Role of frequency mismatch in neuronal communication through coherence

Abstract

Neuronal gamma oscillations have been described in local field potentials of different brain regions of multiple species. Gamma oscillations are thought to reflect rhythmic synaptic activity organized by inhibitory interneurons. While several aspects of gamma rhythmogenesis are relatively well understood, we have much less solid evidence about how gamma oscillations contribute to information processing in neuronal circuits. One popular hypothesis states that a flexible routing of information between distant populations occurs via the control of the phase or coherence between their respective oscillations. Here, we investigate how a mismatch between the frequencies of gamma oscillations from two populations affects their interaction. In particular, we explore a biophysical model of the reciprocal interaction between two cortical areas displaying gamma oscillations at different frequencies, and quantify their phase coherence and communication efficiency. We observed that a moderate excitatory coupling between the two areas leads to a decrease in their frequency detuning, up to ∼6 Hz, with no frequency locking arising between the gamma peaks. Importantly, for similar gamma peak frequencies a zero phase difference emerges for both LFP and MUA despite small axonal delays. For increasing frequency detunings we found a significant decrease in the phase coherence (at non-zero phase lag) between the MUAs but not the LFPs of the two areas. Such difference between LFPs and MUAs behavior is due to the misalignment between the arrival of afferent synaptic currents and the local excitability windows. To test the efficiency of communication we evaluated the success of transferring rate-modulations between the two areas. Our results indicate that once two populations lock their peak frequencies, an optimal phase relation for communication appears. However, the sensitivity of locking to frequency mismatch suggests that only a precise and active control of gamma frequency could enable the selection of communication channels and their directionality.

Neuronal gamma-band synchronization regulated by instantaneous modulations of the oscillation frequency

“…The observed role of detuning is in agreement with a previous study in the rat hippocampus (Akam et al, 2012), in which optogenetic entrainment strength and phase of gamma rhythms were dependent on the frequency-detuning. The results also agree with theoretical conceptions on oscillatory interactions (Ermentrout and Kopell, 1984; Hoppensteadt and Izhikevich, 1998; Sancristóbal et al, 2014; Tiesinga and Sejnowski, 2010). We suggest that small detuning values (mainly <Δ10Hz) reported in the present study and much larger shifts in the gamma frequency-range (25-50Hz to 65-120Hz) reported in the rat hippocampus and cortex (Colgin et al, 2009) represent different but complementary mechanisms for controlling gamma synchronization.…”

Collective excitability in a mesoscopic neuronal model of epileptic activity

“…While in many cases coupling merely coordinates dynamical regimes that are already present in the isolated elements, in others it underlies the emergence of novel behaviors that would not exist in the absence of interaction between the elements [4,5]. In the brain, examples of such emergent behavior exist at the microscopic scale of networks of neurons, in the form of, for instance, collective oscillations arising from a balance between excitation and inhibition [6,7] and recurrent up/down dynamics [8]. Much less is known, however, about the emergent behavior of the brain at the mesoscopic level of coupled brain areas.…”

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

“…Thus, 44 communication through coherence (CTC ) not only provides the means to communicate 45 from one network to another, but it also provides the means to control the 46 communication, because only networks with an appropriate phase synchrony with the 47 sender network can tune in to the spiking activity they receive. However, there are two 48 clear limitations to this mechanism: (1) Experimentally observed oscillations are not 49 stable over long enough times to support their role in communicating spiking 50 activity [22], and (2) The mechanisms by which two networks can generate coherent 51 oscillations have so far remained quite obscure (however, see [23,24]).…”

“…(pulse packets) which can be quantified by the number of spikes in the volley 23 (α = 50 − 100 spikes) and their temporal dispersion (σ ≈ 1 − 10 ms), measuring the 24 degree of synchronization of the spiking activity in the volley [18,20]. Several studies 25 have demonstrated that the downstream effect of a pulse packet depends both on α and 26 σ (see [7] for a review).…”