respiration and olfaction and the presence of olfactory deficits in PD patients, we then investigated the odor-evoked sniffing response in PD rats, using an odor habituation/cross-habituation paradigm. No substantial differences were observed in olfactory abilities between the two groups, as assessed through sniffing frequency. These results corroborate the hypothesis that respiratory impairments in 6-OHDA rats are vigilance-dependent. Our results also shed light on the importance of considering the behavioral state as an impacting factor when analyzing respiration.

Respiratory rhythm modulates membrane potential and spiking of nonolfactory neurons.

In recent years, several studies have shown a respiratory drive of the local field potential (LFP) in numerous brain areas so that the respiratory rhythm could be considered as a master clock promoting communication between distant brain locations. However, outside of the olfactory system, it remains unknown whether the respiratory rhythm could shape membrane potential (MP) oscillations. To fill this gap, we co-recorded MP and LFP activities in different nonolfactory brain areas, medial prefrontal cortex (mPFC), primary somatosensory cortex (S1), primary visual cortex (V1), and hippocampus (HPC), in urethane-anesthetized rats. Using respiratory cycle-by-cycle analysis, we observed that respiration could modulate both MP and spiking discharges in all recorded areas during episodes that we called respiration-related oscillations (RRo). Further quantifications revealed that RRo episodes were transient in most neurons (5 consecutive respiratory cycles in average). RRo development in MP was largely correlated with the presence of respiratory modulation in the LFP. By showing that the respiratory rhythm influenced brain activities deep to the MP of nonolfactory neurons, our data support the idea that respiratory rhythm could mediate long-range communication between brain areas.NEW & NOTEWORTHY In this study, we evidenced strong respiratory-driven oscillations of neuronal membrane potential and spiking discharge in various nonolfactory areas of the mammal brain. These oscillations were found in the medial prefrontal cortex, primary somatosensory cortex, primary visual cortex, and hippocampus. These findings support the idea that respiratory rhythm could be used as a common clock to set the dynamics of large-scale neuronal networks on the same slow rhythm.

A Unifying Method to Study Respiratory Sinus Arrhythmia Dynamics Implemented in a New Toolbox.

Respiratory sinus arrhythmia (RSA), the natural variation in heart rate synchronized with respiration, has been extensively studied in emotional and cognitive contexts. Various time or frequency-based methods using the cardiac signal have been proposed to analyze RSA. In this study, we present a novel approach that combines respiratory phase and heart rate to enable a more detailed analysis of RSA and its dynamics throughout the respiratory cycle. To facilitate the application of this method, we have implemented it in an open-source Python toolbox called physio This toolbox includes essential functionalities for processing electrocardiogram (ECG) and respiratory signals, while also introducing this new approach for RSA analysis. Inspired by previous research conducted by our group, this method enables a cycle-by-cycle analysis of RSA providing the possibility to correlate any respiratory feature to any RSA feature. By employing this approach, we aim to gain a more accurate understanding of the neural mechanisms associated with RSA.

Respiratory response to an odor throughout development in rats.

Odor-induced sniffing has proven to be a useful behavioral readout for assessing olfactory performance in adult rats. However, little is known about how the respiratory response changes throughout ontogeny. Thus, this study aimed at characterizing respiratory response to an odor in rats using paradigms suitable to infants, juveniles, and adults. We first analyzed the respiratory response to a neutral, novel odor. Then the value of the odor was changed either through its repeated presentation (odor habituation), or its association with a foot-shock (odor fear). In the habituation task, we found that the first presentation of the novel odor induced a clear sniffing response at all 3 ages, but the peak respiratory frequency was higher in adults than in juveniles and infants. When the odor was presented repeatedly, the sniffing response gradually faded and the younger the animal, the faster the fading of the response. In the fear conditioning task, the odor induced an increase in respiratory rate that persisted until the end of the session in adults and infants, but not in juveniles. In another group for which the odor was explicitly unpaired with the foot-shock, the respiratory response to the odor did not last as long over the session than in the paired condition at all 3 ages. Finally, we observed that shock delivery induced a similar respiratory response at the 3 investigated ages in the paired and unpaired conditions. Collectively, these data show that the respiratory response constitutes a faithful index to assess rat's olfactory abilities throughout ontogeny.

Respiratory influence on brain dynamics: the preponderant role of the nasal pathway and deep slow regime.

As a possible body signal influencing brain dynamics, respiration is fundamental for perception, cognition, and emotion. The olfactory system has recently acquired its credentials by proving to be crucial in the transmission of respiratory influence on the brain via the sensitivity to nasal airflow of its receptor cells. Here, we present recent findings evidencing respiration-related activities in the brain. Then, we review the data explaining the fact that breathing is (i) nasal and (ii) being slow and deep is crucial in its ability to stimulate the olfactory system and consequently influence the brain. In conclusion, we propose a possible scenario explaining how this optimal respiratory regime can promote changes in brain dynamics of an olfacto-limbic-respiratory circuit, providing a possibility to induce calm and relaxation by coordinating breathing regime and brain state.

The deep and slow breathing characterizing rest favors brain respiratory-drive.

A respiration-locked activity in the olfactory brain, mainly originating in the mechano-sensitivity of olfactory sensory neurons to air pressure, propagates from the olfactory bulb to the rest of the brain. Interestingly, changes in nasal airflow rate result in reorganization of olfactory bulb response. By leveraging spontaneous variations of respiratory dynamics during natural conditions, we investigated whether respiratory drive also varies with nasal airflow movements. We analyzed local field potential activity relative to respiratory signal in various brain regions during waking and sleep states. We found that respiration regime was state-specific, and that quiet waking was the only vigilance state during which all the recorded structures can be respiration-driven whatever the respiratory frequency. Using CO2-enriched air to alter respiratory regime associated to each state and a respiratory cycle based analysis, we evidenced that the large and strong brain drive observed during quiet waking was related to an optimal trade-off between depth and duration of inspiration in the respiratory pattern, characterizing this specific state. These results show for the first time that changes in respiration regime affect cortical dynamics and that the respiratory regime associated with rest is optimal for respiration to drive the brain.

New Insights from 22-kHz Ultrasonic Vocalizations to Characterize Fear Responses: Relationship with Respiration and Brain Oscillatory Dynamics.

Fear behavior depends on interactions between the medial prefrontal cortex (mPFC) and the basolateral amygdala (BLA), and the expression of fear involves synchronized activity in θ and γ oscillatory activities. In addition, freezing, the most classical measure of fear response in rodents, temporally coincides with the development of sustained 4-Hz oscillations in prefrontal-amygdala circuits. Interestingly, these oscillations were recently shown to depend on the animal's respiratory rhythm, supporting the growing body of evidence pinpointing the influence of nasal breathing on brain rhythms. During fearful states, rats also emit 22-kHz ultrasonic vocalizations (USVs) which drastically affect respiratory rhythm. However, the relationship between 22-kHz USV, respiration, and brain oscillatory activities is still unknown. Yet such information is crucial for a comprehensive understanding of how the different components of fear response collectively modulate rat's brain neural dynamics. Here, we trained male rats in an odor fear conditioning task, while recording simultaneously local field potentials (LFPs) in BLA, mPFC, and olfactory piriform cortex (PIR), together with USV calls and respiration. We show that USV calls coincide with an increase in delta and gamma power and a decrease in theta power. In addition, during USV emission in contrast to silent freezing, there is no coupling between respiratory rate and delta frequency, and the modulation of fast oscillations amplitude relative to the phase of respiration is modified. We propose that sequences of USV calls could result in a differential gating of information within the network of structures sustaining fear behavior, thus potentially modulating fear expression/memory.

Odorant features differentially modulate beta/gamma oscillatory patterns in anterior versus posterior piriform cortex.

Oscillatory activity is a prominent characteristic of the olfactory system. We previously demonstrated that beta and gamma oscillations occurrence in the olfactory bulb (OB) is modulated by the physical properties of the odorant. However, it remains unknown whether such odor-related modulation of oscillatory patterns is maintained in the piriform cortex (PC) and whether those patterns are similar between the anterior PC (aPC) and posterior PC (pPC). The present study was designed to analyze how different odorant molecular features can affect the local field potential (LFP) oscillatory signals in both the aPC and the pPC in anesthetized rats. As reported in the OB, three oscillatory patterns were observed: standard pattern (gamma + beta), gamma-only and beta-only patterns. These patterns occurred with significantly different probabilities in the two PC areas. We observed that odor identity has a strong influence on the probability of occurrence of LFP beta and gamma oscillatory activity in the aPC. Thus, some odor coding mechanisms observed in the OB are retained in the aPC. By contrast, probability of occurrence of different oscillatory patterns is homogeneous in the pPC with beta-only pattern being the most prevalent one for all the different odor families. Overall, our results confirmed the functional heterogeneity of the PC with its anterior part tightly coupled with the OB and mainly encoding odorant features whereas its posterior part activity is not correlated with odorant features but probably more involved in associative and multi-sensory encoding functions.

High beta rhythm amplitude in olfactory learning signs a well-consolidated and non-flexible behavioral state.

Beta rhythm (15-30 Hz) is a major candidate underlying long-range communication in the brain. In olfactory tasks, beta activity is strongly modulated by learning but its condition of expression and the network(s) responsible for its generation are unclear. Here we analyzed the emergence of beta activity in local field potentials recorded from olfactory, sensorimotor and limbic structures of rats performing an olfactory task. Rats performed successively simple discrimination, rule transfer, memory recall tests and contingency reversal. Beta rhythm amplitude progressively increased over learning in most recorded areas. Beta amplitude reduced to baseline when new odors were introduced, but remained high during memory recall. Intra-session analysis showed that even expert rats required several trials to reach a good performance level, with beta rhythm amplitude increasing in parallel. Notably, at the beginning of the reversal task, beta amplitude remained high while performance was low and, in all tested animals, beta amplitude decreased before rats were able to learn the new contingencies. Connectivity analysis showed that beta activity was highly coherent between all structures where it was expressed. Overall, our results suggest that beta rhythm is expressed in a highly coherent network when context learning - including both odors and reward - is consolidated and signals behavioral inflexibility.

Adaptive quantization of local field potentials for wireless implants in freely moving animals: an open-source neural recording device.

OBJECTIVE: Modern neuroscience research requires electrophysiological recording of local field potentials (LFPs) in moving animals. Wireless transmission has the advantage of removing the wires between the animal and the recording equipment but is hampered by the large number of data to be sent at a relatively high rate. APPROACH: To reduce transmission bandwidth, we propose an encoder/decoder scheme based on adaptive non-uniform quantization. Our algorithm uses the current transmitted codeword to adapt the quantization intervals to changing statistics in LFP signals. It is thus backward adaptive and does not require the sending of side information. The computational complexity is low and similar at the encoder and decoder sides. These features allow for real-time signal recovery and facilitate hardware implementation with low-cost commercial microcontrollers. MAIN RESULTS: As proof-of-concept, we developed an open-source neural recording device called NeRD. The NeRD prototype digitally transmits eight channels encoded at 10 kHz with 2 bits per sample. It occupies a volume of 2 × 2 × 2 cm3 and weighs 8 g with a small battery allowing for 2 h 40 min of autonomy. The power dissipation is 59.4 mW for a communication range of 8 m and transmission losses below 0.1%. The small weight and low power consumption offer the possibility of mounting the entire device on the head of a rodent without resorting to a separate head-stage and battery backpack. The NeRD prototype is validated in recording LFPs in freely moving rats at 2 bits per sample while maintaining an acceptable signal-to-noise ratio (>30 dB) over a range of noisy channels. SIGNIFICANCE: Adaptive quantization in neural implants allows for lower transmission bandwidths while retaining high signal fidelity and preserving fundamental frequencies in LFPs.

In vivo beta and gamma subthreshold oscillations in rat mitral cells: origin and gating by respiratory dynamics.

In mammals, olfactory bulb (OB) dynamics are paced by slow and fast oscillatory rhythms at multiple levels: local field potential, spike discharge, and/or membrane potential oscillations. Interactions between these levels have been well studied for the slow rhythm linked to animal respiration. However, less is known regarding rhythms in the fast beta (10-35 Hz) and gamma (35-100 Hz) frequency ranges, particularly at the membrane potential level. Using a combination of intracellular and extracellular recordings in the OB of freely breathing rats, we show that beta and gamma subthreshold oscillations (STOs) coexist intracellularly and are related to extracellular local field potential (LFP) oscillations in the same frequency range. However, they are differentially affected by changes in cell excitability and by odor stimulation. This leads us to suggest that beta and gamma STOs may rely on distinct mechanisms: gamma STOs would mainly depend on mitral cell intrinsic resonance, while beta STOs could be mainly driven by synaptic activity. In a second study, we find that STO occurrence and timing are constrained by the influence of the slow respiratory rhythm on mitral and tufted cells. First, respiratory-driven excitation seems to favor gamma STOs, while respiratory-driven inhibition favors beta STOs. Second, the respiratory rhythm is needed at the subthreshold level to lock gamma and beta STOs in similar phases as their LFP counterparts and to favor the correlation between STO frequency and spike discharge. Overall, this study helps us to understand how the interaction between slow and fast rhythms at all levels of OB dynamics shapes its functional output. NEW & NOTEWORTHY In the mammalian olfactory bulb of a freely breathing anesthetized rat, we show that both beta and gamma membrane potential fast oscillation ranges exist in the same mitral and tufted (M/T) cell. Importantly, our results suggest they have different origins and that their interaction with the slow subthreshold oscillation (respiratory rhythm) is a key mechanism to organize their dynamics, favoring their functional implication in olfactory bulb information processing.

Significance of sniffing pattern during the acquisition of an olfactory discrimination task.

Active sampling of olfactory environment consists of sniffing in rodents. The importance of sniffing dynamics is well established at the neuronal and behavioral levels. Patterns of sniffing have been shown to be modulated by the physicochemical properties of odorants, particularly concentration and sorption. Sniffing is also heavily impacted by higher processing related to the behavioral context, emotion and attentional demand. However, how the pattern of sniffing evolves over the course of learning of an experimental olfactory conditioning is still poorly understood. We tested this question by monitoring sniffing activity, using a whole-body plethysmograph, on rats performing a two-alternative choice odor discrimination task. We followed sniff variations at different learning stages (naïve, well-trained, expert). We found that during the acquisition of an odor discrimination task, rats acquired a global sniffing pattern, independent of the odor pair used. This pattern consists of a longer sampling duration, a higher sniffing frequency, and a larger amplitude. In parallel, subtle differences of sniffing between the two odors of a pair were also observed. This sniffing behavior was not only associated with a better and faster acquisition of the discrimination task but was also transferred to other odor sets and refined after a long-term pause so as to reduce the sampling duration and maintain a specific sniffing frequency. Our results provide additional arguments that sniffing is a complex sensorimotor act that is strongly affected by olfactory learning.

Competing Mechanisms of Gamma and Beta Oscillations in the Olfactory Bulb Based on Multimodal Inhibition of Mitral Cells Over a Respiratory Cycle.

Gamma (∼40-90 Hz) and beta (∼15-40 Hz) oscillations and their associated neuronal assemblies are key features of neuronal sensory processing. However, the mechanisms involved in either their interaction and/or the switch between these different regimes in most sensory systems remain misunderstood. Based on in vivo recordings and biophysical modeling of the mammalian olfactory bulb (OB), we propose a general scheme where OB internal dynamics can sustain two distinct dynamic states, each dominated by either a gamma or a beta regime. The occurrence of each regime depends on the excitability level of granule cells, the main OB interneurons. Using this model framework, we demonstrate how the balance between sensory and centrifugal input can control the switch between the two oscillatory dynamic states. In parallel, we experimentally observed that sensory and centrifugal inputs to the rat OB could both be modulated by the respiration of the animal (2-12 Hz) and each one phase shifted with the other. Implementing this phase shift in our model resulted in the appearance of the alternation between gamma and beta rhythms within a single respiratory cycle, as in our experimental results under urethane anesthesia. Our theoretical framework can also account for the oscillatory frequency response, depending on the odor intensity, the odor valence, and the animal sniffing strategy observed under various conditions including animal freely-moving. Importantly, the results of the present model can form a basis to understand how fast rhythms could be controlled by the slower sensory and centrifugal modulations linked to the respiration. Visual Abstract: See Abstract.

Two distinct olfactory bulb sublaminar networks involved in gamma and beta oscillation generation: a CSD study in the anesthetized rat.

A prominent feature of olfactory bulb (OB) dynamics is the expression of characteristic local field potential (LFP) rhythms, including a slow respiration-related rhythm and two fast alternating oscillatory rhythms, beta (15-30 Hz) and gamma (40-90 Hz). All of these rhythms are implicated in olfactory coding. Fast oscillatory rhythms are known to involve the mitral-granule cell loop. Although the underlying mechanisms of gamma oscillation have been studied, the origin of beta oscillation remains poorly understood. Whether these two different rhythms share the same underlying mechanism is unknown. This study uses a quantitative and detailed current-source density (CSD) analysis combined with multi-unit activity (MUA) recordings to shed light on this question in freely breathing anesthetized rats. In particular, we show that gamma oscillation generation involves mainly the upper half of the external plexiform layer (EPL) and superficial areas of granule cell layer (GRL). In contrast, the generation of beta oscillation involves the lower part of the EPL and deep granule cells. This differential involvement of sublaminar networks is neither dependent on odor quality nor on the precise frequency of the fast oscillation under study. Overall, this study demonstrates a functional sublaminar organization of the rat OB, which is supported by previous anatomical findings.

Sniff adjustment in an odor discrimination task in the rat: analytical or synthetic strategy?

A growing body of evidence suggests that sniffing is not only the mode of delivery for odorant molecules but also contributes to olfactory perception. However, the precise role of sniffing variations remains unknown. The zonation hypothesis suggests that animals use sniffing variations to optimize the deposition of odorant molecules on the most receptive areas of the olfactory epithelium (OE). Sniffing would thus depend on the physicochemical properties of odorants, particularly their sorption. Rojas-Líbano and Kay (2012) tested this hypothesis and showed that rats used different sniff strategies when they had to target a high-sorption (HS) molecule or a low-sorption (LS) molecule in a binary mixture. Which sniffing strategy is used by rats when they are confronted to discrimination between two similarly sorbent odorants remains unanswered. Particularly, is sniffing adjusted independently for each odorant according to its sorption properties (analytical processing), or is sniffing adjusted based on the pairing context (synthetic processing)? We tested these hypotheses on rats performing a two-alternative choice discrimination of odorants with similar sorption properties. We recorded sniffing in a non-invasive manner using whole-body plethysmography during the behavioral task. We found that sniffing variations were not only a matter of odorant sorption properties and that the same odorant was sniffed differently depending on the odor pair in which it was presented. These results suggest that rather than being adjusted analytically, sniffing is instead adjusted synthetically and depends on the pair of odorants presented during the discrimination task. Our results show that sniffing is a specific sensorimotor act that depends on complex synthetic processes.

The relationship between respiration-related membrane potential slow oscillations and discharge patterns in mitral/tufted cells: what are the rules?

BACKGROUND: A slow respiration-related rhythm strongly shapes the activity of the olfactory bulb. This rhythm appears as a slow oscillation that is detectable in the membrane potential, the respiration-related spike discharge of the mitral/tufted cells and the bulbar local field potential. Here, we investigated the rules that govern the manifestation of membrane potential slow oscillations (MPSOs) and respiration-related discharge activities under various afferent input conditions and cellular excitability states. METHODOLOGY AND PRINCIPAL FINDINGS: We recorded the intracellular membrane potential signals in the mitral/tufted cells of freely breathing anesthetized rats. We first demonstrated the existence of multiple types of MPSOs, which were influenced by odor stimulation and discharge activity patterns. Complementary studies using changes in the intracellular excitability state and a computational model of the mitral cell demonstrated that slow oscillations in the mitral/tufted cell membrane potential were also modulated by the intracellular excitability state, whereas the respiration-related spike activity primarily reflected the afferent input. Based on our data regarding MPSOs and spike patterns, we found that cells exhibiting an unsynchronized discharge pattern never exhibited an MPSO. In contrast, cells with a respiration-synchronized discharge pattern always exhibited an MPSO. In addition, we demonstrated that the association between spike patterns and MPSO types appeared complex. CONCLUSION: We propose that both the intracellular excitability state and input strength underlie specific MPSOs, which, in turn, constrain the types of spike patterns exhibited.

Faster, deeper, better: the impact of sniffing modulation on bulbar olfactory processing.

A key feature of mammalian olfactory perception is that sensory input is intimately related to respiration. Different authors have considered respiratory dynamics not only as a simple vector for odor molecules but also as an integral part of olfactory perception. Thus, rats adapt their sniffing strategy, both in frequency and flow rate, when performing odor-related tasks. The question of how frequency and flow rate jointly impact the spatio-temporal representation of odor in the olfactory bulb (OB) has not yet been answered. In the present paper, we addressed this question using a simulated nasal airflow protocol on anesthetized rats combined with voltage-sensitive dye imaging (VSDi) of odor-evoked OB glomerular maps. Glomerular responses displayed a tonic component during odor stimulation with a superimposed phasic component phase-locked to the sampling pattern. We showed that a high sniffing frequency (10 Hz) retained the ability to shape OB activity and that the tonic and phasic components of the VSDi responses were dependent on flow rate and inspiration volume, respectively. Both sniffing parameters jointly affected OB responses to odor such that the reduced activity level induced by a frequency increase was compensated by an increased flow rate.

Individual and synergistic effects of sniffing frequency and flow rate on olfactory bulb activity.

Is faster or stronger sniffing important for the olfactory system? Odorant molecules are captured by sniffing. The features of sniffing constrain both the temporality and intensity of the input to the olfactory structures. In this context, it is clear that variations in both the sniff frequency and flow rate have a major impact on the activation of olfactory structures. However, the question of how frequency and flow rate individually or synergistically impact bulbar output has not been answered. We have addressed this question using multiple experimental approaches. In double-tracheotomized, anesthetized rats, we recorded both the bulbar local field potential (LFP) and mitral/tufted cells' activities when the sampling flow rate and frequency were controlled independently. We found that a tradeoff between the sampling frequency and the flow rate could maintain olfactory bulb sampling-related rhythmicity and that only an increase in flow rate could induce a faster, odor-evoked response. LFP and sniffing were recorded in awake rats. We found that sampling-related rhythmicity was maintained during high-frequency sniffing. Furthermore, we observed that the covariation between the frequency and flow rate, which was necessary for the tradeoff seen in the anesthetized preparations, also occurred in awake animals. Our study shows that the sampling frequency and flow rate can act either independently or synergistically on bulbar output to shape the neuronal message. The system likely takes advantage of this flexibility to adapt sniffing strategies to animal behavior. Our study provides additional support for the idea that sniffing and olfaction function in an integrated manner.

Stability of fast oscillations in the mammalian olfactory bulb: experiments and modeling.

In the rat olfactory bulb (OB), fast oscillations of the local field potential (LFP) are observed during the respiratory cycle. Gamma-range oscillations (40-90 Hz) occur at the end of inspiration, followed by beta-range oscillations (15-30 Hz) during exhalation. These oscillations are highly stereotypical, and their frequencies are stable under various conditions. In this study, we investigate the effect of stimulus intensity on activity in the OB. Using a double-cannulation protocol, we showed that although the frequency of the LFP oscillation does depend on the respiratory cycle phase, it is relatively independent of the intensity of odorant stimulation. In contrast, we found that the individual firing rate of mitral OB cells dramatically changed with the intensity of the stimulation. This suggests that OB fast oscillation parameters, particularly frequency, are fully determined by intrinsic OB network parameters. To test this hypothesis, we explored a model of the OB where fast oscillations are generated by the interplay between excitatory mitral/tufted cells and inhibitory granule cells with graded inhibition. We found that our model has two distinct activity regimes depending on the amount of noise. In a low-noise regime, the model displays oscillation in the beta range with a stable frequency across a wide range of excitatory inputs. In a high-noise regime, the model displays oscillatory dynamics with irregular cell discharges and fast oscillations, similar to what is observed during gamma oscillations but without stability of the oscillation frequency with respect to the network external input. Simulations of the full model and theoretical studies of the network's linear response show that the characteristics of the low-noise regime are induced by non-linearities in the model, notably, the saturation of graded inhibition. Finally, we discuss how this model can account for the experimentally observed stability of the oscillatory regimes.