Edited by: Pascal Belin, University of Glasgow, UK
Reviewed by: Kristina Simonyan, Mount Sinai School of Medicine, USA; Jonas Obleser, Max Planck Institute for Human Cognitive and Brain Sciences, Germany; Yukiko Kikuchi, Newcastle University Medical School, UK
*Correspondence: Joji Tsunada, Department of Otorhinolaryngology: Head and Neck Surgery, University of Pennsylvania School of Medicine, 3400 Spruce – 5 Ravdin, Philadelphia, PA 19104, USA. e-mail:
This article was submitted to Frontiers in Auditory Cognitive Neuroscience, a specialty of Frontiers in Psychology.
This is an open-access article subject to a non-exclusive license between the authors and Frontiers Media SA, which permits use, distribution and reproduction in other forums, provided the original authors and source are credited and other Frontiers conditions are complied with.
Adaptive behavior depends on an animal’s ability to ignore uninformative stimuli, such as repeated presentations of the same stimulus, and, instead, detect informative, novel stimuli in its environment. The primate prefrontal cortex (PFC) is known to play a central role in this ability. However, the neural mechanisms underlying the ability to differentiate between repeated and novel stimuli are not clear. We hypothesized that the coupling between different frequency bands of the local field potential (LFP) underlies the PFC’s role in differentiating between repeated and novel stimuli. Specifically, we hypothesized that whereas the presentation of a novel-stimulus induces strong cross-frequency coupling, repeated presentations of the same stimulus attenuates this coupling. To test this hypothesis, we recorded LFPs from the ventrolateral PFC (vPFC) of rhesus monkeys while they listened to a novel vocalization and repeated presentations of the same vocalization. We found that the cross-frequency coupling between the gamma-band amplitude and theta-band phase of the LFP was modulated by repeated presentations of a stimulus. During the first (novel) presentation of a stimulus, gamma-band activity was modulated by the theta-band phase. However, with repeated presentations of the same stimulus, this cross-frequency coupling was attenuated. These results suggest that cross-frequency coupling may play a role in the neural computations that underlie the differentiation between novel and repeated stimuli in the vPFC.
A key characteristic of adaptive behavior is the ability of animals to ignore repeated “uninformative” stimuli (i.e., repeated presentations of the same stimulus) and, instead, devote neural resources to the detection of novel, and likely more informative, stimuli in the environment. Given the key role of the prefrontal cortex (PFC) in adaptive behavior and executive functions (Miller and Cohen,
One possible means by which PFC activity could encode novel versus repeated stimuli is through the interplay between the spiking activity of different neural populations with different tuning profiles (Desimone,
However, since spiking activity and gamma-band oscillations are often coupled with other frequency bands (Bragin et al.,
Here, we tested whether gamma-band oscillations were coupled to other frequency bands and whether this coupling was modulated by the presentation of novel and repeated stimuli. To test this hypothesis, we recorded LFPs from the ventrolateral PFC (vPFC) of rhesus monkeys while they listened to a vocalization (i.e., the novel stimulus) that was followed by two to four repeated presentations of the same vocalization. We found that the cross-frequency coupling between the gamma-band amplitude and theta-band phase of LFP was attenuated by repeated presentations of a stimulus. These results suggest that the cross-frequency coupling between the gamma and theta oscillations may contribute to a mechanism underlying the brain’s capacity to distinguish between novel and repeated stimuli.
Local field potentials and single-unit activity were recorded from two adult male rhesus macaque monkeys (
The stimuli were auditory- and visual-communication signals from a rhesus monkey. These stimuli were provided by Dr. Asif Ghazanfar and detailed information of these stimuli was provided in previous studies (Lewkowicz and Ghazanfar,
For the visual-communication signals, we presented a silent movie of the monkey producing one of the three vocalizations (the
We developed the passive-listening and -viewing paradigm, which is a modified version of the classic oddball paradigm (Näätänen,
Specifically, each monkey participated in the passive-listening and -viewing paradigm (Figure
In each recording session, all the three vocalizations (i.e., the
Neurophysiological recordings were performed with a tungsten microelectrode (1–2 MΩ at 1 kHz; Frederick Haer & Co.) that was seated in a stainless-steel guide tube. The electrode and guide tube were advanced into the brain with a hydraulic microdrive (MO-95, Narishige). The neural signal was sampled at 24 kHz and band-pass filtered between 2.2 Hz and 6 kHz with a pre-amplifier (RA16PA, Tucker-Davis Technologies) and an amplifier (RZ2, Tucker-Davis Technologies) and stored for offline analysis.
Neural activity was recorded from the left vPFC of Monkey T and from the right vPFC of Monkey G. All recording sites were guided by pre- and post-operative magnetic resonance images of each monkey’s brain. The vPFC was identified by its anatomical location (i.e., anterior to the arcuate sulcus and Area 8a and below the principal sulcus; areas 45 and 12) and its neurophysiological properties (Romanski and Goldman-Rakic,
Electrodes were advanced into the vPFC until single-unit activity was identified (∼1–3 mm from the cortical surface). After single-unit activity was well isolated from a recording site, the monkeys participated in the passive-listening and -viewing paradigm. Neural activity was recorded while the recording properties of the site remained stable; typically, we were able to collect data from ≥75 trials of the paradigm in each recording session.
The LFPs were extracted using methods similar to those described in previous studies (Ghazanfar et al.,
Neural activity was first low-pass filtered with a 300-Hz cut-off frequency using a four-pole bidirectional Butterworth filter. Next, the filtered signals were resampled at 1 kHz. When necessary, 60-Hz (line) noise was removed from the signals. We analyzed three different types of LFP signals: the “total”-LFP power, stimulus-evoked LFP, and induced LFP. Total-LFP power, which reflects both phase-locked and non-phase-locked neural signals, was obtained by calculating the power spectrum from individual trials with a Morlet-wavelet decomposition (Sinkkonen et al.,
The LFPs were divided into different analysis periods. The “baseline” period was a 250-ms period that preceded the onset of the first vocalization of each trial. For the
The amplitudes of the band-limited signals of the induced and total LFPs were tested by first applying Butterworth filters with different pass-bands (4–10 Hz filter for the theta band; 10–15 Hz filter for the alpha band; 15–25 Hz filter for the beta band; and 25–50 Hz filter for the gamma band) and then by applying a Hilbert transformation to calculate the envelope of these band-limited signals (Chandrasekaran and Ghazanfar,
The phase-locking of the total-LFP across trials was calculated using two different approaches (Tallon-Baudry et al.,
In the second approach, we were interested only in the phase-locking of the band-limited signals (e.g., the theta band). To test the phase-locking of these signals, for each trial, we first applied Butterworth filters to the LFPs with different pass-bands and then applied a Hilbert transformation to these filtered signals to obtain a unit vector at each time point. We then averaged these trial-by-trial unit vectors. The phase-locking value was calculated using a procedure analogous to the one described above.
To quantify the effect that the repeated presentations of a vocalization had on the LFPs, we calculated a modulation index. The general form of the index was: (P1st − P2nd)/(P1st + P2nd). P was either the peak amplitude or the peak phase-locking value of the band-limited signals. The superscript “1st” or “2nd” refers to whether these values were obtained from the first or second presentation of a vocalization, respectively. The value of the index ranges between −1 and 1. If the index value is >0, it indicates that the P1st value was greater than the P2nd. If the index value is <0, it indicates that the P2nd value was greater than the P1st.
Next, the hypothesis that a novel-stimulus induces cross-frequency coupling between the gamma-band amplitude and the phase of other frequency bands of the LFPs was tested (Bragin et al.,
In a second analysis, we tested the cross-frequency coupling between the frequency bands of the induced LFP. On a trial-by-trial basis, we calculated the amplitude of the gamma band during the first 300 ms following stimulus onset. Next, on a site-by-site basis, we formed distributions of the gamma-band amplitude as a function of the phase of the theta, alpha, or beta bands; similar to the previous analysis, the amplitude distributions were divided amongst six phase bins. A permutation test examined whether our measured amplitude distribution was reliably different from a shuffled distribution. We used a permutation test that is analogous to the one described above except that we used amplitude instead of percentage of trials.
We recorded the LFPs from 168 recording sites in the vPFC of two rhesus monkeys (monkey T: 103 sites, monkey G: 65 sites). Neural data were recorded while the monkeys listened to three to five repeated presentations of a vocalization. Of the 168 sites, we found that total-LFP power increased at 79 sites (monkey T: 59 sites, monkey G: 20 sites) during the presentation of at least one type of vocalization (i.e., auditory-modulated sites).
Figure
Figures
To quantify these observations, we calculated the mean and peak amplitude of different band-limited LFPs as a function of the number of repeated vocalizations. We found that both the mean and the peak amplitude of the gamma-, as well as the theta-, alpha-, and beta-bands, decreased reliably between the first and second vocalization presentations (one-way ANOVA with
Next, we tested the hypothesis that the phase-locking of the LFPs attenuated with repeated presentations of a vocalization (Jansen et al.,
To quantify these observations, we calculated the peak phase-locking value as a function of the number of repeated vocalizations. We found that the peak phase-locking value of the gamma-, as well as the theta-, alpha-, and beta-bands, decreased reliably between the first and second vocalization presentations (one-way ANOVA with
Next, we tested the hypothesis that the peak amplitude of the gamma band occurred at specific phases of other LFP frequency bands. Moreover, we hypothesized that this relationship between the gamma and other frequency bands was strongest during the first-stimulus (i.e., the novel stimulus) presentation and decreased with repeated presentations of a vocalization.
To test these hypotheses, we analyzed the cross-frequency coupling between frequency bands of the total LFP. Figure
The cross-frequency coupling between the peak amplitude of the gamma band and theta-band phase may not simply be a function of stimulus presentation. Indeed, we speculated that the theta-band phase may also covary with the magnitude of the gamma band’s peak amplitude (Lakatos et al.,
However, since the aforementioned analyses used the total LFP (which includes components of both the stimulus-evoked and induced LFP), our findings may simply reflect the inherent phase-locking that occurs with stimulus-evoked activity and the neural habituation that normally occurs with repeated-stimulus presentations. To further explore this issue, we tested the cross-frequency coupling between bands of the induced LFP. Similar to the results from the total LFP, we found that the gamma-band amplitude of the induced-LFP decreased with stimulus repetition (Figure
Next, we examined the cross-frequency coupling between bands of the induced LFP. Figure
We found that the magnitude and the phase-locking of the LFPs in the vPFC decreased with repeated presentations of a vocalization. Additionally, the cross-frequency coupling between the gamma and theta bands was attenuated by repeated presentations of a vocalization. Specifically, during the first (novel) presentation of a stimulus, the peak of the gamma-band activity occurred most often at specific theta-band phases, and these specific phases induced large-amplitude gamma oscillations, relative to other phases; see Figures
In the vPFC, the theta, alpha, beta, and gamma bands of the LFPs were modulated by all three vocalizations (Figure
Similar to our results in the vPFC, previous studies have found that vocalizations modulate low-frequency (<20 Hz) LFPs in both the auditory cortex and the superior temporal sulcus (Ghazanfar et al.,
In general, our finding that repeated presentations of an auditory stimulus reduces the power and phase-locking of the LFP is consistent with a large body of literature that has examined this issue in early regions of the auditory cortex (Rosburg,
We found that during the first presentation of a vocalization, the peak of the gamma band of the total-LFP coincided with specific phases of the theta oscillation (Figure
The high degree of phase-locking of the theta band (see Figures
Importantly, this relationship does not simply reflect the simultaneous occurrence of stimulus-evoked gamma- and theta-band phase-locked activity because the gamma-theta coupling was also seen in the induced LFP, which reflects non-phase-locking activity (Figure
Finally, similar to previous studies (Lakatos et al.,
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
We would like to thank Asif Ghazanfar for providing stimuli and helpful comments on previous versions of the manuscript, Chandramouli Chandrasekaran for help with the analyses and Jung Hoon Lee, Maria Geffen, and Heather Hersh for helpful comments on the preparation of this manuscript. Yale E. Cohen was supported by grants from the NIDCD-NIH.
1
2