Limits...
Gamma power is phase-locked to posterior alpha activity.

Osipova D, Hermes D, Jensen O - PLoS ONE (2008)

Bottom Line: Our findings show that high-frequency gamma power (30-70 Hz) is phase-locked to alpha oscillations (8-13 Hz) in the ongoing MEG signals.The topography of the coupling was similar to the topography of the alpha power and was strongest over occipital areas.Interestingly, gamma activity per se was not evident in the power spectra and only became detectable when studied in relation to the alpha phase.

View Article: PubMed Central - PubMed

Affiliation: Donders Institute for Brain, Cognition and Behaviour, Radboud University Nijmegen, Nijmegen, The Netherlands. daria.osipova@fcdonders.ru.nl

ABSTRACT
Neuronal oscillations in various frequency bands have been reported in numerous studies in both humans and animals. While it is obvious that these oscillations play an important role in cognitive processing, it remains unclear how oscillations in various frequency bands interact. In this study we have investigated phase to power locking in MEG activity of healthy human subjects at rest with their eyes closed. To examine cross-frequency coupling, we have computed coherence between the time course of the power in a given frequency band and the signal itself within every channel. The time-course of the power was calculated using a sliding tapered time window followed by a Fourier transform. Our findings show that high-frequency gamma power (30-70 Hz) is phase-locked to alpha oscillations (8-13 Hz) in the ongoing MEG signals. The topography of the coupling was similar to the topography of the alpha power and was strongest over occipital areas. Interestingly, gamma activity per se was not evident in the power spectra and only became detectable when studied in relation to the alpha phase. Intracranial data from an epileptic subject confirmed these findings albeit there was slowing in both the alpha and gamma band. A tentative explanation for this phenomenon is that the visual system is inhibited during most of the alpha cycle whereas a burst of gamma activity at a specific alpha phase (e.g. at troughs) reflects a window of excitability.

Show MeSH

Related in: MedlinePlus

Coupling between alpha and gamma activity.A. Cross-frequency interaction in ongoing human MEG signals during eyes closed. The highlighted area indicates increased coherence between alpha activity (along the x-axis) and the power of the gamma activity (along the y-axis). B. Grand-average (black line) and standard deviation (red line) of log-transformed power. While a strong peak could be observed in the alpha band there was no detectable peak in the gamma band. C. Topography of cross-frequency coupling (the highlighted area in A). D. Topography of the log–transformed alpha power (9–11 Hz). E. Topography of the log–transformed gamma power (30–70 Hz). A to E are calculated from an average of 6 subjects. F. Correlation over 14 subjects between alpha power and the cross-frequency coupling in the gamma band. Mainly subjects with higher alpha power had significant cross-frequency interactions (shown in red).
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2602598&req=5

pone-0003990-g001: Coupling between alpha and gamma activity.A. Cross-frequency interaction in ongoing human MEG signals during eyes closed. The highlighted area indicates increased coherence between alpha activity (along the x-axis) and the power of the gamma activity (along the y-axis). B. Grand-average (black line) and standard deviation (red line) of log-transformed power. While a strong peak could be observed in the alpha band there was no detectable peak in the gamma band. C. Topography of cross-frequency coupling (the highlighted area in A). D. Topography of the log–transformed alpha power (9–11 Hz). E. Topography of the log–transformed gamma power (30–70 Hz). A to E are calculated from an average of 6 subjects. F. Correlation over 14 subjects between alpha power and the cross-frequency coupling in the gamma band. Mainly subjects with higher alpha power had significant cross-frequency interactions (shown in red).

Mentions: The cross-frequency measure was applied to the sensor data in each subject to investigate coherence between a low frequency signal and the time-course of the power at higher frequencies. Interaction was observed between alpha and gamma bands. The sensors with the strongest alpha-gamma coupling were identified subject by subject. The human gamma-band activity reported in different studies varies from 30 to 150 Hz (for a review, see [2]), possibly depending on a cognitive task, imaging method used and/or intersubject differences. Therefore, the boundaries of the gamma-band of a single subject may somewhat differ from those of the grand average. Cross-frequency coupling was significant in six subjects (four subjects: p<0.01; two subjects: p<0.05). Data from these subjects were subjected for further analysis. The cross-frequency representations were averaged showing coupling between the phase of alpha (8–12 Hz) and the power of gamma activity (30–70 Hz) (Fig. 1A). Note that the 10 to 20 Hz coupling is likely to be explained by the first harmonic of the alpha activity. The power spectra for those sensors were also averaged over subjects. Interestingly, while the alpha activity resulted in a strong 10 Hz peak, the gamma activity was not apparent as a peak in the spectrum (Fig. 1B). Then we extracted mean cross-frequency values for all sensors for a 8–12 Hz by 30–70 Hz tile (white rectangle in Fig. 1A). The corresponding topography averaged over sensors is shown in Fig. 1C. The posterior distribution resembled the topography of the ∼10 Hz alpha power (Fig. 1D). Next, we tested if the strength of the alpha power correlated with degree of cross-frequency coupling in all 14 subjects (Fig. 1F). This resulted in a significant correlation suggesting that strong alpha power (Spearman r = 0.68, p<0.01) is a prerequisite for the cross-frequency coupling. Although the differences in alpha-gamma coupling and alpha power between subjects cannot be attributed to the amount of data used in the analysis, it remains hard to dissociate whether inter-individual differences in the observed effect are due to differences in signal-to-noise or reflect an underlying physiological phenomenon. It is important to note, however, that the amount of data used in the present study resembles the amount of data used in a typical cognitive paradigm (100 trials of 1 s). Thus there is a realistic chance to detect cross-frequency coupling even in the absence of an extended recording.


Gamma power is phase-locked to posterior alpha activity.

Osipova D, Hermes D, Jensen O - PLoS ONE (2008)

Coupling between alpha and gamma activity.A. Cross-frequency interaction in ongoing human MEG signals during eyes closed. The highlighted area indicates increased coherence between alpha activity (along the x-axis) and the power of the gamma activity (along the y-axis). B. Grand-average (black line) and standard deviation (red line) of log-transformed power. While a strong peak could be observed in the alpha band there was no detectable peak in the gamma band. C. Topography of cross-frequency coupling (the highlighted area in A). D. Topography of the log–transformed alpha power (9–11 Hz). E. Topography of the log–transformed gamma power (30–70 Hz). A to E are calculated from an average of 6 subjects. F. Correlation over 14 subjects between alpha power and the cross-frequency coupling in the gamma band. Mainly subjects with higher alpha power had significant cross-frequency interactions (shown in red).
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2602598&req=5

pone-0003990-g001: Coupling between alpha and gamma activity.A. Cross-frequency interaction in ongoing human MEG signals during eyes closed. The highlighted area indicates increased coherence between alpha activity (along the x-axis) and the power of the gamma activity (along the y-axis). B. Grand-average (black line) and standard deviation (red line) of log-transformed power. While a strong peak could be observed in the alpha band there was no detectable peak in the gamma band. C. Topography of cross-frequency coupling (the highlighted area in A). D. Topography of the log–transformed alpha power (9–11 Hz). E. Topography of the log–transformed gamma power (30–70 Hz). A to E are calculated from an average of 6 subjects. F. Correlation over 14 subjects between alpha power and the cross-frequency coupling in the gamma band. Mainly subjects with higher alpha power had significant cross-frequency interactions (shown in red).
Mentions: The cross-frequency measure was applied to the sensor data in each subject to investigate coherence between a low frequency signal and the time-course of the power at higher frequencies. Interaction was observed between alpha and gamma bands. The sensors with the strongest alpha-gamma coupling were identified subject by subject. The human gamma-band activity reported in different studies varies from 30 to 150 Hz (for a review, see [2]), possibly depending on a cognitive task, imaging method used and/or intersubject differences. Therefore, the boundaries of the gamma-band of a single subject may somewhat differ from those of the grand average. Cross-frequency coupling was significant in six subjects (four subjects: p<0.01; two subjects: p<0.05). Data from these subjects were subjected for further analysis. The cross-frequency representations were averaged showing coupling between the phase of alpha (8–12 Hz) and the power of gamma activity (30–70 Hz) (Fig. 1A). Note that the 10 to 20 Hz coupling is likely to be explained by the first harmonic of the alpha activity. The power spectra for those sensors were also averaged over subjects. Interestingly, while the alpha activity resulted in a strong 10 Hz peak, the gamma activity was not apparent as a peak in the spectrum (Fig. 1B). Then we extracted mean cross-frequency values for all sensors for a 8–12 Hz by 30–70 Hz tile (white rectangle in Fig. 1A). The corresponding topography averaged over sensors is shown in Fig. 1C. The posterior distribution resembled the topography of the ∼10 Hz alpha power (Fig. 1D). Next, we tested if the strength of the alpha power correlated with degree of cross-frequency coupling in all 14 subjects (Fig. 1F). This resulted in a significant correlation suggesting that strong alpha power (Spearman r = 0.68, p<0.01) is a prerequisite for the cross-frequency coupling. Although the differences in alpha-gamma coupling and alpha power between subjects cannot be attributed to the amount of data used in the analysis, it remains hard to dissociate whether inter-individual differences in the observed effect are due to differences in signal-to-noise or reflect an underlying physiological phenomenon. It is important to note, however, that the amount of data used in the present study resembles the amount of data used in a typical cognitive paradigm (100 trials of 1 s). Thus there is a realistic chance to detect cross-frequency coupling even in the absence of an extended recording.

Bottom Line: Our findings show that high-frequency gamma power (30-70 Hz) is phase-locked to alpha oscillations (8-13 Hz) in the ongoing MEG signals.The topography of the coupling was similar to the topography of the alpha power and was strongest over occipital areas.Interestingly, gamma activity per se was not evident in the power spectra and only became detectable when studied in relation to the alpha phase.

View Article: PubMed Central - PubMed

Affiliation: Donders Institute for Brain, Cognition and Behaviour, Radboud University Nijmegen, Nijmegen, The Netherlands. daria.osipova@fcdonders.ru.nl

ABSTRACT
Neuronal oscillations in various frequency bands have been reported in numerous studies in both humans and animals. While it is obvious that these oscillations play an important role in cognitive processing, it remains unclear how oscillations in various frequency bands interact. In this study we have investigated phase to power locking in MEG activity of healthy human subjects at rest with their eyes closed. To examine cross-frequency coupling, we have computed coherence between the time course of the power in a given frequency band and the signal itself within every channel. The time-course of the power was calculated using a sliding tapered time window followed by a Fourier transform. Our findings show that high-frequency gamma power (30-70 Hz) is phase-locked to alpha oscillations (8-13 Hz) in the ongoing MEG signals. The topography of the coupling was similar to the topography of the alpha power and was strongest over occipital areas. Interestingly, gamma activity per se was not evident in the power spectra and only became detectable when studied in relation to the alpha phase. Intracranial data from an epileptic subject confirmed these findings albeit there was slowing in both the alpha and gamma band. A tentative explanation for this phenomenon is that the visual system is inhibited during most of the alpha cycle whereas a burst of gamma activity at a specific alpha phase (e.g. at troughs) reflects a window of excitability.

Show MeSH
Related in: MedlinePlus